Indoor Air Quality Book 2021

Indoor Air Quality
Edited by
Dikaia E. Saraga
Printed Edition of the Special Issue Published in Applied Sciences
www.mdpi.com/journal/applsci
Indoor Air Quality
Indoor Air Quality
Editor
Dikaia E. Saraga
MDPI • Basel • Beijing • Wuhan • Barcelona • Belgrade • Manchester • Tokyo • Cluj • Tianjin
Editor
Dikaia E. Saraga
National Centre for Scientific Research “Demokritos”
Greece
Editorial Office
MDPI
St. Alban-Anlage 66
4052 Basel, Switzerland
This is a reprint of articles from the Special Issue published online in the open access journal
Applied Sciences (ISSN 2076-3417) (available at: https://www.mdpi.com/journal/applsci/special
issues/Indoor Quality).
For citation purposes, cite each article independently as indicated on the article page online and as
indicated below:
LastName, A.A.; LastName, B.B.; LastName, C.C. Article Title. Journal Name Year, Volume Number,
Page Range.
ISBN 978-3-03943-703-0 (Hbk)

ISBN 978-3-03943-704-7 (PDF)
Cover image courtesy of Dikaia E. Saraga.

c 2020 by the authors. Articles in this book are Open Access and distributed under the Creative
Commons Attribution (CC BY) license, which allows users to download, copy and build upon
published articles, as long as the author and publisher are properly credited, which ensures maximum
dissemination and a wider impact of our publications.
The book as a whole is distributed by MDPI under the terms and conditions of the Creative Commons
license CC BY-NC-ND.
Contents
About the Editor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
Dikaia E. Saraga
Special Issue on Indoor Air Quality
Reprinted from: Appl. Sci. 2020, 10, 1501, doi:10.3390/app10041501 . . . . . . . . . . . . . . . . . 1
Myung Eun Cho and Mi Jeong Kim

Residents’ Perceptions of and Response Behaviors to Particulate Matter—A Case Study in
Seoul, Korea
Reprinted from: Appl. Sci. , 9, 3660, doi:10.3390/app9183660 . . . . . . . . . . . . . . . . . . . . . 5
Ioannis Sakellaris, Dikaia Saraga, Corinne Mandin, Yvonne de Kluizenaar, Serena Fossati,
Andrea Spinazzè, Andrea Cattaneo, Tamas Szigeti, Victor Mihucz,
Eduardo de Oliveira Fernandes, Krystallia Kalimeri, Paolo Carrer and John Bartzis
Personal Control of the Indoor Environment in Offices: Relations with Building Characteristics,
Influence on Occupant Perception and Reported Symptoms Related to the Building—The
Officair Project
Reprinted from: Appl. Sci. 2019, 9, 3227, doi:10.3390/app9163227 . . . . . . . . . . . . . . . . . . . 21
Jongseong Gwak, Motoki Shino, Kazutaka Ueda and Minoru Kamata

An Investigation of the Effects of Changes in the Indoor Ambient Temperature on Arousal
Level, Thermal Comfort, and Physiological Indices
Katja Tähtinen, Sanna Lappalainen, Kirsi Karvala, Marjaana Lahtinen and Heidi Salonen
Probability of Abnormal Indoor Air Exposure Categories Compared with Occupants’
Symptoms, Health Information, and Psychosocial Work Environment
Reprinted from: Appl. Sci. 2019, 9, 99, doi:10.3390/app9010099 . . . . . . . . . . . . . . . . . . . . 63
Michał Piasecki and Krystyna Barbara Kostyrko

Combined Model for IAQ Assessment: Part 1—Morphology of the Model and Selection of
Substantial Air Quality Impact Sub-Models
Zhengguo Yang, Yuto Lim and Yasuo Tan

An Accident Model with Considering Physical Processes for Indoor Environment Safety
Reprinted from: Appl. Sci. 2019, 9, 4732, doi:10.3390/app9224732 . . . . . . . . . . . . . . . . . . 115
Xiaogang Cheng, Bin Yang, Kaige Tan, Erik Isaksson, Liren Li, Anders Hedman,
Thomas Olofsson and Haibo Li
A Contactless Measuring Method of Skin Temperature based on the Skin Sensitivity
Index and Deep Learning
Anna Mainka and Elwira Zajusz-Zubek

Keeping Doors Closed as One Reason for Fatigue in Teenagers—A Case Study
Yu-Chuan Yen, Chun-Yuh Yang, Kristina Dawn Mena, Yu-Ting Cheng and Pei-Shih Chen
Cooking/Window Opening and Associated Increases of Indoor PM2.5 and NO2 Concentrations
of Children’s Houses in Kaohsiung, Taiwan
v
Sungroul Kim, Sujung Park and Jeongeun Lee
Evaluation of Performance of Inexpensive Laser Based PM2.5 Sensor Monitors for Typical
Indoor and Outdoor Hotspots of South Korea
Mehmet Taştan and Hayrettin Gökozan

Real-Time Monitoring of Indoor Air Quality with Internet of Things-Based E-Nose
Gonçalo Marques and Rui Pitarma

An Internet of Things-Based Environmental Quality Management System to Supervise the
Indoor Laboratory Conditions
Thomas Maggos, Vassilios Binas, Vasileios Siaperas, Antypas Terzopoulos, Panagiotis

Panagopoulos and George Kiriakidis
A Promising Technological Approach to Improve Indoor Air Quality
Junjie Li, Shuai Lu, Qingguo Wang, Shuo Tian and Yichun Jin
Study of Passive Adjustment Performance of Tubular Space in Subway Station
Building Complexes
Pietro Grisoli, Marco Albertoni and Marinella Rodolfi

Application of Airborne Microorganism Indexes in Offices, Gyms, and Libraries
Anna M. Marcelloni, Alessandra Chiominto, Simona Di Renzi, Paola Melis, Annarita Wirz,
Maria C. Riviello, Stefania Massari, Renata Sisto, Maria C. D’Ovidio and Emilia Paba
How Working Tasks Influence Biocontamination in an Animal Facility
vi
About the Editor
Dikaia E. Saraga is an Associate Researcher at the Institute of Nuclear & Radiological Sciences
and Technology, Energy & Safety, NCSR “Demokritos”, Greece. She holds a BSc in Physics (University
of Athens, Greece), an MSc in Environmental Physics (University of Athens, Greece) and a PhD
in Indoor Air Pollution (University of West Macedonia, Greece). Her full operating skills include,
among others: indoor/outdoor field air quality measurements, application of reference methods
for PM and VOC determination, ionic and OC/EC analysis; receptor modeling application for
source apportionment studies. Her research interests also include: air pollution and climate change
mitigation, population exposure, air quality sensors, big data treatment. She has participated in
more than 20 European (Horizon 20202, FP6, FP7, Life+) and national research projects in the field
of air quality. She is a member of the management board of COST17136 INDAIRPOLLNET action
and a member of the FAIRMODE group. She is a recognized reviewer of 18 scientific journals and
author of 44 publications in peer-reviewed journals, and more than 80 conference proceedings and
scientific reports.
vii
applied
sciences
Editorial
Special Issue on Indoor Air Quality
Dikaia E. Saraga
National Center for Scientific Research “Demokritos”, Aghia Paraskevi, 15310 Athens, Greece;
[email protected]
Received: 8 February 2020; Accepted: 11 February 2020; Published: 22 February 2020
1. Introduction
It is a fact that people in developed countries spend almost 90% of their time indoors, where
they experience their greatest exposures. However, regulation of air pollution focuses on outdoor air,
as indoor environment is less well-characterized and recognized as a potential location for exposure to
air pollution.
What makes indoor air intrinsically more interesting than outdoor air from a scientific point
of view? Some sources are undoubtedly uniquely building-related (e.g., cleaning agents, emissions
from building materials and personal care products), while some contaminant dynamics operate
only in buildings (e.g., the distribution of particles and gases by mechanical ventilation systems and
photochemical reactions, and the infiltration of soil gases). Besides this, air pollutant concentrations
are often higher indoors than outdoors, particularly following activities such as cleaning and cooking
(with a greater source strength indoors than outdoor on a per area basis), while it has already been
proven that many indoor air pollutants are harmful to human health.
Another issue is ventilation. While indoor microenvironments are a microcosm of most urban
settings, the effective air exchange and renewal in buildings is much lower than outdoors, even in
urban areas. This concern is amplified by the fact that energy efficiency measures, driven by climate
change awareness, have made modern buildings more airtight, further degrading the quality of indoor
air. Therefore, a person is significantly more likely to inhale a harmful chemical molecule if it is emitted
indoors rather than outdoors.
Monitoring of indoor air pollutants in a spatio-temporal basis is challenging. A key element is
the access to local (i.e., indoor residential, workplace, or public building) exposure measurements.
Unfortunately, the high cost and complexity of most current air pollutant monitors results in a lack
of detailed spatial and temporal resolution. Therefore, individuals of vulnerable groups (children,
pregnant, elderly, and sick people) have little insight into their personal exposure levels. This
becomes significant in cases of hyper-local variations and short-term pollution events such as instant
indoor activity (e.g., cooking, smoking, and dust resuspension). Advances in sensor miniaturization
have encouraged the development of small, inexpensive devices capable of estimating pollutant
concentrations. This new class of sensors presents new possibilities for indoor exposure monitoring.
Furthermore, indoor air chemistry models typically account for air exchange with outdoors
through ventilation, deposition on indoor surfaces, and photochemical reactions. Surface chemistry on
furnishings, building materials, and human bodies is becoming increasingly recognized as being of
crucial importance.
In light of the above, this Special issue on ‘Indoor Air Quality’ was introduced to collect latest
research and address challenging issues in the areas of the triptych: Indoor environment quality
monitoring, indoor air modeling, and exposure to indoor air pollution.
2. From Indoor Environment Quality Perception to Indoor Air Quality Monitoring and Control
In this Special Issue, 24 papers were submitted, and 16 were accepted for publication (67%
acceptance rate). 63% of the published studies originate from Europe, while 37% of them were
Appl. Sci. 2020, 10, 1501; doi:10.3390/app10041501 1 www.mdpi.com/journal/applsci

Appl. Sci. 2020, 10, 1501
conducted in Asiatic countries. Various topics have been addressed in the contributed articles: Indoor
air quality (IAQ) monitoring and modelling, occupants’ comfort related to indoor environment
parameters as well as innovative techniques for IAQ monitoring and improvement. When looking
back, it can be concluded that the majority of the studies can be distinguished into two main groups.
The first one refers to occupants’ perception for the quality of the indoor environment as well as their
comfort inside a building. The second group focuses on new techniques of monitoring and controlling
the parameters determining the quality of indoor air. Finally, a quite smaller group includes studies
performed in indoor environments of special characteristics.
To be more specific, nine papers discuss the issues of IAQ perception and control as well as
thermal comfort. The study of M. Cho and M. Kim, [1] titled ‘Residents’ Perceptions of and Response
Behaviors to Particulate Matter—A Case Study in Seoul, Korea’ aimed at understanding the perception
of 171 people in Seoul for indoor air quality based on domestic particulate matter levels. In a
European-scale study titled ‘Personal Control of the Indoor Environment in Offices: Relations with Building
Characteristics, Influence on Occupant Perception and Reported Symptoms Related to the Building—The Officair
Project’, Sakellaris I. et al. [2] focused on revealing the complex relationship between office employees’
control over various indoor environment parameters and their comfort, health and productivity. 7441
occupants of 167 recently built or retrofitted office buildings in eight European countries participated
in an online survey about personal/health/work data as well as physical/psycho-social information.
In another study titled ‘An Investigation of the Effects of Changes in the Indoor Ambient Temperature on
Arousal Level, Thermal Comfort, and Physiological Indices’, Gwak J. et al. [3] aimed to design a thermal
environment that improves both the arousal level and thermal comfort of the occupants. To this end,
they investigated the relationships between the physiological indices, subjective evaluation values,
and task performance under several conditions of changes in the indoor ambient temperature. The
study ‘Probability of Abnormal Indoor Air Exposure Categories Compared with Occupants’ Symptoms, Health
Information, and Psychosocial Work Environment’ authored by Tähtinen K. et al. [4] aimed at (i) evaluating
the relation between the four-level categorized probability of abnormal indoor air exposure and
employees’ work environment-related symptoms, group-level health information, and psychosocial
work environment, (ii) assessing the relation between ventilation system deficiencies and employees’
work environment-related symptoms and evaluating the impact of prolonged IAQ problem solution
processes on perceived IAQ. The study ‘Combined Model for IAQ Assessment: Part 1—Morphology of the
Model and Selection of Substantial Air Quality Impact Sub-Models’ of Piasecki M. and Kostyrko K.B. [5]
provided an overview of models defining occupants’ comfort and satisfaction with IAQ. Specifically,
subcomponents of three potential IAQ models were classified according to their application potential:
IAQ quality index, IAQ comfort index, and an overall health and comfort index. The authors provide
a method for using the combined IAQ index to determine the indoor environmental quality index,
IEQ and a practical case study which provides IAQ and IEQ model implementation for a large
office building assessment. The study titled ‘An Accident Model with Considering Physical Processes for
Indoor Environment Safety’ authored by Yang Z. et al. [6] also deals with thermal comfort in an indoor
environment. In particular, authors presented an extension of Systems-Theoretic Accident Model
and Process (STAMP) while considering physical processes in an indoor environment as temperature
changes. Thermal comfort was also the subject of the study of Cheng X. et al. [7] titled ‘A Contactless
Measuring Method of Skin Temperature based on the Skin Sensitivity Index and Deep Learning’. In this study,
a skin sensitivity index was proposed to describe individual sensitivity of thermal comfort, and the
index was combined with skin images for deep learning network training. Also, a novel contactless
measuring algorithm (NISDL) based on SSI was proposed, with two different frameworks of NISDL
having been designed for real-time thermal comfort measurements. Finally, a deep learning algorithm
without SSI was also generated and trained. Two more studies have worked on the evaluation of the
indoor environment quality while focusing on a specific population group: Children and teenagers.
The study of Mainka A. and Zajusz-Zubek E., [8] named ‘Keeping Doors Closed as One Reason for Fatigue
in Teenagers—A Case Study’ investigated the variability of CO2 concentration in naturally ventilated
2
Appl. Sci. 2020, 10, 1501
bedrooms of teenagers in Polland, by correlating bedroom door opening during the night with CO2
concentration and thermal comfort. In the study ‘Cooking/Window Opening and Associated Increases of
Indoor PM2.5 and NO2 Concentrations of Children’s Houses in Kaohsiung, Taiwan’, Yen Y. et al. [9] attempted
to assess the influence of window opening and cooking activity to measured air pollutants levels in 60
children homes in an industrial city in Taiwan.
The second sub-group of papers published in this Special Issue, includes five studies which feature
the introduction of new developments in technology and computational science to the field of indoor
environment monitoring and control. The study of Kim S. et al. [10] titled ‘Evaluation of Performance
of Inexpensive Laser Based PM2.5 Sensor Monitors for Typical Indoor and Outdoor Hotspots of South Korea’
presents the results of the evaluation of a low-cost real-time PM monitor under indoor testing with
common PM2.5 sources of Korea (frying pork in a pan or smoking). In another study, ‘Real-Time
Monitoring of Indoor Air Quality with Internet of Things-Based E-Nose’, authored by Tastan M. [11] and
Gokozan H., an ‘e-nose’, a real-time mobile air quality monitoring system with various air parameters
such as CO2 , CO, PM10, NO2 temperature and humidity was presented and evaluated. The proposed
e-nose is produced with an open source, low cost, easy installation and do-it-yourself approach. An
environmental quality solution based on IoT to supervise Laboratory Environmental Conditions (LEC)
named iAQ+ was introduced in the paper titled ‘An Internet of Things-Based Environmental Quality
Management System to Supervise the Indoor Laboratory Conditions’ authored be Marques C. and Pitarma
R [12]. This low-cost wireless solution for indoor environment quality supervision incorporates mobile
computing technologies for data consulting, easy installation, significant notifications for enhanced
living conditions, and laboratory activities. Further, Cheng X. et al. [7] study (presented in the previous
paragraph) belongs to this sub-group as it included training of deep learning network. Finally, the
study ‘A Promising Technological Approach to Improve Indoor Air Quality’ authored by Maggos T. et al. [13]
presents an innovative paint material which exhibits intense photocatalytic activity under direct and
diffused visible light for the degradation of air pollutants, suitable for indoor use. This innovative
photo-paint was tested under laboratory and real scale conditions.
Last but not least, three papers published in this Special Issue include studies of air quality in
indoor environments of special characteristics. The first one, ‘Study of Passive Adjustment Performance of
Tubular Space in Subway Station Building Complexes’ by Li J. et al. [14] focused on the various tubular space
forms in subway station building complexes with the scope of proposing an improvement of the indoor
environment in terms of comfort and energy consumption. The second one, ‘Application of Airborne
Microorganism Indexes in Offices, Gyms, and Libraries’ authored by Grisoli P. et al. [15] quantified the
levels of microorganisms present in the air in different places such as offices, gyms, and libraries. The
third one, ‘How Working Tasks Influence Biocontamination in an Animal Facility’ authored by Marcelloni
M. et al. [16] aimed to determine what factors could be associated with a high level of exposure to
biological agents in an animal facility, through the measuring and characterization of airborne fungi,
bacteria, endotoxin, (1,3)-β-d-glucan and animal allergens.
3. Conclusions
As a final point, the papers in this Special Issue have pinpointed two thematic areas of IAQ that
researchers currently focus on, and basically, answer two questions: (i) How do people perceive the
quality of air inside their home, office, school etc.? and (ii) what are the state of the art tools (both
instrumentational and computational) to monitor, control and improve indoor air quality? While this
Special Issue has been closed, further research towards these directions is expected in the very near
future. There are still several challenging research questions to be answered. Manuscripts addressing
challenging future research for Indoor Air Quality are invited in the second volume, named ‘New
Challenges for Indoor Air Quality’ launched by Applied Sciences, MDPI.
Author Contributions: D.E.S. performed the papers’ review and wrote the editorial. The author has read and
agreed to the published version of the manuscript.
3
Appl. Sci. 2020, 10, 1501
Acknowledgments: This issue would not be possible without the contributions of various talented authors,
hardworking and professional reviewers, and dedicated editorial team of Applied Sciences. Congratulations
to all authors—no matter what the final decisions of the submitted manuscripts were, the feedback, comments,
and suggestions from the reviewers and editors which substantially helped the authors to improve their papers.
Finally, I place on record my gratitude to the editorial team of Applied Sciences, and special thanks to Daria Shi,
Assistant Editor from MDPI Branch Office, Beijing.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Cho, M.; Kim, M. Residents’ perceptions of and response behaviors to particulate matter—A case study in
Seoul, Korea. Appl. Sci. 2019, 9, 3660. [CrossRef]
2. Sakellaris, I.; Saraga, D.; Mandin, C.; de Kluizenaar, Y.; Fossati, S.; Spinazzè, A.; Cattaneo, A.; Szigeti, T.;
Mihucz, V.; de Oliveira Fernandes, E.; et al. Personal control of the indoor environment in offices: Relations
with building characteristics, influence on occupant perception and reported symptoms related to the
building—The officair project. Appl. Sci. 2019, 9, 3227. [CrossRef]
3. Gwak, J.; Shino, M.; Ueda, K.; Kamata, M. An investigation of the effects of changes in the indoor ambient
temperature on arousal level, thermal comfort, and physiological indices. Appl. Sci. 2019, 9, 899. [CrossRef]
4. Tähtinen, K.; Lappalainen, S.; Karvala, K.; Lahtinen, M.; Salonen, H. Probability of abnormal indoor air
exposure categories compared with occupants’ symptoms, health information, and psychosocial work
environment. Appl. Sci. 2019, 9, 99. [CrossRef]
5. Piasecki, M.; Kostyrko, K. Combined model for IAQ assessment: Part 1—Morphology of the model and
selection of substantial air quality impact sub-models. Appl. Sci. 2019, 9, 3918. [CrossRef]
6. Yang, Z.; Lim, Y.; Tan, Y. An accident model with considering physical processes for indoor environment
safety. Appl. Sci. 2019, 9, 4732. [CrossRef]
7. Cheng, X.; Yang, B.; Tan, K.; Isaksson, E.; Li, L.; Hedman, A.; Olofsson, T.; Li, H. A Contactless measuring
method of skin temperature based on the skin sensitivity index and deep learning. Appl. Sci. 2019, 9, 1375.
[CrossRef]
8. Mainka, A.; Zajusz-Zubek, E. Keeping doors closed as one reason for fatigue in teenagers—A case study.
Appl. Sci. 2019, 9, 3533. [CrossRef]
9. Yen, Y.; Yang, C.; Mena, K.; Cheng, Y.; Chen, P. Cooking/window opening and associated increases of indoor
PM2.5 and NO2 concentrations of children’s houses in kaohsiung, Taiwan. Appl. Sci. 2019, 9, 4306. [CrossRef]
10. Kim, S.; Park, S.; Lee, J. Evaluation of performance of inexpensive laser based PM2.5 sensor monitors for
typical indoor and outdoor hotspots of south Korea. Appl. Sci. 2019, 9, 1947. [CrossRef]
11. Taştan, M.; Gökozan, H. Real-time monitoring of indoor air quality with internet of things-based E-nose.
Appl. Sci. 2019, 9, 3435. [CrossRef]
12. Marques, G.; Pitarma, R. An internet of things-based environmental quality management system to supervise
the indoor laboratory conditions. Appl. Sci. 2019, 9, 438. [CrossRef]
13. Maggos, T.; Binas, V.; Siaperas, V.; Terzopoulos, A.; Panagopoulos, P.; Kiriakidis, G. A promising technological
approach to improve indoor air quality. Appl. Sci. 2019, 9, 4837. [CrossRef]
14. Li, J.; Lu, S.; Wang, Q.; Tian, S.; Jin, Y. Study of passive adjustment performance of tubular space in subway
station building complexes. Appl. Sci. 2019, 9, 834. [CrossRef]
15. Grisoli, P.; Albertoni, M.; Rodolfi, M. Application of airborne microorganism indexes in offices, gyms, and
libraries. Appl. Sci. 2019, 9, 1101. [CrossRef]
16. Marcelloni, A.M.; Chiominto, A.; Di Renzi, S.; Melis, P.; Wirz, A.; Riviello, M.C.; Massari, S.; Sisto, R.;
D’Ovidio, M.; Paba, E. How working tasks influence biocontamination in an animal facility. Appl. Sci. 2019,
9, 2216. [CrossRef]
© 2020 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).
4
applied
sciences
Article
Residents’ Perceptions of and Response Behaviors to
Particulate Matter—A Case Study in Seoul, Korea
Myung Eun Cho and Mi Jeong Kim *
School of Architecture, Hanyang University, 222 Wangsimni-ro, Seongdong-gu, Seoul 04763, Korea
* Correspondence: [email protected]; Tel.: +82-2-2220-1249
Received: 31 May 2019; Accepted: 27 June 2019; Published: 4 September 2019
Abstract: This study is interested in understanding the particulate matter perceptions and response
behaviors of residents. The purpose of this study was to identify indoor air quality along with the
response behaviors of residents in Seoul, to ascertain whether there is a difference in behaviors when
particulate matter is present, according to the characteristics of residents and to grasp the nature
of this difference. A questionnaire survey of 171 respondents was conducted. The questionnaire
measured the indoor air quality perceived by residents, the health symptoms caused by particulate
matter, residents’ response behaviors to particulate matter and the psychological attributes affecting
those response behaviors. Residents of Seoul were divided into college students in their twenties,
male workers in their thirties and forties and female housewives in their thirties and forties. The data
were calibrated by SPSS 23 using a one-way analysis of variance (ANOVA) and multiple regression
analyses. The results show that most people found particulate matter to be an important problem
but were unable to do sufficient mitigation action to prevent its presence. Residents showed greater
psychological stress resulting in difficulty going out than physical symptoms. The most influential
factor on response behaviors was psychological attributes. Participants were aware of the risks of
particulate matter but believed it to be generated by external factors; thus, they felt powerless to do
anything about it, which proved to be an obstacle to response behaviors.
Keywords: particulate matter; perception; response behavior; psychological attribute
1. Introduction
In recent years, particulate matter (PM) has emerged as a big problem in Korea. According to the
Organization for Economic Cooperation and Development (OECD)’s annual report on the concentration
of ultra-particulate matter in countries by 2017, the mean population was exposed to PM2.5 and with
pollution at 25.1 μg/m3 , Korea was the second worst of the member countries [1], with a level twice as
high as the average OECD member countries (12.5 μg/m3 ) and 2.5 times higher than the World Health
Organization (WHO)’s annual average recommended concentration (10 μg/m3 ). Based on Korea’s PM
forecast, the number of “bad” (36–75 μg/m3 ) and “very bad” (more than 76 μg/m3 ) days in metropolitan
areas increased from 62 in 2015 to 77 in 2018 [2]. In early March 2019, Korea experienced the most
severe PM situation. In Seoul, an 8-day ultra-particulate matter warning (PM2.5 with a time-averaged
concentration of more than 75 μg/m3 for 2 h) and 2 days (March 5 and 6) with an alert level (an average
PM2.5 of more than 150 μg/m3 for 2 h) [3]. As a result, the PM levels became hazardous to health.
PM is a WHO Level 1 carcinogen that has negative effects on health, contributing to cardiovascular
and respiratory diseases [4,5]. Choe and Lee [6] investigated the effect of particulate emissions on
specific diseases in Seoul and found that the number of hospitalizations for various respiratory diseases
increased as the amount of ultra-PM increased. Korea’s increase in PM is related to rapid economic
growth. Large cities, such as Seoul, have high levels of energy use resulting from the concentration of
population and economic activity and their direct emission of air pollutants is high. Further, since its

Appl. Sci. 2019, 9, 3660
geographical location is on the mid-latitude westerly wind area, seasonal influx of PM from neighboring
China also affects the increase of PM in Korea [7].
According to the survey data on the perception of environmental problems among Koreans aged 13
and over, conducted by the Korea National Statistical Office (KNSO), 82.5% of respondents experience
anxiety about PM [8]. Kim et al. [9] observed that Koreans regard PM as the most serious social risk factor.
Because national concern about PM has been increasing, the government introduced a comprehensive
plan for PM in 2017. In 2018, it attempted to reduce PM emissions by enforcing a Special Act on
Particulate Matter in major cities across the country, including the capital region [10]. The Ministry of
Environment, in consideration of atmospheric environmental standards and health effects, produced
a PM forecasting system, which presents the levels of PM as well as countermeasures [11]. However,
despite the various risk indicators for PM and notwithstanding the government measures, the residents
of Seoul are notably passive in protecting individuals and society from PM despite viewing it as
a threat [12]. To prevent and reduce the PM generated by anthropogenic rather than natural factors,
public efforts must be accompanied by measures at the national level. Without ensuring that residents
understand PM, it is predicted that reduction measures will be ineffective. Therefore, this study aims
to identify levels of awareness of PM, recognition of indoor air quality, symptoms of PM exposure
experienced by residents and coping behavior in relation to PM in Seoul. Specific research questions
are as follows:
First, how do residents perceive the indoor air quality, how do they feel PM symptoms and how
do they behave in response to PM?
Second, do the different characteristics of residents produce any variations in behavior in response
to PM?
Third, what are the impediments to the proper responses to PM by residents?
In this study, we have sought to understand the perceptions of PM and the subsequent response
behaviors of residents, as well as to identify the causes of these behaviors. It is critical to understand
and solve the barriers to public engagement to avoid the worst consequences of PM. The results of this
study are expected to be used as basic data for effective governmental measures to reduce PM.
2. Status and Risks of Particulate Matter in Korea
2.1. Characteristics of Particulate Matter Generation in Seoul

To identify the characteristics of PM generation in Seoul, we investigated the annual average
concentration of PM along with its highest levels of concentration. As shown in Figure 1, the average
concentrations of PM10 and PM2.5 in Seoul were calculated using Seoul’s atmospheric environment
information from the past 10 years. From 2009 to 2018, the average annual concentration of PM10 in
Seoul was 53.8 μg/m3 and the average annual concentration of PM2.5 was 27.2 μg/m3 [3]. The average
annual concentration of PM10 decreased from 76 μg/m3 in 2002 to 41 μg/m3 in 2012 but this decline has
since slowed. Since the government enacted the Special Act on the Improvement of the Air Quality in
the Seoul Metropolitan Area and established and performed its Basic Plan for the Management of Air
Quality in the Seoul Metropolitan Area in 2003, air pollution as well as PM concentrations have been
reduced [7]. However, no further improvement has occurred since 2012.
6
Appl. Sci. 2019, 9, 3660
Figure 1. Annual mean particulate matter (PM) concentration in Seoul (2009–2018).
The results of a survey of “bad” and “very bad” days exceeding a PM2.5 of 35 μg/m3 for an average
of 24 h showed that 44 days in 2015, 73 days in 2016, 64 days in 2017 and 61 days annually on average
were recorded as “bad” and “very bad”. This means that the number of days when the concentration
of PM was significant is very large. This can be seen by comparing the number of PM warnings and
days of alarm in Seoul. Figure 2 shows the number of PM and ultra-PM warning days using Seoul’s
atmospheric environment information. If there is a PM10 of 150 μg/m3 or a PM2.5 of 75 μg/m3 for more
than 2 h, a warning is issued. If there is a PM10 of 300 μg/m3 or a PM2.5 of 150 μg/m3 for more than 2 h,
an alarm is issued. In 2018, there were 17 PM warning days, 1 PM alarm day and 7 ultra-PM warning
days. Seoul residents were exposed to extremely high dust concentrations on these days.
(a) (b)
Figure 2. High PM concentration days: (a) days exceeding a PM2.5 of 35 μg/m3 on average for 24 h;
(b) PM and ultra-PM warning and alarm days (Source: Seoul Atmospheric Environment Information).
This result is closely related to the monsoon season experienced in the geographical location
of Korea. Korea’s winter atmospheric circulation is affected by the northwestern winds associated
with the winter monsoon in East Asia and the location of the barometer over the Korean peninsula.
Inflow from China and Mongolia in winter greatly affects the concentration of PM in Korea [13].
Figure 3 shows the changes in the concentrations of PM and ultra-PM from March 2018 to February
2019 [3]. It shows that the levels of PM are high in winter and spring and low in summer. Specifically,
they are lowest in September and highest in January. The atmospheric environmental standards for
domestic PM are below 50 μg/m3 on average for 24 h and below 25 μg/m3 on average per year. Korea’s
levels exceed these standard values in all seasons except summer.
7
Appl. Sci. 2019, 9, 3660
Figure 3. Seasonal dust concentrations in Seoul.
The air quality in Korea deteriorates considerably as the number of days with seasonally high
concentrations of PM increase [14].
2.2. Domestic Particulate Matter Standards and Health Protection

The Ministry of Environment [2] proposed enhanced environmental standards for PM in
2018, taking into consideration national health effects, international standards, pollution status
and achievability. Table 1 shows the WHO recommendations along with the standards in other regions.
The concentration of PM in the domestic environment of higher than 50 μg/m3 far exceeds the WHO’s
recommended level of an annual average PM10 of 20 μg/m3 and is much greater than that of other
regions. In addition, the rate of achieving the annual PM environmental standard was around 60% in
2015 and the rate of achieving the 24 h environmental standard for PM2.5 and PM10 was very low at 4%
and 10.7% respectively.
Table 1. Comparison of air quality standards for PM (Source: Ministry of Environment).
PM Standards Standard Achievement

Category Standard Time
Korea WHO USA EU Rate (2015)
yearly 15 10 12, 15 25 65.0%

PM2.5 (μg/m3 )
24 h 35 25 35 - 4.0%
yearly 50 20 - 40 65.6%
PM10 (μg/m3 )
24 h 100 50 150 50 10.7%
Considering public health, the Ministry of Environment has been conducting PM forecasting
from February 2014 and ultra-PM forecasting with a warning system from January 2015 [2]. The PM
forecasting system provides forecasts of the PM concentration four times a day for the present day, the
following day and the day after that. The PM forecast is graded in four stages: “good”, “normal”,
“bad” and “very bad” (Table 2). The PM warning system works to promptly inform the public when
a high concentration of PM occurs and to reduce the damage. It is issued when the air quality is
harmful to health. Air pollution alarms are classified into two stages—warning and alarm—as shown
in Table 3. At the time of a PM alarm, there are eight countermeasures for people to take: remaining
indoors; wearing a health mask; reducing external activities; washing the body after returning home;
drinking water and eating fruits and vegetables; undertaking ventilation and indoor water cleaning;
managing indoor air quality; and restricting air pollution inducing activities.
8
Appl. Sci. 2019, 9, 3660
Table 2. Particulate matter forecast grade (2018).
PM Concentration (μg/m3 , Average for 24 h) Good Normal Bad Very Bad

PM10 0–30 31–80 81–150 over 151
PM2.5 0–15 16–35 36–75 over 76
Table 3. Particulate matter warning and alarm issuing grade (2018).
Category Warning Issuing Alarm Issuing

PM10 hourly average concentration over PM10 hourly average concentration over
PM10
150 μg/m3 for 2 h 300 μg/m3 for 2 h
PM2.5 hourly average concentration over PM2.5 hourly average concentration over
PM2.5
75 μg/m3 for 2 h 150 μg/m3 for 2 h
3. Materials and Methods
3.1. Participants and Questionnaire Design

This study surveyed the residents of Seoul. To identify differences in the response to PM associated
with gender and age, respondents were divided into three groups. A total of 171 respondents were
used for the analysis. The groups were 20-year-old college students (N = 70, gender = 32 male and
38 female, mean age = 21.88, SD = 2.33), 30- to 40-year-old male workers (N = 51, mean age = 41.11,
SD = 5.45) and 30- to 40-year-old housewives (N = 50, mean age = 37.06, SD = 4.33). The questionnaire
comprised five main parts. The first part included questions about the quality of indoor air perceived
by residents in the home. The second part addressed the health of the residents in terms of objective
symptoms, subjective symptoms and health behaviors concerning the symptoms. The third section
elicited the residents’ responses to PM, dividing them into mitigating behavior, adaptive behavior and
behavior intentions. The fourth part sought to ascertain the psychological causes of interference with
the response behaviors of the residents to PM and the final content consisted of questions investigating
the residents’ overall knowledge of PM.
3.2. Measures
3.2.1. Measuring Response Behaviors to Particulate Matter

Mitigating behavior and adaptive behavior are countermeasures to deal with risk. Swart and
Raes [15] define “mitigation” as an anthropogenic intervention to reduce the sources of air pollution,
whereas “adaptation” is an adjustment in natural or human systems in response to climatic stimuli.
Mitigation is a way to reduce the cause of hazards associated with PM. Usually, the benefits of
mitigating behavior are not seen in the short term, so it is considered a long-term countermeasure [16].
Mitigating behavior is a personal effort to reduce the generation of PM, which includes “using public
transportation to reduce atmospheric gas generation”, “not using electricity and heating to restrain
unnecessary energy use”, “using kitchen utensils that generate less harmful gas” and other similar
measures. Adaptation refers to controlling the damage caused by PM.
Adaptive behavior can reduce the risks associated with PM by prophylactic and post-exposure
measures that minimize the negative effects of PM [17]. In contrast to mitigating behavior, from
which long-term effects arise, the effects of adaptive behavior are immediate and are characterized
by the matching of subject and beneficiary [18]. Adaptive behavior is action to prevent the damage
caused to individuals by high concentrations of dust. Behavioral intentions signify an individual’s
specific willingness to act in response to PM and include two positive intentions. To achieve sustainable
development policies, it is necessary to integrate mitigating actions that directly reduce the concentration
of PM and adaptive prevention actions that reduce the impact of the existing dust risk. In this study,
mitigating behavior, adaptive behavior and the behavioral intentions of Seoul residents are explored,
as shown in Table 4.
9
Appl. Sci. 2019, 9, 3660
Table 4. Responsive behaviors to particulate matter.
Question: What do you do when Particulate Matter Occurs?

MB1 Using public transportation to reduce atmospheric gas generation
Mitigating
MB2 Refraining from using electricity and heating for the reduction of unnecessary energy use
behavior
MB3 Using kitchen appliances that generate less harmful gas
AB1 Checking the levels of particulate matter concentration every day
AB2 Wearing a particulate matter mask on high particulate matter density days
Adaptive AB3 Refraining from going out when the concentration of particulate matter is high
behavior AB4 Turning on an air purifiers to reduce particulate matter concentration
AB5 Washing the whole body thoroughly in running water after returning home
AB6 Wiping dust from the floor with a damp cloth and undertaking indoor water cleaning
BI1 Being ready to suffer immediate damage or inconvenience to reduce particulate matter
concentrations.
Behavioral BI2 Being willing to participate in actions to reduce particulate matter
intentions BI3 Not feeling any urgency to change my behavior to prevent particulate matter pollution
BI4 Thinking that changes in my behavior do not affect particulate matter reduction
BI5 Not knowing what concrete action may be taken to reduce particulate matter
3.2.2. Measuring Psychological Attributes that Hinder Reactions to Particulate Matter

Gilfford [19] observed that most people find environmental sustainability to be an important
issue but that psychological barriers prevent them from engaging in sufficient actions to address it.
These psychological barriers impede the behavioral choices that would facilitate mitigation, adaptation
and environmental sustainability [20]. Risk perception is a factor that affects human attitudes and
behavioral intentions and that is very important in making individual decisions about a behavior [21].
Brewer et al. [22] explained three dimensions of risk perception. The first is the likelihood that one
will be harmed by the hazard; the second is susceptibility, referring to an individual’s constitutional
vulnerability to a hazard; and the third is severity, indicating the extent of harm a hazard may cause.
Psychological distance describes how individuals participate in future events [23]. The perceived
distance of events indicates how they are mentally construed. As the perceived distance increases, events
are interpreted as more abstract, decontextualized and conceived in generalized terms. When events
become closer, they use more specific, contextualized and detailed features [24]. The psychological
distance for PM was evaluated using the four distance domains of geography, temporality, socialization
and awareness. In addition, this research investigated various barriers increasing the perceived concern
of the public, such as information distrust, externalizing responsibility and uncertainty about the causes
of PM pollution. In this study, to understand the effects of response behaviors to PM, the psychological
attributes of the residents were classified into nine concepts within three domains: “risk perception”,
“psychological distance”, and “perceived concerns” (Table 5).
Table 5. Psychological attributes.
Question: What Are your Thoughts in Response to Particulate Matter?

1. Contact with particulate matter is a health hazard
Likelihood
Risk perception
2. If I continue to be exposed to particulate matter, I will be damaged in a few years

3. Particulate matter is an important problem for me
Susceptibility
4. I am more affected by the risk of particulate matter than other risks
5. The risks of particulate matter for health are very serious
Severity
6. Even short-term contact with particulate matter can increase the likelihood of
cancer and early death
10
Appl. Sci. 2019, 9, 3660
Table 5. Cont.
Question: What Are your Thoughts in Response to Particulate Matter?

Geographic and 7. The particulate matter condition is more serious in Korea than in other countries
distance
Psychological
Temporal 8. The risk of particulate matter is present very often or constantly

9. Particulate matter will have a significant impact on me and my family
Social
10. Particulate matter is sure to be a serious social problem
11. I am very concerned about particulate matter
Awareness
12. I am interested in issues related to particulate matter
Information 13. I cannot trust the information about particulate matter
distrust 14. The risk of particulate matter is greatly exaggerated
concerns
Perceived
Externalizing 15. The problem of particulate matter is beyond my ability to solve

responsibility 16. The solution to the excess of particulate matter should be provided by the
government rather than individuals
17. The damage caused by particulate matter is unclear
Uncertainty 18. It is difficult to measure health damage arising from particulate matter
19. Particulate matter does not affect me right now
3.3. Methodology
This research used a questionnaire to collect its data; the survey was conducted online in March 2019
and the data were analyzed using IBM’s SPSS Statistics Program 23. The question concerning residents’
perceptions of indoor air quality comprised positive and negative language (the semantic differential
method). We distinguished between positive vocabulary and negative vocabulary over five levels,
asking residents to assign one level to each question according to their perceptions. The vocabulary
used to measure the indoor air quality of residents included “bad”, “good”, “stuffy”, “refreshed”,
“unpleasant”, and “comfortable”. A five-point Likert scale was used to assess the symptoms, behaviors
and psychological responses of residents, ranging from “very unlikely” (1 point), to “unlikely” (2 points),
to “average” (3 points), to likely (4 points), to “very likely” (5 points). Questions measuring residents’
knowledge of PM could be answered as “right” () or “wrong” (X) and the “right” answers were
added to produce a score.
An ANOVA was used for the statistical analysis to determine whether there were differences in
behavior as a response to PM, according to the characteristics of residents. Participants were divided
into three groups based on their genders and ages: college students in their twenties, male workers
in their thirties and forties and female housewives in their thirties and forties. In addition to the
response behavior, we compared differences between the groups in terms of environmental exposures
to PM, perceived symptoms and psychological attributes. Further, post-hoc tests were conducted to
analyze variations between groups. Multiple regression analysis was conducted to identify the factors
influencing response behavior.
4. Results
4.1. Perceived Indoor Air Quality and Overall Satisfaction

The perceived indoor air quality and the satisfaction of residents in Seoul were measured using
a semantic differential approach. Residents assessed their perceptions of indoor air quality by comparing
two opposing pairs of vocabulary in five steps. For example, the evaluation score for the first question
on indoor air quality was divided into five levels, spanning two opposing experiences: “a lot of
particulate matter” and “no particulate matter”. The more PM respondents perceived, the closer to
1 point they scored and the less PM they observed, the closer they scored to 5 points.
A total of 171 residents were analyzed and the results are shown in Table 6. Respondents’ answers
to the four questions about the quality of indoor air produced an average score of approximately three
(2.93–3.25) points. This score comprises the median value of 1 and 5, indicating that the residents of
11
Appl. Sci. 2019, 9, 3660
Seoul understand the indoor air quality to be normal. The evaluations of overall indoor air quality also
averaged 3.21 points. However, the average score for the question, “Do you think seriously about or
are you interested in the quality of indoor air where you live?” was 4.17 points (where 1 point indicated
“not at all” and 5 points denoted “very much”), implying that respondents were interested in the air
quality of their living spaces.
Table 6. Perceptions of and satisfaction with air quality.
Perceptions of
Mean Std. Deviation
1 Point 5 Points
very bad very good 3.26 0.91
Indoor air a lot of particulate matter little particulate matter 2.94 0.98
quality stuffy refreshed 2.93 1.02
unpleasant comfortable 3.25 0.92
dissatisfaction satisfaction 3.21 1.10
Satisfaction
no thought or interest much thought and interest 4.17 1.10
Two phrases corresponded to 1 point and 5 points respectively and between these two phrases were expressions
valued at 2 points (signifying the perception of slightly more PM), 4 points (signifying the perception of slightly
less PM) and 3 points (signifying the perception of an average amount of PM).
4.2. Particulate Matter Environmental Exposures and Perceived Symptoms

When measuring the outdoor activity times of respondents, it was found that they stayed outdoors
for an average of 4.54 h (SD = 4.31) per day and the analysis of variance showed a statistically significant
difference (p < 0.01) depending on the subject group. The results are shown in Table 7. The group of
20 college students and the group of 30- to 40-year-old male workers remained outside for an average of
5 h a day—specifically 5.12 h and 5.07 h respectively—while the group of 30- to 40-year-old housewives
stayed outdoors for 3 h a day on average.
Table 7. Average daily external activity time.
Group 1 Group 2 Group 3

F (p-Value)
Mean Std Deviation Mean Std Deviation Mean Std Deviation
5.12 h 3.43 5.07 h 5.49 3.20 h 3.81 3.54 (0.031)
Group 1: 20-year-old college students; Group 2: 30- to 40-year-old male workers; Group 3: 30- to 40-year-old
housewives.
The physical symptoms and health problems experienced by residents as a result of PM were
verified through objectively judged symptoms, such as cough, allergy and headache, as well as
subjectively judged psychological symptoms, such as anxiety and stress. In addition, the extent of
any actual treatment at the hospital because of these symptoms was investigated. The respondents’
answers ranged from “not at all” (1 point) to “very strongly” (5 points) and the results are shown in
Table 8.
The average score for the response, “Thinking that particulate matter has a negative effect on my
health” was 4.39, the highest and the average score for the response “Thinking that particulate matter
causes problems in my life” was 4.00, the second highest. Symptoms such as respiratory problems and
complications with nose, skin and eyes rated 3.69 and 3.66 respectively, revealing that the average
scores for the objective symptoms of residents with health problems arising from PM were normal.
These findings indicate that the psychological symptoms (Mean = 4.39) were more significant than the
objective symptoms (Mean = 3.75); further, the incidence of direct treatment in hospitals (M = 3.15)
for such symptoms was not significant (Figure 4). However, ANOVA analysis showed statistically
significant differences according to subject groups. A Tukey post-hoc test was conducted to analyze
the differences between the groups and the results showed that the perceived symptoms were greater
12
Appl. Sci. 2019, 9, 3660
among the group of housewives in their thirties and forties than in the group of college students in
their twenties or in the group of male workers in their thirties and forties. Notably, the differences in
psychological symptoms were pronounced among these groups (Table 8).
Table 8. Health-related symptoms experienced by residents as a consequence of particulate matter.
Mean
Perceived F (p-Value)
Group 1 Group 2 Group 3 Total
Suffering from cough, hoarseness,
bronchitis, respiratory distress and 3.54 3.50 4.08 3.69 3.33 (0.038 *)
Objective symptom
difficulty breathing
Experiencing clogged nose, rhinitis,
sneezing, allergy and atopic symptoms, 3.51 3.43 4.10 3.66 3.76 (0.025 *)
such as itchy or swollen skin or dry eyes
Suffering from headache, arthritis and
2.61 2.92 3.38 2.92 4.38 (0.014 *)
other pain
Experiencing the deterioration of
2.42 2.88 3.38 2.84 7.57 (0.001 **)
an already present illness
Suffering increased stress and fatigue 3.34 3.50 4.14 3.62 7.23 (0.001 **)
Subjective symptom
Feeling worried and anxious 3.18 3.60 4.30 3.63 12.84 (0.000 ***)
Thinking that particulate matter has
4.21 4.27 4.78 4.39 8.55 (0.000 ***)
a negative effect on my health
Thinking that particulate matter causes
3.71 3.68 4.72 4.00 17.55 (0.000 ***)
problems in my life
Thinking that particulate matter reduces
3.20 3.31 4.02 3.47 6.80 (0.001 **)
my performance and concentration
I have recently been to the hospital for
1.94 2.15 3.00 2.31 8.96 (0.000 ***)
any of the above symptoms
Remedy
I have recently been taking medication

2.12 2.31 3.04 2.45 6.19 (0.003 **)
for some of the above symptoms
I have recently taken a break at home
2.51 2.39 3.42 2.74 7.80 (0.001 **)
because of some of the above symptoms
* p < 0.05, ** p < 0.01, *** p < 0.001. Group 1: 20-year-old college students; Group 2: 30- to 40-year-old male workers;
Group 3: 30- to 40-year-old housewives.
Figure 4. Physical symptoms of particulate matter experienced by residents.
13
Appl. Sci. 2019, 9, 3660
4.3. Behavioral Changes in Response to Particulate Matter

Residents’ response behaviors to the occurrence of PM were analyzed and the results indicate
that adaptive behaviors (M = 3.66, SD = 0.95) controlling the damage caused by PM were more
prevalent than mitigating behaviors (M = 2.98, SD = 1.03) to reduce the fundamental causes of PM,
as shown in Table 9. The most common PM adaptive response behaviors were “AB1: Checking
the PM concentration every day” (4.00 points) and “AB5: Washing your whole body thoroughly
in running water after returning home” (3.78 points). Regarding behavioral intentions, the most
common PM response behaviors were “BI2: Being willing to participate in actions to reduce particulate
matter” (4.00 points) and “BI1: Being ready to experience immediate damage or inconvenience to
reduce the particulate matter concentration” (3.69 points). For other responses, scores ranging from
2.50 to 3.50 were obtained, indicating that residents’ responses to PM were not significant. However,
there were statistically significant differences among the groups (Table 9). In particular, the group of
30- to 40-year-old housewives showed higher scores for adaptive behavior and behavioral intentions
than the other groups (Figure 5).
Table 9. Residents’ response behaviors to particulate matter.
Mean
Behavior Type F (p-Value)
MB1 4.04 2.82 2.60 3.25 19.63 (0.000 ***)
Mitigating behavior MB2 2.80 3.00 3.00 2.91 0.49 (0.613)
MB3 2.55 2.94 3.12 2.83 2.91 (0.057)
AB1 3.55 3.80 4.84 4.00 19.63 (0.000 ***)
AB2 2.77 3.35 4.28 3.38 20.36 (0.000 ***)
AB3 2.82 3.41 4.32 3.43 19.53 (0.000 ***)
Adaptive behavior
AB4 2.77 3.92 4.32 3.56 19.03 (0.000 ***)
AB5 3.48 3.76 4.24 3.78 5.55 (0.005 **)
AB6 2.91 3.35 3.98 3.35 13.37 (0.000 ***)
BI1 3.48 3.45 4.24 3.69 12.11 (0.000 ***)
BI2 3.91 3.86 4.28 4.00 3.72 (0.026 *)
Behavioral intentions BI3 1 −2.72 −2.82 −1.82 −2.49 12.44 (0.000 ***)
BI4 1 −2.54 −2.82 −2.12 −2.50 4.88 (0.009 **)
BI5 1 −3.05 −2.94 −2.86 −2.96 0.42 (0.657)
* p < 0.05, ** p < 0.01, *** p < 0.001, 1 Reverse scored for negative perceptions. Group 1: 20-year-old college students;
Group 2: 30- to 40-year-old male workers; Group 3: 30- to 40-year-old housewives.
Figure 5. Residents’ response behaviors to particulate matter.
14
Appl. Sci. 2019, 9, 3660
4.4. Barriers to Mild Behavioral Changes as a Response to Particulate Matter

We investigated the residents’ understandings of PM and their psychological causes as obstacles
to behavioral changes to reduce and prevent PM.
4.4.1. Lack of Knowledge

The items indicating the residents’ knowledge of PM consisted of 10 authentic scales (true/false)
based on data from experts and the Ministry of Environment. The respondents selected “correct” or
“incorrect” in answer to each item and answers of “correct” were treated as single points, producing
a total score between 0 and 10 points. Table 10 shows the results for the 171 respondents. The average
knowledge of PM was 6.56 (SD = 1.44) and the results of the analysis are shown in Figure 6. Knowledge
of PM did not vary among the groups.
Table 10. Residents’ knowledge of particulate matter.
Number of Correct
Knowledge
Answers (%)
1. The influx of particulate matter is mainly caused by artificial actors, such as boilers,
138 (80.7)
automobiles and power generation facilities
2. More than 80% of particulate matter is from abroad, as a result of the yellow dust
39 (22.8)
and smog from China
3. Preliminary reduction measures against emissions of particulate matter are
43 (25.1)
implemented nationwide and are targeted based on particulate matter concentration
4. Ultra-particulate matter comprises very thin, small particles but causes deterioration
of visibility in places where the flow of air is stagnant, creating obstacles to traffic and 118 (69.0)
navigation
5. When there is a particulate matter alarm, schools can prohibit outdoor classes, adjust
151 (88.3)
the times of travel to and from school and temporarily shut down
6. Domestic particulate matter concentrations are similar to those in major cities of
141 (82.5)
other OECD countries, such as New York and London
7. Even when exposure occurs for only a short time, particulate matter can penetrate
119 (69.6)
directly into the alveoli, resulting in asthma, lung disease and even death
8. The mask is a quasi-drug certified by the Food and Drug Administration that can
133 (77.8)
only prevent exposure to particulate matter if the product is marked “KF94” or “KF80”
9. Generally, high concentrations of particulate matter occur in spring and summer 89 (52.0)
10. It is helpful to eat fruits and vegetables that are rich in water and vitamin C to
152 (88.9)
combat the effects of high dust concentrations
Figure 6. Residents’ total knowledge of particulate matter.
15
Appl. Sci. 2019, 9, 3660
4.4.2. Psychological Attributes of Particulate Matter

The results of the survey of the psychological influences on behavioral responses to PM show
that risk perception (Mean = 4.01, SD = 0.76) and psychological distance (Mean = 4.32, SD = 0.73)
achieved a higher than average score of 4.0 points (where 1 point = “not at all” and 5 points =
“very much”). Table 11 shows the results of the analysis. Question 2 related to likelihood when
considering residents’ risk perception—“If you continue to be exposed to particulate matter, you will
be affected in a few years” (Item 4.5)—as did Question 1, “Contact with particulate matter is harmful
to health” (4.35 points)’; both of these questions received the highest scores. The score for the residents’
perception of the psychological distance with which they regarded PM was higher than that for their
risk perception. All questions concerning psychological distance were rated at greater than 4.0 and
among them, the scores for Question 10, relating to the social implications of PM—”particulate matter
is sure to be a serious social problem” (4.57 points)—and for Question 8, pertaining to the geographic
and temporal significance of PM—“the risk of particulate matter is present very often or consistently”
(4.47 points)—were the highest. Questions about perceived concerns, such as information distrust,
the externalizing of responsibility and uncertainty, were reverse scored. Among them, Question 16,
relating to the externalizing of responsibility—“The solution to the excess of particulate matter should be
provided by the government rather than individuals” (−4.35 points)—and Question 15—“The problem
of particulate matter is beyond my ability to solve” (−4.02 points)—received the highest scores.
There was a statistically significant difference between the groups (Table 11). The group of housewives
aged in their thirties and forties showed higher scores for risk perception and psychological distance
than the other groups (Figure 7).
Table 11. Psychological influences on residents’ response behaviors.
Mean
Psychological Attributes F (p-Value)
1 4.08 4.33 4.74 4.35 8.58 (0.000 ***)
Likelihood
perception
Risk
2 4.44 4.29 4.82 4.50 6.22 (0.002 **)

3 4.10 4.19 4.62 4.28 5.11 (0.007 **)
Susceptibility
4 2.84 3.07 3.68 3.15 6.70 (0.002 **)
5 4.10 4.15 4.58 4.27 4.69 (0.010 *)
Severity
6 3.34 3.41 3.86 3.51 3.10 (0.048 *)
Geographic 7 4.18 4.19 4.60 4.30 2.99 (0.053)
distance
Psychological
and temporal 8 4.41 4.25 4.80 4.47 8.13 (0.000 ***)

9 3.92 4.19 4.70 4.23 11.01 (0.000 ***)
Social
10 4.44 4.43 4.92 4.57 10.55 (0.000 ***)
11 4.02 4.13 4.76 4.27 9.96 (0.000 ***)
Awareness
12 3.70 4.00 4.60 4.05 11.86 (0.000 ***)
Information 13 −2.97 −3.62 −3.80 −3.40 10.84 (0.000 ***)
distrust 14 −2.40 −2.43 −1.52 −2.15 13.65 (0.000 ***)
concerns 1
Perceived
Externalizing 15 −4.04 −3.96 −4.08 −4.02 0.30 (0.741)

responsibility 16 −3.98 −4.52 −4.70 −4.35 11.91 (0.000 ***)
17 −2.68 −2.88 −2.78 −2.77 0.36 (0.692)
Uncertainty 18 −2.97 −3.01 −2.88 −2.95 0.15 (0.857)
19 −2.40 −2.50 −1.88 −2.28 4.08 (0.019 *)
* p < 0.05, ** p < 0.01, *** p < 0.001, 1 Reverse scored for negative perceptions. Group 1: 20-year-old college students;
Group 2: 30- to 40-year-old male workers; Group 3: 30- to 40-year-old housewives.
16
Appl. Sci. 2019, 9, 3660
Figure 7. Residents’ psychological influences.
4.5. Factors Affecting Response Behaviors to Paticulate Matter

A multiple regression analysis was conducted to determine the factors with the greatest effect on
residents’ response behaviors. Knowledge of PM, psychological attributes and PM-related symptoms
were selected as factors influencing response behaviors and the analytical result was statistically
significant (p < 0.05; F = 25.306, p-value = 0.000). Table 12 shows the evaluation of the contribution and
statistical significance of the individual independent variables. Psychological attributes were the most
influential factor on residents’ responses to PM, followed by PM-related symptoms.
Table 12. Multiple regression analysis of the response behaviors to particulate matter.
Variable B Std Error Beta t Value p-Value

Constant 8.103 3.577 2.265 0.025
Knowledge of PM 0.630 0.443 0.095 1.424 0.156
Psychological attributes 0.373 0.082 0.375 4.555 0.000
PM-related symptoms 0.191 0.070 0.219 2.737 0.007
Note: R = 0.55, R2 = 0.31, Adj. R2 = 0.30.
5. Discussion and Conclusions

In Seoul, it is inadvisable to go out on any day, except in summer, as a result of the PM concentration
and its effect on human health. When the WHO standard is applied, the concentration of PM in Seoul
is revealed to be remarkably high and cases of excessive concentration exceeding the daily average
atmospheric environment standard are also frequent. The government has been pursuing various
policies to reduce PM concentration but the concentration of PM has not improved to the extent that
people recognize any change. Research has shown that most people believe PM to be an important
problem but do not undertake enough mitigating behaviors to prevent its occurrence. Reliable research
should be preceded by an analysis of cause and effect before the preparation of PM countermeasures.
We sought to understand residents’ perceptions of PM, their response behaviors and the psychological
causes of these behaviors; the main findings of this study are as follows.
First, to identify the perceived indoor air quality of the residents, the questionnaire was distributed
at a time when the concentration of PM had been significant throughout the year; thus, residents’
interest in indoor air quality was great. However, despite high outdoor concentrations of PM, the indoor
air quality was not considered poor or dissatisfactory. The symptoms that residents attributed to PM
were not physical, such as allergies or bronchitis. They more commonly experienced psychological
stress or anxiety, feeling that PM had a negative effect on health and was interfering more frequently
with daily life, in that residents were going out less often to avoid it.
17
Appl. Sci. 2019, 9, 3660
Second, the problem with the PM-related behaviors of residents in Seoul was that many people
did not take actual action despite their willingness to do so. Mitigating behavior to try to minimize
PM generation was less common than adaptive behavior to protect individuals from the risks of
PM. This may have been because of the nature of mitigating behavior, which often entails loss or
inconvenience. Even if the risks associated with PM were largely recognized, it would be necessary
to compensate for the loss when the habitual behavior of individuals was abandoned in favor of
following the recommended behavior. Unlike preventive actions to protect individuals from PM,
compensation for mitigating behavior is only likely to be achieved in the distant future and even
then, such compensation would provide a social rather than an individual benefit. This suggests
that more persuasive strategies should be employed to stimulate changes in individual behaviors for
collective interests.
Third, the psychological characteristics of individuals can either hinder or promote action for
environmental sustainability. The greater the perceived risk and the more psychological that risk,
the more likely it is that an active response will occur. Analysis of the psychological factors affecting
residents in Seoul revealed that they perceived the risk of PM to be very close (in terms of psychological
distance). However, despite revealing significant levels of risk perception and psychological distance,
the results of this study did not show active response behaviors. Attention must be paid to the
externalizing of responsibility among the psychological causes of this lack of action. This study found
distrust of and uncertainty about PM information to be largely unrecognized as causes of apathy;
on the contrary, understanding PM saturation to be externally produced was the main reason for
inaction. If external factors are emphasized as the cause of the problem, it is possible that individuals
perceive the risk as uncontrollable, which may nullify their will to respond. The results of this study
also suggest that residents’ psychological barriers have a negative effect on their response behaviors
for the reduction of PM.
Fourth, the results of the correlation analysis show that the most influential factor in PM-related
behavior was psychological attributes but that physical symptoms also affect response behaviors to
PM. This result is meaningful in that it indicates desirable directions for PM policy. Based on the
analysis of psychological factors, it was found that, because PM is also produced by individuals, active
prevention is likely to occur if the importance of personal responses is conveyed. While knowledge of
PM was not found to significantly affect response behaviors, residents in Seoul showed an average
score of 6.56 out of 10 for PM knowledge. Lack of knowledge and understanding of the PM problem
hinders appropriate response behavior. It is necessary to provide accurate and correct PM education
and continuously available reliable information to enhance the knowledge of PM. In addition,
There was significant variation in residents’ perceptions, response behaviors and psychological
attributes concerning PM based on gender and age. Conversely, the time during which residents
undertook external activities exposed to PM did not significantly affect PM-related symptoms or
behavior. In contrast to college students in their twenties and male workers in their thirties and forties
who remained outdoors for 5 h on average, women in their thirties and forties who stayed outdoors for
only 3 h on average responded more sensitively to PM-related symptoms and their commitment to PM
reduction and mitigating actions was much greater. Similarly, the results of the psychological factor
analysis for PM showed that the group of women in their thirties and forties were more aware of the
risks than the other groups and felt those risks to be closer psychologically. However, it is the younger
generations who will become the subjects of society in future and will need to face these problems and
solve them. There is an urgent need for plans to increase awareness of the risks and the will of the
public to act.
Government efforts to reduce PM have focused primarily on expensive communication campaigns.
However, the results of this study suggest that the mere encouragement of attitudinal change is not
effective; public engagement in terms of accurate PM knowledge, health education and the study of
various personal characteristics and psychological causes is required.
18
Appl. Sci. 2019, 9, 3660
This study identified residents’ thoughts, response behaviors and psychological factors relating
to PM and showed that personal characteristics, cognition and emotions affect behavioral intentions.
Unlike previous studies, the significance of this research is that it seeks to understand residents’
behaviors in a multidimensional way. Psychological factors are expected to be used effectively to
motivate individuals to protect themselves from the risks of PM exposure. The residents of Seoul
involved in this study were well aware of the risks of PM but felt that its generation and any possible
solutions were external and therefore beyond their control. This situation is likely to prevent public
action to protect individuals and the society. It is recommended that we shift our focus from risk and
attention to participation and action. It is expected that a proper direction for efficient PM reduction
can be established using these results.
Author Contributions: Conceptualization, M.E.C. and M.J.K.; methodology, M.E.C. and M.J.K.; formal analysis,
M.E.C.; investigation, M.E.C.; data curation, M.J.K.; writing—original draft preparation, M.E.C.; writing—review
and editing, M.J.K.; supervision, M.J.K.; funding acquisition, M.J.K.
Funding: This research was supported by a grant (19AUDP-B127891-03) from the Architecture & Urban
Development Research Program funded by the Ministry of Land, Infrastructure and Transport of the
Korean government.
References
1. OECD. OECD Statistics. 2019. Available online: https://stats.oecd.org/index.aspx?queryid=72722 (accessed
on 24 March 2019).
2. Ministry of Environment. Particulate Matter Status. 2019. Available online: http://www.me.go.kr/cleanair/
sub02.do#2 (accessed on 12 March 2019).
3. Seoul Metropolitan Government. Seoul Atmospheric Environment Information. 2019. Available online:
http://cleanair.seoul.go.kr/air_pollution.htm?method=average (accessed on 29 April 2019).
4. Sacks, J.D.; Stanek, L.W.; Luben, T.J.; Johns, D.O.; Buckley, B.J.; Brown, J.S.; Ross, M. Particulate
matter—Induced health effects: Who is susceptible? Environ. Health Perspect. 2011, 119, 446–454. [CrossRef]
[PubMed]
5. Pope, C.A., III; Burnett, R.T.; Thun, M.J.; Calle, E.E.; Krewski, D.; Ito, K.; Thurston, G.D. Lung cancer,
cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. JAMA 2002, 287, 1132–1141.
[CrossRef] [PubMed]
6. Choe, J.I.; Lee, Y.S. A study on the impact of PM2.5 emissions on respiratory diseases. Korea Environ. Policy
Adm. Soc. 2015, 23, 155–172. [CrossRef]
7. Lee, H.J.; Jeong, Y.M.; Kim, S.T.; Lee, W.S. Atmospheric circulation patterns associated with particulate matter
over south Korea and their future projection. J. Clim. Chang. Res. 2018, 9, 423–433. [CrossRef]
8. Korean Statistical Information Service. Recognition of Environmental Problems (PM Inflow). 2019. Available
online: http://kosis.kr/statisticsList/statisticsListIndex.do? (accessed on 1 April 2019).
9. Kim, Y.; Lee, H.; Kim, H.; Moon, H. Exploring message strategies for encouraging coping behaviors against
particulate matter. Korea J. Commun. Inf. 2018, 92, 7–44.
10. Hwang, I.C. Particulate matter management policy of Seoul: Achievements and limitations. Korea Assoc.
Policy Stud. 2018, 27, 27–51.
11. Han, H.; Jung, C.H.; Kim, H.S.; Kim, Y.P. The revisit on the PM10 reduction policy in Korea. J. Korea Environ.
Policy Adm. 2017, 25, 49–79.
12. Yang, W.H. Changes in air pollutant concentrations due to climate change and the health effect of exposure
to particulate matter. Health Welf. Policy Forum 2019, 269, 20–31.
13. Park, S.; Shin, H. Analysis of the factors influencing PM2.5 in Korea. Korea Environ. Policy Adm. Soc. 2017,
25, 227–248.
14. Kim, Y.P. Air pollution in Seoul caused by aerosols. Korea Soc. Atmos. Environ. 2006, 22, 535–553.
15. Swart, R.O.B.; Raes, F. Making integration of adaptation and mitigation work: Mainstreaming into sustainable
development policies? Clim. Policy 2007, 7, 288–303. [CrossRef]
19
Appl. Sci. 2019, 9, 3660
16. Klein, R.J.T.; Schipper, E.L.F.; Dessai, S. Integrating mitigation and adaptation into climate and development
policy: Three research questions. Environ. Sci. Policy 2005, 8, 579–588. [CrossRef]
17. Brouwer, R.; Schaafsma, M. Modelling risk adaptation and mitigation behaviour under different climate
change scenarios. Clim. Chang. 2013, 117, 11–29. [CrossRef]
18. Wilbanks, T.J.; Sathaye, J. Integrating mitigation and adaptation as responses to climate change: A synthesis.
Mitig. Adapt. Strateg. Glob. Chang. 2007, 12, 957–962. [CrossRef]
19. Gifford, R. The dragons of inaction: Psychological barriers that limit climate change mitigation and adaptation.
Am. Psychol. 2011, 66, 290–302. [CrossRef] [PubMed]
20. Finell, E.; Tolvanen, A.; Pekkanen, J.; Minkkinen, J.; Ståhl, T.; Rimpelä, A. Psychosocial problems, indoor
air-related symptoms, and perceived indoor air quality among students in schools without indoor air
problems: A longitudinal study. Int. J. Environ. Res. Public Health 2018, 15, 1497. [CrossRef] [PubMed]
21. Ferrer, R.; Klein, W.M. Risk perceptions and health behavior. Curr. Opin. Psychol. 2015, 5, 85–89. [CrossRef]
[PubMed]
22. Brewer, N.T.; Chapman, G.B.; Gibbons, F.X.; Gerrard, M.; McCaul, K.D.; Weinstein, N.D. Meta-analysis of the
relationship between risk perception and health behavior: The example of vaccination. Health Psychol. 2007,
26, 136–145. [CrossRef] [PubMed]
23. Menon, G.; Chandran, S. When a day means more than a year: Effects of temporal framing on judgments of
health risk. J. Consum. Res. 2004, 31, 375–389.
24. Jones, C.; Hine, D.W.; Marks, A.D.G. The future is now: Reducing psychological distance to increase public
engagement with climate change. Risk Anal. 2017, 37, 331–341. [CrossRef]
© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
20
applied
sciences
Article
Personal Control of the Indoor Environment in Offices:
Relations with Building Characteristics, Influence on
Occupant Perception and Reported Symptoms
Related to the Building—The Officair Project
Ioannis Sakellaris 1, *, Dikaia Saraga 1,2 , Corinne Mandin 3 , Yvonne de Kluizenaar 4 ,
Serena Fossati 5 , Andrea Spinazzè 6 , Andrea Cattaneo 6 , Tamas Szigeti 7 , Victor Mihucz 7 ,
Eduardo de Oliveira Fernandes 8 , Krystallia Kalimeri 1 , Paolo Carrer 9 and John Bartzis 1, *
1 Department of Mechanical Engineering, University of Western Macedonia, Sialvera & Bakola Str.,
50100 Kozani, Greece
2 Environmental Research Laboratory, INRASTES, National Center for Scientific Research “DEMOKRITOS”,
Aghia Paraskevi Attikis, P.O. Box 60228, 15310 Athens, Greece
3 Université Paris Est, CSTB-Centre Scientifique et Technique du Bâtiment, 84 avenue Jean Jaurès,
77447 Marne-la-Vallée CEDEX 2, France
4 The Netherlands Organization for Applied Scientific Research (TNO), P.O. Box 96800,
2509 JE The Hague, The Netherlands
5 ISGlobal, Institute for Global Health, 08036 Barcelona, Spain
6 Department of Science and High Technology, University of Insubria, Via Valleggio 11, 22100 Como, Italy
7 Cooperative Research Centre for Environmental Sciences, Eötvös Loránd University, Pázmány Péter sétány
1/A, H-1117 Budapest, Hungary
8 Institute of Science and Innovation in Mechanical Engineering and Industrial Management,
Rua Dr. Roberto Frias s/n, 4200-465 Porto, Portugal
9 Department of Biomedical and Clinical Sciences-Hospital “L. Sacco”, University of Milan, via G.B. Grassi 74,
20157 Milano, Italy
* Correspondence: [email protected] (I.S.); [email protected] (J.B.); Tel.: +30-246-105-6710 (J.B.)
Received: 11 July 2019; Accepted: 5 August 2019; Published: 7 August 2019
Abstract: Personal control over various indoor environment parameters, especially in the last decades,
appear to have a significant role on occupants’ comfort, health and productivity. To reveal this
complex relationship, 7441 occupants of 167 recently built or retrofitted office buildings in eight
European countries participated in an online survey about personal/health/work data as well as
physical/psycho-social information. The relationship between the types of control available over
indoor environments and the perceived personal control of the occupants was examined, as well as
the combined effect of the control parameters on the perceived comfort using multilevel statistical
models. The results indicated that most of the occupants have no or low control on noise. Half of the
occupants declared no or low control on ventilation and temperature conditions. Almost one-third of
them remarked that they do not have satisfactory levels of control for lighting and shading from sun
conditions. The presence of operable windows was shown to influence occupants’ control perception
over temperature, ventilation, light and noise. General building characteristics, such as floor number
and floor area, office type, etc., helped occupants associate freedom positively with control perception.
Combined controlling parameters seem to have a strong relation with overall comfort, as well as
with perception regarding amount of privacy, office layout and decoration satisfaction. The results
also indicated that occupants with more personal control may have less building-related symptoms.
Noise control parameter had the highest impact on the occupants’ overall comfort.
Keywords: IEQ; perceived comfort; sick building syndrome; health effects

Appl. Sci. 2019, 9, 3227
1. Introduction
Office employees spend a significant part of their time in modern office buildings that are
characterized by sealed facades and complex building systems (e.g., mechanical, electrical, plumbing,
controls and fire protection systems) designed to reduce energy costs through controlled indoor
environment conditions. Beside central control, a wide range of degrees of personal control, such as
local thermostats, windows, personal lights etc., over the indoor environment can be found in modern
office buildings. Personal control has a crucial role in achieving a healthy [1–3], comfortable [4–7] and
productive [8–11] environment, reducing energy consumption in buildings without sacrificing the
comfort of occupants [12,13]. The effect of personal control on occupant satisfaction, especially with
regards to providing well-being and comfort is an important area of study [14,15].
Several studies have focused on estimating occupants’ comfort in offices and the ability to control
indoor environment parameters, producing a better Indoor Environment Quality (IEQ). In 1990,
Paciuk [16] studied if indoor environment comfort was affected by personal control, leading to a
model using thermal control parameters. Three different types of control parameters were involved:
available control (the degree and type of control made available by the environment), exercised control
(the relative frequency in which occupants engage in several types of controls to obtain comfort) and
perceived control (how the different degrees of available and exercised control interact to produce
different levels of perceived control). The model was applied on data from ten offices (511 workstations).
The results pointed out that the occupants’ level of control perception at their workstation enhances
their satisfaction in their working environment.
The possibility to control light satisfaction was surveyed by Collins et al. [17] in 13 office buildings.
The occupants provided with a task light recorded in general a higher satisfaction with light comfort
than those without a task light. In addition, both groups expressed improved light satisfaction when
they had the possibility to control light conditions. In another study [18], the role of personal control in
natural versus mechanical ventilated office buildings was investigated. Personal control on operable
windows, electronic lightning and solar blinds in natural ventilated buildings resulted in higher
levels of perceived control. It was also mentioned that control systems should be simple and in
compliance with the building design, as well as with quick response to alleviate discomfort as soon
as it is experienced. The level of perceived control by occupants had a small influence on the indoor
environment satisfaction, as described by Haghighat and Donnini [19] through a survey in 12 office
buildings. Satisfaction with respect to temperature, air quality, ventilation, air circulation and overall
comfort showed a moderate correlation with perceived control. On the contrary, there was a decline in
prevalence of health symptoms with an increasing amount of control over the indoor environment.
The association of temperature personal control and operable windows with reported health
symptoms and complaints was indicated in the early 1990s [20]. Many years later, Toftum [21] examined
occupants’ comfort perception and symptoms prevalent in mechanical versus natural ventilated
buildings in 24 office buildings located in Denmark. Multiple logistic regression analysis indicated that
the perceived control was more important for the prevalence of symptoms and environment perception
than the type of ventilation.
Zagreus et al. [22] remarked that occupants with a sense of high degree of control over
environmental parameters such as temperature, air movement, air quality and noise, were more
satisfied with the indoor environment. Boerstra et al. [23] examined the impact of perceived control
and access to control options in occupants’ health and comfort, through 64 office buildings in Europe.
The link between perceived comfort and control parameters such as temperature, ventilation, shading
from the sun, light, noise was investigated. The analysis showed that occupants feel more comfortable
when the perceived control over temperature, ventilation and noise is high. No significant correlation
was found between comfort and the different types of access to control like operable windows, type of
thermostats, etc., except for solar radiation.
In a recent study, Kwon et al. [24] tried to identify the relationship between the level of personal
control and users satisfaction within offices. They found that higher controllability leads to more
22
Appl. Sci. 2019, 9, 3227
thermal and visual satisfaction, while the results revealed the psychological impact on the users’
satisfaction by indicating differences among the available control types. The psychological aspect in
personal control was also raised by Luo and Cao [25] and Karjalainen [26].
The physiological and psychological aspect of IEQ satisfaction [27] and more specific thermal
satisfaction makes it harder to control indoor environment conditions and provide optimal results
for everyone in a given space. The ASHRAE standard 55 [27,28] or ISO 7730 [29,30] tries to give a
solution in this complex relation by stating the appropriate conditions that should be met in order to
establish comfort levels in offices. To achieve that, these standards consider both personal factors, such
as metabolic rate and clothing level, and environmental factors, such as air temperature, mean radiant
temperature, air speed and humidity. Apart from thermal comfort, other factors that can influence
comfort levels are usability of a space, acoustics, ventilation, daylight and energy use in a building.
Literature review highlights the role of personal control on the IEQ in office buildings. The objective
of this paper is to provide an updated overview of the personal control in office buildings and the
association with occupants’ perception. To the best of our knowledge, this is the first time that
European employees have participated in a questionnaire survey covering simultaneously in detail
records of comfort, control and health perception in office buildings. This large-scale survey was
performed under the framework of the European FP7-funded project OFFICAIR [31] and included eight
widely distributed across European countries (Finland, France, Greece, Hungary, Italy, Portugal, Spain,
The Netherlands) with different characteristics (e.g., geographical location, climate, socio-economic
status). More specifically, the aim of this study is threefold: (i) to describe the degree of personal
control over indoor environments in office buildings as reported by occupants (perceived personal
control) and the association with access to available controls; (ii) to investigate the associations between
perceived control and building characteristics; (iii) to study the associations between perceived control
and perceived comfort and health of the occupants.
2.1. Data Collection

This study is based on data collected between October 2011 and May 2012 in the OFFICAIR
project in modern office buildings in eight European countries [32–40]. ‘Modern’ buildings, constructed
during the last 10 years, are described by the presence of several sorts of new electronic equipment and
ventilation, heating and cooling systems, making the indoor environment almost unaffected by local
climate [34]. About 19 to 24 modern buildings, selected on a voluntary basis, were investigated per
country, resulting in 167 office buildings.
The protocol for data collection is described elsewhere in detail [34]. After a preliminary
inspection of each building, a checklist was filled in by a local investigator along with a building
manager, gathering information about building characteristics (e.g., presence of solar devices, operable
windows), mechanical systems (e.g., type of mechanical ventilation, heating, cooling), rooms and
activities (e.g., type of work, cleaning schedules). An online survey for the building occupants was
developed in the national language of the participating countries and included questions on personal
control of the indoor environment as perceived by the occupant, perceived comfort and building-related
symptoms. The survey was anonymous and the participants gave their consent prior to participation.
The study was approved by the competent local/national ethics committees. In total, 26,735 email
invitations were sent to the occupants with an average response rate of 41% across the buildings.
Although the questionnaire was online and its length might have influenced the response rate, the
participation rate can be considered satisfying, in line with other recent surveys [34]. The final database
involved 7441 participants—52% were females and 48% were males with an average age of 41 years.
23
Appl. Sci. 2019, 9, 3227
2.2. Characteristics of the Buildings and Access to Control

The available control types in each office building were obtained from the OFFICAIR checklist
(details in Table 1). The checklist included information regarding the presence of solar shading
devices (grouped as: not present, internal, external), type of solar shading device control, type of
temperature control, presence of operable windows, type of main lights control and type of mechanical
ventilation control.
Table 1. Checklist used to investigate the types of controls over the indoor environment available for
occupants in the building in the OFFICAIR project.
Parameter Items
Not present
South side only
One or more other facades
External vertical blinds
External shutters
External roller shutters
External louvers
External screens
Are there solar shading devices present? External window films
Which kind? External horizontal blinds
External awnings/canopies
External overhangs
External vertical fins
Blind between glazing
Internal vertical blinds
Internal louvers
Atrium
Double façade
Other
No control (fixed)
Individual
How are the solar shading devices controlled?
Central down, individual up
Automatic
Manual radiator valve
Local thermostat at radiator/heating unit
Local thermostat (e.g., on wall)
Central sensor
How is the room temperature controlled? Façade sensor(s)—i.e., outside temperature
Zone sensor(s)
Manual control in room(s)
According to occupancy
Other
Yes
Yes, some
Are the windows operable?
Yes, but occupants are not allowed to open them
No
Automatic by time (building/floor/zone)
Automatic with manual end control (building/floor/zone)
How are main lights (e.g., ceiling or wall)
Demand control: Daylight (photocells)
controlled?
Demand control: Occupants (motion sensors)
Manual
Central—Manual (on/off)
Central—Clock/Central—Demand control (temperature,
What type of control system is there for
CO2 , other pollutant, relative humidity)
mechanical ventilation?
Local—Manual (on/off)
Local—Clock/Local—Demand control
(temperature, CO2 , other pollutant, relative humidity)
Recirculation control
24
Appl. Sci. 2019, 9, 3227
2.3. Perceived Control over the Indoor Environment

Five controlling parameters were set for defining the occupants’ evaluation of the perceived
control and IEQ: Personal control of the occupants over temperature, ventilation, shade from the sun,
light and noise was investigated using the following question with a seven-point Likert-like scale
answer (from “1, not at all” to “7, full control”): ‘How much control do you personally have over the
following aspects of your working environment?’. The combined control variables were introduced.
Several combinations were used (e.g., perceived control over temperature and ventilation) and the
overall combined control variable containing all the 5 parameters [23].
2.4. Personal Comfort, Reported Health Symptoms and Self-Assessed Productivity

The satisfaction of the occupants toward the following parameters was evaluated: overall comfort,
temperature (overall, too hot/cold, variation), air movement, air quality (overall air quality satisfaction,
humid or dry air, stuffy or fresh air, odor), light (overall light satisfaction, natural, artificial, glare),
noise (overall noise satisfaction, outside noise, noise from building systems, noise within the building),
vibration, amount of privacy, office layout, office decoration, and view from the windows. The following
question was used: ‘How would you describe the typical indoor conditions in your office environment
during the past month?’ or ‘How would you describe the following in your office?’ A seven-point
Likert-like scale answer (from “1, dissatisfied” to “7, satisfied”) was provided for most of the questions,
except for those questions investigating two extreme conditions in contrast where a seven-point scale
answer ranging from −3 to 3 was adopted (Table 2) and converted to a scale from 1 to 7 as follow:
+/−3 = 1; +/−2 = 3; +/− 1 =5; 0 = 7.
Table 2. Questions used to investigate occupants’ indoor environment quality perception in the
OFFICAIR project.
Parameter Sub-Parameters Type of Answer

Overall Comfort Seven-point Likert-like scale
Overall temperature
Temperature Seven-point Likert-like scale
Too hot/too cold temperature From “−3, too hot” to “+3, too cold”
Temperature variation From “−3, too much” to “+3, not enough”
Air movement From “−3, too draughty” to “+3, too still”
Air quality Overall air quality Seven-point Likert-like scale
Humid/dry air From “−3, too humid” to “+3, too dry”
Stuffy or fresh air Seven-point Likert-like scale
Odor Seven-point Likert-like scale
Light Overall light Seven-point Likert-like scale
Natural Seven-point Likert-like scale
Artificial Seven-point Likert-like scale
Glare Seven-point Likert-like scale
Noise Overall noise Seven-point Likert-like scale
Outside noise Seven-point Likert-like scale
Noise from building systems Seven-point Likert-like scale
Noise within the building Seven-point Likert-like scale
Vibration Seven-point Likert-like scale
Amount of privacy Seven-point Likert-like scale
Office layout Seven-point Likert-like scale
Office decoration Seven-point Likert-like scale
View from the windows Seven-point Likert-like scale
The occupants were also requested to estimate their productivity at the workstation and in other
locations inside the building, considering the influence of environmental conditions on a scale of
7 levels, from +30% to −30%. In addition, they were requested to record building-related health
symptoms. The Personal Symptom Index-5 (PSI-5) was calculated based on the incidence of five
health symptoms—dry eyes, blocked or stuffy nose, dry throat, headache, and tiredness—which
25
Appl. Sci. 2019, 9, 3227
are considered to be the fundamental components of sick building symptoms as mentioned by

Raw et al. [41]. This indicator has a 0 to 5 score, according to the prevalence of the reported symptoms.
The respective question was: ‘Have you ever experienced any of the following symptoms while
working in this building (or workstation) (including today)?’
2.5. Statistical Analysis

The statistical analysis of the dataset was performed in four steps (Figure 1):
i. Descriptive results were obtained.

ii. The correlation between the available control and the perceived control using the Kruskal
Wallis analysis of variance test was investigated [23,42]. Groups with less than five individuals
were not included in the analysis.
iii. The relationship between the occupants’ overall combined perceived control and the general
physical building characteristics as well as the occupant personal characteristics was examined
by applying a multilevel model [43,44], to account for the three-level structure of our data
(level 1-occupant, level 2-building, level 3-country). The ordered logistic regression analysis
was applied using building and country as random effects and the covariates as fixed effects.
Four step-by-step models were applied. The first was an empty model without any variable
and with building and country variance only. In the second model, individual level variables
were imported. The final version of the second model included variables, with p-value
below 0.2, such as gender, age (in four groups, <35, 35–45, 46–55, >55), effort reward ratio,
experience of negative events, use of air fresheners at home, type of job (managerial, professional,
clerical/secretarial, other), type of job contract (full-time, part-time), and job contract duration
(permanent, fixed-term). In the third model, building characteristics were imported iteratively
on the second model to identify significant relations with the perceived control. Variables with
a p-value below 0.2 were selected to be used in the next model. In the fourth model, both
individual and building level variables were imported. The results of the fixed effects were
reported in Odds Ratios (OR) and 95% Confidence Interval (CI). For the random effects, the
explained variance is reported as well as the Proportional Change in Variance (PCV) between
the null model and the final model with the variables.
iv. The potential relations between the occupants’ perceived control and perceived comfort and
reported health symptoms (PSI5) were examined by using the spearman correlation [45].
This study focused on the correlation between the various combined control scores (e.g.,
perceived control over temperature and ventilation) and the perceived occupants’ comfort.
Additionally, the dependent variable (overall comfort satisfaction) and the response-variables
were expressed in values on an ordinal scale; hence, ordinal regression analysis was
employed to determine the impact of the controlling parameters on overall comfort. In the
regression model, the response-dependent variable was the overall comfort satisfaction and the
predictor-independent variables were the satisfaction for each personal controlling parameter
as evaluated by the occupants. The results are presented in the format of OR and its CI95%.
The ORs were used to rank the effect of the personal controlling parameters on overall comfort.
p-values <0.05 were considered as statistically significant. The statistical package IBM SPSS
Statistics [46] was used for the analysis.
26
Appl. Sci. 2019, 9, 3227
Step1: Descriptive Analysis
Step 2: Kruskal Wallis tests. Model 1: Empty model

Available control vs.
perceived control
Step 3: Multilevel model - Model 2: Individual level

Ordered logistic regression variables
(Level 1-occupant, level 2-
building, level 3-country)
Perceived control vs.
general physical building Model 3: Building characteristics.
characteristics & occupant Variables with a p-value <0.2
personal characteristics were selected for the next model
Model 4: Final model with

individual and building level
variables
4.1: Spearman correlation -

Combined control scores vs.
perceived comfort & PSI5
Step 4: Perceived control vs.
perceived comfort & health
(PSI5)
4.2: Ordered logistic regression -
Overall comfort vs. Perceived
control
Figure 1. Schematic overview of the methodology used.
3. Results
3.1. Characteristics of Perceived Personal Control

The responses to the perceived controls (with a response rate above 99% for all control types)
are presented in Figure 2. In general, noise, ventilation and temperature were perceived as poorly
controlled. Regarding shading from the sun and lighting, the occupants reported the perceived control
as moderate. A group of people (up to 35%) declared “no control” can be observed for all control
parameters. For temperature and ventilation control, the rest of the occupants are equally distributed
in the five levels of control degree. Very few occupants seem to be able to control the noise level at
their workstation, where answers follow a descending rate (full control below 5%). On the contrary,
regarding lighting and shading from the sun, after the distinctive “no control” group, the occupants’
answers exhibited an ascending rate, indicating that they feel more able to control the light conditions
(full control up to 20%).
Figure 3 presents the percentages of the office occupants who have no or low perceived control
(values < 4) on their indoor working environment. The majority (63%) has no or low control on noise.
Half of them declared no or low control on ventilation (53%) and temperature (47%). Almost one-third
of them remarked that they do not have satisfactory levels of control for lighting and shading from
sun conditions.
In Figure 4, the responses are presented based on the occupancy per room. Responses were
categorized into five groups. In the first group, which is characterized by personal office rooms,
occupants seem to have higher levels of control, except for ventilation. It is interesting to note that as
the occupancy increases, the degree of personal control becomes lower. Offices with many occupants
(30+), probably open space offices, show the lowest degrees of personal control for all parameters.
Furthermore, in all groups, the occupants characterized noise control as the worst parameter, while
lighting and shading from the sun control gathered higher degrees of control.
27
Appl. Sci. 2019, 9, 3227
Figure 2. Percentage of occupants based on self-reported degree of control (n = 7441).
Noise control
Ventilation control
Temperature control
Shading from the sun control
Lighting control
0% 10% 20% 30% 40% 50% 60% 70%
Figure 3. Percentage of occupants who reported no or low control.
Figure 4. Mean degree of perceived control according to room occupancy (1 = no control, 7 = full control).
The satisfaction comfort towards IEQ parameters was also examined, with regard to their respective
control options. The perceived satisfaction parameters were categorized in three levels of perceived
control: no control (1), low control (2–3), high control (≥4) and are presented in Figure 5. For each IEQ
parameter, the related personal control option was selected, e.g., for parameters about temperature
satisfaction, temperature control was selected; for indoor air quality parameters, ventilation control
28
Appl. Sci. 2019, 9, 3227
was selected, etc. Occupants with high degrees of personal control over their working environment
reported higher levels of satisfaction on average. In all cases, occupants with high perceived control
reported higher satisfaction levels. Occupants without personal control evaluated satisfaction of the
IEQ with lower levels. Only in some cases, ‘none’ and ‘low personal’ control options are reversed, e.g.,
in the case of satisfaction with odor and reflection.
7
Perceived comfort
6
5
4
3
2
1
0
No perceived control (1) Low perceived control (2-3)
Hi perceived control (ǃ4) Neutral level
Figure 5. Perceived comfort towards indoor environment quality according to three levels of
perceived control.
Figure 6 shows the degree of personal control vs. self-reported productivity at the workstation
(n = 7289) and in other places in the building (n = 7154). In both cases, occupants with higher levels
of personal control reported higher levels of productivity (Kruskal Wallis: p < 0.001). Regarding the
workstations, a small increase in the low control area results in a clear increase in productivity.
Figure 6. Self-reported productivity vs. degree of control.
3.2. Relationship between Perceived Control and Available Control

The statistically significant (p < 0.05) correlations between the available control and perceived
control are presented in Figures 7–11.
Regarding the availability of operable windows (Figures 7–10), the scores of perceived control over
temperature, ventilation, light and noise differ significantly (p < 0.001, p < 0.001, p = 0.001, p < 0.001,
respectively) among buildings with operable windows, some operable windows, operable windows
that people are not allowed to open, and no operable windows. The mean occupants’ score over
29
Appl. Sci. 2019, 9, 3227
temperature in buildings with operable windows was 1.3 points higher than in buildings without
operable windows (p < 0.001). The corresponding difference in the mean score for ventilation, light
and noise was 2 (p < 0.001), 1 (p < 0.001) and 0.8 (p = 0.001), respectively.
Figure 7. Perceived control over temperature vs. availability of operable windows (n is the number of
office buildings).
Figure 8. Perceived control over ventilation vs. availability of operable windows (n is the number of
office buildings).
Figure 9. Perceived control over light vs. availability of operable windows (n is the number of
office buildings).
30
Appl. Sci. 2019, 9, 3227
Figure 10. Perceived control over noise vs. availability of operable windows (n is the number of
office buildings).
With regard to the presence of solar shading devices, the buildings were grouped into three
categories based on the solar devices set up: internal, external, none (Figure 11). Perceived control over
temperature varied significantly through the different types of the solar devices (p = 0.028). The highest
score of perceived control over temperature was in buildings with no solar devices. This might
be explained by the fact that the design of these office buildings provides adequate control over
temperature without the use of solar shading devices; the multilevel regression analysis in the next
step provides additional insights. No significant differences were observed between perceived control
over shading and lighting in relation to the types of solar shading devices (p = 0.635 and p = 0.255,
respectively, Table 3).
Figure 11. Perceived control over temperature vs. availability of solar shading devices (n is the number
of office buildings).
Table 3 presents the remaining outcomes of the statistical tests performed between the available
control and occupants’ perceived control. Regarding temperature control, the analysis showed no
significant differences (p = 0.166) between the perceived control over temperature and the various
types of temperature controls (e.g., manual radiator valve and central sensor). The different types
of operable windows did not affect the perceived control of the shading from the sun (p = 0.100).
Subsequently, the type of control on the solar shading devices (fixed, individual, central control,
automatic) was examined (Table 3). Unlike the various types of solar devices, analysis of the available
different control types of the shading devices showed that there are no significant differences among
the scores of the perceived control over shading, lighting and temperature (p = 0.147, p = 0.710 and
p = 0.755, respectively). As far as the available controls of lights are concerned, no significant difference
was found for perceived control over lighting (p = 0.328). Finally, the several types of mechanical
ventilation control were examined. The occupants’ perceived control over ventilation and temperature
did not significantly differ (p = 0.105 and p = 0.768, respectively) from the available control types.
31
Table 3. Available control at building level and occupants’ perceived control (from 1 = no control at all to 7 = full control) with p > 0.05.
Perceived Control Perceived Control Perceived Control Perceived Control

over Temperature over Shading over Light over Ventilation
Type of Available Temperature Controls
Manual radiator valve 3.38
Local thermostat at radiator/heating unit 3.96
Appl. Sci. 2019, 9, 3227
Local thermostat (e.g., on wall) 3.53

Central sensor 3.52
Facade sensor(s)—i.e., outside temperature 4.12
Zone sensor(s) 4.03
Manual control in room(s) 3.70
p-value (Kruskal-Wallis test) 0.166
Availability of shading devices
None 4.09 4.69
Internal 4.43 4.43
External 4.22 4.44
p-value (Kruskal-Wallis test) 0.635 0.255
Type of control of the available solar shading devices
No control (fixed) 3.45 3.90 4.53
Individual 3.60 4.43 4.43
Central down, individual up 3.25 4.46 3.76
Automatic 3.77 4.18 4.61
32
p-value (Kruskal-Wallis test) 0.755 0.147 0.710
Type of operable windows
Yes 4.42
Yes, some 4.15
Yes, but occupants are not allowed to open them 4.81
No 4.11
Type of available light controls
Manual 4.52
Demand control: Occupants (motion sensors) 4.42
Demand control: Daylight (photocells) 4.26
Automatic with manual end control (building/floor/zone) 4.91
Automatic by time (buildings/floor/zone) 4.54
Type of available mechanical ventilation controls
Central—Manual (on/off) 3.81 3.33
Central—Clock 3.62 3.23
Central—Demand control (temperature, CO2 , other pollutant, relative humidity) 3.58 3.27
Local—Manual (on/off) 4.11 4.27
Local—Clock 3.74 3.78
Local—Demand control (temperature, CO2 , other pollutant, relative humidity) 4.42 3.73
p-value (Kruskal-Wallis test) 0.768 0.105
Appl. Sci. 2019, 9, 3227
3.3. Combined Perceived Control Versus Building Characteristics

The variances of country and building level according to Model 1 (the null model) are 0.38 and
1.17, respectively. The respective variances in Model 2, which includes the individual variables (gender,
age groups, effort reward ratio, experience of negative events, use of air fresheners at home, type of job,
type of job’s contract, job’s contract duration), were 0.52 and 0.94. The relations between the building
characteristics, which were added in the model iteratively, and the combined perceived control with
p-values < 0.2 are presented in Table 4 (Model 3). The strongest associations with a p-value < 0.001 were
found for the building’s location (suburban area), the maximum number of occupants, the documented
complaints procedure, the number of people in office, as well as for the office type and availability of
opening windows. Significant associations with a p-value < 0.05 were found for the number of adjacent
facades with other buildings, glares from neighboring buildings, total floor area, pesticide treatment
plan, smoking permission, use of portable air cleaner, floor of the workspace, partition in the offices,
and noise source of occupants—distracting conversations and the location of air and exhaust devices.
Table 4. Relations between combined overall perceived control and building characteristics (Model 3).
Characteristics n1/N1 n2/N2 OR (CI 95%) p-Value

General building characteristics
Location
Mixed industrial/residential area (vs. industrial area) 696/7441 14/167 1.68 (0.76–3.69) 0.199
Commercial area (vs. industrial area) 789/7441 18/167 0.95 (0.46–1.95) 0.883
Mixed commercial/residential area (vs. industrial area) 2279/7441 50/167 1.71 (0.94–3.09) 0.076
City center, densely packed housing (vs. industrial area) 1344/7441 30/167 1.35 (0.7–2.63) 0.374
Town, with no or small gardens (vs. industrial area) 328/7441 8/167 3.42 (1.33–8.75) 0.010
Suburban, with larger gardens (vs. industrial area) 864/7441 22/167 3.88 (1.92–7.82) <0.001
Village in a rural area (vs. industrial area) 24/7441 1/167 0.82 (0.09–7.46) 0.859
Rural area with no or few other homes nearby (vs. industrial area) 192/7441 6/167 1.27 (0.45–3.57) 0.646
Façades with adjacent buildings
2 façades (vs. 1 façade) 1932/4664 37/102 0.6 (0.38–0.95) 0.030
3 façades (vs. 1 façade) 288/4664 9/102 1.48 (0.68–3.25) 0.322
Density of nearby obstructions
Moderately dense (vs. very dense) 2610/7441 58/167 1.31 (0.69–2.5) 0.407
Few buildings (vs. very dense) 2992/7441 58/167 1.43 (0.74–2.76) 0.285
Free standing (vs. very dense) 1031/7441 26/167 1.95 (0.93–4.09) 0.076
Neighboring buildings with glass facades or
light-colored facades causing glare
Yes, in summer (vs. No) 215/7441 4/167 0.4 (0.13–1.27) 0.120
A little in summer (vs. No) 654/7441 12/167 0.5 (0.25–1) 0.049
Yes, in winter (vs. No) 102/7441 2/167 0.35 (0.07–1.71) 0.195
A little in winter (vs. No) 217/7441 5/167 0.24 (0.09–0.68) 0.007
Maximum number of occupants in the building
Between 101 and 250 (vs. ≤100) 2288/7260 56/163 0.79 (0.52–1.2) 0.272
≥251 (vs. ≤100) 3544/7260 51/163 0.39 (0.25–0.61) <0.001
Total floor area (m2 )
Between 1441 and 3210 (≤1440) 2300/7234 53/160 0.78 (0.49–1.24) 0.295
≥3211 (≤1440) 3341/7234 54/160 0.51 (0.31–0.83) 0.007
Activities in the building besides office work
Underground car park (vs. No) 2678/7441 44/167 0.73 (0.46–1.16) 0.178
Print shop (vs. No) 428/7441 7/167 1.92 (0.78–4.73) 0.156
Pesticide treatment plan (vs. No) 7025/7441 99/164 0.59 (0.39–0.88) 0.010
Documented complaints procedure for occupants (vs. No) 5112/7349 99/165 0.41 (0.28–0.61) <0.001
Smoking permission
Only outside the building (vs. No) 3916/7441 88/167 0.55 (0.33–0.93) 0.026
Only in separately ventilated rooms (vs. No) 973/7441 13/167 0.29 (0.13–0.63) 0.002
Percentage of office furniture is less than one year old and made of MDF 7025/7441 162/162 0.99 (0.99–1) 0.200
Portable air cleaner (vs. No) 225/7441 5/167 0.33 (0.12–0.93) 0.035
33
Appl. Sci. 2019, 9, 3227
Table 4. Cont.
Characteristics n1/N1 n2/N2 OR (CI 95%) p-Value

Office characteristics
Floor surface of the office 7410/7441 167/167 1.05 (1.02–1.09) 0.001
Number of people in the office on average
Private (vs. 30+) 1501/7352 144/167 7.65 (6.19–9.45) <0.001
2–4 (vs. 30+) 2134/7352 152/167 5.15 (4.24–6.26) <0.001
5–9 (vs. 30+) 1019/7352 117/167 2.62 (2.15–3.21) <0.001
10–29 (vs. 30+) 1492/7352 100/167 1.48 (1.25–1.76) <0.001
Type of office
Shared private office (vs. single person private office) 2236/7417 150/167 0.56 (0.48–0.65) <0.001
Open space with partitions (vs. single person private office) 991/7417 115/167 0.21 (0.17–0.26) <0.001
Open space without partitions (vs. single person private office) 2602/7441 139/167 0.21 (0.17–0.25) <0.001
Other (vs. single person private office) 90/7417 60/167 0.39 (0.24–0.65) <0.001
Partitions within the offices (vs. No) 4112/7441 91/167 0.67 (0.45–0.99) 0.043
Noise source of occupants–distracting conversations (vs. No) 2653/7441 64/167 0.66 (0.45–0.98) 0.040
PC or laptop monitor position
In front of windows (vs. Not) 3018/7441 69/167 1.44 (1–2.08) 0.052
Printer/copy machines location
In the offices (vs. on the corridor) 2976/7441 75/167 0.62 (0.39–1) 0.051
In a separate printing room (vs. on the corridor) 2501/7441 47/167 0.64 (0.39–1.05) 0.075
Operable windows
Yes (vs. No) 3724/7441 85/163 5.51 (3.14–9.67) <0.001
Yes, some (vs. No) 1769/7441 40/163 3.08 (1.72–5.53) <0.001
Yes, but occupants are not allowed to open them (vs. No) 913/7441 20/163 1.54 (0.79–2.98) 0.204
Location of air supply devices inside offices—ceiling (vs. No air supply) 5783/7441 123/167 0.63 (0.4–0.99) 0.046
Location of air exhaust devices inside offices
High (vs. None) 6189/7089 128/154 0.46 (0.25–0.87) 0.017
Low (vs. None) 382/7089 10/154 0.63 (0.24–1.63) 0.339
p-values in bold are significant at 5%. Adjusted for gender, age groups, effort reward ratio, experience of negative
events, use of air fresheners at home, type of job, type of job’s contract, job’s contract duration. Characteristics
with a p value lower than 0.20 are presented. Level 1—Occupant level, Level 2—Building level, Level 3—Country
level. OR: Odd Ratio, CI: Confidence interval 95%, n1/N1: Occupants’ answers/total number of occupants, n2/N2:
building observations/total number of buildings.
All individual and building characteristics from Models 2 and 3 were imported to Model 4; the
relevant results are presented in Table 5. The variance between buildings was equal to 0.14 in Model 4.
PCV shows that 88% of the initial variance in overall perceived control was explained by the included
variables. Buildings with larger total floor areas were positively associated with perceived overall
control (OR 1.99, 95% CI: 1.22–3.25). The workspace floor was also positively associated with control
perception (OR 1.06, 95% CI: 1.02–1.10). The existence of any kind of operable windows (OR 3.64 up
to 6.53) and private and semi-private offices had the highest impact on the perceived overall control
(OR 1.58 up to 3.73). On the other hand, the building’s location (OR 0.24, 95% CI: 0.06–0.96) and the
buildings’ number of adjacent facades (OR 0.65, 95% CI: 0.44–0.97) with other buildings tended to
negatively associate with the perceived control. Both high number of occupants in the building (OR
0.67, 95% CI: 0.46–0.99) and the type of office (OR from 0.23 to 0.36) had a significant negative effect on
the perceived overall control. Indoor noise from distracting conversations was associated negatively
with the perceived overall control (OR 0.69, 95% CI: 0.49–0.97).
Concerning individual characteristics, in addition to gender (OR 1.21, 95% CI: 1.06–1.38) and age
group (OR 0.78, 95% CI: 0.61–0.99), strong association with the overall perceived control was found for
the ERI (OR 0.67, 95% CI: 0.46–0.99) and the occupants’ habit to use air fresheners at home (OR 1.26,
95% CI: 1.11–1.44).
34
Appl. Sci. 2019, 9, 3227
Table 5. Associations between combined overall perceived control, building characteristics and
individual characteristics (Model 4).
Factors OR (CI 95%) p-Value

Individual characteristics/personal activities
Female (vs. Male) 1.21 (1.06–1.38) 0.006
Age
<35 (vs. 55+) 0.89 (0.69–1.14) 0.345
35–45 (vs. 55+) 0.78 (0.61–0.99) 0.040
46–55 (vs. 55+) 0.95 (0.74–1.23) 0.720
Effort-reward ratio 0.31 (0.25–0.38) <0.001
Experience of negative events (vs. No) 0.94 (0.83–1.07) 0.367
Type of job
Managerial (vs. Other) 0.97 (0.75–1.27) 0.844
Professional (vs. Other) 0.98 (0.78–1.23) 0.873
Clerical-secretarial (vs. Other) 0.95 (0.77–1.17) 0.625
Type of job’s contract–Full-time (vs. Part-time) 1.12 (0.89–1.41) 0.344
Job’s contract duration–Permanent (vs. Fixed-term) 1.13 (0.91–1.39) 0.263
Air fresheners use in home (vs. No) 1.26 (1.11–1.44) <0.001
Building characteristics
Location
Mixed industrial/residential area (vs. industrial area) 0.64 (0.31–1.29) 0.208
Commercial area (vs. industrial area) 1.59 (0.83–3.07) 0.164
Mixed commercial/residential area (vs. industrial area) 1.05 (0.61–1.83) 0.849
City Centre, densely packed housing (vs. industrial area) 1.15 (0.62–2.1) 0.660
Town, with no or small gardens (vs. industrial area) 2.98 (0.95–9.39) 0.062
Suburban, with larger gardens (vs. industrial area) 1.73 (0.63–4.72) 0.285
Village in a rural area (vs. industrial area) 0.24 (0.06–0.96) 0.043
Rural area with no or few other homes nearby (vs. industrial area) 1.02 (0.24–4.45) 0.974
Density of nearby obstructions
Moderately dense (vs. Very dense) 1.11 (0.63–1.95) 0.729
Few buildings (vs. Very dense) 1.14 (0.65–2.02) 0.648
Free standing (vs. Very dense) 0.99 (0.44–2.19) 0.972
Maximum number of occupants in the building
Between 101 and 250 (vs. ≤100) 0.67 (0.46–0.99) 0.045
≥251 (vs. ≤100) 0.92 (0.56–1.52) 0.745
Total floor area
Between 1441 and 3210 (≤1440) 1.99 (1.22–3.25) 0.006
≥3211 (≤1440) 1.44 (0.81–2.57) 0.217
Façades with adjacent buildings
2 façades (vs. 1 façade) 0.65 (0.44–0.97) 0.033
3 façades (vs. 1 façade) 1 (0.51–1.95) 0.992
Neighboring buildings with glass facades or light-colored facades causing glare
Yes, in Summer (vs. No) 3.16 (1.15–8.66) 0.025
A little in Summer (vs. No) 1.52 (0.74–3.14) 0.256
Yes, in Winter (vs. No) 0.24 (0.06–0.94) 0.040
A little in Winter (vs. No) 0.49 (0.17–1.39) 0.180
Operable windows
Yes (vs. No) 4.81 (2.85–8.14) <0.001
Yes, some (vs. No) 6.53 (3.76–11.34) <0.001
Yes, but occupants are not allowed to open them (vs. No) 3.64 (1.82–7.27) <0.001
Activities in the building besides office work
Underground car park (vs. No) 0.79 (0.52–1.19) 0.258
Print shop (vs. No) 1.4 (0.6–3.27) 0.435
Portable air cleaner (vs. No) 0.79 (0.31–2) 0.622
Pesticide treatment plan (vs. No) 1.33 (0.74–2.38) 0.336
Smoking permission
Only outside the building (vs. No) 0.82 (0.46–1.46) 0.492
Only in separately ventilated rooms(vs. No) 0.66 (0.32–1.36) 0.260
Documented complaints procedure for occupants (vs. No) 0.77 (0.5–1.17) 0.221
Percentage of office furniture is less than one year old and made of MDF 1 (0.99–1.01) 0.800
35
Appl. Sci. 2019, 9, 3227
Table 5. Cont.
Factors OR (CI 95%) p-Value

Office characteristics
Floor surface of the office 1.06 (1.02–1.1) 0.002
Number of people in the office on average
Private (vs. 30+) 2.51 (1.59–3.97) <0.001
2–4 (vs. 30+) 3.73 (2.78–4.99) <0.001
5–9 (vs. 30+) 2.3 (1.79–2.96) <0.001
10–29 (vs. 30+) 1.58 (1.29–1.93) <0.001
Type of office
Shared private office (vs. Single person private office) 0.36 (0.23–0.55) <0.001
Open space with partitions (vs. Single person private office) 0.23 (0.14–0.36) <0.001
Open space without partitions (vs. Single person private office) 0.26 (0.17–0.4) <0.001
Other (vs. Single person private office) 0.57 (0.25–1.29) 0.178
Partitions within the offices (vs. No) 1.02 (0.64–1.61) 0.935
Noise source of Occupants–distracting conversations (vs. No) 0.69 (0.49–0.97) 0.034
PC or laptop monitor position
In front of windows (vs. Not) 1.32 (0.94–1.86) 0.103
Printer/copy machines location
In the offices (vs. on the corridor) 0.99 (0.63–1.55) 0.956
In a separate printing room (vs. on the corridor) 1.23 (0.76–1.99) 0.396
Location of air supply devices inside offices—ceiling (vs. No air supply) 0.81 (0.51–1.28) 0.365
Location of air exhaust devices inside offices
High (vs. None) 1.43 (0.61–3.34) 0.408
Low (vs. None) 0.93 (0.37–2.35) 0.879
County level σ2/PCV (%) 0.42/10
Building level σ2/PCV (%) 0.14/88
3.4. Impact of Perceived Control on Perceived Comfort and Health
3.4.1. Combined Perceived Control vs. Perceived Comfort and Health—Bivariate Analysis
The relationships between perceived control and perceived comfort are presented in Table 6.
Perceived overall comfort correlated significantly and positively with perceived control over all control
parameters. When the scores of the perceived control were combined, the correlation with the overall
comfort increased, reaching the highest value of r = 0.465. The combined correlation values are stronger
than the correlations between the single perceived control parameters. The strongest correlation for
perceived overall temperature was indicated with control over temperature (r = 0.420) and not with the
combined controls. The satisfaction with extreme hot or cold temperature conditions seems to have the
strongest correlation only with the control over temperature (r = 0.281), while the combined perceived
controls show equal effect. On the other hand, temperature variation was found to be more related
to the combined control over temperature and shading from the sun controls (r = 0.209). Perceived
air quality about dry or humid air and odor did not show strong correlations, either with single or
with combined perceived control parameters, indicating that these parameters were more difficult
to be controlled by the occupants. However, perceived overall air quality satisfaction and perceived
satisfaction with fresh air were found to be more correlated with the combined control perception over
temperature, ventilation and shading from the sun (r = 0.380 and 0.320, respectively).
Perceived comfort regarding natural light correlated positively with perceived control over shading
from the sun and over light (r = 0.370). Moreover, artificial light perception showed correlation with
the perceived control over light (r = 0.377). This means that occupants with higher degrees of shading
and lighting controls feel more satisfied with the light levels in their offices. Perceived satisfaction
with glare did not show strong correlation. In general, combined controls do not significantly affect
light satisfaction.
36
Table 6. Correlations between combined perceived control and perceived comfort and health.
Perceived Control Perceived Control over

over Temperature Temperature, Ventilation,
over Temperature and Temperature, Ventilation
and Ventilation Shading from the Sun,
Shading from the Sun and Shading from the
Combined [2: Light and Noise
Combined [2: None at Sun Combined [3: None
None at all-14: Combined [5: None at
all-14: Full Control] at all-21: Full Control]
Full Control] all-35: Full Control]
Appl. Sci. 2019, 9, 3227
0.367 0.415 0.415 0.465

Overall comfort [1: Unsatisfactory-7: Satisfactory]
<0.001 <0.001 <0.001 <0.001
0.258 0.257 0.255 0.245
Temperature [1: Too hot/cold- 7: Satisfactory]
<0.001 <0.001 <0.001 <0.001
0.154 0.209 0.186 0.188
Temperature [1: Varies too much/not enough variation-7: Satisfactory]
<0.001 <0.001 <0.001 <0.001
0.403 0.414 0.416 0.43
Temperature [1: Unsatisfactory-7: Satisfactory]
<0.001 <0.001 <0.001 <0.001
0.175 0.164 0.178 0.182
Air movement [1: Draughty/Still- 7: Satisfactory] Air movement
<0.001 <0.001 <0.001 <0.001
0.143 0.142 0.147 0.162
Air quality [1: Dry/humid-7: Satisfactory]
<0.001 <0.001 <0.001 <0.001
0.287 0.312 0.322 0.342
Air quality [1: Stuffy- 7: Fresh]
37
<0.001 <0.001 <0.001 <0.001
0.129 0.183 0.17 0.195
Air quality [1: Smelly-7: Odorless]
<0.001 <0.001 <0.001 <0.001
0.35 0.363 0.38 0.405
Air quality [1: Unsatisfactory-7: Satisfactory]
<0.001 <0.001 <0.001 <0.001
0.247 0.356 0.342 0.361
Natural light [1: Unsatisfactory-7: Satisfactory]
<0.001 <0.001 <0.001 <0.001
0.245 0.281 0.284 0.331
Artificial light [1: Unsatisfactory-7: Satisfactory]
<0.001 <0.001 <0.001 <0.001
0.123 0.145 0.142 0.166
Reflection or glare [1: Glare|-7: No glare]
<0.001 <0.001 <0.001 <0.001
0.295 0.364 0.361 0.401
Light overall [1: Unsatisfactory-7: Satisfactory]
<0.001 <0.001 <0.001 <0.001
−0.004 0.077 0.048 0.077
Noise from outside the building [1: Unsatisfactory-7: Satisfactory]
0.707 <0.001 <0.001 <0.001
Noise from building systems (e.g., heating, plumbing, ventilation, air conditioning) [1: 0.172 0.191 0.198 0.226
Unsatisfactory-7: Satisfactory] <0.001 <0.001 <0.001 <0.001
Noise from within the building other than from building systems (e.g., phone calls, 0.28 0.302 0.304 0.367
colleagues chatting, photocopiers, etc.) [1: Unsatisfactory-7: Satisfactory] <0.001 <0.001 <0.001 <0.001
Table 6. Cont.

over Temperature Temperature, Ventilation,
over Temperature and Temperature, Ventilation
and Ventilation Shading from the Sun,
Shading from the Sun and Shading from the
Combined [2: Light and Noise
Combined [2: None at Sun Combined [3: None
None at all-14: Combined [5: None at
all-14: Full Control] at all-21: Full Control]
Full Control] all-35: Full Control]
Appl. Sci. 2019, 9, 3227
0.286 0.322 0.321 0.389

Noise overall [1: Unsatisfactory-7: Satisfactory]
<0.001 0.000 0.000 0.000
0.108 0.177 0.16 0.193
Vibration [1: Unsatisfactory-7: Satisfactory]
<0.001 <0.001 <0.001 <0.001
0.333 0.385 0.379 0.433
Amount of privacy [1: Unsatisfactory-7: Satisfactory]
<0.001 <0.001 <0.001 <0.001
0.31 0.367 0.361 0.403
Layout [1: Do not like at all-7: Like very much]
<0.001 <0.001 <0.001 <0.001
0.22 0.295 0.278 0.315
Decoration [1: Do not like at all-7: Like very much]
<0.001 <0.001 <0.001 <0.001
0.147 0.256 0.233 0.246
View from the windows [1: Do not like at all-7: Like very much]
<0.001 <0.001 <0.001 <0.001
−0.249 −0.251 -0.263 −0.289
PSI5
38
<0.001 <0.001 <0.001 <0.001
Rho and p-value of Spearman correlation. Significant at 5% is in bold.
Appl. Sci. 2019, 9, 3227
Perceived comfort over outdoor noise seems to have low correlation with perceived controls.
The occupants were not able to control the noise levels coming from outside probably due to inefficient
available control types. Noise from building systems and noise within the building correlated positively
with perceived control over noise (r = 0.205 and 0.377). Moreover, the occupants were more satisfied
with the noise from building systems when they had higher degree of all combined controls (r = 0.226).
The highest correlation was observed (r = 0.408) between the overall noise satisfaction and perceived
noise control. This indicates that the occupants are more comfortable in buildings where they perceive
high degree of control over noise.
The amount of privacy was found to have significant correlation with all control parameters,
both single and combined, while the maximum correlation was recorded with all combined controls
(r = 0.433). This implies that high degrees of personal control improve the privacy that the occupant
needs. This positive relation was also observed for office layout and decoration perceived satisfaction
with r = 0.403 and 0.315, respectively.
As far as perceived control and the presence of health symptoms assessed through the PSI5 are
concerned, a significant negative correlation was found for all control parameters, both single and
combined. The combined perceived control scores exhibited a stronger negative correlation with
the PSI5.
3.4.2. Perceived Control vs. Perceived Overall Comfort—Regression Analysis

The relation between the overall comfort and five perceived control variables has been examined
by applying the ordinal regression analysis. The results are presented in Table 7. The maximum OR
value (1.28) corresponded to the perceived control over noise. The results showed that if the perceived
control over noise increases by one unit in the 1–7 point scale, there is a 1.28 time likelihood that the
overall comfort will increase by one unit. Perceived noise control was the parameter with the lowest
score according to the occupants’ recordings, as shown in Table 2. The impact on the overall comfort of
the control over noise, lighting, temperature and shading from the sun, was almost equivalent (ORs:
1.16, 1.14 and 1.12, respectively). Lastly, the perceived control over ventilation, despite the fact that it
was ranked as the second worst parameter, had the lowest impact on overall comfort (OR 1.03, 95% CI:
1.00–1.06).
Table 7. Relations between perceived control and perceived overall comfort.
Parameters OR CI (95%) p-Value

Perceived control over noise 1.28 (1.25–1.32) p < 0.001
Perceived control over lighting 1.16 (1.13–1.19) p < 0.001
Perceived control over temperature 1.14 (1.11–1.18) p < 0.001
Perceived control over shade 1.12 (1.10–1.15) p < 0.001
Perceived control over ventilation 1.03 (1.00–1.06) p < 0.05
Moreover, occupants were separated in two groups: those who had no or low control and those
who had high control for the sum of the five control parameters. The results are presented in Table 8.
The ranking remains the same as in the previous case, but two remarks can be highlighted. First, in the
“low control” group, occupants were less sensitive to controlling parameters, with lower OR values.
In addition, the perceived control over ventilation was not significant (p > 0.05). Second, in the “high
control” group, occupants were more sensitive to all controlling parameters with increased OR values.
The perceived control over lighting showed an almost equal impact on overall comfort as the perceived
control over noise.
39
Appl. Sci. 2019, 9, 3227
Table 8. Relations between perceived control in low and high control groups and perceived
overall comfort.
Parameters— Parameters—
OR CI (95%) p-Value OR CI (95%) CI (95%)
Low Control Group High Control Group
Perceived control Perceived control
1.21 (1.15–1.27) p < 0.001 1.37 (1.32–1.43) p < 0.001
over noise over noise
1.14 (1.11–1.17) p < 0.001 1.32 (1.25–1.39) p < 0.001
over lighting over lighting
1.11 (1.06–1.15) p < 0.001 1.26 (1.20–1.31) p < 0.001
over temperature over temperature
1.10 (1.07–1.13) p < 0.001 1.19 (1.14–1.25) p < 0.001
over shade over shade
0.99 (0.95–1.04) p > 0.05 1.08 (1.04–1.12) p < 0.001
over ventilation over ventilation
4. Discussion
4.1. Occupants’ Personal Control

The current study analyzed data from 7441 occupants in 167 European modern office buildings,
sustaining and reinforcing findings from previous studies, revealing the importance of the perceived
personal control over indoor environment parameters.
The overall occupants’ comfort should be examined by investigating the role of every aspect
of personal control. Paciuk [16] highlighted three aspects of personal control; (i) available control,
(ii) exercised control; (iii) perceived control. Available control can be described by the degree and
type of control made available by the environment. It can be defined by ‘the degree of manipulation
of thermostats and other manual controls as well as the existence of operable windows, blinds,
sunshades, ventilation vanes, doors and HVAC system components’. Exercised control is defined by
‘the relative frequency in which occupants engage in several types of thermal-related behaviors in
order to obtain thermal comfort when needed’. The occupants’ perceived control is being produced as
the interaction of the different degrees of available (building controls) and exercised control (frequency
of use). OFFICAIR database does not include information about the occupants’ exercise control so in
that study this aspect was not examined. The main idea is that personal control, both available and
perceived, works as a moderator affecting the occupants’ satisfaction on IEQ parameters and on the
overall comfort.
Paciuk [16] examined the perceived control over temperature, ventilation, shading from the
sun, lighting and noise. The results showed that occupants have moderate or low control on these
parameters with a significant group of occupants, who declared that they had no access to the control
of their environment. This was also remarked by Haghighat and Donnini [19], where almost 60%
of the responders had no control access. Furthermore, around 30% of the participants were very
dissatisfied with the level of control. An interesting relationship between the perceived productivity
in offices and perceived control has emerged, where higher productivity levels were recorded when
occupants perceived higher degrees of personal control. Boardass et al. [18] also pointed out the issue
of productivity versus personal control, concluding in the same finding.
In this analysis, only a limited number of available control types were found to be associated with
the perceived control. Perceived control over temperature, ventilation, light and noise was found to be
associated only with the presence of operable windows. The effect of operable windows in the indoor
environment on occupants’ comfort was also raised by Brager et al. [47]. The scores of perceived
temperature control differ significantly with the different types of solar shading devices. However,
after importing these available control types along with various general building characteristics
in the multilevel model, only operable windows remained significant. Similar findings were also
40
Appl. Sci. 2019, 9, 3227
reported in another study [48] where opening windows were the most frequent behavior of controlling
thermal conditions. The limited effect of the available control on the perceived control was pointed
out in previous studies [23,24], which found no association between available and perceived control.
This could be explained by the fact that occupants could identify the available controls, but they were not
allowed to use them or did not know how to set them up or the control types could not respond rapidly.
The multilevel model, containing both individual and building characteristics (both general and
available control types), showcased 88% variance in perceived control between buildings. A Building’s
location and the distance with adjacent facades from other neighboring buildings seem to be significant.
The area and number of floors were also found to affect occupant control perception. The number of
occupants inside the building showed strong effect, especially inside the office, as well as the type of
office (private or open space with partitions). Gou et al. [49] remarked that the provision of personal
control in open-plan work environments is an important means to alleviating adverse perceptions.
Noise due to conversations and phone calls negatively affected the controlling perception, which was
expected as people cannot control themselves the level of noise. Around one-third of the occupants
reported negative perception of lighting control, which is similar to findings by Moore et al. [50].
Boardass et al. [18] and Toftum [21] also indicated that the perceived degree of controlling parameters
(temperature, ventilation, lighting) decreases with the increased number of occupants inside offices.
To some extent, it seems that occupants perceive controlling capabilities better when providing them
with the feeling of freedom inside the office buildings. Effort-Reward Imbalance (ERI), which is
a critical psychological parameter in office environments, significantly affected control perception
and the occupants’ perceived comfort inside offices [36]. Moreover, occupants who habitually use
air fresheners in their home seem to be more sensitive to adjusting their indoor work environment.
The psychological aspect of perceived control is reinforced, as observed by Luo et al. [25].
A significant positive but weak correlation was observed between perceived control and perceived
comfort. This finding is in compliance with Boerstra et al. [23], Haghighat and Donnini [19] and
Roulet et al. [51]. This work includes a deeper analysis of the appreciation of the perceived control.
It should be noted that the analysis using the combined control parameters is not widely used in this
type of studies. Combined scores of the controlling parameters were positively correlated with IEQ
satisfaction, as Boerstra et al. [23] also found. However, in our study, the combined control perception
affected with a stronger correlation the overall comfort, in contrast to the findings of Boerstra et al. [23].
This could be explained by the fact that occupants with higher combined control perception perceive
overall comfort with more satisfaction. Combined control perception also resulted in occupants being
more satisfied with the overall air quality as well as with the amount of privacy and office layout.
In general, occupants who are more able to adjust their environment feel more satisfied.
Regarding building-related health symptoms, the combined scores unveil higher negative
correlations with the PSI5. Office occupants reported less health symptoms related to buildings
where they perceived higher degree of personal controls.
4.2. Strengths and Limitations

This study has several strengths: a relatively large sample size, a survey performed in different
geographical areas across Europe, and the use of standardized procedures (questionnaire and checklist).
Data on socio-demographics, psycho-social work environment, and perceived environmental quality
were collected by a validated questionnaire. IEQ was assessed using both crude IEQ items (satisfaction
with perceived control over temperature, ventilation, shading from the sun, light and noise, as
well as satisfaction with overall thermal comfort, noise, light, and indoor air quality), and with
14 detailed indoor environmental parameters (e.g., layout, noise within the building, noise from
building systems, noise outside the building, air movement). Some limitations, however, should
be noted. Caution is needed when interpreting results, because data on IEQ were self-reported.
Consequently, a potential (recall) bias cannot be excluded and this type of surveys do not always
capture IEQ issues. A combination of objective and subjective measurements would be useful for
41
Appl. Sci. 2019, 9, 3227
assessing IEQ. Another limitation is the cross-sectional study design. Therefore, no causality of the
identified relations can be confirmed.
5. Conclusions
The following conclusions can be drawn from the present study:
1. In general, occupants have a moderate or low access in the control of their indoor working
environment. They have rated with a moderate control perception the lighting and shading from
the sun parameters. Noise was the worst control parameter, while temperature and ventilation
control were rated in the low control scale. Among the five control parameters, there was a
significant group of occupants who were not able to control their environment at all. Nearly all
occupants have no or low control on noise. Half of the occupants declared no or low control on
ventilation and temperature conditions. Almost one-third of them remarked that they do not
have satisfactory levels of control for lighting and shading from sun conditions. It is noteworthy
that as the occupancy within the offices increases, the degree of personal control becomes lower.
2. Occupants with a higher level of personal control were reported to be more productive within
their working environment. Moreover, occupants who declared high degrees of personal control
reported higher levels of IEQ satisfaction.
3. Some significant correlations were found between the available controls within the building and
the perceived control. The availability of operable windows had a higher impact on the occupants’
control perception over temperature, ventilation, light and noise than floor area or occupancy.
Perceived control over temperature differed significantly with the different types of solar devices.
4. Occupants’ perceived control is related to psychological aspects. General building characteristics,
such as floor number and floor area, office type, operable windows, etc., which help occupants
feel freedom were positively associated with the perceived control. ERI remains a significant
parameter of the controlling perception in office environments.
5. Concerning the impact of perceived control on perceived comfort, the results of the present study
are in general agreement with the literature. More specifically:
• The perceived combined control over all studied parameters is positive on the overall comfort
• The combined control over all parameters seems to have a stronger effect on overall comfort
than the single controls
• Temperature variation seems to be more related to the combined control of temperature and
shading from the sun control
• Combined control perception over temperature, ventilation and shading positively affects
the occupants and gives them the feeling of fresher air and an overall satisfaction with
air quality
• Noise from the buildings systems seems to be more affected by the combined control
perception of all parameters rather than single controls
• The combined perceived control of all parameters was found to affect the occupants’ privacy,
office layout and decoration satisfaction more
6. Regarding reported health symptoms, office occupants with a higher degree of personal controls
reported less building-related health symptoms.
Author Contributions: Conceptualization, I.S., D.S., C.M. and J.B.; data curation, I.S.; formal analysis, I.S.; funding
acquisition, J.B.; investigation, I.S., D.S., C.M., Y.d.K., S.F., A.S., A.C., T.S. and V.M.; methodology, I.S., C.M., Y.d.K.
and J.B.; project administration, J.B.; resources, K.K.; supervision, D.S. and J.B.; visualization, I.S.; writing-original
draft, I.S.; writing-review & editing, D.S., C.M., Y.d.K., S.F., A.S., A.C., T.S., V.M., E.d.O.F., K.K., P.C. and J.B.
Funding: This research was funded by the European Union 7th 427 Framework, grant number Agreement 265267
under Theme: ENV.2010.1.2.2-1.
42
Appl. Sci. 2019, 9, 3227
Acknowledgments: This work was supported by the project “OFFICAIR” (On the Reduction of Health Effects
from Combined Exposure to Indoor Air Pollutants in Modern Offices).
References
1. Huang, Y.-H.; Robertson, M.M.; Chang, K.-I. The Role of Environmental Control on Environmental Satisfaction,
Communication, and Psychological Stress: Effects of Office Ergonomics Training. Environ. Behav. 2004, 36,
617–637. [CrossRef]
2. Gee, G.C.; Payne-Sturges, D.C. Environmental Health Disparities: A Framework Integrating Psychosocial
and Environmental Concepts. Environ. Health Perspect. 2004, 112, 1645–1653. [CrossRef]
3. Kalimeri, K.K.; Saraga, D.E.; Lazaridis, V.D.; Legkas, N.A.; Missia, D.A.; Tolis, E.I.; Bartzis, J.G. Indoor air
quality investigation of the school environment and estimated health risks: Two-season measurements in
primary schools in Kozani, Greece. Atmos. Pollut. Res. 2016, 7, 1128–1142. [CrossRef]
4. Lee, S.Y.; Brand, J.L. Effects of control over office workspace on perceptions of the work environment and
work outcomes. J. Environ. Psychol. 2005, 25, 323–333. [CrossRef]
5. Huizenga, C.; Abbaszadeh, S.; Zagreus, L.; Arens, E.A. Air quality and thermal comfort in office buildings:
Results of a large indoor environmental quality survey. Proceeding Healthy Build. 2006, 3, 393–397.
6. Brager, G.; Baker, L. Occupant satisfaction in mixed-mode buildings. Build. Res. Inf. 2009, 37, 369–380.
[CrossRef]
7. Langevin, J.; Wen, J.; Gurian, P.L. Relating occupant perceived control and thermal comfort: Statistical
analysis on the ASHRAE RP-884 database. HVAC&R Res. 2012, 18, 179–194.
8. Vine, E.; Lee, E.; Clear, R.; DiBartolomeo, D.; Selkowitz, S. Office worker response to an automated Venetian
blind and electric lighting system: A pilot study. Energy Build. 1998, 28, 205–218. [CrossRef]
9. Lee, S.Y.; Brand, J.L. Can personal control over the physical environment ease distractions in office workplaces?
Ergonomics 2010, 53, 324–335. [CrossRef]
10. Samani, S.A. The Impact of Personal Control over Office Workspace on Environmental Satisfaction and
Performance. J. Soc. Sci. Humanit. 2015, 1, 163–172.
11. Xuan, X. Study of indoor environmental quality and occupant overall comfort and productivity in LEED-
and non-LEED–certified healthcare settings. Indoor Built Environ. 2018, 27, 544–560. [CrossRef]
12. Yun, G.Y. Influences of perceived control on thermal comfort and energy use in buildings. Energy Build. 2018,
158, 822–830. [CrossRef]
13. Zhao, Z.; Amasyali, K.; Chamoun, R.; El-Gohary, N. Occupants’ Perceptions about Indoor Environment
Comfort and Energy Related Values in Commercial and Residential Buildings. Procedia Environ. Sci. 2016,
34, 631–640. [CrossRef]
14. Al horr, Y.; Arif, M.; Katafygiotou, M.; Mazroei, A.; Kaushik, A.; Elsarrag, E. Impact of indoor environmental
quality on occupant well-being and comfort: A review of the literature. Int. J. Sustain. Built Environ. 2016, 5,
1–11. [CrossRef]
15. Frontczak, M.; Schiavon, S.; Goins, J.; Arens, E.; Zhang, H.; Wargocki, P. Quantitative relationships between
occupant satisfaction and satisfaction aspects of indoor environmental quality and building design. Indoor Air
2012, 22, 119–131. [CrossRef]
16. Paciuk, M. The Role of Personal Control of the Environment in Thermal Comfort and Satisfaction at the Workplace;
University of Wisconsin-Milwaukee: Milwaukee, WI, USA, 1989.
17. Collins, B.L.; Fisher, W.; Gillette, G.; Marans, R.W. Second-Level Post-Occupancy Evaluation Analysis. J. Illum.
Eng. Soc. 1990, 19, 21–44. [CrossRef]
18. Bordass, B.; Bromley, K.; Leaman, A. User and Occupant Controls in Office Buildings; BRE: Brussels, Belgium,
1993; p. 13.
19. Haghighat, F.; Donnini, G. Impact of psycho-social factors on perception of the indoor air environment
studies in 12 office buildings. Build. Environ. 1999, 34, 479–503. [CrossRef]
20. Zweers, T.; Preller, L.; Brunekreef, B.; Boleij, J.S.M. Health and Indoor Climate Complaints of 7043 office
Workers in 61 Buildings in the Netherlands. Indoor Air 1992, 2, 127–136. [CrossRef]
21. Toftum, J. Central automatic control or distributed occupant control for better indoor environment quality in
the future. Build. Environ. 2010, 45, 23–28. [CrossRef]
43
Appl. Sci. 2019, 9, 3227
22. Zagreus, L.; Huizenga, C.; Arens, E.; Lehrer, D. Listening to the occupants: A Web-based indoor environmental
quality survey. Indoor Air 2004, 14, 65–74. [CrossRef]
23. Boerstra, A.; Beuker, T.; Loomans, M.; Hensen, J. Impact of available and perceived control on comfort and
health in European offices. Archit. Sci. Rev. 2013, 56, 30–41. [CrossRef]
24. Kwon, M.; Remøy, H.; van den Dobbelsteen, A.; Knaack, U. Personal control and environmental user
satisfaction in office buildings: Results of case studies in the Netherlands. Build. Environ. 2019, 149, 428–435.
[CrossRef]
25. Luo, M.; Cao, B.; Ji, W.; Ouyang, Q.; Lin, B.; Zhu, Y. The underlying linkage between personal control and
thermal comfort: Psychological or physical effects? Energy Build. 2016, 111, 56–63. [CrossRef]
26. Karjalainen, S. Thermal comfort and use of thermostats in Finnish homes and offices. Build. Environ. 2009,
44, 1237–1245. [CrossRef]
27. De Dear, R.; Brager, G.S. Developing an Adaptive Model of Thermal Comfort and Preference. ASHRAE Trans.
1998, 104, 1–18.
28. ASHRAE Standard 55 Thermal Environmental Conditions for Human Occupancy; ANSI/ASHRAE: Atlanta, GA,
USA, 2013.
29. International Organization for Standardization ISO 7730:2005. Available online: http://www.iso.org/cms/
render/live/en/sites/isoorg/contents/data/standard/03/91/39155.html (accessed on 26 July 2019).
30. Olesen, B.W.; Parsons, K.C. Introduction to thermal comfort standards and to the proposed new version of
EN ISO 7730. Energy Build. 2002, 34, 537–548. [CrossRef]
31. OFFICAIR Project. Available online: http://www.officair-project.eu/ (accessed on 6 August 2019).
32. Nørgaard, A.W.; Kofoed-Sørensen, V.; Mandin, C.; Ventura, G.; Mabilia, R.; Perreca, E.; Cattaneo, A.;
Spinazzè, A.; Mihucz, V.G.; Szigeti, T.; et al. Ozone-initiated Terpene Reaction Products in Five European
Offices: Replacement of a Floor Cleaning Agent. Environ. Sci. Technol. 2014, 48, 13331–13339. [CrossRef]
33. Mihucz, V.G.; Szigeti, T.; Dunster, C.; Giannoni, M.; de Kluizenaar, Y.; Cattaneo, A.; Mandin, C.; Bartzis, J.G.;
Lucarelli, F.; Kelly, F.J.; et al. An integrated approach for the chemical characterization and oxidative potential
assessment of indoor PM2.5. Microchem. J. 2015, 119, 22–29. [CrossRef]
34. Bluyssen, P.M.; Roda, C.; Mandin, C.; Fossati, S.; Carrer, P.; de Kluizenaar, Y.; Mihucz, V.G.; de Oliveira
Fernandes, E.; Bartzis, J. Self-reported health and comfort in ‘modern’ office buildings: First results from the
European OFFICAIR study. Indoor Air 2016, 26, 298–317. [CrossRef]
35. De Kluizenaar, Y.; Roda, C.; Dijkstra, N.E.; Fossati, S.; Mandin, C.; Mihucz, V.G.; Hänninen, O.; de Oliveira
Fernandes, E.; Silva, G.V.; Carrer, P.; et al. Office characteristics and dry eye complaints in European
workers-The OFFICAIR study. Build. Environ. 2016, 102, 54–63. [CrossRef]
36. Sakellaris, I.; Saraga, D.; Mandin, C.; Roda, C.; Fossati, S.; de Kluizenaar, Y.; Carrer, P.; Dimitroulopoulou, S.;
Mihucz, V.; Szigeti, T.; et al. Perceived Indoor Environment and Occupants’ Comfort in European “Modern”
Office Buildings: The OFFICAIR Study. Int. J. Environ. Res. Public Health 2016, 13, 444. [CrossRef]
37. Szigeti, T.; Dunster, C.; Cattaneo, A.; Cavallo, D.; Spinazzè, A.; Saraga, D.E.; Sakellaris, I.A.; de Kluizenaar, Y.;
Cornelissen, E.J.M.; Hänninen, O.; et al. Oxidative potential and chemical composition of PM2.5 in office
buildings across Europe—The OFFICAIR study. Environ. Int. 2016, 92–93, 324–333. [CrossRef]
38. Campagnolo, D.; Saraga, D.E.; Cattaneo, A.; Spinazzè, A.; Mandin, C.; Mabilia, R.; Perreca, E.; Sakellaris, I.;
Canha, N.; Mihucz, V.G.; et al. VOCs and aldehydes source identification in European office buildings—The
OFFICAIR study. Build. Environ. 2017, 115, 18–24. [CrossRef]
39. Mandin, C.; Trantallidi, M.; Cattaneo, A.; Canha, N.; Mihucz, V.G.; Szigeti, T.; Mabilia, R.; Perreca, E.;
Spinazzè, A.; Fossati, S.; et al. Assessment of indoor air quality in office buildings across Europe—The
OFFICAIR study. Sci. Total Environ. 2017, 579, 169–178. [CrossRef]
40. Szigeti, T.; Dunster, C.; Cattaneo, A.; Spinazzè, A.; Mandin, C.; Le Ponner, E.; de Oliveira Fernandes, E.;
Ventura, G.; Saraga, D.E.; Sakellaris, I.A.; et al. Spatial and temporal variation of particulate matter
characteristics within office buildings—The OFFICAIR study. Sci. Total Environ. 2017, 587–588, 59–67.
[CrossRef]
41. Raw, G.J.; Roys, M.S.; Whitehead, C.; Tong, D. Questionnaire design for sick building syndrome: An empirical
comparison of options. Environ. Int. 1996, 22, 61–72. [CrossRef]
42. Theodorsson-Norheim, E. Kruskal-Wallis test: BASIC computer program to perform nonparametric one-way
analysis of variance and multiple comparisons on ranks of several independent samples. Comput. Methods
Programs Biomed. 1986, 23, 57–62. [CrossRef]
44
Appl. Sci. 2019, 9, 3227
43. Gelman, A.; Hill, J. Data Analysis Using Regression and Multilevel/Hierarchical Models; Cambridge University
Press: Cambridge, UK, 2007; ISBN 978-0-521-68689-1.
44. Goldstein, H. Multilevel Statistical Models; John Wiley & Sons: West Sussex, UK, 2011; ISBN 978-1-119-95682-2.
45. Mukaka, M. A guide to appropriate use of Correlation coefficient in medical research. Malawi Med. J. 2012,
24, 69–71.
46. IBM SPSS Statistics for Windows; Version 22.0; IBM: Armonk, NY, USA, 2013.
47. Brager, G.; Paliaga, G.; de Dear, R.; Olesen, B.; Wen, J.; Nicol, F.; Humphreys, M. Operable Windows, Personal
Control, and Occupant Comfort. ASHRAE Trans. 2004, 110, 17–35.
48. Raja, I.A.; Nicol, J.F.; McCartney, K.J.; Humphreys, M.A. Thermal comfort: Use of controls in naturally
ventilated buildings. Energy Build. 2001, 33, 235–244. [CrossRef]
49. Gou, Z.; Zhang, J.; Shutter, L. The Role of Personal Control in Alleviating Negative Perceptions in the
Open-Plan Workplace. Buildings 2018, 8, 110. [CrossRef]
50. Moore, T.; Carter, D.; Slater, A. User attitudes toward occupant controlled office lighting. Light. Res. Technol.
2002, 34, 207–216. [CrossRef]
51. Roulet, C.-A.; Johner, N.; Foradini, F.; Bluyssen, P.; Cox, C.; De Oliveira Fernandes, E.; Müller, B.; Aizlewood, C.
Perceived health and comfort in relation to energy use and building characteristics. Build. Res. Inf. 2006, 34,
45
applied
sciences
Article
An Investigation of the Effects of Changes in the
Indoor Ambient Temperature on Arousal Level,
Thermal Comfort, and Physiological Indices †
Jongseong Gwak 1, *, Motoki Shino 2 , Kazutaka Ueda 3 and Minoru Kamata 2
1 Institute of Industrial Science, The University of Tokyo, Tokyo 153-8505, Japan
2 Department of Human and Engineered Environment Studies, Graduate School of Frontier Sciences,
The University of Tokyo, Chiba 277-8563, Japan; [email protected] (M.S.);
[email protected] (M.K.)
3 Department of Mechanical Engineering, The University of Tokyo, Tokyo 113-8656, Japan;
[email protected]
† This paper is an extended version of paper published in the 2015 IEEE International Conference on Systems,
Man, and Cybernetics, held in Hong Kong, 9–12 October 2015.
Received: 30 January 2019; Accepted: 26 February 2019; Published: 3 March 2019
Abstract: Thermal factors not only affect the thermal comfort sensation of occupants, but also
affect their arousal level, productivity, and health. Therefore, it is necessary to control thermal
factors appropriately. In this study, we aim to design a thermal environment that improves both the
arousal level and thermal comfort of the occupants. To this end, we investigated the relationships
between the physiological indices, subjective evaluation values, and task performance under several
conditions of changes in the indoor ambient temperature. In particular, we asked subjects to
perform a mathematical task and subjective evaluation related to their thermal comfort sensation
and drowsiness levels. Simultaneously, we measured their physiological parameters, such as
skin temperature, respiration rate, electroencephalography, and electrocardiography, continuously.
We investigated the relationship between the comfort sensation and drowsiness level of occupants,
and the physiological indices. From the results, it was confirmed that changes in the indoor ambient
temperature can improve both the thermal comfort and the arousal levels of occupants. Moreover,
we proposed the evaluation indices of the thermal comfort and the drowsiness level of occupants
using physiological indices.
Keywords: thermal comfort; arousal level; physiological indices; electroencephalography;

electrocardiography
1. Introduction
The control of indoor environmental quality (IEQ, which consists of visual elements, olfactory
elements, and thermal factors, etc.) is important for improvements of the comfort and productivity
of occupants. Among the factors of IEQ, thermal factors, such as ambient temperature, radiant
temperature, humidity, and air velocity, are especially related with not only the thermal sensation and
thermal comfort of occupants, but also productivity and health. With the high technology of heating,
ventilation, and air conditioning systems (HVAC systems), the relationship between thermal factors
and thermal sensation and thermal comfort has received attention from several researchers worldwide.
In previous research, indices of thermal comfort, such as the predicted mean vote (PMV), predicted
percent dissatisfied (PPD), and standard new effective temperature (SET*), have been proposed
based on the relationship between thermal factors and the subjective evaluation of thermal comfort

Appl. Sci. 2019, 9, 899
sensation [1,2]. Many studies have analyzed the indoor thermal quality based on the evaluation of
thermal factors using these indices [3–7].
In addition, improvements of the productivity of occupants who are working, studying, and
driving in indoor spaces, such as offices, classrooms, and vehicles, are needed. It is necessary to
maintain a high arousal state of the occupants to improve their productivity. In previous studies,
it was shown that productivity improved when the arousal level of occupants was high, and this
was related to the indoor ambient temperature that occupants felt as cool or cold [8,9]. From the
results of previous studies, the conditions of the ambient temperature to improve task performance
were different from the conditions of the ASHRAE (American Society of Heating, Refrigerating and
Air-Conditioning Engineers)’s thermal comfort zone [10]. Therefore, the conditions of the constant
ambient temperature were not appropriate to improve both the task performance and thermal comfort
levels. In previous studies, it was shown that the thermal comfort of occupants was immediately
changed in accordance with changes in the indoor ambient temperature [11]. The arousal level was
increased by outer stimulation and was maintained at a high state for several minutes [12,13]. Based on
these results of previous studies, we hypothesized that thermal stimulation due to cooling can improve
the arousal level of occupants. Furthermore, thermal comfort can be improved while maintaining
high arousal levels due to the removal of thermal stimulation. When considering the possibility that
changes in thermal factors can improve both arousal and thermal comfort levels, the physiological
effects associated with such changes in thermal factors, and how they affect the arousal level and
thermal comfort of occupants are not clear. Therefore, the design requirements for changes in ambient
temperature to improve both the arousal level and thermal comfort are also not clear. Therefore,
a continuous and quantitative evaluation of the thermal comfort and arousal level of occupants using
indices, which can be measured both continuously and quantitatively, such as physiological signals,
is needed to clarify the design requirements of changes in the indoor ambient temperature to improve
both the arousal level and thermal comfort.
In this study, we aimed to investigate the characteristics of changes in the arousal level and
feelings of thermal comfort of occupants, and the relationship between them when thermal factors
are changed. There are many factors that affect the thermal comfort and arousal level of occupants,
such as the ambient temperature, indoor air velocity, mean air radiant temperature, and metabolic
activity [5–7]. Especially, we focused on the changes in the indoor ambient temperature as a
fundamental investigation in this study. In addition, to propose evaluation indices that can evaluate
the thermal comfort and arousal level of occupants continuously and quantitatively, we investigated
the relationship between the arousal level, feelings of thermal comfort, and physiological indices,
which can be measured continuously and quantitatively.
2. Strategy
2.1. Investigation of the Effects of Changes in the Indoor Ambient Temperature on Arousal Levels, Thermal
Comfort, and Task Performance
To verify the hypothesis that thermal stimulation, due to cooling, can improve the arousal level of
occupants, and that thermal comfort can be improved while maintaining high arousal levels due to the
removal of thermal stimulation, several thermal conditions were set. Subjects were asked to conduct a
mathematical task and periodically evaluate their sensation values (using a subjective sensation vote)
in relation to their arousal level and thermal comfort. We attempt to clarify the effect of changes in the
indoor ambient temperature on the arousal level, thermal comfort, and task performance of occupants
by conducting an analysis of the results of the subjective sensation vote and task performance.
2.2. Investigation of the Relation between the Subjective Evaluation Value and Physiological Parameters, and a
Recommendation for Evaluation Indices
Indices that can evaluate the arousal and thermal comfort levels of occupants continuously and
quantitatively are necessary to clarify the design requirements of a temperature control that can
48
Appl. Sci. 2019, 9, 899
improve the arousal and thermal comfort levels of occupants. However, as it was not possible to
perform a subjective sensation voting continuously, we considered clarifying the design requirements
for thermal environments using changes over time. This relied on the possibility to evaluate the
arousal level and feelings of thermal comfort using physiological parameters that could be measured
quantitatively and continuously. Previous studies have shown that the physiological indices measured
utilizing electroencephalograms (EEG), electrocardiograms (ECG), and respiration rates were effective
for the evaluation of arousal levels [14,15], and the skin temperature, EEGs, and ECGs of occupants
were effective for the evaluation of the thermal comfort [16,17]. Therefore, in this study, we assumed a
flow of physiological responses when the occupant was stimulated by thermal factors, as shown
in Figure 1, based on the above-mentioned previous studies to clarify the characteristics of the
physiological parameters under the condition of changes in both the arousal level and thermal comfort.
A thermal stimulation is transmitted from the sensory organ, such as warm spots and cold spots of
skin, to the central nervous system, which affects thermal and comfort sensation. This change in
comfort sensation then affects the autonomic nervous system, and the response is then conducted
through the locomotive organ. Based on this process, we selected the skin temperature as an indicator
of heat transfer between the environment and the human body, EEGs as an indicator of the reaction of
the central nervous system, and ECGs as indicators of the reaction of the autonomic nervous system.
Next, we attempted to search for indices that could separately evaluate the arousal level and thermal
comfort by performing a multiple regression analysis utilizing the physiological parameters and
subjective evaluation values of the arousal level and feelings of thermal comfort. In Figure 1, the flow
of physiological responses with the thermal stimulus is shown.
Figure 1. Flow of physiological responses with the thermal stimulus.
3. Methodology
3.1. Subjects
Ten subjects with vital statistics of a height of 173.5 ± 4.4 cm, weight of 61 ± 4.8 kg, age of 22.1 ±
1.2 years old, and who were right-handed participated in the experiment. The experimental contents
and procedures, which were approved by the ethics committee of the University of Tokyo, were
explained to the subjects before conducting the experiment. The subjects were then asked to avoid
intense physical activity, alcohol, and caffeine for 24 h prior to the experimental session.
3.2. Experimental Task

The subjects were asked to conduct mental arithmetic tasks known as “MATH”, which is based on
the algorithm proposed by Tuner et al. [18]. A 1–3-digit addition or subtraction question is displayed
for 2 s on the monitor, and then “equals to” is displayed for 1.5 s. Lastly, the answer is displayed for 1
s, and the next question is displayed after 0.5 s. Subjects had to determine if the answer was correct or
incorrect when the answer was displayed and click the left mouse button if the answer was correct
or the right mouse button if the answer was incorrect. The levels of questions in the original version
of MATH consisted of 1–5 levels. In this study, the beginning level was 3 (which is 2-digit addition
or subtraction), and the level of the next question was raised if the responses of the subjects were
correct, and was reduced if answers were incorrect. We deduced that changing the levels affected
49
Appl. Sci. 2019, 9, 899
the physiological indices. Thus, we did not change the level of the question, and fixed it to level 3.
The MATH task included 50 questions for 250 seconds per set.
3.3. Experimental Conditions and Experimental Procedure

To evaluate the effects of the duration and degree of thermal stimulation, three environmental
conditions (A–C) were set as follows.
• Condition A: The indoor ambient temperature was maintained at 27 ◦ C.

• Condition B: The indoor ambient temperature was decreased from 27 ◦ C to 20 ◦ C, and then
increased from 20 ◦ C to 27 ◦ C.
• Condition C: The indoor ambient temperature was decreased from 27 ◦ C to 20 ◦ C, and then
maintained at 20 ◦ C.
We set condition A as a thermally comfort condition [10], and condition C as a high arousal
condition [8,9]. Condition B was set to verify the hypothesis that thermal stimulation due to cooling
can improve the arousal levels of occupants, after which thermal comfort can be improved while
maintaining high arousal levels due to the removal of thermal stimulation. There are many parameters
which affect the thermal comfort and arousal levels of occupants, such as the indoor velocity, mean air
radiant temperature, metabolic activity, and amount of clothing. In this study, we focused on only the
indoor ambient temperature as a fundamental investigation, thus those parameters except the indoor
ambient temperature were controlled in the experiment. The subjects wore short sleeves and short
pants and were asked to remain in the pre-room. The room temperature was set at 27 ± 0.5 ◦ C for
approximately 1 h so that subjects could adjust to the thermal environment; they were asked to practice
the “MATH” at least twice during this time. Sensors were then attached to the bodies of the subjects to
measure their physiological indices. One set of tasks consisted of completing the subjective sensation
vote and the “MATH” task. The subjects were asked to perform seven sets of tasks at time intervals of
10 min and to rest between each task for approximately 3 min. Physiological indices were measured
from the start time (0 min) till the end time (70 min). The experimental procedure and environmental
conditions are shown in Figure 2.
Figure 2. Experimental conditions and environmental procedure.
3.4. Measurement
3.4.1. Subjective Evaluation of the Drowsiness Level and the Thermal Comfort Sensation
Before completing the “MATH” task, the subjects were asked to complete a questionnaire related
to their thermal sensation, thermal comfort sensation, and arousal level. The scale of thermal sensation
was based on ASHRAE/ISO (International Organization for Standardization) [10], and was denoted
using integral numbers from −3 to 3 (where −3, −2, −1, 0, 1, 2, and 3 are the meanings of cold, cool,
slightly cool, neutral, slightly warm, warm, and hot, respectively.). The scale of comfort sensation
was based on ISO10551 [19], and was denoted using integral numbers from −3 to 0 (where −3,
50
Appl. Sci. 2019, 9, 899
−2, −1, and 0 are the meanings of very uncomfortable, uncomfortable, slightly uncomfortable, and
comfortable, respectively.). The scale of the arousal level was based on the drowsiness level of Zilberg’s
indicators [20], and was denoted using integral numbers from 0 to 4 (where 0, 1, 2, 3, and 4 are the
meanings of alert, slightly drowsy, moderately drowsy, significantly drowsy, and extremely drowsy,
respectively.).
3.4.2. Physiological Indices
(a) EEG
EEGs were recorded using an EEG-measuring instrument (EEG-1200, Nihonkohden Co., Japan) at
a sampling rate of 500 Hz. EEG electrodes were attached on 16 channels (based on the internationally
accepted 10–20 system, Fp1, Fp2, F7, F3, F4, F8, T7, C3, C4, T8, P7, P3, P4, P8, O1, O2). Next, raw data
were processed using the Fourier transform method, and the spectral power of each frequency band,
such as the content of theta wave (4–8 Hz), low-alpha wave (8–10 Hz), high-alpha wave (10–13 Hz),
low-beta wave (13–20 Hz), high-beta wave (20–30 Hz), and SMR (12–15 Hz) bands, was calculated in
addition to the values of the beta per alpha and alpha per high-beta for each channel.
(b) ECG
ECGs were recorded using an ECG-measuring instrument (WEB-7000 and ECG picker,
Nihonkohden Co., Japan) at a sampling rate of 1000 Hz. Three electrodes were attached to the
chests of subjects using the precordial leads method. The R-R interval (RRI) was calculated from the
ECG waveform using MATLAB (Mathwork Co.) programs. The values of the mean of the RRI and the
coefficient of variance of RRI [CVRR (100*SD/Mean of RRI)] were calculated from the RRI data. In
addition, the spectral power of each frequency band, such as the very low frequency (VLF, 0.001–0.04
Hz), low frequency (LF, 0.04–0.15 Hz), and high frequency (HF, 0.15–0.45 Hz) bands, was calculated
from the time series of the RRI using the fast Fourier transform method.
(c) Respiration
A thermal picker (WEB-7000, Nihonkohden Co., Japan) was used to measure the temperature of
the breath of subjects. The peak values in the time series graph of the temperature were detected, and
the mean and standard deviation of the respiratory cycle time were calculated.
(d) Skin Temperature

Thermocouples were attached at 7 places on the bodies of subjects based on the Hardy–Du
Bois method [16] to measure the skin temperature (LT8, GRAM Co., Japan). Finally, the mean skin
temperature was calculated using following Equation (1):
MST = 0.07 T1 + 0.14 T2 + 0.05 T3 + 0.35 T4 + 0.19 T5 + 0.13 T6 + 0.07 T7 (1)
where MST is the mean skin temperature based on the Hardy-DuBois method [16], and T1 , T2 , T3 ,
T4 , T5 , T6 , and T7 are the temperatures of the forehead, forearms, hands, abdomen, thighs, legs, and
feet, respectively.
3.4.3. Facial Expression and Task Performance

The activities of subjects were recorded using the video camera, and the drowsiness levels were
recorded by an observer at intervals of 10 s based on Zilberg’s criteria [20]. After the experiment
was completed, the mean of the data recorded when subjects were performing the MATH task was
calculated. The task performance of MATH included the number of correct answers and the mean of
the reaction time taken to solve 50 questions. Figure 3 shows a participant completing the experiment.
Figure 3 shows photos of the field of experimental scene.
51
Appl. Sci. 2019, 9, 899
Figure 3. Field photos of the experimental scene.
4. Results and Discussion
4.1. Characteristics of the Arousal Level and Thermal Comfort Corresponding to Changes in the Indoor Ambient
Temperature
4.1.1. Subjective Evaluation Value

The results of the subjective evaluation value corresponding to the drowsiness levels and the
thermal comfort sensation vote of subjects under conditions A, B, and C are shown in Figures 4 and 5.
We used multiple comparison based on the Bonferroni method to investigate significant differences
of the subjective evaluation value due to changes in the indoor ambient temperature. There was a
small change in the comfort sensation vote, but the drowsiness level increased under condition A,
where the indoor temperature was maintained at 27 ◦ C. Under condition B, the comfort sensation vote
decreased to an uncomfortable state when the indoor ambient temperature dropped, and increased to
a comfortable state when the indoor ambient temperature increased. Under condition C, the comfort
sensation vote decreased to an uncomfortable state when the indoor temperature decreased and
was then maintained in this uncomfortable state. Under conditions B and C, the drowsiness levels
decreased and were maintained at an alert state, corresponding to the drop in the indoor temperature.
After the completion of set 4, the comfort sensation vote increased, corresponding to the increase
in the indoor temperature under condition B, but was maintained at an uncomfortable state under
condition C. The drowsiness level was maintained at a low state under both conditions. According
to this result, the arousal level was maintained at a high state even when the temperature increased,
and subjects felt comfortable even when the temperature dropped, leading to an increased arousal
level. This suggests that there is not always a dependence relationship between the arousal levels and
feelings of thermal comfort. Figures 4 and 5 show that when condition B is applied in sets 6 and 7, it
becomes possible to improve both the arousal level and thermal comfort of occupants by changing the
indoor ambient temperature. These results show that the hypothesis that thermal stimulation due to
cooling can improve the arousal level of occupants, after which thermal comfort can also be improved
while maintaining a high arousal levels due to the removal of thermal stimulation was verified for
condition B. Figure 4 shows the result of the thermal comfort sensation vote. Figure 5 shows the result
of the drowsiness level sensation vote.
52
Appl. Sci. 2019, 9, 899
Figure 4. Results of the value of the thermal comfort sensation vote.
Figure 5. Results of the value of the drowsiness level sensation vote.
4.1.2. Rating Value of Zilberg’s Drowsiness Level by Observer

The rating values of Zilberg’s drowsiness level for each condition are shown in Figure 6, and it is
evident from the figure that the results are similar to those of the subjective sensation vote. Multiple
comparison based on the Bonferroni method was conducted to investigate significant differences of
the drowsiness level due to the changes in the indoor ambient temperature. After the completion of
set 3, the drowsiness levels under conditions B and C decreased significantly compared to that under
condition A. From the result of condition B, it was confirmed that the drowsiness level was maintained
at a low state after the completion of set 3.
Figure 6. Drowsiness level reached when calculating MATH (analysis of facial expressions based on
Zilberg’s method [20]).
4.1.3. Task Performance of MATH

The results of the MATH score under each condition are shown in Figure 7, and the results of the
response time taken to calculate MATH questions under each condition are shown in Figure 8. Multiple
53
Appl. Sci. 2019, 9, 899
comparison based on the Bonferroni method was conducted to investigate significant differences in
the task performance due to changes in the indoor ambient temperature. It can be seen from Figures 7
and 8 that the MATH score was high after set 4, and the response time taken to calculate MATH
questions increases in the order of conditions C, B, and then A. Differences were observed between the
conditions in the results of the subjective sensation vote and the rating value of Zilberg’s drowsiness
level. However, significant differences were observed only between the conditions in the result of
the response time taken to calculate MATH for set 4, and the result of the MATH score for set 5. It
was assumed that the MATH work was easy enough for the subjects to perform, even at a low state
of arousal. Figure 6 shows the drowsiness level reached when calculating MATH (analysis of facial
expressions based on Zilberg’s method [20]). Figures 7 and 8 show the result of the MATH score and
the response time from the display of the MATH question to the participant’s click.
Figure 7. Result of MATH score.
Figure 8. Result of response time.
4.1.4. Consideration of Thermal Stimulation for High Arousal Levels

In the previous studies of Yoshida et al. [12] and Mohri et al. [13], to maintain a high arousal
level, the effects of several stimulation methods, such as the use of a fragrance, alarm, and shoulder
oscillation, were confirmed. As a result, the sleeping rebound occurred 10 min after the stimulation in
those studies. As in our results, high arousal states were maintained for over 10 min after the removal
of thermal stimulation. Though further investigation of these comparisons of the time to rebound is
necessary, our results showed that thermal stimulation is effective in improving the arousal levels of
occupants. Especially, our method is useful in spaces where air conditioning systems are installed,
because the installation of additional devices in the space to make stimulations, such as fragrances,
alarms, and shoulder oscillation, is unnecessary.
54
Appl. Sci. 2019, 9, 899
4.2. Relationship between the Subjective Evaluation Value and Physiological Indices
Three hundred and twenty-nine physiological indices were calculated by using the data of
the EEGs, ECGs, respiration rate, and mean skin temperature in the experiment (320 indices from
EEGs, 7 indices from ECGs, 2 indices from respiration rate, and 1 index from skin temperature). By
considering the result of the subjective evaluation value as the standard for the perception of state,
we used correlative analysis to investigate the relationship between the subjective evaluation value
and the physiological indices. According to the flow of the physiological response shown in Figure 1,
changes in both the arousal level and thermal comfort affect the physiological response. To find
indices (out of the 329 physiological parameters) that corresponded to the arousal level, we used data
from sections when the arousal level changed, but the value of the thermal comfort sensation was
maintained constantly. On the other hand, to find indices that corresponded to thermal comfort, we
used data from sections when the value of the thermal comfort sensation changed and the arousal
level was maintained constantly.
4.2.1. Relationship between the Drowsiness Level and Physiological Indices

As a result of the correlation analysis, the correlation coefficients between values of the drowsiness
level vote and physiological indices were calculated using data from 57 sets in which the value of the
drowsiness level vote varied, but the value of the comfort sensation vote remained at 0 (comfortable).
There was a significant correlation between the drowsiness level and 102 physiological indices (data
N = 57, r > 0.339, p < 0.01). The indices in relation to the ECGs were not included in these parameters;
however, the indices in relation to the EEGs were included. Thus, this result showed that the
relationship between the reaction of the central nervous system and changes in arousal levels is
more significant than that between the reaction of the autonomic nervous system and changes in
arousal levels.
4.2.2. Relationship between Thermal Comfort and Physiological Indices

As a result of the correlation analysis, the correlation coefficients between the values of the comfort
sensation vote and physiological indices were calculated using the data from 98 sets in which the value
of the comfort sensation vote varied, but the value of the drowsiness level vote was between 0 (alert)
and 1 (slightly drowsy). There was a significant correlation between the value of the comfort sensation
vote and seven physiological indices in relation to skin temperature, EEGs, and ECGs (data N = 98,
r > 0.259, p < 0.01). The change in the mean skin temperature is related to thermal stimulation from
the outer environment, which is transmitted to the central nervous system and perceived as thermal
comfort, thus affecting the parameters of the EEGs. It was observed that changes in the thermal
comfort affected the autonomic nervous system, which in turn affected the indices in relation to the
ECGs. It appears that this process affected the result of the correlation analysis in this experiment.
4.2.3. Calculation of the Evaluation Index Using Multiple Regression Analysis

To propose an index that evaluates the changes in each arousal level and thermal comfort states,
multiple regression analysis was performed using the physiological indices that have a significant
correlation with the value of subjective evaluation. It was found that 102 parameters had a significant
correlation with the drowsiness level vote, and seven parameters with the comfort sensation vote.
These were thus considered as explanatory variables. Because an explanatory variable should have a
high correlation with the subjective sensation vote, to obtain a high multiple regression coefficient, we
selected an explanatory variable for use in the multiple regression analysis using the following process.
• Indices that had a significant correlation with the value of subjective evaluation were sorted
according to their correlation coefficient from high to low (x1, x2, . . . x102).
• x1 was selected as the explanatory variable, since it had the highest correlation coefficient with
the value of subjective evaluation.
55
Appl. Sci. 2019, 9, 899
• x2 was selected as the explanatory variable if there was no significant correlation between x2
and x1.
• xn was selected as the explanatory variable if there was no significant correlation between xn and
all the parameters selected previously as explanatory variables.
(a) Evaluation Index indicating the Arousal Level

The result of the multiple regression analysis 1 (Yd) is shown in Table 1, and the regression
equation obtained from the result is expressed in Equation (2). The variables of the regression equation
include the indices of the EEGs. As a result, the coefficient of determination, R2 , in relation to Yd
was 0.750.
Yd = −3.31 + 0.134 Xd1 + 2.520 Xd2 + 0.088 Xd3 − 0.268 Xd4 + 0.271 Xd5 (2)
where Xd1 , Xd2 , Xd3 , Xd4 , and Xd5 are the high alpha content of T7, beta per alpha content of F7, beta
content of F3, low beta content of F7, and alpha content of Fp1, respectively.
Table 1. Result of multiple regression 1 (Yd).
Model Summary
R 0.8657 Std. Error 0.7358
R2 0.7495 data N 57
Adjusted R2 0.7249
Index Coefficient Std.Error t p-value
Intercept −3.3123 0.4959 −4.6631 2.28 × 10−5
T7_High-alpha 0.1338 0.0201 6.6478 1.96 × 10−8
F7_Beta/Alpha 2.5196 0.3073 8.1996 7.09 × 10−11
F3_Beta 0.0878 0.0245 3.5817 0.0008
F7_Low beta −0.2684 0.0481 −5.5791 9.29 × 10−7
Fp1_Alpha 0.2708 0.0315 8.5969 1.72 × 10−11
We defined Yd as the index of the arousal level. To confirm the possibility that Yd can evaluate
the arousal level even when the thermal comfort changed, we calculated Yd from the data obtained
from five subjects for whom the arousal level and thermal comfort changed considerably, and then
performed a correlation analysis between Yd and the evaluation value of the drowsiness level. As a
result of the correlation analysis, the correlation coefficient (R) was calculated to be 0.726, and there
was a significant correlation between Yd and the evaluation value of the drowsiness level. Therefore,
this result suggests that Yd can be used to evaluate the drowsiness level, even when there are changes
in the thermal comfort.
(b) Evaluation Index Indicating Thermal Comfort

The result of the multiple regression analysis 2 (Yc) is shown in Table 2, and the regression
equation obtained from the result is expressed in Equation (3). The variable for the regression equation
included the indices of the skin temperature and EEGs. As a result, the coefficient of determination,
R2 , in relation to Yc was 0.528.
Yc = −23.372 + 0.697 Xc1 + 0.172 Xc2 + 0.142 Xc3 (3)
where Xc1 , Xc2 , and Xc3 are the MST (Mean Skin Temperature), alpha per high-beta content of C4, and
alpha per high beta content of P4, respectively.
56
Appl. Sci. 2019, 9, 899
Table 2. Result of multiple regression 2 (Yc).
Model Summary
R 0.7267 Std. Error 0.6859
R2 0.5276 data N 98
Adjusted R2 0.5125
Index Coefficient Std.Error t p-value
Intercept −23.3721 2.5461 −9.1795 1.02 × 10−14
MST 0.6967 0.0765 9.1119 1.42 × 10−14
C4_Alpha/HB 0.1718 0.0586 2.9295 0.0043
P4_Alpha/HB −0.1418 0.0296 −4.7885 6.25 × 10−6
We defined Yc as the index of thermal comfort. To confirm the possibility that Yc can evaluate
thermal comfort even when there were changes in the drowsiness level, we calculated Yc from the data
of nine subjects, excluding the data of participant D whose thermal comfort vote and drowsiness level
vote changed slightly. We thus confirmed the relationship between Yc and the value of thermal comfort.
As a result, the correlation coefficient (R) was calculated to be 0.728, and there was a significant
correlation between Yc and the value of thermal comfort (p < 0.01). Therefore, this result suggests that
Yc can be used to evaluate thermal comfort even when there are changes in the drowsiness level.
Table 1 shows the result of the multiple regression 1 (Yd), and Table 2 shows the result of the
multiple regression 2 (Yc). Figure 9 shows the relationship between the drowsiness level vote and
Yd(a), and the relationship between the comfort sensation and Yc (b)
(a) (b)
Figure 9. The relationship between the drowsiness level and Yd (a), the comfort sensation and Yc (b).
(c) Review of Continuous Evaluation using Yd and Yc

As mentioned in Section 2.2., indices that can continuously evaluate the drowsiness level and
thermal comfort of occupants are needed to clarify the design requirements related to thermal factors
that can improve these states of the occupants. Therefore, we conducted a time-series analysis with Yd
and Yc to confirm the possibility that Yd and Yc can continuously evaluate arousal levels and thermal
comfort. We calculated the time series of Yd using data from subject C, who seemed to be drowsy
frequently throughout the experiment, and compared these data with those of the drowsiness level,
which was recorded at intervals of 10 s by analyzing the facial expressions of the subject based on
Zilberg’s method [20]. The time series analyses of the normalized Yd and the drowsiness level of
subject C are shown in Figure 10a,b. The figures clearly show the trend of changes in both indicators. It
can be seen that Yd and values of the drowsiness level show a similar trend. For example, the red circles
in Figure 10a,b indicate that values of the drowsiness level and the normalized Yd are higher than 3
(significantly drowsy) and 0, respectively, at the same time. Furthermore, it can be seen from Figure 11
57
Appl. Sci. 2019, 9, 899
that there is a similar trend between Yc and the value of the comfort sensation vote. In Figure 11a,b,
the value of the comfort sensation drops to −3 (very uncomfortable) at 30 min, and at the same time,
the value of Yc drops and becomes a negative number. Thereafter, the value of the comfort sensation
vote remains at an uncomfortable state (−3: Very uncomfortable, −2: Uncomfortable), and similarly,
the value of Yc remains at a negative value. These results suggest that Yd and Yc can be used to
evaluate the drowsiness level and thermal comfort of occupants continuously. Figures 10 and 11 show,
respectively, changes in the normalized Yd (a) and drowsiness level (b), (subject C) and changes in Yc
(a) and the comfort sensation vote (b) (subject B).
(a)
(b)
Figure 10. Changes in the normalized Yd (a) and drowsiness level (b), (subject C): (a) Time series of
normalized Yd, (b) Time series of the value of the drowsiness level.
(a)
(b)
Figure 11. Changes in Yc (a) and the value of the comfort sensation vote (b): Time series of the value of
the comfort sensation vote, (subject B).
58
Appl. Sci. 2019, 9, 899
(d) Review of the Utility and Validity of the Evaluation Index

To review the utility and validity of the evaluation index, as an example, we set the threshold
value in relation to Yd and the drowsiness level. When the threshold value of Yd is sets as 2.3, as shown
in Figure 12, Yd can be used as the classifier, which can classify the drowsiness level between level 0,
1, 2, or more with an 86.7% accuracy. This result suggests that there is a significant difference in the
physiological reaction between the drowsiness level of 1 (slightly drowsy) and 2 (moderately drowsy),
and is a similar trend to the results of a previous study about the relation between the arousal levels
and physiological indices of a driver [21], though the task of the subjects is different from that of the
previous study (arithmetic task vs driving). This result also suggests that Yd is a valid indicator of the
drowsiness level of occupants to classify between the drowsiness level of 1 (slightly drowsy) and 2
(moderately drowsy). Figure 12 shows the setting of the threshold value in relation to Yd.
Figure 12. Setting of the threshold value in relation to Yd.
5. Conclusions
In this study, we aimed to design a thermal environment that can improve the indicators of both
the arousal level and feelings of thermal comfort. We hypothesized that thermal stimulation due to
cooling can improve the arousal level of occupants, after which thermal comfort can be improved while
maintaining high arousal levels due to the removal of thermal stimulation. To verify the hypothesis,
we measured physiological indices, values of subjective evaluation, and performances of arithmetic
tasks throughout several thermal conditions in which the indoor ambient temperature was changed.
In addition, we investigated the relationships between them to identify the indices that can be used to
evaluate the arousal levels and thermal comfort of occupants. As a result, the following findings were
noted:
• When the indoor ambient temperature decreased and then increased, both the arousal level and
thermal comfort of occupants remained at high levels. This result suggests that the hypothesis of
this study was verified and changes in the indoor ambient temperature can be used to improve
both thermal comfort and the arousal level of occupants.
• We proposed the evaluation indices of thermal comfort and the drowsiness level of occupants. It
was observed that the drowsiness level and thermal comfort of occupants can be evaluated
quantitatively and continuously using Yd and Yc, which were obtained from the equation
consisting of physiological indices in relation to EEGs, ECGs, and skin temperature.
59
Appl. Sci. 2019, 9, 899
In future work, we will investigate the relationship between the change patterns of the indoor
ambient temperature, thermal comfort, and arousal levels of occupants. After that, we aim to design a
novel thermal environment, considering all comfort parameters based on these findings, with the aim
of improving both the arousal levels and feelings of thermal comfort of occupants, and then carry out
an evaluation of the validity of the designed thermal environment.
Author Contributions: Conceptualization, J.G., M.S. and M.K.; methodology, J.G., M.S., K.U. and M.K.; software,
J.G.; validation, J.G.; formal analysis, J.G.; investigation, J.G.; resources, M.S., K.U. and M.K.; data curation, J.G.;
writing—original draft preparation, J.G.; writing—review and editing, J.G., M.S., K.U. and M.K.; visualization,
J.G.; supervision, M.S. and M.K.; project administration, M.S. and M.K.; funding acquisition, M.S. and M.K.
Funding: This research received no external funding.
Formula Symbols
Index Meaning Unit

MST Mean skin temperature ◦C
T1 Temperature of the forehead ◦C
T2 Temperature of the forearms ◦C
T3 Temperature of the hands ◦C
T4 Temperature of the abdomen ◦C
T5 Temperature of the thighs ◦C
T6 Temperature of the legs ◦C
T7 Temperature of the feet ◦C
Yc Index of thermal comfort

Yd Index of drowsiness level
Xd1 High alpha content of T7
Xd2 Beta content per alpha content of F7
Xd3 Beta content of F3
Xd4 Low beta content of F7
Xd5 Alpha content of Fp1
Xc1 Mean skin temperature ◦C
Xc2 Alpha content per high beta content of C4

Xc3 Alpha content per high beta content of P4
References
1. ISO7730:2005. Ergonomics of the thermal environment-Analytical determination and interpretation of
thermal comfort using calculation of the PMV and PPD indices and local thermal comfort criteria
2. De Dear, R.; Brager, G.S. Developing an adaptive model of thermal comfort and preference. ASHRAE Trans.
1998, 104, 145–167.
3. Yang, Y.; Li, B.; Liu, H.; Tan, M.; Yao, R. A study of adaptive thermal comfort in a well-controlled climate
chamber. Appl. Therm. Eng. 2015, 76, 283–291. [CrossRef]
4. Cannistraro, M.; Cannistraro, G.; Restivo, R. Some Observations on the Influence on Exchanges Radiative.
Int. J. Heat Techonol. 2015, 33, 115–122. [CrossRef]
5. Cannistraro, M.; Lorenzini, E. The Applications of the New Techonologies E-Sensing in Hospitals. Int. J. Heat
Techonol. 2016, 34, 551–557. [CrossRef]
6. Cannistraro, M.; Cannistraro, G.; Restivo, R. Smart Controll of Air Climatization System in Function on the
Values of the Mean Local Radiant Temperature. Smart Sci. 2015, 3, 157–163. [CrossRef]
7. Liu, J.; Yao, R.; McCloy, R. A method to weight three categories of adaptive thermal comfort. Energy Build.
2012, 47, 312–320. [CrossRef]
8. McIntyre, D.A. Indoor Climate; Applied science publishers: London, UK, 1980.
60
Appl. Sci. 2019, 9, 899
9. Gohara, T.; Iwashita, G. Disscussion of relationship between indoor environmental comfort and performance
in simple task. J. Environ. Eng. 2003, 572, 75–80. [CrossRef]
10. ASHRAE-55. Thermal Environmental Conditions for Human Occupancy; American Society of Heating,
Refrigerating and Air-conditioning Engineers: Atlanta, GA, USA, 2004.
11. Kondo, E.; Kurazumi, Y.; Horikoshi, T. Human physiological and psychological reactions to thermal
transients with air temperature step changes: A Case of Climacteric Aged Females in Summer. J. Hum. Living
Environ. Jpn. 2014, 21, 75–84.
12. Yoshida, M.; Kato, C.; Kakamu, Y.; Kawasumi, M.; Yamasaki, H.; Yamamoto, S.; Nakano, T.; Yamada, M.
Study on Stimulation Effects for Driver Based on Fragrance Presentation. In Proceedings of the MVA2011
IAPR Conference on Machine Vision Applications, Nara, Japan, 13–15 June 2011; pp. 332–335.
13. Mohri, Y.; Kawaguchi, M.; Kojima, S.; Yamada, M.; Nakano, T.; Mohri, K. Arousal retention effect of magnetic
stimulation to car drivers preventing drowsy driving without sleep rebound. Trans. Inst. Electr. Eng. Jpn.
2016, 136, 383–389. [CrossRef]
14. Mardi, Z.; Naghmeh, S.; Ashtiani, M.; Mikaili, M. EEG-based drowsiness detection for safe driving using
chaotic features and statistical tests. J. Med. Signals Sens. 2011, 1, 130–137. [PubMed]
15. Furman, G.D.; Baharav, A.; Cahan, C.; Akselrod, S. Early detection of falling asleep at the wheel: A heart rate
variability approach. Comput. Cardiol. 2008, 35, 1109–1112.
16. Hardy, J.D.; Dubois, E.F. The technic of measuring radiation and convection. Nutrition 1938, 5, 461–475.
[CrossRef]
17. Yao, Y.; Lian, Z.; Liu, W.; Jiang, C.; Liu, Y.; Lu, H. Heart rate variation and electroencephalograph-the
potential physiological factors for thermal comfort study. Indoor Air 2009, 19, 93–101. [CrossRef] [PubMed]
18. Turner, J.R.; Hewitt, J.K.; Morgan, R.K.; Sims, J.; Carroll, D.; Kelly, K.A. Graded mental arithmetic as an active
psychological challenge. Int. J. Psychophysiol. 1986, 3, 307–309. [CrossRef]
19. SO10551. First edition 1995-05-15, Ergonomics of the thermal environment–Assessment of the influence of
the thermal environment using subjective judgement scales.
20. Zilberg, E.; Burton, D.; Xu, Z.M.; Karrar, M. Methodology and initial analysis for development of non-invasive
and hybrid driver drowsiness detection systems. In Proceedings of the 2nd IEEE International Conference
Wireless Broadband and Ultra Wideband Communications, AusWireless, Sydney, Australia, 27–30 August
2007; p. 16.
21. Gwak, J.S.; Shino, M.; Hirao, A. Eary detection of driver drowsiness using machine learning based on
physiological signals, behavioral measures and driving performance. In Proceedings of the 21th IEEE
International Conference on Intelligent Transportation Systems, Maui, Hawaii, USA, 4–7 November 2018;
pp. 1794–1800.
61
applied
sciences
Article
Probability of Abnormal Indoor Air Exposure
Categories Compared with Occupants’ Symptoms,
Health Information, and Psychosocial
Work Environment
Katja Tähtinen 1,2, *, Sanna Lappalainen 2 , Kirsi Karvala 2 , Marjaana Lahtinen 2 and
Heidi Salonen 1
1 Department of Civil Engineering, Aalto University, 02150 Espoo, Finland; heidi.salonen@aalto.fi
2 Finnish Institute of Occupational Health, Healthy Workspaces, P.O. Box 18, 00032 Työterveyslaitos, Finland;
sanna.lappalainen@ttl.fi (S.L.); kirsi.karvala@ttl.fi (K.K.); marjaana.lahtinen@ttl.fi (M.L.)
* Correspondence: katja.tahtinen@ttl.fi; Tel.: +358-40-584-3360
Received: 29 November 2018; Accepted: 21 December 2018; Published: 28 December 2018
Abstract: Indoor air problems are complicated and need to be approached from many perspectives.
In this research, we studied the association of four-level categorisation of the probability of
abnormal indoor air (IA) exposure with the work environment-related symptoms, group-level
health information and psychosocial work environment of employees. We also evaluated the
multiprofessional IA group assessment of the current indoor air quality (IAQ) of the hospital
premises. We found no statistical association between the four-level categorisation of the probability
of abnormal IA exposure and the employees’ perceived symptoms, health information, and perceived
psychosocial work environment. However, the results showed a statistical association between
perceived symptoms and man-made vitreous fibre sources in ventilation. Furthermore, extensive
impurity sources in the premises increased the employees’ contact with health services and
their perceived symptoms. The employees perceived stress and symptoms in all categories of
abnormal IA exposure, which may be related to IAQ or other factors affecting human experience.
Prolonged process management may influence users’ experiences of IAQ. The results suggest that an
extensive impurity source in premises does not always associate with the prevalence of perceived
symptoms. We conclude that indoor air questionnaires alone cannot determine the urgency of the
measures required.
Keywords: perceived indoor air quality; building research; indoor air questionnaires; psychosocial
work environment; categorisation; ventilation; mould; moisture; man-made mineral fibres
1. Introduction
To assess the health significance, urgency, and extent of required indoor air quality (IAQ) measures,
property owners and occupational health and safety professionals need reliable information on the
buildings’ conditions and impurity sources. Information is also needed regarding the experiences
and health of the users of the premises, and on the cooperation in indoor air (IA) solution processes.
When all the factors affecting the IAQ problem have been properly assessed, the degree, timing,
and possible prioritisation of measures can be decided on. Properly timed and targeted measures
have important implications for the economy, health, and well-being. IAQ problems can be controlled
and good IAQ achieved if (i) the factors affecting the indoor environment are under control, (ii) the
indoor environment is perceived as good and healthy [1–3], and (iii) good practices are in place for
maintaining the indoor environment and solving indoor air (IA) problems [3,4].

Appl. Sci. 2019, 9, 99
IAQ problems are often the result of many different factors and their interaction or complex
combination. In addition to moisture and mould damage, several other factors and their interactions,
such as material emissions [5], ventilation deficiencies [6], system impurities [7], outdoor and soil
impurities [8], human activities in the premises [9], IA temperature, and dry air [10] can cause IAQ
problems. Office-type buildings have no established measurement methods for all IA pollutants,
and no health-based limit values for most of them [2,11].
Several IA pollutant sources can cause symptoms and harm to the users of premises. A recent
review concluded that the greater the presence of moisture and mould damage in the premises,
the greater the risk of respiratory health effects [11]. Man-made vitreous fibres (MMVF) in the
ventilation system may cause upper respiratory irritation and skin symptoms among users of the
premises [5,12–14]. Volatile organic compounds (VOC) from building materials may cause sensory
irritation [5,7]. IAQ problems may also affect sick leaves and work efficiency [15].
It has been estimated that predictive property management reduces the IAQ-related symptoms
of premises users [16], and some evidence shows that repairing moisture and mould damage and
removing contaminants from buildings can reduce respiratory symptoms [17,18] and improve work
efficiency [19].
In addition to IAQ factors, several other factors at workplaces, such as stress, poor cooperation,
heavy workload, and individual factors may also affect perceived IAQ and play a role in IAQ
problems [20,21]. These problems should be examined from a wider perspective, and experience
of the users of the premises and the psychosocial environment should be considered [20–23]. It has
been suggested that good practices for solving complex or prolonged IAQ problems are well organised
and involve long-term multiprofessional cooperation between experts [23,24].
The aim of this study was to test the use of the holistic approach in determining the urgency of
the measures required from the perspective of building health. It can be divided into the following
sub-aims: (i) to evaluate the relation between the four-level categorised probability of abnormal
IA exposure and employees’ work environment-related symptoms, group-level health information,
and psychosocial work environment, (ii) to assess the relation between ventilation system deficiencies
and employees’ work environment-related symptoms and (iii) to evaluate the impact of prolonged
IAQ problem solution processes on perceived IAQ.
This paper uses the term probability of abnormal IA exposure, which means a comprehensive
method of categorising the results of building and ventilation system research. The method used to
assess the probability of abnormal IA exposure is presented in our earlier study [25].
2.1. Materials
This study is based on two research and development projects conducted at the Finnish Institute
of Occupational Health (FIOH). These projects were carried out between February 2013 and April 2014,
in 27 hospital buildings (studied area altogether about 130,000 m2 ), which form a unified building
complex located in two Finnish hospital districts. Background information on the buildings’ earlier
history and documents revealed that parts of the buildings had IAQ problems. We investigated
altogether 111 building floors or sections and selected forty building floors or sections on which to
focus in more detail, from the premises in which both the IA questionnaire and the assessment of
probability of abnormal IA exposure were carried out. We also conducted the building investigations
and abnormal IA exposure assessments were still carried out in building premises that were not
workplaces or were not in use. The oldest building was built around 1902 and the newest in 2010.
One half of the buildings (48%) were built between the 1940s and 1950s. All of them were of stone or
different combinations of stone materials and were mostly multistorey and divided between many
hospital department areas and hospital functions. Some of the buildings had been repaired in several
different stages and these renovations varied greatly. The ventilation system of the buildings was
64
Appl. Sci. 2019, 9, 99
mostly mechanical extract and supply ventilation. However, several different ventilation systems and
machines served different parts of the building. The maintenance, repair, reliability, and age of the
ventilation systems varied considerably across the floors or sections of even one building.
The same IA researcher group conducted all the building research and ventilation system
assessments. All the data were analysed by the same multi-professional group of experts,
which comprised IA researchers, a civil engineer, an occupational health physician, a microbiologist,
and a ventilation and building health specialist.
2.2. Assessment of Probability of Abnormal IA Exposure

We carried out systematic building examinations that covered (i) structural and architectural
plan surveys, (ii) maintenance staff interviews, (iii) examinations and openings of high-risk building
structures, (iv) moisture- and mould-damaged range and severity authentications, (v) assessments of
ventilation systems, (vi) assessments of air leaks from or through damaged structures, (vii) assessments
of air pressure differences, and (viii) assessments of other IA pollutants or pollutant sources in the
buildings [25] (Figure 1). We collected building investigation and IAQ measurement results and used
a four-level categorisation method to assess the probability of abnormal IA exposure [25] (Figure 1).
Figure 1. Process for assessing the probability of abnormal indoor air (IA) exposure [25].
We collected categorised parameters for the final assessment of the probability of abnormal IA
exposure. This probability was categorised as: (1) probability of abnormal IA exposure unlikely,
(2) probability of abnormal IA exposure possible, (3) probability of abnormal IA exposure likely, and
(4) probability of abnormal IA exposure very likely [25] (Table 1). In cases of moisture and mould
damage, air leaks from or through damaged structures to IA must be examined simultaneously
with indoor negative pressure. In the main criteria for assessing the probability of abnormal IA
exposure, the predominant IA impurity source is a determining criterion. The probability of abnormal
IA exposure arises when the national limit values (IAQ, material samples, ventilation) are exceeded,
structures or systems have found damaged, or IAQ pollutant sources that are known to affect indoor air
quality and building health are found [25]. The national maximum limit values for IA concentrations,
microbial growth on building material, MMVF and asbestos in dust, and specific detailed methods for
evaluating building and ventilation conditions and IAQ are presented in our earlier research [25].
65
Appl. Sci. 2019, 9, 99
Table 1. Main criteria and categories for assessing probability of abnormal IA exposure in buildings.
Categories Main Criteria for Assessing Probability of Abnormal IA Exposure in Buildings.

No moisture or mould damage in structures. No air leaks from or through damaged
structures. Ventilation system can be controlled by indoor pressure difference from the
Unlikely building envelope. Room acoustic materials and ventilation system have no man-made
vitreous fibres (MMVF) sources. Indoor air quality corresponds to national reference
values and guidelines set for the premises.
Mould-damaged structure type is not widespread in building and repairs are easily
definable (less than 1 m2 ). A few or single air leaks from or through damaged
structures or from surrounding premises. Room acoustic materials or ventilation
system have MMVF sources and fibres may end up in the indoor air or on surfaces.1
Possible
Concrete floor has extensive moisture, which can cause water vapour damage to
permeable floor coating (emissions).1 Indoor air quality does not correspond to
national reference values or the guidelines set for the premises, and an indoor air
impurity source has been identified.1
Building or premises have widespread mould-damaged structure. Repairs are
significant and affect a large part of the (one) structure of the building or premises, e.g.,
whole base floor structure. There is recurrent damage in the type of structure. Air leaks
from or through damaged structures or from surrounding premises and moisture or
mould-damaged materials are regular and recurrent in the structure, occasionally there
Likely
is negative pressure in the premises and/or air-tightness is risky. Indoor air quality
does not correspond to national reference values or the guidelines set for the premises,
and an indoor air impurity source has been identified.1 Creosote has been used in the
structure and air leaks into the indoor air from the structure. There is also a notable
smell of creosote (e.g., naphthalene) in the indoor air.1
The building or premises has a great deal of extensive mould damage in several
structures. The extent of repairs is significant and affects several structures in the
building or premises e.g., whole façade and whole base floor. There is recurrent
damage in the type of the structures. Air leaks from or through damaged structures are
regular and recurrent, negative pressure is significant in the premises and/or
air-tightness is very risky. Indoor air quality does not correspond to national reference
Very likely values or the guidelines set for the premises, and an indoor air impurity source has
been identified.1 Creosote has been used in the structures and air leaks into the indoor
air from the structures. In addition, concentrations of polycyclic aromatic
hydrocarbons (PAH) or separate components exceed the set national values and
guidelines.1 Dust sample tests have found asbestos fibres in the premises, and the
pollution source has been defined.1 Indoor radon concentrations exceed the set
national values and guidelines (400 Bq/m3 [26]).1
1 The assessment must take into account the extent and impact of the problem and impurity source.
2.3. Employees’ Experiences of Indoor Air Quality and Psychosocial Work Environment
To study the users’, i.e., employees’ experiences of the work premises, work environment-related
symptoms and psychosocial work environment, we used FIOH’s validated and frequently used IA
questionnaire, which is based on Örebro’s [27] indoor climate questionnaire [28,29]. To study perceived
stress, we used a validated single-item measure of stress symptoms [30]: “Stress means a situation in
which a person feels tense, restless, nervous or anxious or is unable to sleep at night because his/her
mind is troubled all the time. Do you feel this kind of stress these days?”. The response options were:
(1) not at all, (2) just a little, (3) some, (4) quite a lot, and (5) very much. In the analyses, we combined
the levels (1) not at all and (2) just a little into one level, and levels (4) quite a lot and (5) very much
into one level.
We sent the questionnaire to 3608 hospital employees, of whom 2669 responded. The total
response rate was 74%, with a range of 51% to 93%. The surveys were conducted in the spring from
February to April and in the autumn in November, in 2013. We selected 40 IA questionnaire groups for
the study, totaling 1558 respondents. The selected IA questionnaire groups were in premises in which
66
Appl. Sci. 2019, 9, 99
the probability of abnormal IA exposure assessment had already been performed. The employees
did not know the results of the assessment prior to responding to the IA questionnaires. This is a
questionnaire-based study, in which participation was voluntary and performed no intervention on
individuals, according to Finnish legislation it did not require ethics committee handling.
2.4. Group-Level Information from Occupational Health Services and Multiprofessional Indoor Air Group
We obtained information on the assessment of the group-level health of employees from
occupational health services (OHS). The information covered employees’ health from 43 building
sections or floors. The group-level information from the OHS did not contain information on how
many employees had work environment-related health symptoms in the building sections or on the
floors. In the survey, we used short forms to ask about the following issues in relation to employees’
health: (i) case of new onset asthma or aggravation of existing asthma, (ii) having to change workroom
because of IAQ and work environment-related symptoms, (iii) increased amount of employee visits
to OHS due to IAQ-related issues, and (iv) increased sickness absences due to respiratory symptoms.
The hospital’s multiprofessional IA group also provided information on the estimated duration of the
IAQ problem solution on every building floor or section.
2.5. Statistical Analyses

Statistical analyses were carried out using IBM SPSS Statistics program 25.0 with a statistically
significant level of p < 0.05. The statistical analysis used weighted averages of group response
rates. The Mann–Whitney U test studied the differences between the probability of abnormal IA
exposure categories (unlikely, possible, likely, and very likely) and the employees’ complaints about their
work environment-related symptoms and psychosocial work environment. This test also compared
the difference between the two groups’ (yes/no) ventilation adequacy, ventilation MMVF sources,
ventilation moisture problems, and expired ventilation lifespan and the employees’ complaints about
their work environment-related symptoms. Fisher’s exact test studied the relation between the weekly
work environment-related symptoms experienced by the employees and the categorised group-level
information on employee health. The group-level health information was categorised as ‘yes’ and ‘no’
as follows: case of new onset asthma or aggravation of existing asthma, having to change workroom
or workplace because of work environment-related symptoms, increased amount of employee visits to
OHS, and increased sickness absences due to respiratory symptoms.
3. Results
Probability of Abnormal IA Exposure and Employees’ Experience

All building floors or sections (total 111) were investigated and we were able to assess the
probability of abnormal IA exposure on or in 95 building floors or sections. In the case of forty building
floors or sections, the assessment of the probability of abnormal IA exposure and IA questionnaire
could both be conducted in the same areas (the IA questionnaire group was located in the area that was
assessed as belonging to an abnormal IA exposure category). The probability of abnormal IA exposure
was assessed as unlikely for 5% (n = 2), possible for 40% (n = 16), likely for 45% (n = 18), and very likely for
10% (n = 4) of the selected forty floors or sections of the buildings. In the likely and very likely categories,
these floors or sections had wide moisture and mould damage in their structures together with air leaks
from damaged materials to the IA and often had a detected MMVF source in the ventilation system
as well as ventilation deficiencies (Table 2 and criteria in Table 1). These categories also had other
impurity sources (Table 1), but moisture and mould damage in the building structures were dominant.
The higher (more abnormal) the assessed category of probability of abnormal IA exposure, the more
insufficient the ventilation was, the more often the lifespan of the ventilation system was exceeded
and the more often MMVF sources were detected in the ventilation system from the categories likely
and very likely (and not from unlikely category) (Table 2).
67
Appl. Sci. 2019, 9, 99
Table 2. Ventilation survey findings are included in all the sections or floors in which probability of
abnormal IA exposure was assessed (95 floors or sections). (All building floors or sections (111) were
investigated and we were able to assess the probability of abnormal IA exposure on or in 95 building
floors or sections.).
Assessed Probability of Lifespan of

MMVF (Man-Made Moisture Problem
Abnormal IA Exposure Ventilation Insufficient
Vitreous Fibres) Source in Ventilation
on/in Building Floors or System Had Been Ventilation
in Ventilation System System
Sections Exceeded n (%)
n (%) n (%)
(n = 95) n (%)
Unlikely (n = 7) 7 (100) 3 (43) 3 (43) 0 (0)
Possible (n = 39) 23 (59) 22 (56) 12 (31) 6 (15)
Likely (n = 37) 26 (70) 24 (65) 26 (70) 15 (41)
Very likely (n = 12) 10 (83) 9 (75) 9 (75) 1 (8)
The perceived weekly work environment-related symptoms among the employees exceeded the
corresponding number of weekly work environment-related symptoms in FIOH’s reference data [28,29]
(Figure 2). Even in premises in which the building research showed no source of contamination, some
employees perceived weekly symptoms more than those in FIOH’s reference data [28,29] (Figure 3).
An analysis of the differences between the probability of abnormal IA exposure categories (unlikely,
possible, likely, very likely) and the employees’ weekly perceived work environment-related symptoms
(headaches, concentration difficulties, irritation of the eyes and nose, irritation of the skin on the face
and hands, hoarse throat, coughing, coughing at night, shortness of breath, wheezing, fever or chills,
joint pain, and muscular pain revealed no statistically significant differences (Table 3).
Figure 2. In the buildings assessed for probability of abnormal IA exposure, weekly work
environment-related symptoms (n = 1558) were perceived more often than in the Finnish Institute of
Occupational Health’s (FIOH) comparable reference data [28].
68
Appl. Sci. 2019, 9, 99
(a) (b)
(c) (d)
Figure 3. Perceived work environment-related symptoms of users of premises in assessment of
probability of abnormal IA exposure categories, compared to FIOH reference data [28] (a) unlikely
(n = 61), (b) possible (n = 618), (c) likely (n = 760), (d) very likely (n = 118).
Table 3. An analysis of the differences between the probability of abnormal IA exposure

categories (unlikely, possible, likely, very likely) and the employees’ weekly perceived work
environment-related symptoms.
Unlikely 1 Possible 2 Likely 3

Very Likely 4
Variable (Weekly Symptoms) n=2 n = 16 n = 18
n=4
(p-Value) (p-Value) (p-Value)
Fatigue 1 vs. 2 (0.673) 2 vs. 3 (0.220) 3 vs. 4 (0.609)
1 vs. 3 (0.257) 2 vs. 4 (0.570)
1 vs. 4 (0.355)
Headache 1 vs. 2 (0.778) 2 vs. 3 (0.157) 3 vs. 4 (0.481)
1 vs. 3 (0.208) 2 vs. 4 (0.777)
1 vs. 4 (0.643)
Concentration difficulties 1 vs. 2 (0.672) 2 vs. 3 (0.233) 3 vs. 4 (1.000)
1 vs. 3 (0.165) 2 vs. 4 (0.507)
1 vs. 4 (0.064)
Irritation of the eyes 1 vs. 2 (0.888) 2 vs. 3 (0.262) 3 vs. 4 (0.125)
1 vs. 3 (0.378) 2 vs. 4 (0.777)
1 vs. 4 (0.814)
Irritation of the nose 1 vs. 2 (0.779) 2 vs. 3 (0.073) 3 vs. 4 (0.061)
1 vs. 3 (0.130) 2 vs. 4 (0.925)
1 vs. 4 (0.643)
69
Appl. Sci. 2019, 9, 99
Table 3. Cont.
Unlikely 1 Possible 2 Likely 3

Very Likely 4
Variable (Weekly Symptoms) n=2 n = 16 n = 18
n=4
(p-Value) (p-Value) (p-Value)
Hoarse, dry throat 1 vs. 2 (0.399) 2 vs. 3 (0.101) 3 vs. 4 (0.061)
1 vs. 3 (0.130) 2 vs. 4 (0.508)
1 vs. 4 (1.000)
Coughing 1 vs. 2 (0.481) 2 vs. 3 (0.147) 3 vs. 4 (0.287)
1 vs. 3 (0.378) 2 vs. 4 (0.636)
1 vs. 4 (0.355)
Coughing at night 1 vs. 2 (0.941) 2 vs. 3 (0.652) 3 vs. 4 (0.237)
1 vs. 3 (0.792) 2 vs. 4 (0.289)
1 vs. 4 (0.411)
Irritation of skin on face 1 vs. 2 (0.260) 2 vs. 3 (0.133) 3 vs. 4 (0.551)
1 vs. 3 (1.000) 2 vs. 4 (0.508)
1 vs. 4 (1.000)
Irritation of skin on hands 1 vs. 2 (0.888) 2 vs. 3 (0.152) 3 vs. 4 (0.349)
1 vs. 3 (0.378) 2 vs. 4 (0.705)
1 vs. 4 (1.000)
Shortness of breath 1 vs. 2 (0.562) 2 vs. 3 (0.508) 3 vs. 4 (0.663)
1 vs. 3 (0.509) 2 vs. 4 (0.502)
1 vs. 4 (0.355)
Wheezing 1 vs. 2 (0.374) 2 vs. 3 (0.342) 3 vs. 4 (0.275)
1 vs. 3 (0.178) 2 vs. 4 (0.600)
1 vs. 4 (0.480)
Fever or chills 1 vs. 2 (0.352) 2 vs. 3 (0.308) 3 vs. 4 (1.000)
1 vs. 3 (0.736) 2 vs. 4 (0.562)
1 vs. 4 (0.623)
Muscular pain 1 vs. 2 (1.000) 2 vs. 3 (0.161) 3 vs. 4 (0.932)
1 vs. 3 (0.44) 2 vs. 4 (0.298)
1 vs. 4 (1.000)
Joint pain 1 vs. 2 (0.324) 2 vs. 3 (0.283) 3 vs. 4 (0.898)
1 vs. 3 (0.900) 2 vs. 4 (0.570)
1 vs. 4 (0.643)
Other work environment- 1 vs. 2 (0.176) 2 vs. 3 (0.099) 3 vs. 4 (0.831)
related symptoms 1 vs. 3 (0.074) 2 vs. 4 (0.288)
1 vs. 4 (0.060)
Statistically significant level of p < 0.05.
As regards the IA questionnaire results, most of the employees (88%) felt that their work was often
stimulating and interesting, 74% believed they would receive help from their colleagues if needed,
21% often had the opportunity to influence their own work and working conditions, and 53% had no
feelings of stress. A heavy workload was reported by 14% of the employees, which is below FIOH’s
reference value [28]. Stress was reported by 16% of the employees and 4% believed that they would
not get help from colleagues if needed. Stress and lack of help from colleagues were perceived more
often than in FIOH’s comparable reference data [28]. Stress was perceived more often than in FIOH’s
reference data in every category of abnormal IA exposure (Table 4). Most often, stress was perceived
in premises in which the probability of abnormal IA exposure was estimated as being unlikely (Table 4).
An analysis of the differences between the probability of abnormal IA exposure categories (unlikely,
possible, likely, very likely) and the employees’ perceived psychosocial work environment and stress
revealed no statistically significant differences.
70
Appl. Sci. 2019, 9, 99
Table 4. Perceived psychosocial work environment and stress according to probability of abnormal IA
exposure categories (unlikely, possible, likely, very likely).
FIOH’s
Question on Unlikely Possible Likely Very Likely
Reference
Psychosocial Work n = 61 n = 619 n = 763 n = 118
Value
Environment (%) (%) (%) (%)
%
Do you regard your work as interesting and stimulating?
Yes, often 52 (88.1) 540 (87.5) 684 (89.9) 102 (86.4) 82
Yes, sometimes 7 (11.9) 66 (10.7) 61 (8.0) 11 (9.3) 16
No, seldom or rarely 0 (0) 11 (1.8) 16 (2.0) 5 (4.2) 0
Do you have too much work?
Yes, often 12 (20.7) 95 (15.5) 98 (12.9) 15 (12.7) 20
Yes, sometimes 36 (62.1) 360 (58.7) 489 (64.4) 71 (60.2) 64
No, seldom or rarely 10 (17.2) 158 (25.8) 172 (22.7) 32 (27.1) 16
Do you have opportunities to influence your working conditions?
Yes, often 38 (32.8) 120 (19.5) 161 (21.2) 38 (32.8) 21
Yes, sometimes 61 (52.6) 335 (54.3) 379 (49.9) 61 (52.6) 51
No, seldom or rarely 17 (14.7) 162 (26.3) 220 (29.0) 17 (14.7) 28
Do your fellow workers help you with work-related problems?
Yes, often 33 (55.9) 463 (75.2) 569 (74.8) 83 (70.3) 79
Yes, sometimes 24 (40.7) 129 (20.9) 171 (22.5) 28 (23.7) 19
No, seldom or rarely 2 (3) 24 (3.9) 21 (2.8) 7 (5.9) 2
Do you feel stress?
Quite a lot or very much 15 (26.3) 92 (15.1) 105 (15.0) 16 (14.0) 10
Some 12 (21.1) 181 (29.7) 243 (32.4) 40 (35.1) 28
Not at all or just a little 30 (52.6) 337 (55.2) 403 (53.7) 58 (50.9) 63
In addition, an analysis of the differences between the probability of abnormal IA exposure

categories’ (unlikely, possible, likely, very likely) and the employees’ group-level health information
(obtained from OHS) revealed no statistically significant differences. However, the more abnormal the
probability of the IA exposure category, the more employees contacted OHS due to IAQ-related issues
from the categories possible, likely and very likely (Table 5).
Table 5. IAQ-related group-level health information on employees (obtained from OHS), according to
building floors or sections in which probability of abnormal IA exposure categories (unlikely, possible,
likely, very likely) were assessed.
Unlikely Possible Likely Very likely

IAQ-Related Health Information n=2 n = 16 n = 20 n=5
(%) (%) (%) (%)
Some employees have new asthma or aggravation of previous asthma 1
Yes 2 (100) 5 (31) 5 (25) 3 (60)
No 0 (0) 10 (63) 14 (70) 2 (40)
No information 0 (0) 1 (6) 1 (5) 0 (0)
Some employees have changed work premises or work places due to IAQ-related symptoms 1
Yes 1 (50) 4 (25) 4 (20) 2 (40)
No 1 (50) 11 (69) 16 (80) 3 (60)
No information 0 (0) 1 (6) 0 (0) 0 (0)
The amount of employee contacts with OHS due to IAQ-related issues has increased 1
Yes 2 (100) 5 (31) 10 (50) 3 (60)
No 0 (0) 10 (63) 10 (50) 2 (40)
No information 0 (0) 1 (6) 0 (0) 0 (0)
The amount of sickness absence due to respiratory symptoms has increased 1
Yes 1 (50) 4 (25) 8 (40) 1 (20)
No 0 (0) 7 (44) 11 (55) 3 (60)
No information 1 (50) 5 (31) 1 (5) 1 (20)
1Group-level health information does not contain information on how many employees have IAQ-related health
symptoms on building floors or in sections or the assessed categories. IAQ: indoor air quality; OHS: occupational
health services.
71
Appl. Sci. 2019, 9, 99
The results show a statistical association between detected MMVF sources in ventilation
systems and perceived work environment-related symptoms and a statistical association between
ventilation system age and perceived work environment-related symptoms (Table 6). The hospital’s
multiprofessional IA group estimated the duration of the IAQ problem solution process on every
building floor or in each section, and the attempts to solve the indoor air problems in the unlikely,
possible, likely, and very likely categories had lasted a year or more (Table 7). The hospital’s
multiprofessional IA group also estimated ‘No IAQ problems’ in premises in which the research
group had assessed the probability of abnormal IA exposure as being possible and likely (Table 7).
Table 6. Statistical significance of differences (p-value) between weekly reported work

environment-related symptoms and ventilation factors (yes/no) studied.
Insufficient
Technical
Moisture Ventilation or
Perceived Work Respondents Lifespan of MMVF Source
Problem in Ventilation System
Environment-Related N = 1558 Ventilation in Ventilation
Ventilation Did Not Match
Symptoms Weekly n System Had System
System Purposes of
Been Exceeded
Facilities
Fatigue 532 NS NS 0.005 NS
Headache 266 0.035 NS 0.002 NS
Concentration difficulties 164 0.049 NS 0.005 NS
Irritation of the eyes 559 0.014 NS 0.006 NS
Irritation of the nose 656 0.003 NS 0.001 NS
Hoarse, dry throat 475 0.001 NS 0.001 0.025
Coughing 225 0.016 NS 0.022 NS
Coughing at night 50 0.008 NS NS NS
Irritation of skin on face 373 NS NS NS NS
Irritation of skin on hands 508 0.035 NS 0.003 NS
Shortness of breath 64 0.045 NS NS NS
Wheezing 32 0.001 NS NS 0.044
Fever or chills 42 NS NS NS NS
Muscular pain 160 0.027 NS 0.003 NS
Joint pain 193 0.039 NS NS NS
Other work
environment-related 75 0.005 NS 0.001 0.027
symptoms
NS: not significant. Statistically significant level of p < 0.05.
Table 7. Estimated duration of IAQ problems on each building floor or in each section studied.
Estimated Duration of IAQ Problems: Unlikely Possible Likely Very Likely

Number of Cases (n = 2) (n = 16) (n = 18) (n = 4)
No IAQ problems 1 7 6 0
Duration of IAQ problems less than a year 0 2 4 0
Duration of IAQ problems one year or more 1 7 8 4
For each building floor or section, we looked at all the collected data (results of assessed probability
of abnormal IA exposure, IA questionnaires results, group-level health information and IA group
information) at the same time. The most urgent measures were required for floors or sections
on/in which the probability of abnormal IA exposure was likely or very likely, the health information
(Table 5) and IA questionnaire results (Figure 3a–d) indicated health problems, and solutions were
delayed (Table 7).
72
Appl. Sci. 2019, 9, 99
4. Discussion
The detailed examination of the buildings and the four-level categorisation of the probability
of abnormal IA exposure made it easier to organise the outcomes and obtain a clear picture of
the many factors affecting the IAQ of the buildings and premises. The strength of our research
was its multiprofessional approach, which took into account the employees’ perceived work
environment-related symptoms and health information, multiprofessional IA group information
and the results of the technical building investigations.
According to previous studies, observed indoor mould and moisture damage indicates an
increased health risk, and the greater the mould and moisture damage, the more prevalent respiratory
symptoms in adults [11]. This study found no statistically significant differences between the
four-level categories and employees’ perceived work environment-related symptoms, perceived
psychosocial environment, or OHS group-level information on employees’ health. However, extensive
impurity sources in the premises increased some employees’ perceived work-related symptoms.
Our earlier study [31] shows that all the symptoms perceived by the employees were very similar
to the work-related symptoms examined in this study. The results showed a statistical association
between MMVF sources in ventilation systems and perceived IAQ and work environment-related
symptoms. Other studies have achieved similar results [5,14]. In addition, the age of the ventilation
system was associated with perceived symptoms. The more the probability of abnormal IA exposure
was estimated to differ, the more prevalent were ventilation system deficiencies and the MMVF sources
in the ventilation system. Therefore, IAQ problems were usually affected by impurities from both the
building and ventilation, which shows that many IAQ factors can affect perceived symptoms.
Employees perceived more work environment-related symptoms and stress and lack of social
support from colleagues than those reported in FIOH’s comparative data from (damaged and
nondamaged) hospital buildings, a result which may be related to the poor condition of the buildings or
other factors affecting the human experience. On the other hand, the amount of employee contact with
OHS increased due to IAQ and work environment-related issues in all the categories of the probability
of abnormal IA exposure. An earlier study has shown an association between symptoms and work
strain [21], the psychosocial work environment [4,32], and individual factors (e.g., gender, age) [21,33].
The risk of experiencing the workplace as harmful has shown to be higher among employees who
report mould problems than those who report ventilation problems in workplaces [34].
In our research, symptoms were common, stress was high, and the amount of contact with
OHS in IAQ and work environment-related issues was great on floors or in sections in which the
category of abnormal IA exposure was assessed as being unlikely and building floors or sections
that were undamaged. Workers in non-damaged buildings have also shown to have IAQ and work
environment-related symptoms [29,33]. Overall, the employees perceived stress in every category of
abnormal IA exposure more than the amount reported in FIOH’s reference data [28]. On the other
hand, the employees often perceived their work as stimulating and interesting in all categories of the
probability of abnormal IA exposure. An earlier study has also shown that hospital employees find
their work more interesting and stimulating than office workers [28,32]. As the buildings and the
factors affecting IAQ have been carefully studied, the building technology and IAQ alone may not
explain the prevalence of the perceived work environment-related symptoms and the stress and the
amount of IAQ- and work environment-related contact with OHS. These issues may also be partly
affected by other factors influencing human experience, such as work-related and organizational
factors not investigated in this study, or individual human factors.
Based on the IA group’s information, attempts to solve the indoor air problems in several buildings
had lasted more than a year. The IA group’s evaluation of no IAQ problems in the premises often
contradicted our building investigation results. One reason for these prolonged problem-solving
processes may be that building-related problems or impurity sources have been unclear or building
investigations incomplete. An earlier study also had similar findings [4]. Prolonged or unclear
problem-solving processes may have increased health concerns and distrust of the problem-solving
73
Appl. Sci. 2019, 9, 99
process and influenced the IAQ-related experience of employees. Thus, careful decision-making
procedures are important, especially when people feel threatened by IAQ-related risks [34].
The employees did not know the results of the assessment of the probability of abnormal IA
exposure prior to responding to the IA questionnaires, which contributes to the reliability of the study.
The differences between previous studies and our research results may result from differences in the
building research methods. Many studies are based on observations of visible damage or indications
of damage, and hidden damage has remained unclear [11]. Our study was very detailed, and we
also investigated hidden damage, in addition other pollutant sources. Possible limitations of our
research may be that our data concerned only a small number of buildings with no IAQ problems
or impurities affecting IAQ, which may have affected data distribution. Due to this, we had no
reliable comparison survey of damaged and nondamaged buildings or premises. The method for
assessing the probability of abnormal IA exposure is very pragmatic and is always based on strong
technical expertise in building technology and IAQ. In addition, the criteria for the probability of
abnormal IA exposure recommend taking into account many impurity sources in a building, based
on national instructions, regulations and limit values in the field of the built environment and IAQ.
However, the assessment also involves a researcher’s subjective view. OHS’ information was collected
at group-level, and as the results were analysed categorically, they may not have provided sufficiently
accurate results. In addition, the questions concerning the psychosocial work environment were quite
limited in the IA questionnaire. They covered qualitative and quantitative workload, opportunities to
influence one’s working conditions, and social support at work. Although these are essential factors in
the light of the stress theory, they fail to provide a comprehensive picture of the psychosocial work
environment. The questionnaire did not include, for instance, factors such as organizational changes
and questions concerning leadership.
The probability of abnormal IA exposure provides us with a holistic picture of the many factors
affecting the IAQ of buildings and premises. In this case, the most urgent measures could have been
identified more easily and holistically. Moreover, in the premises that needed the most urgent repairs,
employee contact with OHS was increased, and the employees’ perceived work environment-related
symptoms indicated poor IAQ. On the other hand, in the premises in which no technical problems
were found, the employees still perceived more work environment-related symptoms and stress
than those reported in FIOH’s reference data. Thus, IAQ problems should always be analysed from
many perspectives; (i) the building’s technical condition, (ii) perceived IAQ and psychosocial work
environment, (iii) OHS information, and (iv) measures for solving IAQ problems, which may all
affect the experience of IAQ problems. Indoor air questionnaires can serve as a parallel method with
technical investigations in the building.
5. Conclusions
The four-level categorisation of the probability of abnormal IA exposure provides a comprehensive
and systematic way of ranking building sectors from the perspective of building health. The method is
based on national instructions for building and ventilation investigations, building codes, and limit
values, and is therefore systematic and partially established. Thus, it can be applied in different
environments. The method may also be used in other countries (with similar environments to that
in Finland), if the national instructions, limit values, and building codes are taken into account and
applied. The method enabled the holistic identification of the most urgent measures. This may
help property owners allocate resources for proper repairs and also help OHS identify employees’
IAQ and work environment-related symptoms. The results suggest that the extensive impurity
source in premises does not always associate with the prevalence of perceived IAQ and work
environment-related symptoms. Therefore, the solution to the IAQ problem is more specific when
technical survey results, the health and experience information of the users of the premises, as well as
the problem solution process are taken into account. The results also suggest that IA questionnaires
alone cannot determine the urgency of the measures required. Possible limitations of our research
74
Appl. Sci. 2019, 9, 99
are that the study was only conducted in hospital buildings and premises. Limitations may be that
our data concerned only a small number of buildings with no IAQ problems or impurities affecting
IAQ: this may have affected data distribution. The method for assessing the probability of abnormal
IA exposure is, however, very pragmatic and always based on strong technical expertise in building
technology and IAQ. Further studies should assess the probability of abnormal IA exposure in different
work environments (e.g., offices), and the associations between the probability of abnormal IA exposure
categories with perceived IAQ and the health of employees. They should also assess the impact of the
IAQ solution process perceived IAQ. The results of further studies may possibly validate the method.
Author Contributions: K.T. wrote the paper and analysed the data; S.L. and K.K. contributed to the study design
and data analysis and participated in writing the paper; M.L. contributed to the data analyses and participated in
writing the paper; H.S. was the principal supervisor of the work and participated in writing the paper.
Acknowledgments: The authors acknowledge the contribution of Leena Aalto, Veli-Matti Pietarinen; Ulla-Maija
Hellgren, Jouko Remes and Pauliina Toivio and all co-workers in building research.
Abbreviations
The following abbreviations are used in this manuscript:
FIOH Finnish Institute of Occupational Health

IA Indoor air
IAQ Indoor air quality
OHS Occupational health service
PAH Polycyclic aromatic hydrocarbon
VOC Volatile organic compound
References
1. Frontczak, M.; Wargocki, P. Literature survey on how different factors influence human comfort in indoor
environments. Build. Environ. 2011, 46, 922–937. [CrossRef]
2. WHO. Dampness and Mould—Who Guidelines for Indoor Air Quality; World Health Organisation: Geneva,
Switzerland, 2009.
3. Lappalainen, S.; Lahtinen, M.; Palomaki, E.; Korhonen, P.; Niemelä, R.; Reijula, K. Comprehensive procedure
for solving indoor environment problems. In Proceedings of the NAM Nordic Work Environment Meeting,
Espoo, Finland, 31 August–2 September 2009; p. 93.
4. Lahtinen, M.; Huuhtanen, P.; Kahkonen, E.; Reijula, K. Psychosocial dimensions of solving an indoor air
problem. Indoor Air 2002, 12, 33–46. [CrossRef] [PubMed]
5. Salonen, H.; Lappalainen, S.; Riuttala, H.; Tossavainen, A.; Pasanen, P.; Reijula, K. Man-made vitreous fibers
in office buildings in the helsinki area. J. Occup. Environ. Hyg. 2009, 6, 624–631. [CrossRef] [PubMed]
6. Sundell, J.; Levin, H.; Nazaroff, W.W.; Cain, W.S.; Fisk, W.J.; Grimsrud, D.T.; Gyntelberg, F.; Li, Y.; Persily, A.K.;
Pickering, A.C.; et al. Ventilation rates and health: Multidisciplinary review of the scientific literature.
Indoor Air 2011, 21, 191–204. [CrossRef] [PubMed]
7. Wolkoff, P.; Wilkins, C.K.; Clausen, P.A.; Nielsen, G.D. Organic compounds in office environments—Sensory
irritation, odor, measurements and the role of reactive chemistry. Indoor Air 2006, 16, 7–19. [CrossRef]
[PubMed]
8. Airaksinen, M.; Pasanen, P.; Kurnitski, J.; Seppänen, O. Microbial contamination of indoor air due to leakages
from crawl space: A field study. Indoor Air 2004, 14, 55–64. [CrossRef] [PubMed]
9. Nevalainen, A.; Täubel, M.; Hyvärinen, A. Indoor fungi: Companions and contaminants. Indoor Air 2015, 25,
10. Nordström, K.; Norbäck, D.; Akselsson, R. Effect of air humidification on the sick building syndrome and
perceived indoor air quality in hospitals: A four month longitudinal study. Occup. Environ. Med. 1994, 51,
75
Appl. Sci. 2019, 9, 99
11. Mendell, M.J.; Kumagai, K. Observation-based metrics for residential dampness and mold with
dose–response relationships to health: A review. Indoor Air 2017, 27, 506–517. [CrossRef]
12. Hellgren, U.-M.; Hyvärinen, M.; Holopainen, R.; Reijula, K. Perceived indoor air quality, air-related
symptoms and ventilation in finnish hospitals. Int. J. Occup. Med. Environ. Health 2011, 24, 48–56. [CrossRef]
13. Redlich, C.; Sparer, J.; Cullen, M.R. Sick-building syndrome. Lancet 1997, 349, 1013–1016. [CrossRef]
14. Schneider, T. Dust and fibers as a cause of indoor environment problems. Scand. J. Work. Environ. Health
2008, 33, 10–17.
15. Milton, D.K.; Glencross, P.M.; Walters, M.D. Risk of sick leave associated with outdoor air supply rate,
humidification, and occupant complaints. Indoor Air 2000, 10, 212. [CrossRef] [PubMed]
16. Cox-Ganser, J.; Rao, C.; Park, J.; Schumpert, J.; Kreiss, K. Asthma and respiratory symptoms in hospital
workers related to dampness and biological contaminants. Indoor Air 2009, 19, 280–290. [CrossRef] [PubMed]
17. Mendell, M.; Brennan, T.; Lee, H.; Odom, J.D.; Offerman, F.J.; Turk, B.H.; Wallingford, K.M.; Diamond, R.C.;
Fisk, W.J. Causes and prevention of symptom complaints in office buildings. Facilities 2006, 24, 436–444.
[CrossRef]
18. Sauni, R.; Verbeek, J.H.; Uitti, J.; Jauhiainen, M.; Kreiss, K.; Sigsgaard, T. Remediating buildings damaged
by dampness and mould for preventing or reducing respiratory tract symptoms, infections and asthma.
Cochrane Database Syst. Rev. 2015, 76, 1. [CrossRef] [PubMed]
19. Wargocki, P.; Lagercrantz, L.; Witterseh, T.; Sundell, J.; Wyon, D.P.; Fanger, P.O. Subjective perceptions,
symptom intensity and performance: A comparison of two independent studies, both changing similarly
the pollution load of an office. Indoor Air 2002, 12, 74. [CrossRef]
20. Brauer, C.; Mikkelsen, S. The influence of individual and contextual psychosocial work factors on the
perception of the indoor environment at work: A multilevel analysis. Int. Arch. Occup. Environ. Health 2010,
83, 639–651. [CrossRef]
21. Magnavita, N. Work-related symptoms in indoor environments: A puzzling problem for the occupational
physician. Int. Arch. Occup. Environ. Health 2015, 88, 185–196. [CrossRef]
22. Finell, E.; Haverinen-Shaughnessy, U.; Tolvanen, A.; Laaksonen, S.; Karvonen, S.; Sund, R.; Saaristo, V.;
Luopa, P.; Ståhl, T.; Putus, T.; et al. The associations of indoor environment and psychosocial factors on the
subjective evaluation of indoor air quality among lower secondary school students: A multilevel analysis.
Indoor Air 2017, 27, 329–337. [CrossRef]
23. Lahtinen, M.; Lappalainen, S.; Reijula, K. Multiprofessional teams resolving indoor-air problems-emphasis
on the psychosocial perspective. Scand. J. Work. Environ. Health 2008, 34, 30–34.
24. Carrer, P.; Wolkoff, P. Assessment of indoor air quality problems in office-like environments: Role of
occupational health services. Int. J. Environ. Res. Public Health 2018, 15, 741. [CrossRef] [PubMed]
25. Tähtinen, K.; Lappalainen, S.; Karvala, K.; Remes, J.; Salonen, H. Association between four-level
categorisation of indoor exposure and perceived indoor air quality. Int. J. Environ. Res. Public Health
2018, 15, 679. [CrossRef] [PubMed]
26. Finnish Ministry of the Environment, Department of Built Environment. D2 the National Building Code of
Finland, Health. Indoor Climate and Ventilation of Buildings, Regulations and Guidelines; Department of Built
Environment: Helsinki, Finland, 2012.
27. Andersson, K. Epidemiological approach to indoor air problems. Indoor Air 1998, 8, 32–39. [CrossRef]
28. Hellgren, U.-M.; Palomaki, E.; Lahtinen, M.; Riuttala, H.; Reijula, K. Complaints and symptoms among
hospital staff in relation to indoor air and the condition and need for repairs in hospital buildings. Scand. J.
Work Environ. Health 2008, 34, 58–63.
29. Reijula, K.; Sundman-Digert, C. Assessment of indoor air problems at work with a questionnaire.
Occup. Environ. Med. 2004, 61, 33–38. [PubMed]
30. Elo, A.-L.; Leppänen, A.; Jahkola, A. Validity of a single-item measure of stress symptoms. Scand. J. Work
Environ. Health 2003, 29, 444–451. [CrossRef]
31. Tähtinen, K.; Lappalainen, S.; Karvala, K.; Salonen, H. A comprehensive approach to evaluating the urgency
of iaq measures. In Proceedings of the 15th Conference of the International Society of Indoor Air Quality &
Climate (ISIAQ), Philadelphia, PA, USA, 22–27 July 2018.
32. Lahtinen, M.; Sundman-Digert, C.; Reijula, K. Psychosocial work environment and indoor air problems:
A questionnaire as a means of problem diagnosis. Occup. Environ. Med. 2004, 61, 143. [CrossRef]
76
Appl. Sci. 2019, 9, 99
33. Bakke, J.; Moen, B.; Wieslander, G.; Norbäck, D. Gender and the physical and psychosocial work
environments are related to indoor air symptoms. J. Occup. Environ. Med. 2007, 49, 641–650. [CrossRef]
34. Finell, E.; Seppälä, T. Indoor air problems and experiences of injustice in the workplace: A quantitative and a
qualitative study. Indoor Air 2018, 28, 125–134. [CrossRef]
77
applied
sciences
Article
Combined Model for IAQ Assessment:
Part 1—Morphology of the Model and Selection of
Substantial Air Quality Impact Sub-Models
Michał Piasecki * and Krystyna Barbara Kostyrko
Department of Thermal Physic, Acoustic and Environment, Building Research Institute, Filtrowa 1,
00-611 Warsaw, Poland; [email protected]
Received: 17 July 2019; Accepted: 9 September 2019; Published: 18 September 2019
Abstract: Indoor air quality (IAQ) is one of the most important elements affecting a building user’s
comfort and satisfaction. Currently, many methods of assessing the quality of indoor air have
been described in the literature. In the authors’ opinion, the methods presented have not been
collected, systematized, and organized into one multi-component model. The application purpose
of the assessment is extremely important when choosing IAQ model. This article provides the
state-of-the-art overview on IAQ methodology and attempts to systematize approach. Sub-models of
the processes that impact indoor air quality, which can be distinguished as components of the IAQ
model, are selected and presented based on sensory satisfaction functions. Subcomponents of three
potential IAQ models were classified according to their application potential: IAQ quality index,
IAQ comfort index, and an overall health and comfort index. The authors provide a method for using
the combined IAQ index to determine the indoor environmental quality index, IEQ. In addition,
the article presents a method for adjusting the weights of particular subcomponents and a practical
case study which provides IAQ and IEQ model implementation for a large office building assessment
(with a BREEAM rating of excellent).
Keywords: indoor environment quality; IEQ; PPD; IAQ; TVOC; BREEAM assessment; occupant satisfaction
1. Introduction
1.1. State-of-the-Art Indoor Air Quality Measurement Systems

Approximately 30 years ago, people began to realize that buildings not only provide them with
a sense of security, but can also significantly affect their health and well-being. This is particularly
important due to the fact that people spend an increasing amount of time in closed indoor environment.
Air quality and ventilation approaches were initially based on the users’ dissatisfaction with the scent
of the human body and, as such, the understanding of indoor air quality (IAQ) had serious limitations.
Large quantities of pollutants and their sources clearly influence the indoor comfort of building
inhabitants, as well as their health. In 1998, Fanger [1] presented an approach to the quantitative
determination of perceived IAQ based on the level of dissatisfaction of residents caused by bad odors
and irritants, smoke, and other sources of pollution. This approach provided two new measures of
IAQ: the olf, which quantifies the pollution generated from a strong source of human bio-pollutant
in the range of the impact of emitted odors on perceived air quality, and the decipol, measuring the
perceived air quality in an indoor space with a source of pollution of one olf at a ventilation rate of 10 l/s.
The number of emitted olfs per floor unit in different types of buildings and the amounts of pollutants
from tobacco smoking (in olfs) can then be determined. Consideration of only the odour of the human
body, without taking into account the influence of pollutants from various other sources (for example
emissions from construction products), was very limited. At this stage, the determination of IAQ did

Appl. Sci. 2019, 9, 3918
not take into account the significant differences among contaminants and did not distinguish their
specific impacts on health or comfort. The study of emissions undetectable by the senses (such as
carbon monoxide and other pollutants that affect health at concentrations below their odor threshold),
together with health-effects thresholds, has become particularly important. The range of air pollutants
that should be considered as IAQ components is very difficult to determine, because the composition
of pollutants constantly changes due to the fact of their dynamic nature, secondary reactions, sorption
processes, and other physical and chemical phenomena occurring in indoor environments, thus it
cannot be inherently determined, as illustrated in Figure 1.
Figure 1. Overview of the physical and chemical processes of potential pollution sources in the indoor
environment of a building. (VOCs—Volatile Organic Compounds, SVOCs—semi Volatile Organic Compounds).
Smoking alone emits more than 7000 different compounds, many of which are harmful [2] for humans
and animals and may transport biological pollutants that can act as allergens. People and household animals
emit gases which are unpleasant, transfer pathogens, and cause diseases. These examples show that there are
many paths for penetration of and exposure to the sources of pollution in indoor environments.
In connection with the growing need to determine levels of indoor air pollution, new centers
performing tests and new methods have been created considering the ability to analyze an increasing
number of harmful substances. Measurement of pollutant concentrations in the air is generally a
task performed by experts mainly in accredited laboratories and the results are published in scientific
journals, technical reports, and, eventually, in guidelines, e.g., those of American Society of Heating
Refrigerating and Air Conditioning Engineers ASHRAE [3]. The presence and concentrations of
pollutants are often detected and measured without careful consideration of the significance of these
measurements, and the pollutants measured may not be the most widespread or the most harmful.
Some emissions are incorrectly grouped together; for example, more than one million volatile organic
compounds (VOCs) are known and their toxicities are generally unknown, but they are often reported
as a single value and referred to as the total VOCs (TVOCs) component. Frequently, carbon dioxide is
used as an indicator of IAQ, although it does not have such a negative effect on the health of residents
in the concentrations in which it is usually found in buildings. In our opinion, CO2 is rather a marker of
human bioeffluents. Examples of different understandings of the set of typical pollutants in an indoor
environment are shown in Table 1 [4]. This table provides recommended values from the results of the
European project HealthVent [5], which aimed to develop health-based ventilation guidelines. Table 1
also includes recommendations provided by the Word Health Organization (WHO) on the acceptable
levels of pollutant concentrations [6,7], as well as recommendations from other organizations, such as
China’s IAQ standard values [8]. Different approaches to the IAQ issue mean that the exposure limits
assumed in the various source materials differ.
80
Table 1. Acceptable levels of pollutant concentrations occurring in indoor air, according to World Health Organization (WHO) recommendations supported by the
results of the EU research program HealthVent, values recommended by the standard EN 16798-1:2019 (replacing EN 15251:2007), values recommended by the
Chinese indoor air quality (IAQ) standard GB/T 18883-2002 and for acceptable levels of pollution for the certification of office buildings in Hong Kong.
WHO Guidelines for

HealthVent Project IAQ Standards for IAQ Certification
Pollutant IAQ with Updates EN 16798-1:2019 [10]
[5,9] China [8,11,12] Hong Kong [13]
Appl. Sci. 2019, 9, 3918
[6,7]
<500 ppm beyond <1000 ppmv
CO2 485 ppm
outdoor level 1800 mg/m3 (8 h)
100 mg/m3 (15’)
100 mg/m3 (15’)
35 mg/m3 (1 h) 10 mg/m3
CO 19 mg/m3 (8 h) 35 mg/m3 (1 h) 1 mg/m3
10 mg/m3 (8 h) 7000 μg/m3 (8 h)
7 mg/m3 (24 h)
7 mg/m3 (24 h)
Formaldehyde
0.1 mg/m3 (30’) 0.03 mg/m3 (30’) 0.1 mg/m3 (30’) 10 μg/m3 <0.1 mg/m3 (8 h)
HCHO
<outdoor concentra No safe level can be
Benzene >0.17 mg/m3 0.11 mg/m3 17 μg/m3
tion determined
81
40 μg/m3
40 μg/m3 (1 year) 40 μg/m3
NO2 (1 year) 10 μg/m3 150 μg/m3 (8 h)
200 μg/m3 (1 h) (1 week)
200 μg/m3 (1 h)
SO2 20 μg/m3 (24 h) 20 μg/m3 (24 h) 20μg/m3 (24 h)
0.01 mg/m3 0.01 mg/m3 0.01 mg/m3
Naphthalene 0.01 mg/m3 (8 h)
0.02 (1 year) (1 year) (1 year)
Trichloroethene >2.3 μg/m3 230 μg/m3 (8 h)
0.25 mg/m3 0.25 mg/m3 0.25 mg/m3
Tetrachloroethene 0.25 mg/m3 (8 h)
(1 year) (1 year) (1 year)
Respirable 15 μg/m3 (1 year)
10.0 μg/m3 10 μg/m3
particulate matter 35 μg/m3 (24 h) 100 μg/m3 (8 h)
(1 year) (1 year)
PM2.5
No safe level can be
PAH * >0.012 ng/m3 1.2 ng/m3 (8 h)
determined
TVOC ** 1000 μg/m3 600 μg/m3 600 μg/m3 (8 h)
* PAH Para-Aminohippuric Acid—cyclic aromatic hydrocarbons; ** TVOC Total Volatile Organic Compounds; Reference [13] provides additional TVOC certification tests for new office
buildings by determining (at the ppbv level) the content of carbon tetrachloride, chloroform, 1,2- and 1,4-dichlorobenzene, ethylbenzene, toluene, and o-, m- and, p-xylene.
Appl. Sci. 2019, 9, 3918
Theoretical work on a combined IAQ model allowing aggregation of the results of the assessment
of components affecting humans [14] is not yet well recognized in the literature. However, studies on
IAQ indicators, which aim to provide a quantitative description of indoor air pollution, have been
conducted since the nineties. In 2003, a significant study by Sekhar et al. [15] was published related to
the standard indoor pollutant index (IPSI), the disease symptom index in the building symptom index
(BSI), and to the often-cited works by Moschandres and Sofuoglu [16,17] on the indoor environmental
index (IEI), indoor air pollution index (IAPI), and the indoor pollutant standard index (IPSI). The IAPI
characterizes air pollution in an office with a single number: the index. The index value ranges
between zero (lowest pollution level, i.e., best indoor air quality) and 10 (highest pollution level i.e.,
worst indoor air quality). The IAPI is a composite index; sub-indices ed are aggregated using the
arithmetic mean in conjunction with a tree-structured calculation scheme. This scheme gives rise to
some reservations, because at the top of the tree-structured calculation scheme is the IEI (calculated as
the arithmetic mean of the IAPI and the IDI (indoor air discomfort index), and the combination of IAQ
sensation and thermal conditions does not appear until later.
While considering the indicators for the quantitative description of pollution, the proposal of the IEA
Working Group named “Defining the Metrics of IAQ” should also be mentioned. This group prepared,
in 2017, the document entitled “In the Search of Indices to Evaluate the Indoor Air Quality of Low-Energy
Residential Buildings” [18]. The group made the following assumption for the categorization of various
indicators: there should be one index per individual pollutant and a dimensionless coefficient should be
specified to evaluate the IAQ, provided that the current (observed) concentrations of a given pollutant
cj are related to the ELVs (exposure limit values) concentration cj,ELV .
cj
IAQindex ( j) = I j = (1)
c j,ELV
The index is calculated for each individual pollutant [18], which is specific only for this exact
pollutant. The report showed that aggregation can be performed by addition, by taking the maximum
value or by other methods, in an attempt to define metrics that can be used to evaluate IAQ.
The assumption was that the reference value usually refers to health risks (accounting for chronic
or acute effects), but other metrics can also be used, (e.g., odor or irritation threshold). There are
two important properties to be considered when aggregating sub-indices: ambiguity and eclipsing.
As a result of the analysis, the authors concluded “that there are problems with model aggregation
methods. In the aggregation model Iagg = I1 + I2 , ambiguity creates a false alarm and in the aggregation
model Iagg = 1/2(I1 + I2 ), eclipsing underestimates the effect” [18]. Therefore, the discussion remains
open [18]. The report also showed how there are large spreads of concentrations of individual pollutants
(up to seven rows), even in the group of pollutants for which sub-indices were built. It determined
the difficulties of building a weighted scheme based on the simplest percentage adjustment of the
concentration shares and, thus, the share of the mass of pollutants to be removed by ventilation.
The current state of knowledge does not provide information authorizing the omission of certain
pollutants. Hence, taking into account the lack of data on the characteristics of each chemical compound
and consideration of the “removal efficiency” [19] requires us to abandon thinking about the adjustment
of many individual pollutants, and to focus only on the creation of a model based on the representative
and target components. In this state of knowledge, there are hopeful studies and proposals with a grey

combined IAQindex model and the grey clustering model for IAQ indicators proposed by Zhu and
Li in 2017 [20] is particularly interesting, especially when the relationships between system factors
and the system’s IAQ behavior and the interrelationships among the factors are uncertain. At first,
all specific indoor air pollutants and related parameters should be measured. However, this is a
very complex and time-consuming process. On the basis of the characteristics and correlations of
the pollutants, the indoor air quality can be characterized by representative indicators. Studies [20]
have pointed out that respirable particulates, CO2 and TVOCs, were the three most representative
and independent environmental parameters which can be used as an evaluation index of indoor air
82
Appl. Sci. 2019, 9, 3918
quality in office buildings. Since each indicator represents a class of pollutants with similar sources and
dissemination characteristics, this index group avoids unreliability due to the fact that these indicators
are “too small” because of critical concentration depression. A data pretreatment method must be
used in the calculation procedure, reflecting the differences in concentration levels among different
pollutants, but also expressing their influence on the comfort and health of the indoor occupants.
Moreover, the measured pollutant concentrations can be used to predict the probable levels of other
parameters, and good agreement was found between the predictions and measured values.
1.2. The Research Questions

The main research question contained in the paper concerned whether it was possible with the
current state of knowledge to create and use in practice an IAQ model that was based on a unified and
coherent approach for input indoor air parameters (such as pollutant concentrations, odor levels, and
moisture content) and provided one output parameter (we proposed occupant satisfaction, IAQindex
(in %)). The authors looked for physical equations for the IAQindex ’s subcomponents and dependencies
for their predicted occupant satisfaction functions with a pollutant concentration cj (PD = f (cj ) in %)
which could be used as a model for subcomponents.
This paper’s intention was to provide an IAQ model with a step-by-step process which can
be used to determine the value of the overall indoor environmental quality index (in %) including
another three components: thermal comfort, acoustic comfort, and lighting quality. The innovative
approach and added value of this article is in the use of the proposed IAQ model in practice and
the relatively simple calculation of the overall IEQ value (with an uncertainty estimation) using the
actual results of measurements in the Building Research Establishment Environmental Assessment
Method BREEAM certified case study office building. The authors also provided occupant satisfaction
functions for CO2 , TVOCs, and formaldehyde HCHO in two variants: with experimental %PD values
taken from the literature and for these pollutants’ %PD values converted from an Air Quality Index
system (see Section 2.2.).
2. Methods
2.1. Research Content and Strategy

The proposed IAQ model is presented in Sections 2.2–2.7. The model is later used to analyze the
case study of an office building described in Section 2.9. Figure 2 presents the subsequent research
steps from theory to practical application. Section 2.8 shows the method for determining indoor
environmental quality IEQindex where IAQ index is a subcomponent/part of the IEQindex model.
In order to determine the IAQ and IEQ, physical measurements of the indoor environment in the
building were conducted using the experimental approach provided in Section 2.10. Based on these
physical indoor measurements the IAQ and IEQ indexes (number of occupants satisfied with the indoor
air and overall indoor quality, respectively) were assessed (see Section 3) and discussed (see Section 4).
83
Appl. Sci. 2019, 9, 3918
Figure 2. To determine the indoor air quality and indoor environmental quality (IEQ) indexes for the
case study. (TC—thermal comfort, L—light quality, ACc—acoustic comfort).
2.2. The IAQ Model Proposal—Basic Assumptions

In the IAQ model construction process (our proposal), the commonly accepted approach is to
transform individual concentrations of pollutants into subcomponents before they are aggregated into
a single index (occupant satisfaction in %). However, summation of sub-indices can lead to situations
in which all are under individual health thresholds, but the final indicator shows when the threshold
has been exceeded. Conversely, the averaging of partial sub-indices can lead to an overall indicator
showing an acceptable IAQ, even though one or more partial indicators are larger than their individual
thresholds. One solution is to use the maximum value of all sub-indices to create the final form of the

IAQindex . Taking these issues into consideration, the authors created the IAQ model with three
complication levels adapted to the purposes of potential applications of the model, as presented in
Figure 3;
i. Certification of a building, e.g., via the BREEAM system using (three sub-indices), called
“quality”;
ii. Design, including perceptible contaminants affecting comfort and using the IAQ index when
calculating the IEQ (five sub-indices), called “comfort”;

iii. Complex design, with the IAQindex representing both comfort and health (seven sub-indices)
called “comfort/health”.

Figure 3. IAQ model has three possible levels (i.e., similar to a Russian doll structure) adapted to the
potential applications of the model.

The simplest one, “quality”, is an inner part of the IAQcomfort model and can be used separately
for simple applications with the main purpose of supporting a green building certification, e.g., via the
84
Appl. Sci. 2019, 9, 3918
BREEAM system using three components, i.e., CO2 , HCHO, and TVOC. This model is used later on
the case study of a BREEAM building.

Figure 1 shows the processes influencing the morphology of the IAQ model and Figure 3
provides a list of the pollutants for which three IAQ submodels were built, containing human-perceived
contaminants (IAQquality and IAQcomfort models), but also the IAQcomfort/health model for both
perceptible and imperceptible pollutants, i.e., those that are not perceptible by humans but affect
health and require additional energy for intensive ventilation for health reasons. There are potential
sub-indices, such as IAQ(VOCnon-odorous ) or IAQ(CO) [5,7]. Dust pollutants may have their sub-index
both in the comfort model (if reliable curves of human sensory perception of PM concentrations are
known) or in the IAQcomfort/health model if their health impact is considered to be the dominant feature.
Considering the types of pollutants harmful to health assigned to the sub-indices of IAQ, we only
consider the most important air pollutants (i.e., target emissions) that were given in the WHO guide in
2010 [5,7]. Submodels of processes that impact on air quality in indoor environments, which can be
distinguished as components of the IAQ model, were based on sensory satisfaction functions (index of
occupant dissatisfaction (PD) with the level of air pollution). Subcomponents of the three potential
IAQ models were classified according to their future potential applications: in the assessment of
environmental quality index IEQ (models IAQquality and IAQcomfort ) or in the design of ventilation
taking into account all possible harmful-to-health pollutants (model IAQcomfort/health ). In our opinion,
such systematization creates order and has a practical dimension as presented later on in the case study.
The following are the target pollutant groups:

i. In the air quality model (IAQ)quality , the IAQ index subcomponents were assigned to the
selected three pollutants. The submodels for the IAQ were CO2 , TVOC, and formaldehyde
HCHO, as recommended by References [5,10,21,22];

ii. In the (IAQ)comfort model, thepreviously provided simplified IAQquality subcomponents for the
three main pollutants were extended with a set of selected compounds VOCodorous , related to the
collection of IAQ sub-indices (VOCodorous ) with an unknown cardinality, increased appropriately
for the number of dominant pollutants. In addition, we provided a conditional deluge of
two more components: (1) calculated using the enthalpy of hot and humid air (high enthalpy
h > 55 kJ/kg [23]), the percentage of persons dissatisfied with respiratory cooling with
humid air at relatively high temperatures and (2) the percentage of persons dissatisfied
with indoor pollution with respect to dust pollution (PM10 and PM2.5 ), measured via panel
tests. The introduction of a dust-pollution subcomponent to the IAQ model may be debatable,
because some experimenters [24,25] underline the unique results of sensory tests of discomfort
from dust, and the influence of “emissions” of respiratory dust particles on satisfaction is still
under-researched. Considering the above, we expected two variants of the comfort model:
with PD (PM10 , PM2.5 ) or without this factor;

iii. In the overall IAQ model, comfort (IAQ)comfort and health risk (IAQ)health indicators were
used, and, hence, this model was called (IAQ)comfort/health . Models for subcomponents of IAQ
not perceived by humans but influencing health, can be borrowed from the index set in the AQI
(air quality index) system [26–28], which was adapted to assess the quality of indoor air based
on, and in accordance with, the concepts of the air quality assessment system used globally by
the American EPA.
Values of AQI indices published on active EPA websites using the air quality index system were
introduced for application in US federal regulations in 1999 [28]. Currently, the AQI system for outdoor
air includes the following pollutants: ozone, particulate pollutants (PM10 and PM2.5 ), carbon monoxide
CO, sulfur dioxide SO2 , and nitrogen dioxide NO2 . To convert a specific air pollutant concentration
to an AQI, the EPA developed a tool called the AQI Calculator, which is an open resource [29].
This system (referring to the index from 2004 [16] and the indoor pollutant standard index (IPSI))
was further developed, and the proposed IAQI for indoor air presented by Wang et al. [30] in 2008
85
Appl. Sci. 2019, 9, 3918
and a newer proposal [27] from 2017 for a similar but narrower set of indices, also for indoor air,
were both modeled on it. The AQI and IAQI indicators showed an increase in the level of impact on
human health with increasing concentrations of air pollution. There are some detected difficulties
here, since “AQI is a piecewise linear function of the pollutant concentration” [27]. The calculated
values of the AQI [31] or IAQI [30] indices, over the entire 0–500 scale calculated from the measured
concentrations of selected contaminants or in the part of the scale corresponding to the IAQ rating,
ranged from “good” to “unhealthy”, and can be converted to PD% for use in the model equation,
(IAQ)comfort/health . Concentrations will be significant when the uncertainties of scale conversions are
estimated. Authors believe that their way of converting the AQI scale to PD% (which is similar to the
method of conversion of the IEQ components’ ordinal scales from the OFFICAIR EC project [32] to
PD% scale; for example, the occupant percentage dissatisfied with noise [33,34] should be accepted in

light of the expected results of a metrological analysis of the reliability of the combined IAQ model.
Subcomponent models (physical functions, PD%) of the IAQ model for all individual air pollutants
are presented later in this section.

2.3. IAQ Model Weighting Scheme Considering Air Pollution Ventilating
To obtain a comprehensive picture of IAQ in a building, it is necessary to measure the number of
pollutants with different individual concentrations. There are methods that weigh sub-indices [21] but
the problem is finding an effective weighting scheme and understanding how to adjust them in the

overall model of all the pollutants in IAQ. For this reason, we proposed an adjustment method for the
weights. In our opinion, provided in detail in References [33,34], and also according to Reference [22],
the best weighting scheme, which would lead to a credibly aggregated model of IAQ composed
of many extractable components (sub-indices), would be a system based on concentration values
(the “excess masses” of pollutants to be regarded as loadings for the ventilation system). Therefore,
the we aimed to determine the individual pollutants assigned to the IAQ model, their concentrations,
cj , as the inputs of the IAQ submodels, and their “excess concentrations” originating from emissions
or determined within indoor environments. Thus, it was possible to determine directly the energy
requirements for ventilation purposes and the required minimum global ventilation rate. Determining
the input concentration value, cj , for each IAQ sub-index enables the determination of the total mass of
pollutants in the air, which is the basis for determining the air change rate N1, . . . 7 (overall air change
rate), assuming that the model includes all significant IAQ pollutants.
Currently, according to References [35–37], the most common assumption made is that pollution
from VOCj compounds arises only from emissions due to the presence of construction or finishing
materials (for j = 1, 2, . . . n) (it can be assumed that the source i of an emission is the entire indoor
environment and then i = 1) from the zero state. The physical model for determining the ventilation
rate in indoor environments polluted with VOC-type pollutants from building materials is given by EN
16798-1:2019 [10], assuming that design parameters for indoor air quality are derived using limit values
for substance concentrations. In accordance with ECA Report Number 11 [36], the design ventilation
rate required to dilute an individual substance emitted from building materials is calculated as:
Gh 1
Qh = × (2)
Ch,j − Ch,0 εv
where Qh is the ventilation rate required for dilution in m3 per second, Gh is the emission rate of the
substance in micrograms per second, Ch,j is the guideline value of the substance in micrograms per
m3 , Ch,o is the concentration of the substance in the supply air in micrograms per m3 , and εv is the
ventilation effectiveness.
In fact, in a building with active “indoor chemistry” (see Figure 1), the use of this formula seems to
be increasing. Taking into account the dynamic nature of the processes of generating various pollutants,
the approach to IAQ and its components should be changed and subcomponents should be treated as
pollution load processes, increasing in number not only due to the emission processes but also due to
86
Appl. Sci. 2019, 9, 3918
the generation of bio-pollution, water evaporation, and even dust infiltration from outside. It is also
possible to set steady-state (initial) concentrations of pollutants and to determine the expected time

courses of removal of these pollutants by means of ventilation (curves cj = f(τ)), at constant values of
−1
air change rate per hour ACH (h ). Such ventilation rate calculations for CO2 were developed in 1997
by Persily [38] and similar ones were provided in 2017 by Gyot [39] at the Berkeley National Laboratory.
These calculations are not very accurate, as shown in the general demonstration graph in Figure 4.
Figure 4. Of CO2 above outdoors levels with two people in the contaminated building (ACH—air
change rate) and in a typical clean office room (ACH*).
Less accurate time-dependent curves of the total minimum ventilation rate ACH needed for
“contaminant exhaustion” can be determined using programs [40] based on generic engineering

equations for the sum of pollutants cj . The generic equation for pollution concentration (the ratio
of the amount of polluting product to the amount of fluid in the space (such as air in a room) can be
calculated from the following equation:
c = q/(n·V·(1 − e−Nt )) (3)
where c is the pollution concentration in the space (or in the room) with perfect mixing (m3 /m3 ) or
(kg/kg), q is the amount of pollution added to the space (m3 /h) (kg/h), N is the air change rate per hour
(h−1 ), V is the volume or mass of the space (m3 ) or (kg), e is the number 2.72, and t is time (h). If the
initial concentration (at t = 0) in the space and the concentration in the supply fluid is zero, after some
time the concentration in the room will stabilize. The ventilation rate graph for an amount of pollution

q = 1 and a volume of space V = 1, shows the values of cj , similar to Figure 4. In order to obtain
more exact values of the VOC concentrations remaining in the room for the air change rate function,
it is possible to use the published dependencies or for the assumed volumes V of ventilated spaces
with determinate concentrations, as they can be determined experimentally. A simplified method for
determining the ventilation rate N from a simple formula for the time, t, course of a trial ventilation
was provided by the Japanese researchers Noguchi et al. [41] in 2016. A description of this method is
worth reading. Based on the temporal changes of the TVOC concentration measured using a PIDTVOC
meter [42], the air change rate N or the ventilation rate F was estimated using the following method.
Assuming perfect mixing of the air in the room and a constant TVOC emission rate E, the concentration
change of TVOC in the room can be expressed by the following equation:

E F E
C(t) = C j + 1 − e− V t = C j + 1 − e−Nt (4)
F F
87
Appl. Sci. 2019, 9, 3918
Finally, Equation (5) can be expressed with one unknown parameter, the air exchange rate N as:

Cst − C j
log = Nt (5)
Cst − C(t)
The initial concentration cj can be determined from the experimental results. After a long time,
when the exponential term in Equation (5) can be assumed to be zero, the concentration C(t) becomes
constant. The steady-state concentration Cst can be determined from the temporal change in the
experimental results where the concentration levels off.
The rule that the IAQ model should include a weighting scheme, referring to the variation in the
share of pollutants in the IAQ, has been noted in References [35,37]. According to the first proposal,
the weighting system is based on the differentiation of coefficients Rj , which are the ratios of the real
concentration values (or mass of pollutants) to the values of reference concentrations, representing the
so-called relative masses of non-eliminated pollutants. This can also be represented by the desirable
reduction in the level of pollution by means of ventilation and, thus, also by the energy requirements.
For one emission source, the proposed system with a coefficient Rj for a given pollutant Cj has the
formula:
yj
R j = , j = 1, 2, . . . m (6)
Ij
where Ij is the ratio of the gas-phase concentration to the reference concentration value. For example,
for the LCI (Lowest Concentration of Interest) value for the jth compound emitted from the building
material, the factor Rj is dimensionless, since yj is the gas-phase concentration for the jth compound in
μg·m−3 , Ij is the lowest concentration of interest (LCI) [41] for the jth compound in μg·m−3 and m is the
number of all elected compounds. The weight coefficients Wj are used as the weights of the equations
in the IAQ index, according to:
Rj
Wj = j = 1, 2, . . . . . . m (7)

m
Rj
j=1

The authors of Reference [35] justified adjusting the coefficients where Rj ≤ 1, but they did not
explain the physical meaning of this condition. We believe that further discussion should include the
issue of whether the “relative mass” of contamination expressed by Equation (6) has a proper place
in the weighting scheme for the equations. The dimensionless quantity (7) does not have a sound
physical meaning [37]. In our opinion, this type of calculation method is debatable.
One should strive to cover all the sub-indices of the combined IAQij model with a weighting
scheme that would give VOCs a share in the total energy requirements for ventilation. The term
“relative mass” should correspond to weights rationally proportional to the energy expenditure for
ventilation of individual pollutants (IAQ sub-indices). Therefore, our future work will focus on
introducing weights for all expressions in the overall IAQ model equation. However, since this
is currently not possible without an adjustment method adapted to weight determination for very
small concentrations, it was decided to present our model as an interim solution. We proposed the
use of weights based on the “excess concentration” values within the pollutant categories only with
similar and comparable orders of concentration values, for example the VOCodorous and VOCnon-odorous
categories. The rules of adjustment with boundary conditions will be provided and justified in the
follow-up article to this report.

2.4. IAQ Model Scheme Morphology
According to the new proposal, for models with air quality sub-indices IAQ(Pj ) (developed with
the standard EN 16798-1:2019 as a reference for IEQ model creation [22,33] and with assumptions
88
Appl. Sci. 2019, 9, 3918
described in Reference [34]), in the case where indoor air has many pollutants, P1 , . . . j, the combined
ΣIAQindex equation is:
ΣIAQindex = WP1 ·IAQ(P1 )index + WP2 IAQ(P2 )index , . . . , WPj IAQ(Pj )index (8)
where the WP1 , . . . Pj weighting system for IAQ components is created on the basis of the arithmetic
mean and the concept of “excess concentration” is introduced only for groups of pollutants with similar
concentration values. There is a difference in concentration Δcj between the observed concentration of
pollutant cj and the reference concentration cref (cELV or cLCI ), which is below the current concentration
in contaminated rooms. Thus, the excess concentration is:
Δcj = cj − cref (9)
The weights W 1, . . . j for all three IAQ models are determined on the basis of arithmetic means or
by adjusting all the values of Δcj in a given model using Equation (10):
Δc j
W j (IAQcom f ort/health ) = (10)
Δc1...7
j=1...7
where the sum of the adjusted weights Wj of all ventilated pollutants described with sub-indices should
be unity. The weight values for a given IAQ model, (e.g., IAQcomfort ) may be different, but the sum of
the sub-index weights must be ≤1.0.
The values of the reference concentrations are the concentration levels that are acceptable or

recommended as limit values for various pollutants Pj . In the case of the IAQquality submodel as
part of the IEQindex model, weights should be used for the VOCodorous (HCHO and TVOC) reference
threshold concentrations of odors. The weights in the weighting system should be adjusted to unity
according to the Equation (11):
Δc j (c j − cre f )
WHCHO = = (11)
Δc2,3 (cTVOC − cre f ) + (cHCHO − cre f )
j=2,3
There are, however, non-typical cases in which the scales have different values. This is the case
for formaldehyde, the concentration of which is many times lower in the building than the threshold
level cth . According to the WHO [7], the admissible value cref is also higher than the concentration
in the building. In this case, the authors recommend taking the reference value as zero. Then, the
weight WHCHO described by Equation (11) (for two pollutants), would not be negative (cHCHO − cref ).
The ASHRAE Guideline 10 (2011) [3] recommends that the IEQ model (and appropriate weights, Wi )
should contain synergy effects of environmental parameters included in the subcomponents and their
sensory perceptions.
Figure 5 shows the extended IAQindex model with its sub-indices treated as components of the
IEQindex , but also with sub-indices of the IAQcomfort/health type, i.e., pollutants that do not belong to
the IEQ model but are important to health and the energy balance of a building with a mechanical
ventilation system. The experimental dependencies of the percentage of persons dissatisfied, %PD,
and the values of the concentrations of pollutants, cj , sensed in indoor air in the appropriate ranges are
of fundamental significance in the sub-indices relevant to the IEQ model [43].
89
Appl. Sci. 2019, 9, 3918
Figure 5. ΣIAQindex model with weighting scheme.
From the dependencies, expressed as the curves for PD(CO2 ) or PD(VOCodorous ), the equations of
the models are derived Equations (12)–(14):
ΣIAQquality = W1 ·IAQ(CO2 )+ W2 ·IAQ(TVOC) + W3 · IAQ(HCHO) (12)

IAQcomfort = W1 ·IAQ(CO2 ) + W2 ·IAQ(TVOC) + W3 ·IAQ(HCHO)
(13)
+ W4 ·IAQ(VOCodorous ) + W5 ·IAQ(h)

IAQcomfort/health = W1 ·IAQ(CO2 ) + W2 ·IAQ(TVOC) + W3 ·IAQ(HCHO) + W4 ·IAQ(VOCodorous )
(14)
+ W5 ·IAQ(h) + W6 ·IAQ(PM2.5 , PM10 )+ W7a ·IAQ(VOCnon-odorous ) + W7b ·IAQ(CO) + W7c ·IAQ(NO2 )

The scheme of the IAQcomfort/health model consists of seven (or more) components or IAQ
submodels and these are models for the various types of pollutants: IAQ(CO2 ), IAQ(TVOC),
IAQ(HCHO), IAQ(VOCodorous ), IAQ(h), IAQ (PM2.5 , PM10 ), and the selected IAQ(VOCnon-odorous ).
The IAQ(VOCodorous ) and IAQ(VOCnon-odorous ) models should be multiplied, depending on the number

of dominant VOC pollutants, and, hence, the IAQcomfort/health model will, in practice, have more than
seven components.
The inputs of each IAQ submodel are unit concentrations in air of a given pollutant,cj (in the case
of IAQ(h). This is the moisture content x in well-known units “g of water vapor (gw ) per kg of dry
air (kga )”, converted to a concentration of cj in μgwater /m3 (or H—absolute humidity in gw /m3 which
is a measure of water-vapor density). In some cases, it is necessary to convert the pollution-derived
parameter to VOC concentration (conversion of the odor intensity OI to VOC concentration is described
later in this section).
90
Appl. Sci. 2019, 9, 3918
From the concentration values, total air pollution can be calculated, and, subsequently, also the
energy needed to ventilate the indoor air pollution. When the concentration levels of pollutants are
variable and are increasing due to the presence of emissions, then the formulas given in Reference [36]
and the amended standard EN 16798-1:2019 are used to calculate the required ACH ventilation rate.
When the level of contamination is set (or quasi-fixed) and the volume and other parameters of the
ventilated room are known, it is possible to calculate ventilation-time curves, i.e., maximum ventilation
curves for ACH ventilation rate to reach concentration levels ELV, LCI or the olfactory threshold level,
according to Reference [40] or another adequate equation.
There are two outputs of each IAQ submodel as described below:

1. The weights of the weighting system for the model IAQquality and hypothetically W 1, . . . 5 for

the model IAQcomfort or W 1, . . . , 7 for the model IAQcomfort/health (in a hypothetical model with
a set adjustment method). These should reflect the energy load of the IAQ expressed by the
theoretically assumed increase in the current concentration of pollutant cj relative to the reference
concentration of cj ,ref , which determines the level of this concentration intended to be obtained
by ventilation.
2. The PD% with the IAQ as a function of air pollution concentration. These values, determined in
panel tests, reflect the impact of the interaction of air with a given pollutant at the actual level of
concentration, estimated via panelists’ sensations/perceptions (PD = f(cj ) in %).
Examples of measurable physical parameters for the purpose of IAQ and IEQ calculations (see case
study) are given in the following section.

For the construction of the combined model IAQindex with a weighting scheme useful for
aggregating sub-indices, we proposed the model presented in Figure 5. In this scheme, the combined
IAQ model is shown as the basic assumption for the aggregation of all sub-indices. First, the model
was cut by a cross-connected vertical connection regarding the inputs of submodels—the calculation

of the sum of the masses of all pollutants cj in the ventilated space, using the values for the inputs
of all submodels of IAQ concentration values of contaminants. The sum of the concentrations of
all air pollutants expressed as mass units of pollution per m3 of volume (which can be read after
multiplication by V (m3 ) as the mass to be displaced by ventilation), is the basis for calculating the “air
change rate per hour” (ACH), the minimum air exchange rate needed to reduce the observed mass
level of air pollution in a ventilated room (see Reference [40]). The second connection concerns the
submodel outputs—the conversion of excess concentration to a dimensionless value, which allows

to for the weighting scheme of the IAQindex combined model and the weights of individual IAQ
submodels to be determined.
Additional assumptions were as follows:

i. The sum Δc1, . . . 7 , which admittedly constitutes an excess mass increase of the sum of pollutants
described by the submodels, was treated only as a “virtual energy load of the building” for
ventilation conducted for the elimination of pollutants, and therefore, when adjusting the
weights, the possibility of dividing the excess concentration by the sum of concentrations
should be considered.
ii. The percentage of persons dissatisfied PD(IAQcomponent ) determined experimentally in sensory
studies using panelists’ sense of air quality during their exposure to an internal environment
deteriorated by a given contamination component (PD = f (cj )), was derived from the literature
or direct experiments.
91
Appl. Sci. 2019, 9, 3918

Values of weights W 1, . . . j in sets of three, five or more components of the three IAQindex models
(see Equations (12)–(14), were adjusted to a value of unity by dividing Δc1, . . . , j of each component by

the sum of excess concentrations Δc1, . . . j in μ/m3 . We proposed the use of cref values, apart from the
values of cLCI [18], cELV , and the threshold values cth for odorous compounds, were as follows.
i. For IAQ(CO2 ) and IAQ(HCHO), the cELV concentrations were derived from EN 16798-1:2019 [10];
ii. For IAQ(TVOC) and IAQ(VOCodorous ), the threshold concentrations, cth , for identified odorous
compounds or mixtures are from Reference [44];
iii. For the IAQ(h), the water-vapor concentration H (gw /m3 ), recalculated from the moisture
content x (gw /kgdry air ) using the gas constant for the water vapor and the actual temperature,
the value of h up to the critical value for “high enthalpy of humid air” was evaluated using
the formula:
h = 1.006ta + x·(2501 + 1.805ta ) (15)
where h is the specific enthalpy of humid air (kJ/kg), which must be >55 kJ/kg. The EN
16798-1:2019 standard [10] recommends a limit for the dehumidification of air of 12gw /kgdry air
(this value must be converted to a c value in gw /m3 ).
i. For IAQ(VOCnon-odorous ), the cELV value in cases where no established LCI values were derived
from the EN 16798-1:2019 standard.
ii. For IAQ(PM2.5 , PM10 ), the cELV values were derived from the WHO [7] or other organizations
(Tables 1 and 2).
The proposed reference values of pollutants forming the sub-indices of the IAQ model are
given in Table 2. With reference to the concentration values of cLCI , it should be noted that
according to References [18], this value is typically acquired by dividing occupational exposure
limits by a safety factor (100 or 1000). Concentration cLCI is taken from Lowest Concentration
of Interest (EU-LCI) from European Commission lists.
However, the model values for exposure limit values (ELVs) of indoor air pollution, in accordance
with the recommendations of the health-based ventilation guidelines [5], should be adopted in
accordance with the current WHO guidelines given in the periodically issued WHO Air Quality
Guidelines [7].
2.5. Selection of Submodels for Pollutant Components

Our overall selection of physical subcomponent equations and dependences for %PD = f(cj ) is
presented in Table 3. The models presented were used to determine the IEQ for the sample building.
The highlighted pollutants were taken into account in the case study building assessment.
92
Table 2. Exemplary pollution concentration reference values for low-pollution building.
Component of Pollution Pj Reference Concentration Reference Value Reference/Recommendation List

CO2 (outdoor 350 ppm) cELV 380 ppm EN 16798-1:2019
cELV <300 μg/m3 EN 16798-1:2019
TVOC
cth 50 μg/m3 MV. Jokl [45]
Appl. Sci. 2019, 9, 3918
30 μg/m3 (30 min) EN 16798-1:2019

cELV
100 μg/m3 (30 min) WHO Guidelines for IAQ (2010)
9 μg/m3 (1 year) IEA-AIVC Report. Annex 68 (2017)
HCHO
or
300 μg/m3 AIHA Odor Thresholds for Chemicals with
cth
60 μg/m3 Established Health Standards (2013)
VOCodorous (as in EN 16516)

AIHA Odor Thresholds for Chemicals
Naphthalene cth 10 μg/m3 (1 year) (2013)
Nitrogen dioxide or 20 μg/m3 (1 year)
Ammonia cELV 70 μg/m3 (1 year) EN 16798-1:2019
Ozone 100 μg/m3 (8 h)
Carcinogenic VOCs 5 μg/m3
Moisture content x in indoor air
93
(enthalpy >55 kJ/kg, ta = 23 ◦ C and > unacceptable value of x > 12 g/kg
EN 16798-1:2019
60% RH) moisture content x at 23 ◦ C. RH > 60%
x [gwater /kgdry air ] ⇒ H [gwater /m3 ]
cELV 20 μg/m3 WHO Air Quality Guidelines (2010)
PM10
EU Directive 2008/50/EC introduced
cELV 10 μg/m3 additional PM2.5 objectives targeting the
PM2.5
cDIRECTIVE 2 25 μg/m3 (1 year) exposure of the population to fine particles.
VOCnon-odorous (EN 16516 includes

pollutants with limit values on
concentration that have been cELV or EN 16798-1:2019
identified): cLCI EN 16516;
Carbon monoxide 7 mg/m3 (24 h)
Carcinogenic 1 VOCs WHO Guidelines for IAQ (2010)
5 mg/m3
1 Limit values and carcinogenic effect: the level of PAHs, particles, benzene, and trichloroethylene should always be kept as low as possible.
Table 3. Selection of physical equations for IAQindex components and dependences for PD = f(cj ).
Sub Input Sensory Equations

References
Component Parameters to Calculate the %PD Value
PDIAQ(CO2 ) = 395·exp(−15.15·CCO2 −0.25 ) [36]
cIAQ(CO2 ) = 55 833·(ln(PD) − 5.98)−4 [45]
Appl. Sci. 2019, 9, 3918
c(CO2 )
CO2 IAQ = 6 × 10−0.07 c2CO − 0.0025 × cCO2 + 1.9416
2
[46]
(in ppm) cco
PMVCO2 = 6.364 × log 4852 [47]
[48]
PD = 100 − 95·exp(−0.03353·PMV 4 − 0.2179·PMV 2 )
cTVOC = 16,000(ln(PD) − 5.988)−4 ;
c(TVOC)
cTVOC (μg/m3 )
⇒ [49]
PD(IAQ)TVOC = 405·exp(−11,3·cTVOC −0.25 )
PD [22,45]
TVOC 1 In Taiwan EPA Indoor Air Quality Index System (IAQI). Nine major indoor air
pollutants are included: PM10, PM2.5, CO2 , CO, O3 , HCHO, TVOC, bacteria,
IAQI⇒ [30]
and fungi. We proposed the use of a calibration curve IAQI = f(cTVOC ) for
PD*
conversion to PD* %
c(HCHO)
[47]
94
PMVHCHO = 2 log cHCHO
0.01 , cHCHO (mg/m3 )
4
PD = 100–95·exp(−0.03353·PMV − 0.2179·PMV 2 ) [48]
(pi)2 ⇒ 1
HCHO PDIAQ(OI) = 1
[50,51]
OI3 ⇒ 1+ exp(2.14·OI−3.81 )
PD ——————————————————
6
In the Taiwan EPA Indoor Air Quality Index System (IAQI). Nine major indoor
c(HCHO) air pollutants are included: PM10, PM2.5, CO2 , CO, O3 , HCHO, TVOC, bacteria
⇒ and fungi. We proposed the use of the calibration curve IAQI = f(cHCHO ) for
conversion 7 to PD*% [30]
IAQI⇒
PD*
Table 3. Cont.

References
[52]
Appl. Sci. 2019, 9, 3918
cVOC Formulae for the conversion of odorant concentration cj to odor intensity OI for
⇒OI4 recalculation (4) of the conversion of odor intensity to odorant concentration.
——————————————————
cVOC ⇒ Conversion of the chemical concentrations cj (mg m−3 ) into odor concentrations
OI3 [53]
cOD (ouE m−3 ) and odor intensities OI; cOD,0 – unity of odor concentration (cOD,0
= 1ouE m−3 )
OIj = kj log cOD,j + 0.5
——————————————————
OI3 ⇒
The mean of acceptability votes as a function of the mean of intensity votes. [54]
ACCVOC
SVOC and VOCodoros ACC = −0.45 OI + 0.93, R2 = 0.979
⇒PD
ACC in scale: from −1 to +1
OI odor intensity
in scale: from 0to 5
exp(−0.18−5.283ACC
[55]
95
PDIAQ(OI) = 1+exp(−0.18−5.283ACC 3100
——————————————————
The perceived intensity in pi units is determined by comparing the intensity of
the sample with different specified intensities of the reference substance, (e.g.,
acetone). Concentrations for 1 to n (pi) follow a linear gradation of the acetone
OI2 in scale (pi) concentration. [51]
recalculation to Confidence intervals should be within
⇒OI3 ±2 pi. Recalculation (pi) to OI
[34]
Table 3. Cont.

References
[10]
Appl. Sci. 2019, 9, 3918
The recommended criteria for dimensioning of humidification and

dehumidification. It is recommended to limit the absolute humidity value to x =
12gw /kga or H(gw /m3 ) [56]
h = 1.006t + 0.622(2501 + 1.84t)·
Moisture x in high 0.01·RH·exp(23.58 −4043/(t+273.15−37.58))
Patm+0.01·RH·exp(23.58 −4043/(t+273.15−37.58))
enthalpy h of moist IAQ acceptability equation
x (gw /kga )
air 5 ACC = ah + b
⇒H (gw /m3 )
(in room tempera where a and b are different for different pollutants
(H-absolute
ture ta when air —————————————————-
humidity)
enthalpy 100
%
a v
PD = 1+exp(−3.58+0.18(30−t )+0.14(42.5−0.01p )
h > 55 kJ/kg)
ACC = 5.63 + 0.46 ln RH—1.32 ln h [23]
uncertainty of acceptability ACC is ±0.12.
[57]
96
exp(−0.18−5.283ACC
PDIAQ(h) = 1+exp(−0.18−5.283ACC
3100
[55]
6
IAQI (Indoor Air Quality index) is the proposal for a system comparable to the
Particulate matter 6
cPM2.5 World AQI EPA system
PM2.5 [30]
cPM10 cPM2.5 ⇒IAQI⇒PD*%
PM10
cPM10 ⇒IAQI⇒PD*%
Air pollutants 7
IAQI 6 value for cVOC was calculated with System IAQI [27] by interpolation and
VOCnon-odorous cVOC [30]
after which the index IAQI must be converted 7 to PD*%
CO, NO2
1 TVOC, according to Reference [45], represents a narrow chromatographic picture that excludes, for example, the lower aldehydes, e.g., formaldehyde. 2 The measurement of the
intensity of the odors in the building in which emissions from construction materials occur can be performed by a panel of participants, where the room is treated as a “test room for
background odor” according to Section 6.8.1 of ISO 16000-28:2013, “Indoor Air—Part 28: Determination of odor emissions from building products using test chambers” (2013) [51]. The
assessment of the 90% confidence level is possible through the use of a 15-pi odor intensity scale with a reading uncertainty of ±2 pi. 3 OI is perceived odor intensity on a six-level scale
from 0 to 5 (no odor = 0, slight odor = 1, moderate odor = 2, strong odor = 3, very strong odor = 4, and overpowering odor = 5). 4 Based on the study Kim and Kim (2014) [52] selected 22
odorants with a similar chemical structure (structural formula) and determined an equation to convert the concentration x in ppm to intensify y, with odor OI on a scale of zero to five.
The study included odors from the distribution of food. These 22 odorants can be divided into five chemical groups: (1) reduced sulfur compounds, (2) carbonyls, (3) nitrogenous
compounds, (4) VOCs, and (5) volatile fatty acids. For example, for the group of reduced sulfur compounds for compound number 1, the conversion equation for H2 S is Y = 0.950logX +
4.14, for the carbonyl compounds group for compound number 10, ammonium NH3 , it is Y = 670logX + 2.38, and for the VOC group for styrene it is Y = 1.420logX + 3.10. 5 If one
considers a cooling system that removes heat from a space but does not remove moisture unless condensation occurs, such as radiant cooling without dehumidification in a ventilation
system, the importance of humidity is very clear. The sensible cooling of air in a room (no change in absolute humidity) from 25 ◦ C and 60% RH to 20 ◦ C (process a–b for x = 0.012
Appl. Sci. 2019, 9, 3918
kgw /kga ).This value, coupled with a high temperature, ta = 25 ◦ C, is accepted as the critical limit value for dehumidification by EN16798-1:2019 [10] and can be converted into an absolute
humidity H (gw /m3 ). These processes are expected to significantly increase thermal comfort and air quality. Nevertheless, the same change in enthalpy of the air h can be achieved by
simply reducing the humidity by 10% RH and keeping the temperature constant (processes a–c). Since IAQ is a function of enthalpy, these are expected to be the same. Here, a change in
humidity of 10% RH at constant temperature is equivalent to a change in temperature of 5 or 6 ◦ C at constant moisture content in air x (kgw /kgdry air ). 6 The IAQI system [30] adopts the
methodology of AQI to set up the range of IAQI values from zero to 500, including 50, 100, 150, 200, 300, and 500. The IAQI values of 100 and 150 correspond to the concentrations in the
Taiwan IAQG standard. Other IAQI values between 50 and 200 correspond to the concentration rankings given by several reference resources including the US EPA AQI system [31]. All
on-site concentrations of indoor air pollutants (HCHO, TVOC, PM10, PM2.5, CO, CO2 , O3 , bacteria, fungi, SO2 , and NO2 ) are combined using the IAQI system. The IAQI values are
calculated on the basis of the concentration value c using the interpolation method for each air pollutant. In the IAQI system, the index range 0–50 is “good” with a significance level of
“little or no risk”, 51–100 is “moderate” where “sensitive persons or those with respiratory symptoms are concerned”, 101–150 is “unhealthy for sensitive groups”, 151–200 is “unhealthy
for all individuals”, 201–300 is “very unhealthy—more serious health effects for everyone for short-term exposure”, and 301–500 is “hazardous” with “health warning of emergency
conditions for everyone”. Therefore, the comfort scale for IAQI is adequate for index values from zero to 200 and our proposal is to use the “hypothetical” scale of PD* of 0–100% in this
97
range, converted from IAQI. 7 The method of conversion of the IAQI scale to the PD* scale is based on experiences gained during research [27]. The IAQI values in a health-risk scale can
be given in % for persons giving a verbal answer of “no risk”, “moderate”, and “unhealthy” as their health risk evaluation. Therefore, when determining, for instance, the concentration
function values of the %PDIAQI (TVOC) and %PDIAQI (HCHO) components, it is necessary to recalculate the scale IAQI = f(c) in the range from zero to 200 to the scale PD* from 0 to 100%.
There are break points in the new PD*% scale: for IAQI values 0–50, PD*% is 0–25, for IAQI values 51–100, PD*% is 26–50, for IAQI values 101–150, PD*% is 51–75, and for IAQI values
151–200, PD*% is 76–100.
Appl. Sci. 2019, 9, 3918
The air quality indexes (i.e., AQI [31] and IAQI [30]) are piecewise linear functions of the pollutant
concentrations. At the boundary between AQI categories, there is a discontinuous jump of one AQI
unit. To convert from concentration cj to Ij (in the converted scale index Ij will be PDj , Equation (16)
is used.
Ij = ((Ihigh − Ilow )/(chigh − clow )) · (cp − clow ) + Ilow (16)
where Ij is the air quality index I in the PD*% scale, clow is the pollutant concentration break point,
which is ≤cj , chigh is the pollutant concentration break point, which is ≥cj , Ilow is the index break point
corresponding to clow , Ihigh is the index break point corresponding to chigh and cp is the truncated (to an
integer) actual concentration for the pollutant. Little data exist on the AQI’s metrological reliability for
AQI and IAQI. Only in the EPA Air Program undertaken at Cornell University [26] is there a previous
review of the quality assurance requirements for AQI.
2.6. The Representative VOCs for Indoor Environment

The time when IAQ studies focused on a class of contaminants referred to as volatile organic
compounds (VOCs) is bygone. The analytical methodology available was the primary basis for this
focus, but the recent broadening of analytical methods has led to growing realization that other
compounds (i.e., SVOCs) beyond traditional VOCs are implicated in IAQ problems. The choice of
VOCs remains a challenge in IAQ assessment. Moreover, VOCs is somewhat vague term, the definition
of which is not universally agreed upon. It has been defined in terms of vapor pressures and boiling
points, as well as molecular chain lengths detectable by chromatographic techniques. Due to the
complexity of VOC emission profiles, it is tempting to simplify the analysis and reporting of emissions
by grouping all detected compounds together. The first problem with this approach is that individual
compounds have highly variable health and/or comfort effects, the result being that concentration
alone is not predictive of IAQ effects. Levels of concern vary by orders of magnitude, so a collective
concentration will not correlate with IAQ. Second, VOC detection and quantification are highly method
dependent. A given sampling and analysis system cannot capture or respond to all the VOCs present
in any indoor environment or in the test chamber for a given emitting material. Thus, the term
“total” is misleading. The important aspect of IAQ submodel selection is the strategy defined by
the US EPA as “VOCs—Total versus Target: Irritancy, Odor and Health Impact”. The representative
90 target VOCs were presented by Canada’s National Research Council’s Institute for Research in
Construction (NRG-IRC) in collaboration with several academic and governmental partners, including
Health Canada. The compounds were selected based on health impact, occurrence in indoor air,
known emission from building materials, as well as suitability for detection and quantification by gas
chromatography-mass spectrometry (GC-MS) or high-performance liquid chromatography (HPLC).
Our list of target VOCs was actually representative for indoor environment and recommended by the
HealthVent project and is provided in Table 1.
2.7. Steps of ΣIAQindex Calculation

After selection of the IAQ model type ( IAQquality or IAQcomfort ), the IAQindex evaluation was

carried out using the complex model IAQ from Figure 5, which should contain the following stages.
(a) Calculation of the total concentration of pollutants in the ventilated space or the total mass of
air pollutants per m3 , the level of which is to be reduced by the ventilation process (taking into
account the ventilated volume of the room and the emissions present).
(b) Selection of the IAQindex model shape from the models defined by Equations (12)–(14), with the
provision that due to the multiplication of the submodels for IAQ(VOCodorous ), the number of
subcomponents of the IAQcomfort model will be more than five.
(c) Processing the input data of the submodels to obtain the concentration value cj , e.g., converting
the measured OI value into a concentration value for a given pollutant cj in μg/m3 .
98
Appl. Sci. 2019, 9, 3918
(d) Calculation of the excess concentration values for each identified contaminant (Table 2)
Δcj = cj −cref .

(e) Calculation of the sum of excess concentrations (see Table 2), Δcj .
(f) Calculation of adjusted weights Wj for the selected model equations. IAQquality and IAQcomfort ,
are determined on the basis of arithmetic means or by adjusting all the values of Δcj in a given
model using Equation (10), only for groups of pollutants with similar concentration values.
(g) Calculation of the value of the ventilating air flow for the environment described in the IAQindex
model in accordance with the requirements of the standard EN 16798-1:2019 (a method using the
criteria for the ventilation required for the individual substance emitted) [10].
(h) Calculation of given IAQ environmental input parameters, including concentrations of pollutants
cj assigned to submodels. The PD values from their sensory equations (Table 3) are presented as
the dependence of the percentage of persons dissatisfied PD = f (cj , . . . ), from one of the formulas
from Table 3, in order to determine this function.

(i) Selection of the IAQquality model equation (with weights W1 , W2 , and W3 ) or the IAQcomfort
model equation (with weights W1 , . . . W5 or more) and calculation from Equation (13) of the
value with adjusted weights, followed by multiplication of IAQ submodels (Equation (8)) and
insertion as a term of the IEQindex in Equation (18) [34].

When selecting the IAQcomfort/health model type, an IAQindex evaluation is carried out using the

combined model IAQ from Figure 5, which should contain the following steps.

(a) Calculation of the total concentration of pollutants cj in the ventilated space or the total mass of
air pollutants per m3 , the level of which is to be reduced by the ventilation process (taking into
account the ventilated volume of the room and the emissions present).
(b) Choosing the IAQindex model (12) from among the models defined by Equations (12)–(14), with the
provision that by multiplying the IAQ(VOCnon-odorous ) submodels, the number of subcomponents
of the IAQcomfort/health model will be more than seven.
(c) Processing of submodel input data to obtain concentration values cj in μg/m−3 .
(d) Calculation of excess concentration values for each pollutant identified (Table 2) using Δcj = cj −cref .

(e) Calculation of the sum of excess concentration values for submodels 1–7 via Δc1 . . . 7
(f) Determination of the adjusted weights W 1, . . . 7 for the equation of the IAQcomfort/health model on
the basis of arithmetic means or by adjusting all the values of Δcj in a given model using Equation
(10), only for groups of pollutants with similar concentration values.
(g) Calculation of the values of the ventilating air stream from the total concentration of indoor air

pollutants cj (for instance, Reference [42]), for the environment described by the IAQindex
model (Figure 5) in accordance with the requirements of the EN 16798-1:2019 standard for the
individual substances emitted and using an alternative method when the concentration in the
room has stabilized.
(h) Calculation of the given IAQ input parameters, including concentrations of all pollutants, cj ,
assigned to the submodels. The PD values are taken from their sensory equations (Table 3)
depending on the percentage of persons dissatisfied, PD = f (cj , . . . ), or selected from the formulas
for determining this function given in Table 3.
(i) Development of IAQPM2.5 , IAQPM10 , and IAQnon-odorous submodels. When it is planned to use
the indoor air quality index scale IAQI or a similar scale, it is necessary to convert these to PD*
values in %, in two steps: (1) by reading from the standard curves of the IAQI = f(cj ) all IAQI
values for the determined (measured) VOCnon-odorous concentration values and using the converted
scale calibration curve, PD = f(IAQI), by reading from the recalibration curve (Equation (16)),
the PD = f(cj ) values on the dissatisfaction rating scale from zero to 100 in %, according to Footnote
7 in Table 3.
99
Appl. Sci. 2019, 9, 3918
(j) Calculation of IAQ values for Pj pollutant submodels from Equation (8) and insertion into the

model equation IAQcomfort/health (14) (with weights W1 , W2, . . . , W7 or more).

2.8. The IEQ Assessment Equation with IAQ As a Subcomponent
The proposed IAQ model can be a substantial component of the IEQ model; for example, in the
case study shown later in the article. The indoor environmental quality index refers to the quality of a
building’s environment with respect to the occupants’ satisfaction in %. The morphology of the IEQ
index model used to assess buildings, to determine as an IEQ component the IAQindex and to determine
other subcomponents TCindex —thermal comfort, ACcindex —acoustic comfort and Lindex —light quality
based on measurements of physical properties in each of the submodels—in accordance with the
scheme of the Piasecki–Kostyrko model, is presented in Figure 6 [22].
Figure 6. The research steps necessary to determine the IEQindex for buildings, including physical and
design parameters of buildings and subcomponent models.
The EN 16798-1:2019 is the reference for IEQ model creation [22,33]. The standard allows complex
indoor information to be presented as one overall indicator of indoor environmental quality of the
building—IEQindex . The model reliability, including the uncertainties of measurements and data for
this model, was discussed clearly in Reference [34], where the authors also presented the internal
incongruity in the IEQ model structure and the justification for using the crude weights method for each
subcomponent. Originally, the IEQ model was expressed as a polynomial equation consisting of four
terms by Wong [43]. The IEQindex is composed of the following subcomponents (SIi ): thermal comfort
(TCindex ), indoor air quality (IAQindex ), acoustics (ACcindex ), and lighting quality (Lindex ). Multiplying
their weights, Wi , leads to Equation (17).
IEQindex = Σ Wi ·SIi (17)
The authors adopted the crude weighting system, where all elements are weighted in the same
way (0.25 for W 1 –W 4 ), as shown in Equation (18).
IEQindex = 0.25 · TCindex + 0.25 · ΣIAQindex + 0.25 · ACcindex + 0.25 · Lindex (18)
100
Appl. Sci. 2019, 9, 3918
As a consequence of the equation, the subcomponents SIi (the predicted percentage of those
satisfied) can be calculated using Equation (19).
SIi = 100 − PD(SIi ) (19)
where PD is the predicted percentage dissatisfied (PPD) and PD(SIi ) is the percentage of persons
dissatisfied with the IEQ subcomponent (SIi ) level. The authors’ simulations for IEQindex sub-indices
and preliminary metrological analysis of the overall IEQ model fitting were performed with Monte
Carlo tests.
It is easy to show that the standard deviations of these values are equal:
SD(SIi ) = SD(PD(SIi )) (20)
2.9. A Case Study of a Building

The experimental part of this study was performed simultaneously with the BREEAM certification
process, including determination of the three primary IAQ pollutants: formaldehyde concentration,
CO2 , and VOCs in the indoor air [22]. The building is a high tower, made of a convex concrete–steel
structure with a glass facade. The basic information on the assessed building is presented in
Table 4. At the time of the test, the building had a standard empty office without furniture (so-called
pre-occupancy stage). The walls were plastered and painted, the suspended ceilings were in place,
and the floors were finished with synthetic carpets. All building installations were active, including
the mechanical ventilation controlled by the Building Management System BMS system with zonal
CO2 concentration sensors. The building was tested a few days after the formal end of finishing works.
The tests were made on the 55th and 47th floor.
Table 4. Information on the building in the case study.
Office Number Net Area IAQ Assessed

Facade View Indoor View Life-Stage
Building of Floors (m2 )
Number
Certificate Area (m2 )
of Floors
BREEA
pre-occupant 49 59,000 3000 2
Mexcellent
ȱ ȱ
Measurement points in the building were determined based on the analysis of frequencies of
designed occupancies of the room and interior finish standards (open spaces). The sampling plan was
prepared with the BREEAM assessors conducting the certification process of the facility. The main
focus was on the IAQ index of open spaces in which the largest number of people may reside, and these
represent the largest occupied usable floor space. According to the detailed design project documents,
the building emphasizes the use of materials with known and low emission levels (BREEAM certified).
2.10. The Equipment, Measurements, and Experimental Approach

Standardized CEN and ISO analytical methods were used to determine the VOC concentrations
and CO2 and formaldehyde concentrations in the indoor air of the building. Selection of the sampling
points was made with the BREEAM assessor in two representative office zones per tested floor and
a minimum of two floors. The building was tested three days after formal final finishing works at
the pre-occupancy stage with no users inside. For this office building, the tests were conducted on
the 55th and 47th floors. Air samples were collected using an active sampling procedure with an
electronic mass flow controller, which controlled the air flow (10 dm3 /h for VOC tests and up to
30 dm3 /h for formaldehyde tests). Indoor samples were set up in selected representative office locations,
approximately 1.5 m above the floor, away from windows, doors, potential emission sources, and direct
101
Appl. Sci. 2019, 9, 3918
sunlight. Air samples were tested in accordance with the ISO 16000-6:2011 and ISO 16000-3:2011
standards. The VOCs were assessed using tubes filled with Tenax adsorbent. Then, they were thermally
desorbed using a thermal desorption apparatus (TD-20, Shimadzu, Tokyo, Japan). The process of
separation and analysis of volatile compounds was achieved using a gas chromatograph equipped
with a mass spectrometer (GC/MS) (model: GCMS-QP2010, Shimadzu, Tokyo, Japan). The following
GC oven temperature program was applied: initial temperature 40 ◦ C for five min, 10 ◦ C per min to
260 ◦ C, and the final temperature of 260 ◦ C for 1 min. The 1:10 split ratio injection mode was applied.
The method used has a limit of quantification of 2 μg/m3 . The volatile compounds were identified
by comparing the retention times of chromatographic peaks with the retention times of reference
compounds and by searching the NIST data base (National Institute of Standards and Technology,
Gaithersburg, MD, USA) mass spectral database. Identified compounds were quantified using a
relative identification factor obtained from standard solution calibration curves. TVOC was calculated
by summing identified and unidentified compounds eluting between n-hexane and n-hexadecane.
In order to determine volatile aldehydes, air samples were taken via cassettes using a solid absorbent
silica gel coated with 2,4-dinitrilophenyl hydrazine (2,4-DNPH), and then subjected to a laboratory
test using high-performance liquid chromatography (HPLC) with UV-Vis detection (Dionex 170S,
Dionex, Sunnyvale, CA, USA) and an isocratic pump (Dionex P580A, Dionex, Sunnyvale, CA, USA).
The described method has a limit of quantification at 2 μg/m3 .
Other IEQindex components were tested as follows. The acoustic tests confirming the designed
values were carried out by the measurement of the equivalent sound levels, LAeq, in the selected
locations. The measurements were carried out during the daytime (starting at 11:00). The following
equipment was used for the measurements: Brüel&Kjær 4231 acoustic calibrator (Brüel&Kjær,
Nærum, Denmark), Nor-121 analyzer (Norsonic, Tranby, Norway), Brüel&Kjær 4165 measuring
microphones (Brüel&Kjær, Nærum, Denmark), analyzer with microphone Norsonic-140 (Norsonic,
Tranby, Norway). Before the tests were carried out, the calibration of the measuring path was conducted
in accordance with the instructions to “check the acoustic measurement channel”. The test results were
evaluated in relation to the requirements considering permissible sound levels A in rooms intended
for human dwellings. Thermal environmental measurements were provided using the microclimate
multifunctional instrument HD32.1 and the tests were in accordance with ISO 7726 and ISO 7730.
VOCs were tested simultaneously at all points. Visual comfort (Hea 01) was confirmed by using a
MAVOLUX 5032C instrument (USB version) with a 3C15683 detector (Gossen, Nürnberg, Germany),
in accordance with EN 12464 provisions.
2.11. Additional Explanations

The adaptation of the IAQ model to a practical casestudy was mainly for illustrative purposes in the
context of the presented IAQ calculation/aggregation method. We did not focus deeply on discussing
the technical or environmental issues of the presented building. Other IEQ subcomponents, such as
thermal, acoustic, and visual satisfaction (in %), used to determine the IEQ index, were experimentally
determined and partly presented in References [22,33]. Authors do not focus on these results in this
article, as they have already been discussed in other papers [22].
3. Results
3.1. Results for theIAQindex and IEQindex Prediction

A previous publication of ours [22] reported on IEQ and IAQ building assessments for a larger

number of BREEAM buildings, where IEQ was assessed without calculating the combined IAQ index.

The combined model of the IAQindex presented in this paper had not yet been previously developed,
and we were limited in determining the IEQindex , thus we only took into account two of the most
well-known pollutants (i.e., CO2 and TVOC) separately. The assessment of the IEQ index was made by
adaptation of the measured parameters (complying with the draft EN 16798-1:2019 standard for indoor
102
Appl. Sci. 2019, 9, 3918
environments) as the input values for the submodels of the IEQindex . The input values for the case
study are presented in Table 5, which provides the input data for determining the IEQindex sub-indices
of thermal comfort (TCindex ), indoor air quality (IAQindex ), acoustics (ACcindex ), and lighting quality
(Lindex ) for an office building (47th floor) three days after completion of the finishing work before users
were allowed in the building (i.e., pre-occupancy stage).
Table 5. Physical parameters 1 and IEQindex results calculated using Equation (9) separately for an
IAQindex with internal air pollution of CO2 and an IAQindex with internal TVOC air pollution, assuming
a realistic uncertainty of parameter measurement for the case study of a building (47th floor; open
space) three days after the completion of finishing works.
Sub-Index (Satisfied)
Sub-Index Sub-Index PD(SIi ) Models Input Values
and ±SD
Icl 0.55 clo
ta 24.0 ◦ C
PMV (Fanger-CBE-ISO 7730)
tr 24.5 ◦ C
TCindex PMV = f(ta , tr , va , pa , M, Icl, ) 90% ± 3.2%
va 0.15 m/s
PDTC = f(PMV)
RH 45%
M 1.1 met
PDIAQ(CO2 ) =
450 ppm 85.2% ± 0.6%
IAQindex 395·exp(−15.15·CCO2 −0.25 )
PDIAQ(TVOC) =
787 μg/m3 52.0% ± 18.0%
405·exp(−11.3·CTVOC −0.25 )
PDACc =
2·(ActualSound_Pressure_Level (dB(A))
ACcindex −DesignSound_Pressure_Level (dB(A))) 80% ± 6.7%
Actual (background) noise level
55 dB(A)
Design sound level
45 dB(A)
PDL = −0.0175 + 1.0361/{1 +
Lindex exp(+4.0835 · (log10 (Emin ) − 450 lux 98.4% ± 9.0%
1.8223))}
IEQ(CO2 ) First variant with cCO2 as an IAQindex parameter IEQCO2 = 92.2% ± 5.8% 1
IEQTVOC Second variant with cTVOC as an IAQindex parameter IEQTVOC = 80.1% ± 10.7%
1 The IEQ and its measurement’s uncertainty (with subcomponent standard deviation values) were calculated for
IEQ physical parameter values, where ta is the air temperature (◦ C), tr is the mean radiant temperature (◦ C), va is
the relative air velocity (m/s), pa is the water-vapor partial pressure (Pa), M is the metabolic rate (met), and Icl, is
the clothing insulation (clo). In addition, cCO2 is the concentration in ppm, cTVOC is the highest observed TVOC
concentration in μg/m3 , actual noise is in dB(A), and Emin is the minimum daylight illuminance (lux).
3.2. Results for the ΣIAQindex and IEQindex Assessment Including Identified Pollutants (CO2, TVOC, and HCHO)

The example of a modified calculation of the collective submodel IAQquality for three basic
pollutants, as a component of the IEQ model for determining one project value for this indicator, is
provided in two variants. The first variant uses sub-indices of IAQ for two pollutants, CO2 and TVOC,
which are described in Table 5, as well the sub-index of the third pollutant, HCHO (according to
Reference [47]), where these differences in the approaches mean one must combine them into one

submodel IAQ in order to be used in IEQ calculation. The second variant uses submodels of IAQ for
TVOC and HCHO pollutants based on the IAQI system [30] and then converts them into percentages

of persons dissatisfied PD* in %. According to the diagram of the model IAQ from Figure 5 and

using Equation (12) of the IAQquality model, the submodel weights are calculated as follows.
WCO2 for the submodel IAQ(CO2 ) = 0.5 is a component of the polynomial:

IAQquality = 0.5·IAQ(CO2 ) + 0.5·IAQ(VOC) (21)
103
Appl. Sci. 2019, 9, 3918
WVOC for submodel IAQ(VOC) = 0.5 is a weight for combined submodel of the polynomial:
IAQ(VOC) = WTVOC ·IAQ(TVOC) + WHCHO ·IAQ(HCHO) (22)
with the terms WTVOC and WHCHO calculated from Equation (14) using the measured values cj (actual
concentration of TVOC and HCHO) and the reference values cref (Table 6).
Table 6. Calculation of the weights of the WTVOC and WHCHO values for the two sub-indices of the
combined IAQ(VOC) model.
Excess
Sub-Index Input Value cj Input 1 Value cref Wj
Concentration Δcj
IAQ(TVOC) 787 μg/m3 300 μg/m3 487 μg/m3 0.96
IAQ(HCHO) 18 μg/m3 0 18 μg/m3 0.04
1 EN 16798-1:2019 for a very low-pollution building.
In our case study, the value of the submodel IAQ weight for the model (HCHO) ought to also
be calculated from the measured value and the reference value cref . However, formaldehyde is an
unusual pollutant because, although it belongs to VOCodorous compounds, the concentrations found in
buildings are many times lower than the HCHO threshold cth = 300 μg/m3 according to the WHO [7]
and lower than the threshold concentrations of HCHO from 60 μg/m3 to 70 μg/m3 issued in 2013 by the
American Industrial Hygiene Association [44]. The permissible value of cref = 100 μg/m3 is also higher
than the formaldehyde concentration found in buildings, according to Reference [5] and the standard
EN16798-1:2019 [10]. Therefore, the authors propose that in such a case (to avoid a negative value of
Δcj ), the value modelling the reference should be taken as zero. Then, the form of the adjusted WHCHO
weight in the model described by Equation (11) for air with three pollutants, would be as follows.
Δc j (cHCHO − 0)
WHCHO = = (23)
Δc2...3 (cTVOC − cELV ) + (cHCHO − 0)
j=2...3

The results of the weights assessment for the two variants of the IAQquality model are presented
in Table 6.

According to the diagram of the model IAQ from Figure 5 and Equation (8), we proposed
sensory equations for the percentage of persons dissatisfied %PD* in two variants.

The submodel IAQ’s first variant includes the following:
1. The IAQ submodels used so far in References [22,33] for CO2 and TVOC pollutants, as shown in
Table 5;
2. The IAQ submodel for formaldehyde, using two types of equations depending on the range of
HCHO concentrations measured in the building. Formaldehyde concentrations in the air with
values above the threshold concentration, cth , for its odor, i.e., above 60 or even 300 μg/m3 , can be
used to create IAQ submodels for rooms with volatile and aromatic VOC compounds as well as
for the HCHO equation [50].
exp (2.14 · OI − 3.81)

PDHCHO = (24)
exp (2.14 · OI − 3.81) + 1
However, in the case study building, the maximum concentration of HCHO was 18 μg/m3 and,
therefore, its concentration in the air was several times lower than the concentration of the odor
threshold, cth [44]. The intensity of the formaldehyde odor was undetectable under these conditions,
and the sensory equation PD = f(OI), which is appropriate for sensory detection of IEQ, is not applicable
for odors below the threshold. Therefore, for small concentrations, we proposed the use of the equation
104
Appl. Sci. 2019, 9, 3918
taken from the work of Zhu and Li [47] based on the analysis of “health effects on the human body”,
derived from “indoor air quality comfort evaluation experiments and the literature”.
cHCHO
PMVHCHO = 2log (25)
0.01
This equation links the value of the new unit “the effect of formaldehyde on human comfort”,
called PMVHCHO , with its cHCHO concentration (μg/m3 ) in the air. It covers the range from 10 μg/m3 to
320 μg/m3 and, as declared by the authors, this value has the same nature as PMV thermal comfort,
which can be converted into a PD% unit according to the formula in Reference [48], experimentally
confirmed for nearly zero energy buildings (NZEBs) by Reference [58].
PDHCHO = 100 − 95·exp(−0.03353·PMV 4 − 0.2179·PMV 2 ) (26)

The submodel IAQ’s second variant includes the following.
i. The IAQ(CO2 ) submodel used so far for CO2 pollution, as shown in Table 5.
ii. The IAQ submodels for TVOC and HCHO types of pollution used as indoor air quality index
ratio values borrowed from the IAQI system [30], which are then converted into percentages of
persons dissatisfied (PD* in %) in the following way
(a) The reference curves of IAQI = f (cj ) [30] for two dependencies of the IAQI index on TVOC and
HCHO contamination values must be reconstructed. On the y-axis are the IAQI index values
from zero to 200 in the range from “no risk” to “unhealthy” and on the x-axis, the cj values
are presented.
(b) In accordance with the measured values of cTVOC and cHCHO , the values IAQITVOC and IAQIHCHO
are determined from the functions IAQITVO c = f(c) and IAQIHCHO = f(c).
(c) Based on the IAQI system parameters [30] given in Footnote 6 of Table 3, which are presented as
data for the functions for indexes, the IAQITVOC and IAQIHCHO values appropriate for the break
pointsi n perceived pollution concentration values are calculated in a range from zero (good) to
200 (unhealthy) using the ordinal scale IAQI = f(c) [27]. This function was converted to IAQITVOC
and IAQIHCHO scales using the concentration function scales PD*(TVOC) and PD*(HCHO) in the
PD* range from 0 to 100% (Figure 8).
(d) The data used for calculation and conversion of IAQI and PD* scales are presented in Table 7.
(e) Based on data determined for the new converted scales (Table 7) for the percentage of persons
dissatisfied (PD*(TVOC) and PD*(HCHO) concentration functions), the PD*(c) working graphs
were drawn (Figure 7). (To harmonize the concentration scale of pollutants on the y-axis, cHCHO
values multiplied by 10 were applied.)
(f) The interpolation of the percentage of persons dissatisfied (PD*(TVOC) and PD*(HCHO)) in %)
for the measured values of pollution concentrations in the air must be made using the graph from
Figure 7 or using Equation (16) [31].
105
Appl. Sci. 2019, 9, 3918
Table 7. Recalculation and conversion of IAQI value scales.
PD* Value IAQI Value Evaluation cTVOC -TVOC (1 h) cHCHO -HCHO (1 h)

% - - ppm μg/m3 ppm μg/m3
0 0 No risk 0 0 0 0
25 50 Good 0.3 1 300 0.01 2 12.3
50 100 Moderate 0.9 900 0.04 49.1
75 150 Unhealthy for sensitive 3.0 3000 0.10 122.8
100 200 Unhealthy 4.6 4600 0.75 921.0
1 Conversion factors TVOC: 0.3 ppm corresponds to 300 μg/m3 [44]. 2 Conversion factors HCHO: 1 ppm corresponds
to 1228 μg/m3 [44].
ϯϮϬϬ
ĐdsKĂŶĚ,,K;JPñDQG
ϮϴϬϬ
ϮϰϬϬ
JPñî
ϮϬϬϬ
ϭϲϬϬ
ϭϮϬϬ
ϴϬϬ
ϰϬϬ
Ϭ
Ϭ Ϯϱ ϱϬ ϳϱ
W;йͿ
ȱ
Figure 7. The percentage of persons dissatisfied function of TVOC (blue line) and HCHO (brown line)
versus concentration.
ϮϬϬ
ϭϴϬ
ϭϲϬ
ϭϰϬ
ϭϮϬ
/Y/
ϭϬϬ
ϴϬ
ϲϬ
ϰϬ
ϮϬ
Ϭ
Ϭ Ϯϱ ϱϬ ϳϱ ϭϬϬ
WΎ;йͿ
Figure 8. Percentage of persons dissatisfied, %PD*, in relation to IAQI values.
In the context of the results for the overall IEQ model index when treating both the main pollutants
CO2 and TVOC separately, we present in Table 8a the transformed IEQ calculations with the sub-indices

IAQ. The results for the individual case study building IAQ subcomponents are taken fromrTable 5

(for cTVOC = 787 μg/m3 and cHCHO = 18 μg/m3 [22]). The IEQ index with IAQquality values was
calculated using two variants—the first conventional and the second borrowed from the IAQI scale [30]
(Table 8a).
Standard deviations for each concentration are provided in Table 8b.
106

Table 8. (a) Physical parameters (Footnote 1 in Table 5) and IEQ results calculated from Equation (9) with IAQ; assuming realistic uncertainty of parameter
measurements for the case study building (47th floor; open space) a few days after completion of the finishing works. (b) Measured pollutant concentrations c and
standard deviations SD(c).
(a)
Sub-Index Sub-Index PD(SI ) Models Input Values Sub-Index (Satisfied) and ±SD
Appl. Sci. 2019, 9, 3918
Icl 0.55 clo

ta 24◦ C
PMV (Fanger-CBE-ISO 7730) tr 24.5◦ C
TCindex 90.0% ± 3.2%
PMV = f(ta ,tr , va , pa , M, Icl,dyn ) va 0.15 m/s
PDTC = f(PMV) RH 45%
M 1.1 met

IAQindex (1) PDIAQ(CO2 ) = 395·exp(−15.15·CCO2 −0.25 ) c = 450 ppm 85.2% ± 0.6%
Sub-indices 52.0% ± 18.0%
PDIAQ(TVOC) = 405·exp(−11.3·CTVOC −0.25 ) c = 787 μg/m3
First variant
PMVHCHO = 2log cHCHO
0.01
PDHCHO = 100 − 95·exp(−0.03353·PMV 4 − 0.2179·PMV 2 ) c = 0.018 mg/m3
65.8% ± 10.7%%
53.0% ± 17.3%
IAQVOC = 0.96·IAQvariant1 (TVOC) + 0.04·IAQvariant1 (HCHO)
IAQindex (1)
107
IAQindex (1) = 0.5·IAQ(CO2 ) + 0.5·IAQvariant 1 (VOC)
69.1% ± 9.0%
PDIAQ(CO2 ) = 395·exp(−15.15·CCO2 −0.25 )
PD*TVOC read from the graph with PD*TVOC = f (cTVOC )
c = 450 ppm 85.2% ± 0.6%
IAQindex (2) or determined from Equation (16)
c = 787μg/m3 ±18% 54% ± 13.8%
Sub-indices 1) PD*HCHO read from the graph with PD*HCHO = f (cHCHO )
Second Variant or determined from Equation (16)
c = 18μg/m3 ±12% 71.1%± 11.0%
SD for cTVOC and cHCHO at
“break points”±12%
IAQVOC = 0.96·IAQvariant2 (TVOC)
54.7% ± 13%
IAQindex (2) + 0.04·IAQvariant2 (HCHO)
IAQindex (2) = 0.5·IAQ(CO2 )
70.0% ± 6.5%
+ 0.5·IAQvariant 2 (VOC)
PDACc = 2·(ActualSound_Pressure_Level (dB(A) −
DesignSound_Pressure_Level (dB(A))
ACcindex 55 dB(A) 80.0% ± 6.7%
Actual (background) noise
Design sound level 45 dB(A)
Lindex PDL = −0.0175 + 1.0361/{1 + exp(+4.0835 ·(log10 (Emin ) − 1.8223))} 450 lux 98.4% ± 9.0%
Table 8. Cont.
Sub-Index Sub-Index PD(SI ) Models Input Values Sub-Index (Satisfied) and ±SD
IEQ
index (1) with 84.4%±3.7%
IAQindex (1) IEQindex ±SD = W1· TCindex + W2 · IAQindex (1) + W3 ·Accindex + W4 ·Lindex
umeas = 2 · 3.7 = ±7.4%
meas. IEQindex (1)
Appl. Sci. 2019, 9, 3918
±SD

overall ±uoverall (IEQ) = ( (SDreal (PD(SIi ))2 + (SDvote PD(SIi ))2 )−2
uoveall = ±16.24%
IEQindex (1)±uoverall 2) IEQindex (1)±uoverall
84.4%±16.24%
IEQ
index (2)with
IAQindex (2) 84.6%±3.3%
measIEQindex (2)±SD IEQindex ±SD =
W1· TCindex + W2 · IAQindex (2) + W3 ·ACcindex + W4 ·Lindex umeas = 2 ·3.3 = ±6.6%
overall
IEQindex (2)±uoverall 2)
±uoverall (IEQ) = ( (SDreal (PD(SIi ))2 + (SDvote PD(SIi ))2 )−2 uoverall = ±16.15%
IEQindex (2)±uoverall 84.6%±16.15%
108
(b)
cmes SD(cmeas ) cH SD(cH ) cL SD(cL ) PD*H SD PD*L SD
TVOC 787 18%⇒141.7 900 12%⇒108 300 12%⇒36 50 12% 25 12%
HCHO 18 12%⇒2.16 49.1 12%⇒5.89 12.3 12⇒1.48 50 12% 25 12%
1) The method of calculation of mean values PD* and ± SD(PD*) for TVOC and HCHO is based on Equation (16) and takes the form (27).

PD∗H − PD∗L
PD∗ = × (cmeas − cL ) + PD∗H (27)
(cH − cL )
Concentrations c and SD(c) are in μg/m3 ; PD*H , PD*L , and SD(PD*) are in %, and standard deviations of the HCHO concentration of 12% was adopted on the basis of reports from IAQ
research conducted as part of BREEAM in 2016 [22]. The assumptions were that cH and PD*H are the coordinates of the upper break point (i.e., high break point) of the converted scale
PD* = f(c), and cL and PD*L are the coordinates of the lower break point (i.e., low break point) of the converted scale PD* = f(c). Standard deviations were assumed for cmeas , cTVOC , and
cHCHO as well as for PD* = ±12%, as this is half of the transformed segment of the scale, which according to Table 7 covers a range of 25% of PD*, with one perceived category of air
quality, e.g.,“no risk” or “moderate”. Therefore, the maximum standard deviation was 12.5% and, according to the literature on AQI and IAQI values, it should be rounded to a total
value. 2) SDvote (PD(SIi )) from the ±uoverall (IEQ) equation was the standard deviation of a probability distribution of an each. (SIi )vote and was calculated primary using the PD(SIi )
equation calibration curve [34].
Appl. Sci. 2019, 9, 3918
4. Discussion

4.1. Discussion of the IAQindex Theoretical Model
For years, the authors, as accredited laboratory personnel, have conducted IAQ pollution tests
in indoor environments for various applications. Based on our experience, it was concluded that
the general approach to assessing combined IAQ has not yet been systematized and that there is a
global tendency to assess individual IAQ parameters separately or to group them without a justified
aggregation method. This is not a good situation from the point of view of building users’ needs.
This, in our opinion, may lead to incorrect IAQ interpretations in specific building situations. In the
context of analyzing the problem in this paper, authors presented a summary of the state-of-the-art
methods and also provided a new approach for solving some of these problems. As presented, it is

possible to create a IAQ index aggregating the results of indoor air analyses, taking into account
various representative pollutants. Three levels of comprehensive air quality assessments (with three,
five or seven subcomponents), depending on the application of the assessment, were proposed, together
with step-by-step procedures. This may be practical, as shown in the evaluation of a case study
on a building. We originally selected the main IAQ subcomponent equations and user satisfaction
dependences, %PD = f(cj ), and provided them all in one place (Table 3). We then proposed and justified
the weighting schemes for the IAQ total equation. In most of the studies in the literature, the weighting
schemes used for IEQ or IAQ assessments are not physically justified or explained. There are known
methods of weighting sub-indices, but the problem that was solved in this paper was an effective
system for weight adjustments. For the construction of the combined model IAQindex , with a weighting
scheme useful for aggregating sub-indices, we proposed the model scheme presented in Figure 5.

According to the results, the advantage of the complex model IAQindex , in which the input quantities
always constitute concentrations of given pollutants, is the ability to use these concentrations to
calculate excess pollution concentrations from Equation (10) and generate weighting schemes W1, . . . n
for all three models by adjusting the weights based on the concentration values of excess air pollutants
to a value ≤1.0 for each IAQindex model. The Δcj values determine the masses of pollutants that must be
removed by ventilation to eliminate the target pollutant effect. They can be determined as differences
between the current concentrations of pollutants and the concentration of pollutants at the reference or
standard level (e.g., cELV or cLCI ), and in the case of VOCodorous , the odor threshold cth. The presented
approach may allow planning of air quality for the building.
As discussed, it is important to identify those VOCs with comfort, health, and impacts and
focus on the IAQ sub-model choice aspect, briefly defined as the strategy “VOCs—Total vs. Target:
Comfort, Irritancy, Odor, and Health Impact”. The model from Figure 5 has uniform inputs, i.e.,
concentration levels cj and two outputs: (1) weighed (adjusted) and (2) sensory equations, PD* = f
(cTVOC ), constituting the IAQ submodel equations. These second outputs of submodels (PD* values)
are coefficients of satisfaction from the comfort sensation or lack of “health risk”. These are the terms

of the equation describing “combined IAQ“, which meets the requirements of the abovementioned
strategy of selecting IAQ sub-models related to IAQ components that have the most impact on the
resulting IEQ perception.
Models for subcomponents of IAQ not perceived by humans but influencing health are
recommended to be used from the index set in the AQI system [26–28], which was adapted by
the authors to assess the quality of indoor air based on, and in accordance with, the concepts of the air
quality assessment system used globally by the American EPA. In the context of the subcomponent of
the TVOC concentration in Figure 7, the authors provided the relationships of PD* = f(cTVOC ) based on
Jokl research [49] and resulting from the IAQI scale [30] as converted by the authors. The relationship
between PD* and TVOC concentration in both approaches is strongly correlated, as shown in Figure 9.
109
Appl. Sci. 2019, 9, 3918
Figure 9. PDTVOC based on conversion from IAQI to the PD* scale and study of Weber-Fechner theory.
The curves obtained from the conversion confirm Jokl’s predictions provided in [49] and previously
accepted idea. However, the final confirmation of these curves will be done experimentally as panel
tests, as planned by the authors for the near future.
4.2. Discussion of Results for the Case Study on a Building

The experimental study was performed in a BREEAM certified building, and included the
determination of formaldehyde concentration, CO2 , and VOCs in the indoor air. The example

calculation of the combined IAQ model for three basic pollutants, as components of the IEQindex
model, is presented in two variants but the calculated PD*TVOC values obtained with both calculation

methodswere very similar. The first variant of IAQ calculation used %PD = f(cj ) curves in %
sub-indices of IAQ for three pollutants, and the differences in approach in Tables 5 and 8a meant

combining them into one IEQ submodel: the IAQ model intended for the IEQ calculation. The second
variant used submodels of IAQ for TVOC and HCHO pollutants based on the IAQI system [30] which
were then converted into percentages of persons dissatisfied (PD* in %).
The first conclusion is that CO2 concentration cannot be used separately for the IAQindex assessment,
especially at the pre-occupancy stage (Table 5). The building was polluted with VOC emissions and
HCHO from the construction products directly after finishing works were completed.
The authors confirmed that all three pollutions should be a simultaneously integrated part of
the IAQ model, because the importance of TVOC is much greater, representing the main source of
pollution—the construction and finishing materials.

According to the results, we recognized two variants of the combined IAQindex calculation.

For the first variant [22], the combined IAQ index of satisfied users was 69.1% and, for the second

variant (new approach with converted AQI index), the IAQ index was 70.0% satisfied. The results
ofthe IEQindex (1) (for Variant 1) were within the interval of combined overall uncertainty,±16.24%,
and the results of the IEQindex (2) (for Variant 2) were that the overall uncertainty was ±16.15%.
Therefore, the result was convergent, which confirms the credibility of the proposed approach.
The results obtained also showed that, in the period immediately after completion of finishing works
in indoor spaces, there may be a temporarily increased concentration of TVOC, which systematically
decreases over time, as we have shown in other papers. In the case of a building, the research showed that
110
Appl. Sci. 2019, 9, 3918
tests carried out immediately after finishing works gave results that significantly exceeded the BREEAM
limits for TVOC at 300 μg/m3 (twice as high). It should be expected that an acceptable level should be
reached after a minimum of one month from the completion of the work.
For correctness of the obtained calculations, the authors are conducting a model credibility analysis
that will be provided in the next article—Indoor Air Quality Model Part II: The Combined Model

IAQindex Reliability Analysis. The model uncertainty estimate may be compromised because the
model reproduces the discomfort level associated with the dominant component.
Author Contributions: Conceptualization, M.P. and K.B.K.; methodology, M.P. and K.B.K.; formal analysis,
M.P. and K.B.K.; investigation, M.P. and K.B.K.; resources, M.P. and K.B.K.; data curation, M.P. and K.B.K.;
writing—original draft preparation, M.P. and K.B.K.; writing—review and editing, M.P. and K.B.K.; visualization,
M.P. and K.B.K.; supervision, K.B.K.; project administration, M.P.; funding acquisition, M.P.
References
1. Fanger, P.O. What is IAQ? Indoor Air 2006, 16, 328–334. [CrossRef] [PubMed]
2. Jones, L.L.; Hashim, A.; McKeever, T.; Cook, D.G.; Britton, J.; Leonardi-Bee, J. Parental and household
smoking and the increased risk of bronchitis, bronchiolitis and other lower respiratory infections in infancy:
Systematic review and meta-analysis. Respir. Res. 2011, 12, 5. [CrossRef] [PubMed]
3. ASHRAE. ASHRAE Guideline No. 10. Interactions Affecting the Achievement of Acceptable Indoor Environments;
ASHRAE: New York, NY, USA, 2011.
4. Kostyrko, K.; Kozicki, M. Trends in odor measurments and detecting MVOC content in the interiors of
buildings. In Conference on Metrology MKM18; The Scientific Papers of Faculty of Electrical and Control
Engineering; Gdańsk University of Technology: Gdansk, Poland, 2018; Volume 59, p. 97.
5. Carrer, P.; de Oliveira Fernandes, E.; Santos, H.; Hänninen, O.; Kephalopoulos, S.; Wargocki, P. On the
development of health-based ventilation guidelines: Principles and framework. Int. J. Environ. Res. Public
Health 2018, 15, 1360. [CrossRef] [PubMed]
6. WHO Regional Office for Europe. Evolution of WHO Air Quality Guidelines: Past, Present and Future; WHO
Regional Office for Europe: Copenhagen, Denmark, 2017.
7. WHO Regional Office for Europe. WHO guidelines for indoor air quality. Nutr. J. 2010, 9, 454.
8. General Administration of Quality Supervision, Inspection and Quarantine of the People’s Republic of China.
GB/T 18883-2002, Indoor Air Quality Standard; General Administration of Quality Supervision, Inspection and
Quarantine of the People’s Republic of China: Beijing, China, 2002. (In Chinese)
9. Asikainen, A.; Carrer, P.; Kephalopoulos, S.; Fernandes, E.D.O.; Wargocki, P.; Hänninen, O. Reducing burden
of disease from residential indoor air exposures in Europe (HEALTH-VENT project). Environ. Health A Glob.
Access Sci. Source 2016, 15 (Suppl. 1), 35.
10. CEN. Part 1: Indoor environmental input parameters for design and assessment of energy performance of
buildings addressing indoor air quality, thermal environment, lighting and acoustics. In EN 16798 Energy
Performance of Buildings–Ventilation of Buildings; CEN: Brussels, Belgium, 2019.
11. Wang, Z.; Bai, Z.; Yu, H.; Zhang, J.; Zhu, T. Regulatory standards related to building energy conservation and
indoor-air-quality during rapid urbanization in China. Energy Build. 2004, 36, 1299–1308. [CrossRef]
12. China National StandardsGB 3095-2012 Ambient Air Quality; The Standardization Administration of China
(SAC): Beijing, China, 2012.
13. The Government of the Hong Kong Special Administrative Region IAQ Management Group. A Guide on
Indoor Air Quality. Certification Scheme for Offices and Public Places; The Government of the Hong Kong Special
Administrative Region IAQ Management Group: Hong Kong, China, 2019.
14. Piasecki, M.; Kostyrko, K. Indoor Environmental Quality Assessment Model IEQ Developed in ITB. Part
1. Choice of the Indoor Environmental Quality Subcomponent Models. Ciepłownictwo Ogrzew. Went.
2018, 49, 223–232.
15. Sekhar, S.C.; Tham, K.W.; Cheong, K.W. Indoor air quality and energy performance of air-conditioned office
buildings in Singapore. Indoor Air 2003, 13, 315–331. [CrossRef]
111
Appl. Sci. 2019, 9, 3918
16. Moschandreas, D.J.; Sofuoglu, S.C. The indoor environmental index and its relationship with symptoms of
office building occupants. J. Air Waste Manag. Assoc. 2004, 54, 1440–1451. [CrossRef]
17. Sofuoglu, S.C.; Moschandreas, D.J. The link between symptoms of office building occupants and in-office air
pollution: The Indoor Air Pollution Index. Indoor Air 2003, 13, 332–343. [CrossRef]
18. IEA. Indoor Air Quality Design and Control in Low-Energy Residential Buildings-Annex 68 Subtask 1: Defining
the Metrics in the Search of Indices to Evaluate the Indoor Air Quality of Low-Energy Residential Buildings; AIVC
Contributed Report, 17; Abadie, M.O., Wargocki, P., Eds.; IEA: Paris, France, 2017.
19. Chen, C.; Chiang, C.; Shao, W.; Huang, L. Hazards of VOCs emissions from building material and ventilation
effectiveness in Taiwan. In Proceedings of the Indoor Air Conference, Copenhagen, Denmark, 17–22 August
2008; Volume 1.
20. Zhu, C.; Li, N. Study on Grey Clustering Model of Indoor Air Quality Indicators. Procedia Eng.
2017, 205, 2246–2253. [CrossRef]
21. Salisa, L.C.R.; Abadiea, M.; Wargocki, P.; Rode, C. Towards the definition of indicators for assessment of indoor
air quality and energy performance in low-energy residential buildings. Energy Build. 2017, 152, 492–502.
[CrossRef]
22. Piasecki, M.; Kozicki, M.; Firlag, S.; Goljan, A.; Kostyrko, K. The approach of including TVOCs concentration
in the indoor environmental quality model (IEQ)- case studies of BREEAM certified office buildings.
Sustainability 2018, 10, 3902. [CrossRef]
23. Toftum, J.; Jørgensen, A.S.; Fanger, P.O. Upper limits of air humidity for preventing warm respiratory
discomfort. Energy Build. 1998, 28, 15–23. [CrossRef]
24. Wang, Z.; Yu, Z. PM2.5 and Ventilation in a Passive Residential Building. Procedia Eng. 2017, 205, 3646–3653.
[CrossRef]
25. Morawska, L.; Ayoko, G.A.; Bae, G.N.; Buonanno, G.; Chao, C.Y.H.; Clifford, S.; Fu, S.C.; Hänninen, O.;
He, C.; Isaxon, C.; et al. Airborne particles in indoor environment of homes, schools, offices and aged care
facilities: The main routes of exposure. Environ. Int. 2017, 108, 75–83. [CrossRef]
26. Cornell’s Legal Information Institute. Available online: https:/www.law.cornell.edu/sfr/text/40/appendix-A_
to_part_58 (accessed on 20 May 2019).
27. Saad, S.M.; Shakaff, A.Y.M.; Saad, A.R.M.; Yusof, A.M.; Andrew, A.M.; Zakaria, A.; Adom, A.H. Development
of indoor environmental index: Air quality index and thermal comfort index. In AIP Conference Proceedings;
American Institute of Physics: Melville, NY, USA, 2017; Volume 1808, p. 020043.
28. Mintz, D. Technical Assistance Document for the Reporting of Daily Air Quality–The Air Quality Index (AQI); U.S.
EPA: Washington, DC, USA, 2013.
29. United States EPA-Environmental Protection Agency Office of Air Quality Planning and Standards Air
Quality Assessment Division Research Triangle Park, NC. Available online: https://airnow.gov/index.cfm?
action=airnow.calculator (accessed on 25 June 2019).
30. Wang, H.; Tseng, C.; Hsieh, T. Developing an indoor air quality index system based on the health risk
assessment. In Proceedings of the Indoor Air Conference, Copenhagen, Denmark, 17–22 August 2008;
Volume 1.
31. United States EPA-Environmental Protection Agency Office of Air Quality Planning and Standards Air
Quality Assessment Division Research Triangle Park, NC. EPA-454/B-18-007- Technical Assistance Document for
the Reporting of Daily Air Quality; United States EPA-Environmental Protection Agency Office of Air Quality
Planning and Standards: Research Triangle Park, NC, USA, 2018.
32. Sakellaris, I.A.; Saraga, D.E.; Mandin, C.; Roda, C.; Fossati, S.; de Kluizenaar, Y.; Carrer, P.;
Dimitroulopoulou, S.; Mihucz, V.G.; Szigeti, T.; et al. Perceived indoor environment and occupants’
comfort in European “modern” office buildings. The OFFICAIR Study. Int. J. Environ. Res. Public Health
2016, 13, 444. [CrossRef]
33. Piasecki, M.; Kostyrko, K.; Pykacz, S. Indoor environmental quality assessment: Part 1: Choice of the indoor
environmental quality subcomponent models. J. Build. Phys. 2017, 41, 264–289. [CrossRef]
34. Piasecki, M.; Kostyrko, K.B. Indoor environmental quality assessment, part 2: Model reliability analysis.
J. Build. Phys. 2018, 42, 288–315. [CrossRef]
35. Ye, W.; Won, D.; Zhang, X. A Simple VOC Prioritization Method to Determine Ventilation Rate for Indoor
Environment Based on Building Material Emissions. Procedia Eng. 2015, 121, 1697–1704. [CrossRef]
112
Appl. Sci. 2019, 9, 3918
36. Bienfait, D. ECA Report No. 11 Guidelines for Ventilation Requirements in Buildings; Commission of the European
Communities: Bruxelles, Belgium, 1992.
37. Ye, W.; Won, D.; Zhang, X. A practical method and its applications to prioritize volatile organic compounds
emitted from building materials based on ventilation rate requirements and ozone-initiated reactions. Indoor
Built Environ. 2017, 92–93, 324–333. [CrossRef]
38. Persily, A.K. Evaluating building IAQ and ventilation with indoor carbon dioxide. ASHRAE Trans. 1997, 103, 4072.
39. Gyot, G.; Sherman, M.H.; Walker, I.S.; Clark, J.D. Residential Smart Ventilation: A Review; LBNL 2001056;
Ernest Orlando Lawrence Berkeley National Laboratory: Berkeley, CA, USA, 2017.
40. Engineering ToolBox. Available online: https://www.engineeringtoolbox.com (accessed on 5 May 2019).
41. Noguchi, M.; Mizukoshi, A.; Yanagisawa, Y.; Yamasaki, A. Measurements of volatile organic compounds in a
newly built daycare center. Int. J. Environ. Res. Public Health 2016, 13, 736. [CrossRef] [PubMed]
42. Haag, W. Rapid continuous TVOC measurements using PPB-level PIDs. In Proceedings of the 10th
International Conference on Indoor Air Quality and Climate, Beijing, China, 4–9 September 2005; pp. 938–941.
43. Wong, L.T.; Mui, K.W.; Hui, P.S. A multivariate-logistic model for acceptance of indoor environmental quality
(IEQ) in offices. Build. Environ. 2008, 43, 1–6. [CrossRef]
44. American Industrial Hygiene Association (AIHA). Odor Thresholds for Chemicals with Established Health
Standards; American Industrial Hygiene Association: Falls Church, VA, USA, 2013.
45. Jokl, M.V. Evaluation of indoor air quality using the decibel concept based on carbon dioxide and TVOC.
Build. Environ. 2000, 35, 677–697. [CrossRef]
46. Tahsildoost, M.; Zomorodian, Z.S. Indoor environment quality assessment in classrooms: An integrated
approach. J. Build. Phys. 2018, 42, 336–362. [CrossRef]
47. Zhu, C.; Li, N. Study on indoor air quality evaluation index based on comfort evaluation experiment.
Procedia Eng. 2017, 205, 2246–2253. [CrossRef]
48. European Committe for Standardization. CEN/CR 1752:1998 Ventilation for Buildings. Design Criteria for the
Indoor Environment; International Organization for Standardization: Geneva, Switzerland, 1998.
49. Jokl, M.V. Indoor Air Quality Assessment Based on Human Physiology–Part 1. New Criteria Proposal.
Acta Polytech. 2003, 43, 31–37.
50. Wargocki, P.; Knudsen, H.N.; Krzyżanowska, J. Some methodological aspects of sensory testing of indoor air
quality. In Proceedings of the CLIMA 2010, Antalya, Turkey, 9–12 May 2010.
51. ISO. ISO 16000-28 Determination of Odor Emissions from Building Products Using Test Chambers; ISO: Geneva,
Switzerland, 2013.
52. Kim, K.H.; Kim, Y.H. Composition of key offensive odorants released from fresh food materials. Atmos. Environ.
2014, 205, 2246–2253. [CrossRef]
53. Wu, C.; Liu, J.; Zhao, P.; Piringer, M.; Schauberger, G. Conversion of the chemical concentration of
odorous mixtures into odor concentration and odor intensity: A comparison of methods. Atmos. Environ.
2016, 127, 283–292. [CrossRef]
54. Yeganeh, B.; Haghighat, F.; Gunnarsen, L.; Afshari, A.; Knudsen, H. Evaluation of building materials
individually and in combination using odor threshold. Indoor Built Environ. 2006, 15, 583–593. [CrossRef]
55. Gunnarsen, L.; Fanger, P.O. Adaptation to indoor air pollution. J. Environ. Int. 1992, 18, 43–54. [CrossRef]
56. Fang, L.; Clausen, G.; Fanger, P.O. Impact of temperature and humidity on the perception of indoor air
quality. Indoor Air 1998, 8, 80–90. [CrossRef]
57. Połednik, B. Zanieczyszczenia a Jakość Powietrza Wewn˛etrznego w Wybranych Pomieszczeniach; Biblioteka Cyfrowa
Politechniki Lubelskiej: Lublin, Poland, 2013. (In Polish)
58. Piasecki, M.; Fedorczak-Cisak, M.; Furtak, M.; Biskupski, J. Experimental Confirmation of the Reliability of
Fanger’s Thermal Comfort Model—Case Study of a Near-Zero Energy Building (NZEB) Office Building.
Sustainability 2019, 11, 2461. [CrossRef]
113
applied
sciences
Article
An Accident Model with Considering Physical
Processes for Indoor Environment Safety
Zhengguo Yang *, Yuto Lim and Yasuo Tan
School of Information Science, Japan Advanced Institute of Science and Technology 1-1 Asahidai, Nomi,
Ishikawa 923-1292, Japan; [email protected] (Y.L.); [email protected] (Y.T.)
* Correspondence: [email protected]
Received: 15 September 2019; Accepted: 1 November 2019; Published: 6 November 2019
Abstract: Accident models provide a conceptual representation of accident causation. They have
been applied to environments that have been exposed to poisonous or dangerous substances that
are hazardous in nature. The home environment refers to the indoor space with respect to the
physical processes the of indoor climate, e.g., temperature change, which are not hazardous in
general. However, it can be hazardous when the physical process is in some states, e.g., a state
of temperature that can cause heat stroke. If directly applying accident models in such a case,
the physical processes are missing. To overcome this problem, this paper proposes an accident
model by extending the state-of-the-art accident model, i.e., Systems-Theoretic Accident Model
and Process (STAMP) with considering physical processes. Then, to identify causes of abnormal
system behaviors that result in physical process anomalies, a hazard analysis technique called
System-Theoretic Process Analysis (STPA) is tailored and applied to a smart home system for indoor
temperature adjustment. The analytical results are documented by a proposed landscape genealogical
layout documentation. A comparison with results by applying the original STPA was made, which
demonstrates the effectiveness of the tailored STPA to apply in identifying causes in our case.
Keywords: STAMP; STPA; physical process; indoor environment safety; smart home systems
1. Introduction
Accident models provide a conceptual representation of accident causation [1]. Their
state-of-the-art development is on the phase of systemic models [1–3], i.e., accident models based
on system theory rather than reliability. They were specifically applied to understand accidents in
industrial areas, e.g., deepwater well control [4], railway [5], and aviation [6]. Some others relate
to places that have been exposed to poisonous or dangerous substances, e.g., oil transportation [7]
and nuclear power plants [8]. These poisonous or dangerous substances are hazards in nature,
which can directly cause harm when leaked or released in workplaces. The workplace is a strictly
managed environment for work. Safety-critical systems are taken as preventative measures for leakage
and release.
The home environment refers to the indoor space with respect to physical processes of indoor
climate, e.g., temperature change, which is different from safety-critical environments in workplaces.
Hereafter, we use physical process to represent the physical process of indoor climate. The home
environment is not a hazardous place in general. However, it can become hazardous when the physical
process transfers from a normal state (e.g., temperature for thermal comfort), then through some
intermediate states (e.g., temperature for thermal discomfort), finally reaching a hazardous state
(e.g., temperature for heat stroke). The home environment is a place for everyday living, and it is not
as strictly managed as that in workplaces. Smart home systems are developed to maintain the home
environment in desired states, not only as preventative measures.

Appl. Sci. 2019, 9, 4732
In systemic accident models [1–3], accidents are the result of the violation of a set of constraints
on the behaviors of the system components, i.e., management, humans, and technology. If directly
applying a systemic accident model to the smart home system, the information of physical processes is
missing. For example, a smart home system violated its constraint on adjusting the indoor temperature
for thermal comfort, and resulted in high temperature for heat stroke. The physical process of
temperature change from one of thermal comfort, to some intermediate states for discomfort, then to
a high-temperature state that can cause, e.g., heat stroke, is missing. This process is important. First,
it assists in understanding how system behaviors could result in accident through intermediate state(s).
Second, we need the anomalies’ information of the physical process and their causes to deploy reactions
and precautionary measures. We can take advantage of hazard analysis techniques to identify the
causes in the system to the anomalies. When an intermediate abnormal state or hazardous state of the
physical process is detected, with considering the corresponding causes in the system, an effective
precautionary or reaction measures can be selected. Generally, if something undesired is not considered
in the very beginning of risk analysis, the causes in the following analysis cannot be identified [2,9].
Therefore, it is necessary to extend the systemic accident model by including the physical process.
A newly developing systemic accident model, i.e., Systems-Theoretic Accident Model and Process
(STAMP) [2], is considered in this paper, as its underlying rationale has been widely acknowledged
by comparing with other systemic accident models [10]. It is based on general system theory for
understanding accident causality of sociotechnical systems. A brief introduction of it is presented
in Section 3.1. We extend it by considering the following facets. First, the home environment is not
inherently hazardous. It can ensure a comfortable life in some physical process states and cause harms
in others. Therefore, we take the physical process into account in understanding accident formation.
Second, the indoor environment is greatly affected by the behaviors of smart home systems. This is
because physical processes are the result of smart home systems and the outdoor climate. However, in a
limited period, e.g., days, the outdoor climate can be considered with no big changes. Third, the role
of people in the home environment. The characteristics of workers in workplaces and occupants in the
home environment are different.
In this paper, we extend the STAMP model with considering physical processes (hereafter denoted
by STAMP-PP) to understand accident formation. Smart home systems interact with the home
environment through its behaviors, e.g., warm up and cool down. Thus, the STAMP-PP connects the
physical world through the behaviors of the systems. Under this consideration, accidents are the result
of the violation of a set of constraints on the behaviors of the systems to cause abnormal changes in
physical processes, and finally result in personal harm. The extended STAMP-PP model demonstrates
accident formation with respect to system behaviors and physical processes. The system behaviors can
be controlled either by the smart home system directly or by occupants indirectly.
The information related to physical processes is important, and the abnormal behaviors of systems
under specific operation scenarios must be known to select the appropriate reactions and precautionary
measures. To this end, hazard analysis techniques [9,11] that can assist in analyzing potential causes
of accidents are required. We adopted a hazard analysis technique to identify causes of abnormal
system behaviors under related operation scenarios, which can result in abnormal changes in physical
processes. A new approach to hazard analysis, called System-Theoretic Process Analysis (STPA) [2,9],
is based on the STAMP model, and it is tailored and applied to the smart home system [12] for adjusting
the indoor temperature, to demonstrate how to identify causes to abnormal system behaviors that
result in physical process anomalies. The STPA can be used to identify unsafe control actions of
a controller and are also the reasons why unsafe control actions can happen under specific scenarios.
As abnormal behaviors of smart home systems that result in intermediate states of physical processes
are also considered, the STPA is then tailored also for identifying causes to these abnormal behaviors.
Landscape Genealogical Layout Documentation (LGLD) is proposed for documenting the analytical
results, and the relations among the results are clearly and straightforwardly represented by comparing
with conventional ways of documentation, i.e., tables and lists. We compared the results with that of
116
Appl. Sci. 2019, 9, 4732
applying the original STPA, which demonstrate the effectiveness of the tailored STPA in identifying
causes of abnormal system behaviors and the LGLD documentation in representing the relations
among the results.
The contributions of this paper are as follows.
• We discussed the characteristics of the smart home in the viewpoint of occupants and the safety
of the home environment.
• The concept of the Performers System, which emphasizes the behaviors performed by various
home appliances, is proposed.
• We propose the STAMP-PP model for understanding accident formation, i.e., abnormal system
behaviors that result in abnormal changes in physical processes and cause hazards.
• We tailored the STPA and applied it to a smart home system to identify inappropriate and unsafe
control actions that cause abnormal system behaviors, which result in abnormal changes in
physical processes, and hazards.
• An LGLD approach is proposed for documenting the STPA analytical results.
This paper extends a conference paper [13] that introduces the STAMP-PP model while
considering physical processes. The application of the tailored STPA and the new way of documenting
the results in this paper are novel.
The rest of this paper is organized as follows. Section 2 discusses some knowledge about smart
home and home environment safety. Section 3 introduces the proposed accident model STAMP-PP.
Then, the hazard analysis technique STPA is tailored and applied to a smart home system for indoor
temperature adjustment in Section 4. In Section 5, a discussion is given. Section 6 introduces the
related work. Finally, Section 7 concludes this paper and points out the future work.
2. Preliminaries
In this section, we discuss the characteristics of the smart home and home environment safety
before introducing the STAMP-PP model and the application of the STPA.
2.1. Smart Home

A home is a place for people like individual or family members, etc. to live. It is the sum of the
place where people live permanently and the social unit—family. Since the 20th century, with the
introduction of electricity, informtion, and communication technologies, great changes have taken
place in the home [14]. One representation of this change is the development of the concept of the
smart home since the 1990s [15]. The primary objective of the smart home is to increase occupants’
comfort and make daily life easier. An example of possible techniques that make a home smart is
machine learning [16]. The smart home has certain characteristics [15,17], e.g., adaptability, connectivity,
controllability, and computability. These are discussed from the viewpoint of technology. This section
discusses the characteristics of the smart home from the viewpoint of occupants.
• Partial Automation: Although home life has been automated a lot more than ever before due to
the development of electricity, electronics, and network technologies, there are still elements of
our lifestyles that have been left unchanged in practice. For example, pots and pans are used for
cooking with gas in everyday meals. Thus, contemporary homes are only partially automated
in practice.
• Application Area: The home is a place where people live. Various off-the-shelf products are
used to improve the quality of life. These are manufactured by different manufacturers for
different purposes. Occupants are not professional in understanding their rationale, particularly
that of high-tech products. Occupants only learn to use them through product instructions or
other occupants.
• Complexity: The complexity of a home is owing to three aspects, i.e., variety of appliances
and products; variety of occupants in terms of age, health condition, knowledge, gender, etc.;
117
Appl. Sci. 2019, 9, 4732
and outdoor environment. Off-the-shelf appliances and products indoors have various purposes
and are produced by different manufacturers. In smart homes, they are connected together by the
smart home network to enable a variety of services [15,17]. Different from workers in workplaces
which rely on skills, rules, and knowledge for a specific job [18], occupants are usually reliant
on their own life experiences to lead their lives with respect to various indoor items. Outdoor
climate, air quality, etc. can affect the indoor environment, which in turn impacts the working of
indoor appliances or devices. All these add to the complexity of the smart home.
2.2. Home Environment Safety

If we refer to a dictionary, the definition of safety could be the condition of being protected from
or unlikely to cause harm, injury or loss. In system safety [2], safety is taken as an emergent property
of systems. It is defined as freedom from conditions that can cause death, injury, occupational illness,
damage to or loss of equipment or property, or damage to the environment [19]. However, it cannot be
freedom from the conditions in practice. Safety is, thus, the condition of risk that has been reduced
to an acceptable level, e.g., as low as reasonably practicable [20]. Home environment safety could be
understood as that the risk of indoor climate has been reduced to a level of no harm to the health
of occupants.
As is known to all, indoor climate is affected by outdoor climate. Bad weather such as heatwaves
has been occurring frequently in recent years due to global warming and weather anomalies around
the world. For example, many places globally experienced intense heat in the summer of 2018. It was
observed that the highest temperature record in Japan was broken and reached a new level of 41.1 ◦ C
(Japan Meteorological Agency [visited 2018.10.03] http://www.jma.go.jp/jma/index.html). Indoor
climate can thus be endangered by outdoor climate anomalies.
The indoor climate is adjusted by dedicated home appliances, e.g., use of air-conditioner to
adjust indoor temperature. Different home appliances are integrated via home networks that yield
value-added integrated services [21,22]. In order to ensure thermal comfort, an indoor temperature
adjustment service is an example of the integrated service, which can potentially adopt a window,
curtain, and air-conditioning unit. Each involved home appliance may have safety instructions.
However, due to the complexity of the smart home, they still could be used in scenarios that cause safety
problems. For example, the heating mode of an air-conditioner was used when it should not be. Since
the smart home is an application area, appliances inside it may be replaced from time to time. Once
an appliance is introduced into the smart home, it also brings about risk. For example, the predefined
integrated service may not be aware of the new item and cause safety problems when using it.
Therefore, home environment safety depends on the proper use of the indoor climate adjustment
service with respect to related home appliances if they were properly designed and manufactured.
To understand home environment safety, we also need to discuss how people relate to the smart
home. This is mostly because home appliances are produced and operated by people. One group of
people are professionals. These relate to the activities of design, manufacture, transport, installation,
and disposal of appliances, home networks, and so on. One distinguishing characteristic is that they
have expertise in a certain field. The other group is occupants who are non-professionals. They are the
customers who use the various home appliances. Both groups of people can affect home environment
safety in different ways. Professionals are responsible for designing, manufacturing, etc. safe systems
and home appliances. Occupants care more about operational safety, since they are more error-prone
in operations. We talk about occupants in this paper.
3. Accident Model
As the smart home is an application area, the STAMP-PP model is discussed with respect to
system operations rather than system development. It aims to understand how accident formation
relates to system behaviors in adjusting the home environment, i.e., how abnormal system behaviors
cause indoor climate anomalies. Concrete information about indoor climate anomalies can be used for
118
Appl. Sci. 2019, 9, 4732
indoor climate anomaly detection [23], and abnormal system behaviors under operation scenarios can
further be used in selecting reactions and precautionary measures when the corresponding indoor
climate anomaly is detected.
The STAMP-PP model starts with system behaviors to describe accident formation. The behaviors
can result in abnormal changes in physical processes, which can cause discomfort or harm.
The connection between smart home systems and the home environment is the system behaviors.
In this section, we first give a brief introduction of the STAMP model, then discuss in detail the
proposed STAMP-PP model.
3.1. STAMP
In systems theory, systems are viewed as interrelated components kept in a state of dynamic
equilibrium through feedback control loops. Figure 1 presents a standard contrl loop [2]. The STAMP
model is based on system theory rather than the reliability that traditional accident models are
grounded on. Safety, in STAMP, is an emergent property of systems. Accidents are the result of
the lack of or inappropriate constraints imposed on the system design and operations. The STAMP
model consists of three building blocks, i.e., safety constraints, a hierarchical safety control structure,
and process models.

Figure 1. A standard control loop.
Safety constraints are a basic concept in the STAMP model. Losses occur only because safety
constraints were not successfully enforced. Systems, in system theory, are viewed as hierarchical
structures, where each level imposes constraints on the activity of the levels beneath it. Constraints
are enforced by control actions of a higher-level system component (controller) to the lower-level one
(controlled process).
The hierarchical safety control structure presents all stakeholders with their internal structures
within the system under analysis, and the control actions and feedbacks that link the independent
stakeholders and their internal components [6]. Control processes operate between levels of a system
to control the processes at lower levels. The feedbacks provide information about how effectively
the control actions ensure the constraints are enforced. The higher level uses the feedbacks to adapt
future controls to more readily achieve its goals. An accident occurs when control processes provide
inadequate control that violates safety constraints. Inadequate control comes from missing constraints,
inadequate safety control commands, commands that were not executed correctly at a lower level,
and inadequately communicated or processed feedback about constraint enforcement.
Process models are used by the controller to determine appropriate control actions. It is up to
the type of the controller. For an automated controller, the process model is embedded in the control
logic. For a human controller, the process model is the mental model. In both situations, it contains
information of the required relationship among the system variables, the current system state, and the
ways the process can change state. There are four conditions required to control a process, i.e., goal,
119
Appl. Sci. 2019, 9, 4732
action condition, observability condition, and model condition. Accidents related to component
interaction can usually be explained in terms of an incorrect process model. The process model used
by the controller does not match the controlled process that results in interaction accidents.
In the STAMP model, safety is achieved when the behaviors of components of a system
appropriately ensured safety constraints. Accidents are the results of flawed processes involving
interactions among people, societal and organizational structures, engineering activities, and physical
system components that lead to violating the system safety constraints. The process leading up to
an accident is described in terms of an adaptive feedback function that fails to maintain safety as
system performance changes over time to meet a complex set of goals and values.
3.2. STAMP-PP
Since the STAMP-PP model focuses on the behaviors of systems to affect physical processes, we
first define the concept of systems that emphasize behaviors, i.e., Performers System. The reason for
choosing the word “performer” is to highlight that the behaviors are performed by the systems. Then,
based on the Performers System, we can describe accident formation considering physical processes.
3.2.1. Performers System

The indoor environment is adjusted by indoor environment adjustment services with respect to
various home appliances. To differentiate the home appliances with other indoor items, e.g., router or
furniture, we define the concept of Performer.
Definition 1 (Performer). A performer is a network-enabled home appliance that can adjust the indoor
environment independently.
There are two points to explain this definition. First, a Performer has networking capability so
that it can be used by indoor environment adjustment services. Second, it has functions of adjusting
the indoor environment, e.g., adjusting indoor temperature, or dehumidification.
There are two types of Performers based on the way it adjusts the indoor environment. One is
direct adjustment, e.g., an air-conditioner which heats or cools indoor air directly; another is indirect
adjustment, e.g., an electric window which adjusts, e.g., indoor temperature by introducing an air flow
or solar radiation of the outdoor environment. In the latter case, the outdoor climate is passively used
to adjust the indoor environment. Then, concept of Performer is used to define the Performers System.
Definition 2 (Performers System). It is a system of all installed Performers in a house that are connected to
the same home network.
By connecting to the same home network, we can ensure that the Performers System can be used
by the same indoor environment adjustment service. The Performers System has a goal prescribed
by the indoor environment adjustment service. The goal is achieved by taking advantage of the
functions of the Performers of the Performers System. Each Performer is taken as a subsystem of the
Performers System. However, there is no need to have all related Performers working at the same
time. Figure 2 shows an example of the Performers System. It consists of an air-conditioner, an electric
window, and an electric curtain, which are taken as Performers. They connect to the same network
and have the ability to adjust the indoor temperature. When adjusting the indoor temperature for
thermal comfort (the goal), the indoor temperature adjustment service could use any combination of
the Performers, but not necessarily all of them. The indoor temperature adjustment service is executed
in the smart home system core. The Performers can of course be operated by occupants who are also
the beneficiaries of the adjustment.
120
Appl. Sci. 2019, 9, 4732

Figure 2. An example of the Performers System.
The Performers System can adjust physical processes with respect to various physical properties,
e.g., temperature and humidity, by utilizing the functions the Performers provide. These physical
processes may have different forms, for example, increasing or decreasing indoor temperature to the
prescribed temperature level for thermal comfort.
3.2.2. Service
In this section, we discuss the behaviors of the Performers System, and when the behaviors can
be taken as Services.
Definition 3 (Behavior of the Performers System). The way the Performers System behaves to adjust the
indoor environment.
The Behaviors of the Performers System are the representation of the functions of related
Performers. For example, the Performers System in Figure 2 has the ability to cool the temperature
down (behavior), which could be achieved by setting a lower temperature level under the cool
mode of the air-conditioner (function). The representation of the Behaviors is the physical process,
e.g., temperature change.
Definition 4 (Service). The Behavior of the Performers System exhibited in order to fulfill occupants’
comfort requirement.
One example of the Service can be the Performers System in Figure 2, which increases the indoor
temperature to 22 ◦ C for thermal comfort. Comfort means psychological and physical satisfaction with
the state of the indoor environment, e.g., thermal comfort. It is the way to evaluate the Behaviors of
the Performers System, and thus implicitly constrain the Behaviors. The comfort has different contents
for different goals of the Performers System, e.g., thermal comfort; comfort in terms of humidity levels.
There are two ways to evaluate comfort. Let us take thermal comfort as an example. The first is
based on the perception of occupants. If occupants feel uncomfortable, one can manually set a desired
temperature level to the Performers System. It is easy and accurate but limited in some scenarios.
For example, babies and elderly people may not be sensitive to temperature change due to their
nervous system not being well developed or being degenerated. The second way is for comfort to
be automatically evaluated by the smart home system core. This depends on various indices [24,25]
for evaluating the physical environment for thermal comfort. For example, the PMV–PPD (Predicted
Mean Vote and Predicted Percentage of Dissatisfied) index [25] is used to evaluate thermal comfort
with respect to environmental factors and personal factors. PMV and PPD are short for predicted mean
vote and predicted percentage of dissatisfaction, respectively, and are calculated based on Fomulas (1)
121
Appl. Sci. 2019, 9, 4732
and (2), where TS denotes the thermal sensation transfer coefficient and is determined by metabolic
rate; MV means internal heat production in the human body and is determined by the metabolic
rate and external work; and HL1, HL2, HL3, HL4, HL5, and HL6 represent heat losses through skin,
sweating, latent respiration, dry respiration, radition and convection, respectively. Then combine with
the ASHRAE (American Society of Heating, Regrigerating and Air-Conditioning Engineers) thermal
sensation scale as shown in Table 1 to determine the comfortableness. The details are referred to in [25].
PMV = TS × (W M − HL1 − HL2 − HL3 − HL4 − HL5 − HL6) (1)
PPD = 100 − 95 × e−0.03353× PMV

4 −0.2179× PMV 2
(2)
Table 1. ASHRAE thermal sensation scale.
Hot Warm Slightly Warm Neutral Slightly Cool Cool Cold

+3 +2 +1 0 −1 −2 −3
Environmental factors like humidity can be acquired through humidity sensors. Personal factors
like metabolic rate can be roughly evaluated based on the occupant’s activities, e.g., sedentary. Another
way is to take advantage of wearable devices to measure personal information and send these data to
the smart home system core to evaluate comfortableness. However, as far as the author is aware, cloth
insulation cannot be evaluated through wearable devices for now. It can only be roughly evaluated
through scenarios like in hot summer, where the value of cloth insulation is small, and an average
value is assigned for the summer season.
Definition 5 (Service Manner). The Service under a specific condition is called a Service Manner.
The condition can be personal or environmental. For a Service, e.g., adjust the indoor temperature
for thermal comfort by an air-conditioner, changes in conditions result in different Service content.
For example, compare a sweating man in summer and a sedentary man in winter. The conditions in
the former are high metabolic rate, high outdoor temperature, etc., and he needs for the temperature
to be cooled down. For the latter, the conditions may be low metabolic rate, low outdoor temperature,
etc., and he needs for the temperature to go up.
Definition 6 (Critical Service Manner). The Service Manner in demand is called Critical Service Manner.
This definition is given from the viewpoint of people who need the Service. In the example from
Definition 5, the adjustment of indoor temperature to achieve a cool environment for thermal comfort
is the Critical Service Manner for the sweating man; the adjustment of indoor temperature to achieve a
warm environment for thermal comfort is the Critical Service Manner for the sedentary man.
3.2.3. Accident Formation

Accidents can be understood as the resilience [26] of the Performers System, i.e., adjusting indoor
environment anomalies to maintain a normal performance has failed and resulted in undesired
consequences. A Service may fail and result in uncomfortableness, then further evolve into hazards
and cause harm to occupants. Accident formation is to describe how Services may fail and further
evolve into a hazard to cause accidents. The causes on the system part, i.e., the Performers System, can
be understood by the STAMP model. The relation between the Behaviors of the Performers System
and the physical processes in accident formation is discussed in this section.
There are some considerations about physical processes from the viewpoint of engineering.
Before a hazard is detected, physical process anomalies should be detected as early as possible so as to
trigger precautionary measures. The information of hazards is also important for triggering reaction
122
Appl. Sci. 2019, 9, 4732
measures. All in all, accident formation, which is shown in Figure 3 focuses on the physical process
anomalies resulting from abnormal Behaviors.

Figure 3. The accident causality model.
Next, let us discuss the terms and related rationales of the accident causality model in Figure 3.
Definition 7 (Defect). Defects are the direct causes of a Mutation.
A Mutation (Definition 8) relates to abnormal Behaviors of the Performers System. Abnormal

Behaviors are due to unexpected control actions of the Performers System. Thus, Defects are
inappropriate control actions (ICAs) and unsafe control actions (UCAs) of the Performers System to the
abnormal Behaviors in adjusting the home environment. ICAs occur when Mutations result in Service
Failures (Definition 9), and UCAs occur when Service Failures evolve into Hazards (Definition 10).
For example, an ICA can be wrongly setting the heating mode of a Performer in hot summer, which
will cause thermal discomfort. To differentiate from causes introduced in the STAMP model, Defects
refer to more superficial reasons. As the home environment is an application area, the deeper causes for
system development defects are not considered here. When physical process anomalies are detected,
the home environment is expected to react to them immediately with respect to Defects. Therefore,
the Defects are the direct causes and should be controllable, e.g., through reconfiguration.
As discussed at the beginning of Section 3, the STAMP-PP model relates to system operations.
The occurrence of Defects is under the scenarios of system operations. The operations can be
categorized into three types. One is operations by occupants. Another is by some controllers,
e.g., Performers. The other is a mixture of by occupants and the controller. To react to a physical
process anomaly efficiently, it is necessary to know the Defect with its corresponding operation scenario.
To identify Defects with respect to the scenarios, we applied the hazard analysis technique STPA [2,9]
in Section 4.
Definition 8 (Mutation). A Mutation is the violation of constraint on the Behavior of the

Performers System.
Abnormal changes in physical processes are the representation of the abnormal Behavior of
the Performers System. Therefore, the constraint on the Behavior of the Performers System is the
prescription of changes in physical processes that satisfy the comfort purpose. Thus, the Mutation is the
change of physical processes under adjustment unacceptably deviating from the prescribed curve(s),
which results in uncomfortableness. For example, the amplitude of temperature fluctuation should
not exceed a threshold value for thermal comfort [23]. The Mutation in this case is the amplitude of
temperature fluctuation exceeding the threshold value.
The Mutation is a state of physical processes between the state that brings about comfort and the
state that results in a hazard. This concept is important not only for the understanding of accident
123
Appl. Sci. 2019, 9, 4732
formation, but also because the information can trigger precautionary measures. First, a Mutation
indicates the current indoor environment adjustment is inefficient. Second, by combining the Mutation
with Defects under certain operation scenarios, one can select appropriate precautionary measures to
restore the state of physical processes that bring about comfort.
Definition 9 (Service Failure). The Behavior of the Performers System exhibits failed to fulfill occupants’
comfort requirement.
The concept of Mutation is related to physical processes, while Service Failure refers to both
physical processes and the perception of occupants. The occurrence of Service Failure is when
occupants perceive the uncomfortableness brought about by the Mutation.
There are two ways to determine whether a Service has failed. One is directly determined by
the perception of occupants. Another one resorts to various indices [24,25], by which the smart home
system core can conclude whether the home environment satisfies the comfort requirement with
occupants on the scene.
Definition 10 (Hazard). It is an indoor environment state that will cause harm to occupants.
A Hazard will harm the health condition of occupants who are on the scene. One or more Service
Failures will form or further evolve into a hazardous situation. For example, a Service Failure for
thermal discomfort evolves into a hazard if the indoor temperature reaches a level that can cause
heat stroke if occupants are present.
Taking thermal related hazards as an example, the evaluation of a hazard depends on heat
stress indices [27] and cold stress indices [28]. These indices are sophisticated techniques to represent
thermal sensations to hot and cold conditions. Heat stress [27] is defined as the net heat load to which
an occupant is exposed from the combined contributions of metabolic heat, environmental factors,
and clothing that results in an increase in heat storage in the body. Cold stress [28] is the climatic
condition under which the body heat exchange is just equal to or too large for heat balance at the
expense of significant heat and sometimes heat debt. Similarly to the PMV–PPD index, they also have
a complex relations with environmental and personal factors, which can be measured in practice.
Definition 11 (Accident). It is an unintentional event where a Hazard results in harm of occupants.
It involves both the home environment and occupants. The Hazard has harmed occupants.
The Accident can be detected by evaluating the Hazard and the health condition of the occupants.
The latter can be measured by taking advantage of wearable devices.
Definition 12 (Harm). Death, physical injury or damage to the health of occupants.
It is the consequence of the Accident. It varies with respect to Hazards and the health conditions
of occupants. For example, it may cause heat illnesses or even death to elderly people due to heat
exposure [29]; it may also affect sleep and the circadian rhythm that can cause cardiac autonomic
response during sleep due to cold exposure [30].
4. Application of STPA
To identify Defects under operation scenarios, the hazard analysis technique STPA is adopted.
In this section, we first introduce the STPA steps, then discuss the way to tailor it and a new way of
documenting the analytical results, and finally illustrate the application results and compare them
with the application of the original STPA.
124
Appl. Sci. 2019, 9, 4732
4.1. STPA
The Systems-Theoretic Process Analysis (STPA) [2,9] is a new hazard analysis technique based on
the STAMP model. The goal is to identify causes that lead to hazards and result in losses so they can be
eliminated or controlled. The STPA has three steps, the latter two of which are taken as the main steps.
The first step is to establish the system engineering foundation. Three things should be done in
this step. The first is to define the interested accident and its related system hazards. They should be
specific and concise. Then, the system level safety requirements and design constraints need to be
specified to prevent system hazards from occurring. The last is to define the safety control structure
that takes the system level safety requirements and design constraints as inputs.
The second step is to identify potentially unsafe control actions (UCAs). Every controller of
systems usually has one or more control actions. System hazards thus result from inadequate control
or enforcement of the safety constraints. Four taxonomies are provided to look for the UCAs:
1. A control action required for safety is not provided or not followed;

2. An unsafe control action is provide;
3. A potentially safe control action is provided too early or too late (at the wrong time or in the
wrong sequence);
4. A control action required for safety is stopped too soon or applied too long.
Then, we translate the identified UCAs into safety requirement and designed constraints on
system component behaviors.
The third step determines how each UCA identified in step two could occur. For each UCA,
examine the parts of the control loop as shown in Figure 4 to see if they can cause the UCA under
some scenarios. Then, design controls and mitigation measures if they do not already exist or evaluate
existing measures.
%# !&% #)%#

#% (# #$$
%# #
"&% %# # $$

#%($#% $$%%
!!# !#% !# $$$ ##% !% #
.%' #$$ "&% #
,% #!%% ##%
%# % $$
*$
%&% #

"&% "&%
!#% !#%
##% #
#% !# '
*
!#% %# # $$ $&#%
&#$
-%
! %&#$
%# # *$
%# % $ $ '#%
# $$!&% # $$ &%!&%

$$ #(# %#&%$%
%, # $*$%+#
&% #
$%&#
Figure 4. A classification of controls flaws leading to system hazards.
4.2. Tailor
The understanding of accident formation within the indoor environment has extended the STAMP
model, i.e., the STAMP-PP model, by also considering physical processes. As disscussed in Section 3.2.3,
abnormal Behaviors result in Service Failures, and abnormal Behaviors are due to inappropriate control
actions of the Performers System. Thus, it is important to know what these inappropriate control
actions are. Therefore, the STPA need to be tailored for this main purpose.
In the first step of STPA, in addition to the definition of accident, system hazard, and safety
requirement, information about the Service Failure and requirements of no failure, i.e., the reliability
of the Performers System to deliver Services, should also be provided. Design constraints are not
required as this work is not for implementing a safe system.
When a Service Failure occurs, which means the corresponding control action is inappropriate,
then precautionary measure(s) should be provided. Thus, in the second step of STPA, ICAs should
125
Appl. Sci. 2019, 9, 4732
be identified with regard to the taxonomies provided in step two of STPA. Namely, given a state of
physical processes, under a specific operation scenario, we need to consider whether a control action
with respect to the taxonomies will cause a Service Failure. Precautionary measures are determined
by considering the context information which consists of the ICAs, operation scenarios, and the
Service Failure. The process of the determination heavily depends on the expertise of concerned
areas, which is directed by the reliability requirements. The precautionary measures are also control
actions. Ineffective precautionary measures will result in the occurrence of a Hazard (then, reaction
measures are required). For each precautionary measure, UCAs are identified by considering the
taxonomies provided in step two of STPA under the conditions of operation scenarios and of a state of
physical processes. Safety requirements to the UCAs also need to be identified to guide the selection of
reaction measures.
The third step of STPA is not necessary. This is because the analysis in our case is to identify
Defects under operation scenarios, which can be utilized in selecting appropriate precautionary and
reaction measures, but not in designing and manufacturing a system.
STPA adopted tables and lists for documenting analytical results [2,9]. After applying the tailored
STPA, the results, i.e., control actions, ICAs, and their related reliability requirements and operation
scenarios; and precautionary measures, UCAs, and their related safety requirements and operation
scenarios, could be documented by that used by the STPA. However, the relations among them are
not clearly represented. In this paper, we propose a Landscape Genealogical Layout Documentation
(denoted as LGLD) for documenting the results, which is illustrated in Figure 5. The ancestor is a
control action. The first generation illustrates ICAs and their related reliability requirements and
operation scenarios. The second generation represents precautionary measures. The third generation
represents UCAs and their related safety requirements and operation scenarios. These results can be
numbered for better reference, e.g., ICA-m for an ICA, which means the inappropriate control action
m. This way of documentation also implies the analysis direction, i.e., from control actions to ICAs,
then to precautionary measures, and finally to UCAs.

Figure 5. Documenting the analytical results by the Landscape Genealogical Layout Documentation
(LGLD) approach.
For each control action, the taxomomies provided in step two of STPA to list the ICAs must be
considered. Every ICA is attached with the reliability requirements and operation scenarios. Then,
each ICA is connected with a precautionary measure. The precautionary measure is to prevent its
connected ICA from changing the home environment from a Service Failure to a Hazard. For each
126
Appl. Sci. 2019, 9, 4732
precautionary measure, it is important to list the UCAs attached with safety requirements and
operation scenarios.
4.3. Results
As discussed in Section 2.2, weather anomalies affect indoor environment, e.g., the heatwave of
summer 2018. Thus, we consider the example of the high-temperature results in heat stroke [31,32].
Heat stroke is clinically diagnosed as a severe elevation in body temperature (a core body temperature
of 40◦ C or higher) that occurs in the presence of central nervous system dysfunction and a history
of environmental heat exposure or vigorous physical exertion. It can be classified into nonexertional
(classic) heat stroke and exertional heat stroke. The former occurs in very young or older people,
or those with chronic illness when the environmental temperature is high. The latter happens to
young fit people and involves prolonged excessive activities like sports. This paper focuses on the
former case.
Heat stroke can be assessed by heat stress indices, among which the most widely accepted and
used one is the wet bulb globe temperature (WBGT) [33]. The Ministry of the Environment of Japan
(Ministry of the Environment, Japan: http://www.wbgt.env.go.jp/en/) has recommended a criterion
for thermal conditions based on the WBGT as shown in Table 2.
Table 2. Recommended criteria for thermal conditions.
WBGT (◦ C) Threat Level

∼21 Almost Safe
21∼25 Caution
25∼28 Warning
28∼31 Severe Warning
31∼ Danger
We applied the tailored STPA to a Performers System that is for indoor temperature adjustment
for thermal comfort. The structure of the system consists of four parts, i.e., home, home gateway,
service intermediary, and service provider [12], as shown in Figure 6. The home is the place that
occupants live in, which is equipped with different kinds of items, e.g., home appliances, to meet
everyday living requirements. The home gateway is the gateway of the networked Performers to
the outside world of the home. The indoor temperature adjustment service is executed here to issue
commands for the control of Performers. The service intermediary aggregates various services from
different service providers and maintains them locally. It can also respond to service subscriptions from
home gateways. The service provider designs, publishes, and updates concrete services to the service
intermediary for future use. We applied STPA to investigate the Behaviors of the Performers System
to elicit possible Defects under related operation scenarios. Thus, we focus on executing the indoor
temperature adjustment service in the home gateway to adjust the indoor temperature. The home
gateway, service intermediary, and service provider can be taken as the smart home system core.

Figure 6. The structure of the Performers System.
127
Appl. Sci. 2019, 9, 4732
In order to apply STPA, an assisting tool, i.e., STAMP Workbench (the STAMP Workbench is an
open-source, free, easy-to-use tool for people who are interested in system safety analysis by using the
STAMP/STPA. It was developed by the IT Knowledge Center of Information-Technology Promotion
Agency, Japan. https://www.ipa.go.jp/english/sec/complex_systems/stamp.html) is adopted for the
analysis. The STAMP Workbench is claimed to include features such as concentration on thinking and
help analysis and is not just an editing tool, or guide analysis procedure, but an unlimited and intuitive
operation. Analytical results can be exported into Excel files and images.
In the first step of STPA, we prepared some concepts for further analysis as shown in Table 3.
For demonstration, Service Failure is defined as when the indoor WBGT temperature is adjusted within
(25,28] ◦ C, and the Hazard is when the indoor WBGT temperature is adjusted over 28 ◦ C. The reliability
requirement corresponds to the Service Failure, which represents the requirement to ensure a Service
will not fail.
Table 3. Preparation for the tailored System-Theoretic Process Analysis (STPA).
Accident Physical harm of occupants due to heat stroke

Service Failure Indoor WBGT temperature is within (25,28] ◦ C
Reliability Requirement Indoor WBGT temperature should be adjusted bellow 25 ◦ C
Hazard Indoor WBGT temperature is over 28 ◦ C
Safety Requirement Indoor WBGT temperature should be adjusted bellow 28 ◦ C
Figure 7 illustrates the safety control structure for indoor temperature adjustment. The Home
Gateway is the controller, which is responsible for executing the indoor temperature adjustment service.
The controlled process is the Home Environment. Performers are taken as actuators. We consider an
air-conditioner and a window here as Performers. Empirically, for energy saving, these two cannot
work at the same time. The control actions are listed on the arrow from the Home Gateway to the
Performers. The control action "set to X ◦ C" means to set Performers to adjust the indoor temperature
to X ◦ C. The feedback to the Home Gateway is the indoor temperature. Indoor temperature can be
adjusted by the indoor temperature adjustment service that is executed in the Home Gateway, or by
occupants to issue commands, i.e., control inputs to the Home Gateway.

Figure 7. The safety control structure for indoor temperature adjustment.
Next, based on the results of the first step, we identifies ICAs and UCAs and elicited their related
requirements and operation scenarios. Part of the results are shown in Figures 8 and 9, which adopted
the LGLD approach introduced in Section 4.2 for the documentation. They illustrate the results
of analyzing the control actions “set OFF” and “set to X ◦ C”. “N/A” denotes the taxonomy is not
applicable to the corresponding control action. In Figure 8, the second “set OFF” can be considered
as a reconfiguration compared with the first one. The precautionary measure “set Cool mode” can
be deployed for the two ICAs that relate to the “set OFF”. One reason could be that a different
configuration has a higher possibility to restore the physical process to a comfortable state, as the “set
128
Appl. Sci. 2019, 9, 4732
OFF” has caused the ICA. For the second operation scenario, not-providing ICA, it is because people
are not sensitive to temperature change and thus did not issue the “set OFF” command manually.
For the results as shown in Figure 9, X satisfies X < 25 ◦ C. There are two reasons for the
not-providing ICA and UCA to say that the Performers System is working in the Fan mode. The first
is if it works in the Warm mode, which in this case is inappropriate or hazardous, and we cannot make
a solemn vow to conclude that not providing “set to X ◦ C” is inappropriate or hazardous. Second,
the other working modes, i.e., Dry and Cool, have a cooling effect based on our experience, which
may not be an ICA or UCA even when “set to X ◦ C” is not provided. One more thing that needs to be
explained is that providing “set to X ◦ C” at the “wrong time” is neither inappropriate nor hazardous.
If it is not provided in time, the home environment would experience Service Failure or Hazard for
some time. However, providing “set to X ◦ C” at a different time should not be inappropriate or
hazardous. Conversely, since “set to X ◦ C” is provided, the home environment could be restored to a
safe level, even though later than expected. In this case, providing “set to X ◦ C” can be thought of as a
precautionary or reaction measure, rather than an inappropriate or hazardous control action.
1()2(%*"&'%+ ,$) $%%')#&')*'

"%$()%0$) (,%'! $ $)
#%

1()2$%)&'%+ ,$) $%%')#&')*'"%$(
)%0$) (,%'! $ $) #%
(*$,'% ((* $1()2%##$

"&%&"%'+'--%*$&%&"'$%)($( ) +)%
)#&')*'$
*&$)(%$%)!$%,%,)%($1()2%##$

*&$)(''"((
1()2(%*"$%)&'%+ ,$) (

,%'! $ $)#%$) $%%')#&')*'"%$()%
/

1()2&'%+ ,$) (,%'! $
$)#%$) $%%')#&')*'"%$()%/
1()2(%*"&'%+ ,$) $%%')#&')*' $'(()%
&&'%0$) (,%'! $ $) #%

1()2$%)&'%+ ,$) $%%')#&')*' $'(()%
()
&&'%0$) (,%'! $ $) #% 1()2(%*"&'%+ ## )"-,$) $%%'
)#&')*'"%$()%/$) (
(*$,'% ((* $1()2%##$ ,%'! $ $) #%
"&%&"%'+'--%*$&%&"'$%)($( ) +)%)#&')*'
$ 1()2&'%+ )')($
*&$)(%$%)!$%,%,)%($1()2%##$ ,%'! $ $) #%%'(%#) #,$) $%%'

*&$)(''"((
)#&')*'"%$()%/

1()2(%*"$%)&'%+ ,$) $%%')#&')*' $'(()%
&&'%/$) (,%'! $ $)#%
1()#%2(%*"&'%+ ,$) $%%')#&')*'"%$(

1()2&'%+ ,$) $%%')#&')*' $'(()% )%/$) (,%'! $ $) #%
&&'%/$) (,%'! $ $)#%

1()#%2$%)&'%+ ,$) $%%')#&')*'"%$(
)%/$) (,%'! $ $) #%
# ()!$"- ((*1()2%##$
*&$)(''"((

()

1()2(%*"&'%+ ## )"-,$) $%%')#&')*'
$'(()%&&'%/$),%'!( $)

()#% %).'%*(
#%"
1()2&'%+ )')($,%'! $ 1()#%2(%*"&'%+ ## )"-,$)

$) #%%'(%#) #,$) $%%')#&')*' $'(( (,%'! $ $)'##%$) $%%')#&')*'
)%&&'%/ "%$()%/

1()#%2&'%+ )')(
$,%'! $ $) #%%'(%#) #,$) $%%'

)#&')*'"%$()%/

Figure 8. The analysis results for the control action “set OFF”.
129
Appl. Sci. 2019, 9, 4732
)!""'*!# $%" " "# )!""'*!# $%" " "#
!!" '"
! !" '"
!%
% " "

)!""(*" $%" " "#
)!""'*" $%" " "#
!!" ("
! !" '"
!%
% " "
! #"$ ! #"$

!!
#"$

#"$

!""'
"& #!

" "
!""'
"& #!

" "

Figure 9. The analysis results for the control action “set to X ◦ C”.
4.4. Comparsion of Results

Originally, the STPA is the only hazard analysis technique was based on the STAMP model [2].
Thus, in this section, let us compare the results presented in Section 4.3 with those by adopting the
original STPA. Since the third step of STPA was not taken into account, the comparison only considers
the results that derived from the first two steps of STPA. For comparability, the temperature issue
discussed in Section 4.3 is still part of our focus when adopting the original STPA.
As discussed, the goal of STPA is to identify causes that lead to hazards and result in losses, so
they can be eliminated or controlled. The causes to be identified are UCAs, and flaws in the control
loop (as shown in Figure 4) under some scenarios which are different from our case. The elimination
or control usually resorts to designing and implementing a safe system, while in our case, the aim is to
select appropriate precautionary and reaction measures which can restore a safe Service delivery.
The system engineering foundation is given first. The prepared definitions are illustrated in
Table 4. The design constraint is system level constraint and is expected to further decompose into
constraints that can be assigned to system components as the analysis evolves. Compared with what is
shown in Table 3, Service Failures and reliability requirements to the system are gone, which indicates
ICAs will not be identified afterward. This is because ICAs are supposed to result in Service Failure,
and reliability requirements of ICAs can be taken as the decomposition of the system level reliability
requirement. The safety control structure as shown in Figure 7 can also be used here.
Table 4. Preparation for the STPA analysis.
Accident Physical harm of occupants due to heat stroke

Hazard Indoor WBGT temperature is over 28 ◦ C
Safety Requirement Indoor WBGT temperature should be adjusted bellow 28 ◦ C
The Performers System is capable of adjusting the indoor
Design constraint
WBGT temperature bellow 28 ◦ C
Next, the UCAs identified in step two of STPA are shown in Tables 5 and 6 for control actions “set
OFF” and “set to X ◦ C”, respectively. Then, for each UCA, safety requirements and design constraints
can be derived. For example, the safety requirement for UCA-1 in Figure 5 could be:
• “set OFF” should be provided when the indoor temperature belongs to (25,28] ◦ C and the
Performers System is working in the Warm mode.
The design constraints for UCA-1 could be:
• The Performers System should be accurately aware of the indoor temperature change;
• “set OFF” should be provided when needed.
130
Appl. Sci. 2019, 9, 4732
Table 5. UCAs for the control action “set OFF”.

Hazard: Indoor WBGT temperature is over 28 ◦ C
Stopped Too Soon
Control Action Not Providing Providing Wrong Timing or Order
or Applied Too Long
UCA-3: “set OFF” provided after the
UCA-1: “set OFF” not provided when UCA-2: “set OFF” provided when the
Performers System has been
the indoor temperature belongs to Performers System is working
set OFF working in the Warm mode for some N/A
(25,28] ◦ C and the Performers System in the Cool mode and the indoor
time when the indoor temperature
is working in the Warm mode temperature belongs to (25,28] ◦ C
belongs to (25,28] ◦ C
Table 6. UCAs for the control action “set to X◦ C”.

Hazard: Indoor WBGT Temperature Is over 28 ◦ C
Wrong Timing Stopped Too Soon
Control Action Not Providing Providing
or Order or Applied Too Long
UCA-4: “set to X ◦ C” not provided when the
set to X ◦ C indoor temperature belongs to (25,28] ◦ C and Not hazardous Not hazardous N/A
the Performers System is working in the Fan mode
There are some differences in the comparison with the results derived in step two of
STPA. The ICAs, operation scenarios for ICAs, reliability requirements, precautionary measures,
and operation scenarios for UCAs cannot be obtained by adopting the original STPA. However,
the design constraints can be derived. The UCAs identified by adopting the original and tailored
STPA are equivalent. The results documented by tables and lists are separated. It is a trivial problem
when checking the relations between the results that were documented by the conventional approach.
The LGLD approach integrated the results and overcame this problem. We discuss the advantages of
the tailored STPA at the end of Section 5.
5. Discussion
As one of our everyday living experiences, physical processes could be in hazardous states that
cause harm to occupants [29,30]. The occurrence of hazardous states can be transferred from a normal
state of a physical process that brings about comfortableness, through some intermediate states for
uncomfortableness. The normal state is maintained by the smart home system, which takes advantage
of home appliances or the outdoor climate. When the behaviors of the smart home system deviate
from expectations, the above transformation will occur. Thus, this living experience validates the
proposed STAMP-PP model.
To validate the terms defined in the STAMP-PP model, let us consider the eight abnormal indoor
temperature situations that were discussed in our previous work [23]. Mutations can be undesired
fluctuation, constantly cooler/warmer than expectation, undesired duration of the temperature that
results in discomfort, and Service Failures when combining the Mutations with the feeling of occupants
on the scene. Hazards can be unbearable hot/cold and undesired duration in hot/cold situations.
Mutations and Hazards provide detectable evidence of physical process anomalies.
The information of Service Failure combines with ICAs under related operation scenarios—e.g.,
as depicted in Figures 8 and 9, one can predict whether Hazards would happen before providing
precautionary measures, because Service Failures and the ICAs under related operation scenarios can
be the context, in which a Hazard could occur if time goes on. This context information can be acquired
from various sensors (e.g., temperature sensor), system states, etc. For example, in Figure 9, we can
know whether the temperature is increasing to approach 25 ◦ C through temperature sensors. Then,
precautionary measures can be selected under the direction of reliability requirements. If a Hazard is
unfortunately detected, reaction measures have to be selected based on the UCAs under operation
scenarios and the Hazard, which are guided by safety requirements.
The purpose of precautionary measures is to restore Service delivery. This is achieved by
reconfiguration of the Performers System. The reconfiguration has two forms. One is to reset the
current working Performer(s), e.g., the precautionary measure “set OFF” (the second one) as shown
131
Appl. Sci. 2019, 9, 4732
in Figure 8, which is achieved by resetting the OFF command to the same Performer. Another is to
reconfigure ready-to-use Performers to achieve the goal, e.g., the “set Cool mode” in Figure 8, which
could be reconfigured to a standby Performer to the Cool mode. The purpose of reactions is to restore
a safe home environment state, which could be a Service Failure (that needs further interference) or the
process of normal Service delivery. It is also achieved by reconfiguring the Performers System, which
has the same contents as those introduced for precautionary measures. The reactions should have
another content that precautionary measures do not have, i.e., a warning mechanism. The warning
mechanism is triggered when a Hazard is detected that implies precautionary measures have failed.
In the very beginning, the warning signals will be sent to occupants who could leave the scene or do
something else to ensure safety. If the reconfiguration in the reaction stage cannot restore the system
to a safe situation, the warning mechanism will inform an emergency department, e.g., a hospital,
through networks.
Accident models can be roughly classified into three categories [1,2], i.e., sequential models,
epidemiological models, and systemic models. Sequential models describe accidents as the result of
time-ordered sequences of discrete events. Epidemiological models view accidents as a combination of
latent and active failures within a system, which is analogous to the spreading of a disease. Latent
conditions, e.g., management practices or organizational culture, can lie dormant within a system for a
long time, which can finally create conditions at a local level to result in active failures. The STAMP
model is based on system theory and so is the STAMP-PP model. It describes physical process
anomalies that cause uncomfortableness and health problems as a result of abnormal Behaviors of
the Performers System. Furthermore, to better understand the causal relation between the Performers
System and physical processes and better represent the time order of physical process evolvement,
the terms in the STAMP-PP model can be taken as events.
Hazard analysis techniques are devised based on accident models. The STPA was invented based
on the STAMP model to identify causes that conventional approaches overlooked [2,9]. We found it
is an efficient tool for assisting in the identification of Defects, i.e., ICAs and UCAs (e.g., the results
shown in Figures 8 and 9). However, hazard analysis heavily depends on expertise in a specific area.
Accident models and hazard analysis techniques are to assist in understanding accidents and guiding
for causal analysis. The STAMP-PP model offers a way to understand physical process anomalies with
respect to the behaviors of smart home systems. To apply the tailored STPA to that in our case (some
results of which are shown in Figures 8 and 9), one may need to possess knowledge in at least the
fields of computer science, software engineering, computer networks, and even physics. This makes
the authors feel that safety research in different areas is like different research fields.
In the first step of STPA, the accidents and system hazards provided mostly depend on the interest
of an organization or the government [2]. For example, in this paper, Accident is defined as physical
harm of occupants due to heat stroke, as illustrated in Table 3. It can also, however, encompass other
symptoms aside from heat stroke, e.g., severe cold. Concrete accidents are different due to the variety
of environments, e.g., workplaces and the home environment. It is also determined by budget, severity
of occurrence, and frequency of occurrence [2].
In step two of STPA, originally, the taxonamies were provided to identify UCAs with respect to
control actions. Physical process anomalies (not Hazards) are the result of abnormal Behaviors of the
Performers System. The behaviors are the representations of the functions which are achieved by the
control actions issued by the Performers Sytem. For example, stopping the Performers System from
working is achieved by the control action “set OFF” as shown in Figure 8. Thus, the taxonomies are
applicable for identifying ICAs. This ensures that all possible ICAs can be effectively identified.
The states of physical process change from a state of comfort to one of discomfort, then to one
of hazard. When a state of discomfort is detected, precautionary measures are adopted to prevent
the physical process from transferring into a state of hazard. Precautionary measures refer to control
actions achieved by reconfigurations. Thus, the taxonomies are also applicable to the precautionary
measures. For example, the control action “set OFF” in Figure 8 can both be the original control action
132
Appl. Sci. 2019, 9, 4732
and the precautionary measure. The same applies to the control action “set to X ◦ C” as shown in
Figure 9.
The purpose of adopting the hazard analysis technique is to identify Defects, so as to select
appropriate precautionary and reaction measures for adjusting home environment anomalies to
maintain a normal performance. By checking the comparison presented in Section 4.4, we found
the tailored STPA can satisfy this purpose, while the original STPA cannot. First, ICAs under
operation scenarios and the reliability requirements of the ICAs can be identified by the tailored
STPA. This is due to Service Failure, and the system level reliability requirement is provided in the first
step of STPA. Naturally, precautionary measures are not necessarily identified in the original STPA,
because precautionary measures are selected for ICAs. Second, even though UCAs can be identified by
both the original and tailored STPA, the operation scenarios have different contents. For the tailored
STPA, the operation scenarios refer to occupants or controllers or both as discussed when introducing
the concept of Defect. For the original STPA, the operation scenarios refer to the control flaws as shown
in Figure 4, which are identified in step three of STPA.
As discussed in Section 4.2, the STPA takes advantage of tables and lists to document the analytical
results (see the results shown in Tables 5 and 6). We proposed the LGLD approach to document the
analytical results. The advantage of comparing tables and lists is that it can clearly represent the
relations among the results in a straightforward way. For example, in Figures 8 and 9, the relations
among control actions, ICAs, precautionary measures, and UCAs are clear. Further, it is clear to see the
(reliability or safety) requirements attached to each ICA and UCA, and the related operation scenarios
under which the ICA and UCA can occur. This kind of relations is not explicitly represented using
tables and lists such as the ones presented in Section 4.4. To build such documentation, one can build
along the way of analysis, because the analysis starts from control actions to identify ICAs under
related operation scenarios, and reliability requirements, then to determine precautionary measures,
and finally to identify UCAs under related operation scenarios, and safety requirements.
When applying the STAMP-PP model to understand accident formation, the Performers System
has to fully control the home environment, or its behaviors may not be the “only” reason to cause
Service Failure, then Hazard. This can be the limitation of the proposed STAMP-PP model.
6. Related Work
This section includes two parts. The first is about accident models based on system theory.
The second discusses safety-related research in the smart home environment.
6.1. Accident Models

The systems approach is considered as the dominant paradigm in safety research [34]. It views
accidents as the unexpected interactions among system components, i.e., technical, social, and human
elements. Among various systemic models, there are three most cited models [10], that is, STAMP [2],
Functional Resonance Analysis Method (FRAM) [3], and Accimap [35,36].
The concept of STAMP has been briefly introduced in Section 3.1. According to [10], over half
of the reviewed papers were STAMP-related, which indicates a pervasive acknowledgement of its
underlying rationale. Its application has attracted researchers from a broad field. The authors of [6]
adopted it to investigate aircraft rapid decompression events. The authors of [4] applied it in the
analysis of deepwater well control safety. In the field of railway, the authors of [5] investigated railway
accidents and accident spreading by taking the China Jiaoji railway accident as the example. It has also
been applied to the field related to poisonous or dangerous substances. The authors of [7] adopted
it in analyzing the China Donghuang oil transportation pipeline leakage and explosion accident.
The authors of [8] applied it to the Fukushima Daiichi nuclear disaster and to promote safety of
nuclear power plants. Smart home systems are generally not considered as safety-critical systems.
However, as weather anomalies, e.g., heatwaves, occur regularly due to global warming, smart home
133
Appl. Sci. 2019, 9, 4732
systems for indoor environment adjustment, in this context, can be taken as safety-critical. Thus,
the STAMP model can be adopted for understanding accident formation in the home environment.
FRAM was developed to act as both an accident analysis and risk assessment tool [10]. The FRAM
model graphically describes systems as interrelated subsystems and functions that will exhibit varying
degrees of performance varation. Accidents result from the fact that the emergent variation produced
from the performance variability of any system component to “resonate” with that of the rest of the
elements is too high to control. It has been discussed that the FRAM and STAMP approaches focus
better on qualitative modeling and description of systemic behavior and accidents [37]. FRAM also
has applications in different fields, e.g., the authors of [38] applied it to railway traffic supervision
to investigate interdisciplinary safety analysis of complex sociotechnological systems. The authors
of [39] extended FRAM by including a framework with steps to support hazard analysis. Some efforts
have been tried to quantify it, e.g., the authors of [40] developed a semiquantitative FRAM based on a
Monte Carlo simulation.
The Accimap method is a graphical representation of a particular accident scenario that relates
to systemwide failures, decisions, and actions [3,36]. Accimap is a generic approach and does not
use taxonomies (that is different than that of the STPA [2,9]) of failures across the different levels of
considered [41]. The Accimap produces less reliable accident analysis results compared to STAMP [42].
The selection of accident analysis techniques depends on the system characteristics,
i.e., manageability and coupling [1]. The systemic approaches are usually adopted by systems with low
manageability and tight coupling. Systemic approaches related to complex sociotechnical systems have
their own strengths. For one such system, it is better to adopt multiple approaches which supplement
each other, even though STAMP is considered much more effective and reliable in understanding
accidents and hazard analysis [10,41,42].
6.2. Smart Home Safety

In the past, safety research inside the home environment used to be based on events or
chains-of-events. With the emergence of the so-called smart homes, safety research inside the
home environment also has new forms. Some of it refers to monitoring the home environment.
With the purpose of detecting safety problems of indoor climate abnormal variations, the authors
of [12] proposed a CPS (cyberphysical system) home safety architecture to support an event-based
detection. The authors of [43] presented a method that maps the real home connection to a virtual
home environment, together with related policies to ensure remote monitoring, to ensure home safety.
Elderly safety in the smart home environment was achieved by analyzing and inferring locations, time
slots, and periods of stay of elderly people [44]. Robot techniques were also employed, e.g., the authors
of [45] developed a robot which can, for example, sense gas leakage and shut off the gas valve. Others
focus on a specific part of the home. The authors of [46] proposed risk analysis and assessment when
cooking to prevent potential risks. This is because the kitchen is also prone to safety problems like
gas leakage and fire accidents. Electricity is also an important risk factor. The authors of [47] adopted
an alert circuit with a voltage level indicator to prevent the smart solar home system from being
overloaded and damaged. With cloud computing techniques becoming pervasive in implementing
smart home systems, risks like cloud service unavailability have also been introduced. To overcome
this, the authors of [48] discussed home resilience in the presence of possible unavailability and
proposed RES-Hub, i.e., a standalone hub to ensure the continuity of required functionalities.
Most of the studies like those discussed above focus on implementing systems to deal with home
safety problems. If not properly designed and implemented, the system itself can be a risk factor.
Thus, requirement elicitation becomes critical. Conventional safety-related techniques are applied to
safety-critical areas, e.g., aviation [6]. Our work employed these techniques to the smart home systems.
134
Appl. Sci. 2019, 9, 4732
7. Conclusions and Future Work

We extended the accident model STAMP by considering physical processes in the home
environment. As the home environment is adjusted by behaviors of the smart home system, we
first proposed the concept of Performers System that emphasizes the behaviors performed by various
home appliances. Then, based on the Behavior of the Performers System, we proposed accident
formation with respect to physical processes going from normal state to some intermediate states that
result in uncomfortableness, and finally to states that cause harm. In order to identify the Defects,
i.e., ICAs and UCAs that result in abnormal system behaviors, the hazard analysis technique STPA was
tailored and applied to the smart home system for indoor temperature adjustment. After comparison
with the results derived by adopting the original STPA, we found that the tailored STPA is an efficient
tool to assist in identifying the Defects. The analytical results of applying STPA were used to adopt
tables and lists for documentation, but the relations among the results were not straight-forward
and unclear. We then proposed the LGLD approach, whose advantages are demonstrated by the
comparison of results.
For future work, we aim to map the context information in which physical process anomalies
occurred in the cyber world, such that precautionary measures and reactions can be effectively
selected in real time. This brings about the problem of how to parameterize the STPA analytical
results, i.e., ICAs, operation scenarios, and reliability requirements; and UCAs, operation scenarios,
and safety requirements.
Author Contributions: Conceptualization, Z.Y., Y.L., and Y.T.; methodology, Z.Y. and Y.T.; validation, Z.Y. and
Y.T.; writing—original draft preparation, Z.Y.; writing—review and editing, Z.Y. and Y.T.; supervision, Y.T.
Acknowledgments: The authors would like to thank Toshiaki Aoki who gave constructive comments.
References
1. Underwood, P.; Waterson, P. Accident Analysis Models and Methods: Guidance for Safety Professionals;
Loughborough University: Loughborough, UK, 2013; p. 28.
2. Leveson, N.G. Engineering a Safer World: Systems Thinking Applied to Safety; The MIT Press: Cambridge, MA,
USA; London, UK, 2011.
3. Hollnagel, E. FRAM: The Functional Resonance Analysis Method: Modelling Complex Socio-technical Systems,
1st ed.; CRC Press: Boca Raton, FL, USA; London, UK; New York, NY, USA, 2012.
4. Meng, X.; Chen, G.; Shi, J.; Zhu, G.; Zhu, Y. STAMP-based analysis of deepwater well control safety. J. Loss
Prev. Process Ind. 2018, 55, 41–52. [CrossRef]
5. Ouyang, M.; Hong, L.; Yu, M.H.; Fei, Q. STAMP-based analysis on the railway accident and accident
spreading: Taking the China—Jiaoji railway accident for example. Saf. Sci. 2010, 48, 544–555. [CrossRef]
6. Allison, C.K.; Revell, K.M.; Sears, R.; Stanton, N.A. Systems Theoretic Accident Model and Process (STAMP)
safety modelling applied to an aircraft rapid decompression event. Saf. Sci. 2017, 98, 159–166. [CrossRef]
7. Gong, Y.; Li, Y. STAMP-based causal analysis of China-Donghuang oil transportation pipeline leakage and
explosion accident. J. Loss Prev. Process Ind. 2018, 56, 402–413. [CrossRef]
8. Daisuke, U. STAMP Applied to Fukushima Daiichi Nuclear Disaster and the Safety of Nuclear Power Plants
in Japan. Master’s Thesis, School of Engineering, Massachusetts Institute of Technology, Cambridge, MA,
USA, 2016.
9. Leveson, N.G.; Thomas, J.P. STPA Handbook; Massachusetts Institute of Technology: Cambridge, MA,
USA, 2018.
10. Underwood, P.; Waterson, P. A critical review of the STAMP, FRAM and Accimap systemic accident analysis
models. In Advances in Human Aspects of Road and Rail Transportation; CRC Press: Boca Raton, FL, USA, 2012;
pp. 385–394.
135
Appl. Sci. 2019, 9, 4732
11. Ericson, C.A., II. Hazard Analysis Techniques for System Safety; John Wiley and Sons, Inc.: Hoboken, NJ, USA,
2005; Chapter 3, pp. 31–54.
12. Yang, Z.; Lim, A.O.; Tan, Y. Event-based home safety problem detection under the CPS home safety
architecture. In Proceedings of the 2013 IEEE 2nd Global Conference on Consumer Electronics (GCCE),
Tokyo, Japan, 1–4 October 2013; pp. 491–495. [CrossRef]
13. Yang, Z.; Lim, Y.; Tan, Y. A risk model for indoor environment safety. In Proceedings of the 2017 IEEE
6th Global Conference on Consumer Electronics (GCCE), Nagoya, Japan, 24–27 October 2017; pp. 1–5.
[CrossRef]
14. Harper, R. Inside the Smart House; Springer: Berlin/Heidelberg, Germany, 2003.
15. Lobaccaro, G.; Carlucci, S.; Löfström, E. A Review of Systems and Technologies for Smart Homes and Smart
Grids. Energies 2016, 9, 1–33. [CrossRef]
16. Djenouri, D.; Laidi, R.; Djenouri, Y.; Balasingham, I. Machine Learning for Smart Building Applications:
Review and Taxonomy. ACM Comput. Surv. 2019, 52, 24:1–24:36. [CrossRef]
17. Toschi, G.M.; Campos, L.B.; Cugnasca, C.E. Home automation networks: A survey. Comput. Stand. Interfaces
2017, 50, 42–54. [CrossRef]
18. Rasmussen, J. Skills, rules, and knowledge; signals, signs, and symbols, and other distinctions in human
performance models. IEEE Trans. Syst. Man Cybern. 1983, SMC-13, 257–266. [CrossRef]
19. MIL-STD-882E. Department of Defense Standard Practice: System Safety; Standard, Department of Defense:
Washington, DC, USA, 2012.
20. Defence Standard 00-56. Safety Management Requirements for Defence Systems, Part 1, Requirements; Standard,
Ministry of Defence: London, UK, 2007.
21. Chemishkian, S. Building smart services for smart home. In Proceedings of the 2002 IEEE 4th International
Workshop on Networked Appliances (Cat. No.02EX525), Gaithersburg, MD, USA, 15–16 January 2002;
pp. 215–224. [CrossRef]
22. Nakamura, M.; Tanaka, A.; Igaki, H.; Tamada, H.; Matsumoto, K. Constructing Home Network Systems
and Integrated Services Using Legacy Home Appliances and Web Services. Int. J. Web Serv. Res. (IJWSR)
2008, 5, 82–98. [CrossRef]
23. Yang, Z.; Aoki, T.; Tan, Y. Multiple Conformance to Hybrid Automata for Checking Smart House Temperature
Change. In Proceedings of the 2018 IEEE/ACM 22nd International Symposium on Distributed Simulation
and Real Time Applications (DS-RT), Madrid, Spain, 15–17 October 2018; pp. 1–10. [CrossRef]
24. Anderson, G.B.; Bell, M.L.; Peng, R.D. Methods to Calculate the Heat Index as an Exposure Metric in
Environmental Health Research. Environ. Health Perspect. 2013, 121, 1111–1119. [CrossRef]
25. ANSI/ASHRAE Standard 55-2010. Thermal Environmental Conditions for Human Occupancy; Standard,
American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.: Atlanta, GA, USA, 2010.
26. Woods, D.D.; Hollnagel, E.; Leveson, N. Resilience Engineering: Concepts and Precepts, 1st ed.; CRC Press:
Boca Raton, FL, USA, 2006.
27. Jacklitsch, B.; Williams, W.J.; Musolin, K.; Coca, A.; Kim, J.H.; Turner, N. NIOSH Criteria for a Recommended
Standard: Occupational Exposure to Heat and Hot Environments; Recommended Standard 2016-106; U.S.
Department of Health and Human Services, Centers for Disease Control and Prevention, National Institute
for Occupational Safety and Health: Cincinnati, OH, USA, 2016.
28. ISO 11079. Ergonomics of the Thermal Environment—Determination and Interpretation of Cold Stress When
Using Required Clothing Insulation (IREQ) and Local Cooling Effects; Standard, International Organization for
Standardization: Geneva, Switzerland, 2007.
29. van Loenhout, J.; le Grand, A.; Duijm, F.; Greven, F.; Vink, N.; Hoek, G.; Zuurbier, M. The effect of high
indoor temperatures on self-perceived health of elderly persons. Environ. Res. 2016, 146, 27–34. [CrossRef]
[PubMed]
30. Okamoto-Mizuno, K.; Mizuno, K. Effects of thermal environment on sleep and circadian rhythm. J. Physiol.
Anthropol. 2012, 31, 14. [CrossRef] [PubMed]
31. Leon, L.R.; Bouchama, A. Heat Stroke. In Comprehensive Physiology; American Cancer Society: Atlanta, GA,
USA, 2015; pp. 611–647, doi:10.1002/cphy.c140017. [CrossRef]
32. Gaudio, F.G.; Grissom, C.K. Cooling Methods in Heat Stroke. J. Emerg. Med. 2016, 50, 607–616. [CrossRef]
[PubMed]
136
Appl. Sci. 2019, 9, 4732
33. Budd, G.M. Wet-bulb globe temperature (WBGT)—Its history and its limitations. J. Sci. Med. Sport
2008, 11, 20–32. [CrossRef] [PubMed]
34. Salmon, P.; Williamson, A.; Lenné, M.; Mitsopoulos-Rubens, E.; Rudin-Brown, C. Systems-Based Accident
Analysis in the Led Outdoor Activity Domain: Application and Evaluation of a Risk Management
Framework. Ergonomics 2010, 53, 927–39. [CrossRef] [PubMed]
35. Svedung, I.; Rasmussen, J. Graphic representation of accident scenarios: Mapping system structure and the
causation of accidents. Saf. Sci. 2002, 40, 397–417. [CrossRef]
36. Rasmussen, J. Risk management in a dynamic society: A modelling problem. Saf. Sci. 1997, 27, 183–213.
[CrossRef]
37. Bjerga, T.; Aven, T.; Zio, E. Uncertainty treatment in risk analysis of complex systems: The cases of STAMP
and FRAM. Reliab. Eng. Syst. Saf. 2016, 156, 203–209. [CrossRef]
38. Belmonte, F.; Schön, W.; Heurley, L.; Capel, R. Interdisciplinary safety analysis of complex socio-technological
systems based on the functional resonance accident model: An application to railway traffic supervision.
Reliab. Eng. Syst. Saf. 2011, 96, 237–249. [CrossRef]
39. Tian, J.; Wu, J.; Yang, Q.; Zhao, T. FRAMA: A safety assessment approach based on Functional Resonance
Analysis Method. Saf. Sci. 2016, 85, 41–52. [CrossRef]
40. Patriarca, R.; Gravio, G.D.; Costantino, F. A Monte Carlo evolution of the Functional Resonance Analysis
Method (FRAM) to assess performance variability in complex systems. Saf. Sci. 2017, 91, 49–60. [CrossRef]
41. Salmon, P.M.; Cornelissen, M.; Trotter, M.J. Systems-based accident analysis methods: A comparison of
Accimap, HFACS, and STAMP. Saf. Sci. 2012, 50, 1158–1170. [CrossRef]
42. Filho, A.P.G.; Jun, G.T.; Waterson, P. Four studies, two methods, one accident—An examination of the
reliability and validity of Accimap and STAMP for accident analysis. Saf. Sci. 2019, 113, 310–317. [CrossRef]
43. Yang, L.; Yang, S.H.; Yao, F. Safety and Security of Remote Monitoring and Control of intelligent Home
Environments. In Proceedings of the 2006 IEEE International Conference on Systems, Man and Cybernetics,
Taipei, Taiwan, 8–11 October 2006; Volume 2, pp. 1149–1153. [CrossRef]
44. Kim, S.C.; Jeong, Y.S.; Park, S.O. RFID-based Indoor Location Tracking to Ensure the Safety of the Elderly in
Smart Home Environments. Pers. Ubiquitous Comput. 2013, 17, 1699–1707. [CrossRef]
45. Lee, K.H.; Seo, C.J. Development of user-friendly intelligent home robot focused on safety and security. In
Proceedings of the ICCAS 2010, Gyeonggi-do, Korea, 27–30 October 2010; pp. 389–392. [CrossRef]
46. Yared, R.; Abdulrazak, B.; Tessier, T.; Mabilleau, P. Cooking Risk Analysis to Enhance Safety of Elderly
People in Smart Kitchen. In Proceedings of the 8th ACM International Conference on PErvasive Technologies
Related to Assistive Environments; ACM: New York, NY, USA, 2015; pp. 12:1–12:4. [CrossRef]
47. Hasan, T.; Nayan, M.F.; Iqbal, M.A.; Islam, M. Smart Solar Home System with Safety Device Low Voltage
Alert. In Proceedings of the 2012 UKSim 14th International Conference on Computer Modelling and
Simulation, Cambridge, UK, 28–30 March 2012; pp. 201–204. [CrossRef]
48. Doan, T.T.; Safavi-Naini, R.; Li, S.; Avizheh, S.; K., M.V.; Fong, P.W.L. Towards a Resilient Smart Home.
In Proceedings of the 2018 Workshop on IoT Security and Privacy, Budapest, Hungary, 20–20 August 2018;
ACM: New York, NY, USA, 2018; pp. 15–21. [CrossRef]
c 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
137
applied
sciences
Article
A Contactless Measuring Method of Skin
Temperature based on the Skin Sensitivity
Index and Deep Learning
Xiaogang Cheng 1,2 , Bin Yang 3,4, *, Kaige Tan 5 , Erik Isaksson 5 , Liren Li 6 , Anders Hedman 5 ,
Thomas Olofsson 4 and Haibo Li 1,5
1 College of Telecommunications and Information Engineering, Nanjing University of Posts and
Telecommunications, Nanjing 210003, China; [email protected] (X.C.); [email protected] (H.L.)
2 Computer Vision Laboratory (CVL), ETH Zürich, 8092 Zürich, Switzerland
3 School of Building Services Science and Engineering, Xi’an University of Architecture and Technology,
Xi’an 710055, China
4 Department of Applied Physics and Electronics, Umeå University, 90187 Umeå, Sweden;
[email protected]
5 KTH Royal Institute of Technology, 10044 Stockholm, Sweden; [email protected] (K.T.); [email protected] (E.I.);
[email protected] (A.H.)
6 School of computer science and technology, Nanjing Tech University, Nanjing 211816, China;
[email protected]
* Correspondence: [email protected]
Received: 16 December 2018; Accepted: 15 March 2019; Published: 1 April 2019
Featured Application: The NISDL method proposed in this paper can be used for real time
contactless measuring of human skin temperature, which reflects human body thermal comfort
status and can be used for control HVAC devices.
Abstract: In human-centered intelligent building, real-time measurements of human thermal comfort

play critical roles and supply feedback control signals for building heating, ventilation, and air
conditioning (HVAC) systems. Due to the challenges of intra- and inter-individual differences and skin
subtleness variations, there has not been any satisfactory solution for thermal comfort measurements
until now. In this paper, a contactless measuring method based on a skin sensitivity index and deep
learning (NISDL) was proposed to measure real-time skin temperature. A new evaluating index,
named the skin sensitivity index (SSI), was defined to overcome individual differences and skin
subtleness variations. To illustrate the effectiveness of SSI proposed, a two multi-layers deep learning
framework (NISDL method I and II) was designed and the DenseNet201 was used for extracting
features from skin images. The partly personal saturation temperature (NIPST) algorithm was use for
algorithm comparisons. Another deep learning algorithm without SSI (DL) was also generated for
algorithm comparisons. Finally, a total of 1.44 million image data was used for algorithm validation.
The results show that 55.62% and 52.25% error values (NISDL method I, II) are scattered at (0 ◦ C,
0.25 ◦ C), and the same error intervals distribution of NIPST is 35.39%.
Keywords: contactless measurements; skin sensitivity index; thermal comfort; subtleness

magnification; deep learning; piecewise stationary time series
1. Introduction
Higher economic growth drives increasing energy consumption, and 50% of housing consumption
is generated by heating, ventilation and air conditioning (HVAC) systems [1,2]. Furthermore, one of
the most important reasons for energy waste is that the actual thermal requirements of indoor

Appl. Sci. 2019, 9, 1375
occupants are ignored, with the result that overheating and overcooling occur often. Fortunately,
real-time thermal comfort perception can provide useful signals to HVAC systems for achieving energy
saving and human-centered intelligent control. Therefore, many researchers have been studying
thermal comfort measurements for indoor environments in recent decades. Many methods were
generated, including the questionnaire survey method [3–5], environmental measurement method [6,7]
and contact measuring method of human body physiological parameters [8–18]. In recent years,
the semi-contact measuring method [19,20] and contactless measuring method [21–23] for human
body physiological parameters were also generated. For example, in references [19,20], an infrared
sensor was fixed on the frame of eyeglasses in order to measure skin temperature. In reference [21],
a normal vision sensor was also used for measuring skin temperature and two non-linear models were
trained. In references [22,23], Kinect was used for recognizing human poses or indoor locations, and
then human thermal comfort and dynamic metabolism were estimated, respectively. All these methods
are meaningful attempts. However, due to the challenges of measuring thermal comfort which are (1)
skin subtleness variation [24], (2) inter-individual differences [14,25] and (3) temporal intra-individual
differences [14,26], there is still no satisfactory method for perceiving human thermal comfort.
To overcome the aforementioned challenges, the skin sensitivity index (SSI) was defined in
this paper. The SSI is strongly related to skin temperature. A contactless measuring method of
skin temperature based on SSI and Deep Learning was proposed, hereinafter referred to as NISDL.
Two different deep learning methods of NISDL have been designed and trained, respectively, which
are NISDL methods I and II. The main difference between them is that the location of SSI participation
in the neural network training is not the same. A total of 1.44 million images were collected for 16
Asian female subjects, and this ‘big data’ was used for algorithm validation.
The main contributions of this paper are:
(1) The skin sensitivity index (SSI) was proposed for describing individual sensitivity of thermal
comfort, and the index was combined with skin images for deep learning network training.
(2) A novel contactless measuring algorithm (NISDL) based on SSI was proposed, with two different
frameworks of NISDL having been designed for real-time thermal comfort measurement.
(3) A deep learning algorithm without SSI was also generated and trained. Two comparisons were
made: (1) comparison between data-driven methods (deep learning) and model-driven methods
(linear models); (2) comparison of measuring effects in the case of SSI participation in training
and non-participation in training.
The rest of this paper is organized as follows. Section 2 introduces the related work about thermal
comfort. In Section 3, the research methods, including SSI computation, subjective experiments and
NISDL methods, are introduced. The results and discussion are shown in Sections 4 and 5. Finally,
Section 6 gives the conclusion.
2. Related Work
Since the 1970s, Fanger has explored human thermal comfort and conducted many kinds of
subjective experiments. Based on this, he eventually established what is known as Fanger’s theory [27].
From then on, many studies about thermal comfort were carried out.
Questionnaire surveys are good as a method to understand the inner feeling of an occupant.
With the development of the internet, online questionnaire surveys can also be generated [3,4].
However, it is inconvenient and also difficult to guarantee that occupants will continue to give feedback
based on their personal thermal feelings [5]. Therefore, the environment measurement method was
also adopted in the building industry [6]. With this kind of method, some objective parameters,
such as indoor temperature, airflow and humidity, are often measured. Unfortunately, the goal of
the environment measurement method is to meet the thermal comfort needs of a majority of indoor
occupants. Therefore, the thermal feelings of a minority were ignored. To overcome this drawback,
a kind of nonlinear autoregressive network, still belonging to the environment measurement method,
140
Appl. Sci. 2019, 9, 1375
was generated to predict indoor temperature [7]. In fact, human thermal comfort is complicated, and
with constant indoor parameters it is difficult to meet each individual’s requirements for thermal
comfort. As such, some researchers study physiological measurement methods, including the contact
measuring method, semi-contact measuring method and contactless measuring method.
For the contact measuring method, skin temperature and heart rate are usually the measured
parameters. Wang and Nakayama [8,9] made early attempts at measuring the skin temperature
around the human body. Liu [10] conducted subjective experiments and a total of 22 subjects were
invited. The data, being local skin temperatures and electrocardiograms, were collected. Based on
these collected data, 26 measuring methods of mean skin temperature were assessed. Takada [11]
presented a multiple regression equation to predict skin temperature in non-steady state. The multiple
regression equation was considered as a function of mean skin temperature. Wrist skin temperatures
and upper extremity skin temperatures were also adopted to estimate human thermal sensation,
respectively [12,13]. Chaudhuri [14] presented a predicted thermal state (PTS) model, and the capture
of peripheral skin temperature. Furthermore, body surface area and clothing insulation were used
for analyzing inter- and intra-individual differences. As for the thermal comfort study using heart
rate, based on physiological experimentation, Yao [15] investigated the relationship between heart rate
variation (HRV) and electroencephalograph (EGG). The results show that HRV and EEG are useful
for thermal comfort studies, but further data validation is needed. Moreover, Dai, Chaudhuri and
Kim [16–18] combined machine learning with contact measurement, and they are all meaningful
attempts. However, as the linear kernel of SVM was used for predicting human thermal sensation in
different experiments, sometimes overfitting can happen.
For the semi-contact measuring method, Ghahramani [19] used an infrared sensor to estimate
skin temperature of different face points. The infrared sensor was mounted on the frame of eyeglasses.
Based on this, a hidden Markov model was constructed to capture personal thermal sensation [20].
In practical application, contact and semi-contact measurement are both difficult to apply widely.
The reason is that an occupant needs to wear a sensor, which is uncomfortable and is also not in line
with the goal of human-oriented intelligent buildings. For this reason, a kind of contactless measuring
method was studied in reference [21]. Based on vision sensors, Cheng [21] extracted the saturation
(S) channel from skin images and constructed two saturation-temperature models to estimate skin
temperature. The two models are the contactless measuring method of thermal comfort based on
saturation-temperature (NIST) and the contactless measuring method of thermal comfort based on
partly saturation-temperature (NIPST). Alan [22] proposed a contactless measuring method based on
human poses. A total of 12 poses of thermal comfort were defined and Kinect was adopted to estimate
human skeleton and poses. Further, Dziedzic [23] also used Kinect to predict human thermal sensation
and dynamic metabolic rates.
With the development of machine learning (ML) and computer vision (CV), some thermal comfort
perception methods based on ML and CV were proposed. Support Vector Machine (SVM) are often
used for analyzing existing databases (RP-884) and captured environmental parameters [24,28,29].
Further, Peng [26] use unsupervised and supervised learning to predict occupants’ behavior, applied
to three types of offices which are single person offices, multi-person offices, and meeting rooms.
The results show that the average energy savings for the entire space is 21% in experimental condition.
Li [30] proposed a fuzzy model to predict thermal sensation, skin temperature and heart rate considered
as objective parameters. For avoiding overheating, Cosma [31] extracted data from multiple local
body parts and analyzed them with four kinds of machine learning algorithms, including SVM,
Gaussian process classifier (GPC), k-neighbors classifier (KNC) and random forest classifier (RFC).
141
Appl. Sci. 2019, 9, 1375
The kinds of machine learning adopted in references [24,26,28–31] are traditional algorithms.
In recent years, the use of deep neural networks is on the rise [32,33]. In addition, a kind of subtleness
magnification technology was presented [34,35]. These provide new directions and opportunities for
the measurement of human thermal comfort. Based on this technology, a novel contactless measuring
method was generated which will be introduced as follows.
3. Research Methods
3.1. Subjective Physiological Experiments
3.1.1. Subjects Data and Chamber Environments

16 human subjects were invited for experiments and the resulting data volume is 1.44 million
images. The experiments were conducted in a chamber with controllable indoor air temperature
and relative humidity. The corresponding dry-bulb air temperature is 22.2 ± 0.2 ◦ C and the relative
humidity is 36.9 ± 2.5%. The resolution of vision sensor used for capturing video is 1280 × 720.
The iButton, model DS192H with uncertainty ±0.125 ◦ C, was used for measuring skin temperature
from the back of subject’s hand. All the subjects are Asian females with an average age of
23.9 ± 3.9 years, average weight of 52.2 ± 6.5 kg, and body mass index (BMI) 19.9 ± 2.2 kg/m2 .
3.1.2. Experimental Procedures

The experiment was conducted during winter in Sweden. There are generally three steps in
subjective physiological experiments. (1) Preparation stage: The indoor environment parameters were
measured and controlled to a suitable level. When the subjects came into the chamber, they should
rest for 10 min for adaptation. At the same time, warm water with constant temperature (45 ◦ C) was
prepared. (2) Thermal stimulus: After 10 min adaptation, subjects were asked to immerse hands into
the water with 45 ◦ C. The whole thermal stimulus process lasted for 10 min. (3) Big data collection:
After 10 min of stimulus, subjects were asked to sit next to the data collection desk and put her
pairs of hands under the vision sensor. The back of the hand is faced up and the data is collected
for 50 min. At the same time, skin temperature sensor (iButton) was attached to the back of one
hand. The corresponding sampling interval is 1 min. It should be noted that, based on piecewise
stationary time series analysis [36], linear interpolation was adopted in this paper, and 11 points were
interpolated into 1 min for real skin temperature captured by iButton.
3.2. Skin Sensitivity Index
3.2.1. SSI Definition

When human body encounters thermal stimuli, blood circulation will change which will also be
reflected in skin’s color and texture. In reference [21], based on the HSV (hue, saturation, value) color
space, the S channel was extracted and a linear ST (saturation-temperature) model was established.
T = ki × S + bi (1)
where i denotes subject number, the k reacts to the change rate of skin temperature, and b denotes the
intercept. S and T are skin saturation and temperature respectively. In this paper, k is defined as the
skin sensitivity index (SSI). SSI is a high weight coefficient in skin temperature changes and SSI reflects
the skin sensitivity level to external thermal stimuli.
142
Appl. Sci. 2019, 9, 1375
3.2.2. SSI Computing

Based on subjective physiological experiments, real skin temperature can be obtained by iButton.
The images were also collected from subjects’ hands. Therefore, the SSI can be calculated. The steps
are as follows: (1) Extracting each frame from captured video; (2) Segment region of interest (ROI);
(3) Extracting S channel from ROI images and computing mean values of S for each ROI image; (4)
Search SSI value based on real skin temperature and S for each subject.
3.3. NISDL Algorithm

In this paper, considering that SSI is a high weight coefficient for contactless thermal comfort
measurement, it will improve the prediction accuracy of skin temperature. Based on SSI, the NISDL
algorithm was introduced in this paper. Furthermore, to validate the effectiveness of SSI, two kinds of
deep learning frameworks (NISDL method I, II) have been constructed. The main difference between
NISDL methods I and II is that the location where the network invokes SSI to participate in the model
training is different. The NISDL algorithm constructed is introduced as follows.
3.3.1. Video Pre-Processing

In fact, skin texture variation is subtle and is difficult to perceive. In this paper, for magnifying
this kind of subtleness variation, an image subtleness magnification technology known as Euler Video
Magnification (EVM) is adopted [34,35]. Based on EVM, let c (x, t) denote skin images which are subtly
varied with time t. Suppose that the variation function is formula (2) [34,35].
C (x, t) = F (x + h (t)) (2)
where, h (t) is variation degree, F is a function which constructs the relationship between C (x, t) and h
(t). If the skin image C (x, t) is magnified, and the first-order Taylor series expansion can be handed to
F (x + h (t)).
C (x, t) = F(x + (1 + ξ) * h(t)) (3)
where ξ is the magnification coefficient which can be set based on practical application. According to
Formula (3), only the variation part was magnified to a magnitude of 1 + ξ, while the other part of skin
texture is not magnified. Therefore, the invisible texture variation is made to be visible.
It should be noted that de-noise processing should be handled before EMV processing. Further,
after the video is magnified, ROI is selected and cropped from each frame. The ROI images are
imported into NISDL method I and II for model training.
3.3.2. NISDL Method I

As shown in Figure 1, SSI values are used as input data and imported into the deep learning
network at the very beginning. The ROI images were also combined with SSI values in the first step.
According to the size of ROI images, the SSI value of each ROI image was expanded into a matrix.
The matrix is considered as a channel and combined with the 3 channels of ROI images.
143
Appl. Sci. 2019, 9, 1375
Figure 1. NISDL method I. (SSI values were combined with skin images firstly.).
The merged data between ROI images and SSI values above is inputted into four convolution
layers, which are used for dimensionality reduction. In NISDL method I, the DenseNet201 [33]
is adopted for features extraction. The last two layers of DenseNet201 are not suitable for skin
temperature measurement, hence the two layers are removed. The reason for this is that the activation
function of the last layers is softmax. Instead of these two layers, an average pooling layer and a fully
connected layer are added behind denseNet201. Based on the deep learning networks designed above,
n ROI images and n SSI values were inputted into NISDL method I. Therefore, n skin temperatures
can be obtained.
3.3.3. NISDL Method II

Figure 2 is the deep learning framework of NISDL method II. In this method, the ROI images
and SSI values were processed for features extraction. Subsequently, the two kinds of features were
combined in the second half of the whole framework. An average pooling layer and the DenseNet201
(excluding the last two layers) were also used for features extraction of ROI images. For SSI
values, a convolution layer and an average pooling layer were adopted for feature extraction and
dimensionality reduction. After features combination, three fully connected layers were constructed in
NISDL method II. Therefore, the skin temperature can be obtained. The algorithm in detail, including
NISDL method I and II, is shown in Table 1.
Figure 2. NISDL method II. (SSI values and skin images were processed by deep learning networks.
Two kinds of features were extracted and the features were combined.).
144
Appl. Sci. 2019, 9, 1375
Table 1. Contactless measuring method of skin temperature based on skin sensitivity index.
Algorithm: the NISDL algorithm

Output: NISDL model (*.h5), skin temperature (°C)
Step:
1. Video pre-processing
(1) De-noise and handle subtleness magnification for captured video.

(2) Magnification coefficient ξ is 10 (formula (3)).
(3) Extracting region of interest (ROI) from each frame of video, the size is 150×150.
2. Making label
(1) Making numerical interpolation for skin temperatures captured by iButton.

(2) Uniform interpolation is adopted, plus 11 points/min.
(3) Establish a correspondence table between ROI images and skin temperatures (after interpolation).
3. Algorithm training
(1) Commonality between NISDL method I and II
1) Training set and test set ratio: 12: 4.

2) Validation set: 500 images.
3) During network training, 32 images/batch, epoch is 8.
4) Training ~30000 images, validate once.
5) Activation function: ReLU
(2) NISDL method I
1) SSI values were concatenated with ROI images in the first steps.
2) Convolutional kernel: 1 × 1.
(3) NISDL method II
1) Features are extracted from SSI values and ROI images, respectively.
2) Concatenating the two kind of features in the second half of network.
3) Convolutional kernel: 3 × 1.
4. Optimizing model parameters
3.3.4. Evaluation Metric

For assessing NISDL algorithm constructed in this paper, the absolute error is adopted.
Error = |Tp (i) - Tr (i)| i = 1, 2, 3, . . . (4)
where Tp (i) is the prediction values of skin temperature and obtained from the proposed NISDL
algorithm. Tr (i) is the real value of skin temperature and captured by iButton. The parameter i denotes
the particular ROI image.
3.3.5. Algorithms for Comparison

Two algorithms are used for comparison in this paper: (1) DL algorithm. The commonality
between NISDL method I and II is that they all use SSI for model training. For validating the
effectiveness of NISDL algorithm (with SSI), we remove the SSI and corresponding hidden layers for
SSI features extraction from NISDL method II, so that it will be another deep learning network (without
SSI) and is named the DL algorithm hereinafter. (2) NIPST algorithm. DL algorithm, NISDL method I
and II are all nonlinear methods and data driven methods with deep learning networks. For further
validating NISDL method I and II, the NIPST algorithm is also used for algorithm comparison, which is
a linear and model driven method.
145
Appl. Sci. 2019, 9, 1375
4. Results
16 subjects were invited for subjective physiological experiments and a total of 1.44 million images
were captured. Based on this, the NISDL algorithm was validated and compared with the NIPST
algorithm and the DL algorithm.
4.1. Hardware Parameters

For this paper, a computer with a GPU was used for images processing and algorithm validation.
The GPU is GeForce GTX TITAN X, the CPU is Intel core i5-4460 [email protected] X 4, the RAM is 16G
and the word size is 64bit.
4.2. Training of NISDL method I

The size of ROI images are n × 150 × 150 × 3, and the size of expanded SSI vlues are n × 150
× 150 × 1. The SSI matrix was considered as a channel and concatenated with ROI images, so that
the result is n × 150 × 150 × 4. The activation function of four convolution layers, shown in Figure 1,
are Rectified Linear Units (ReLU) and the size of convolution kernel is 1 × 1. DenseNet201 was used
for feature extraction and its output is a matrix with size of n × 4 × 4 × 1920. Based on this, two
hidden layers are constructed and the size of the last layer, being a fully connected layer, is 1920 × n.
4.3. Training of NISDL Method II

The size of the expanded SSI matrix is n × 1920 × 1, which differs from that of NISDL method
I. The corresponding convolution kernel is 3 × 1. The features of SSI were extracted by two hidden
layers, and the features of ROI images were extracted by DenseNet201. As shown in Figure 2, in the
second half of framework, the two kinds of features are concatenated. In order to ensure that the SSI
features have a suitable influence on network training (moderate, not too big or too small), the size of
ROI image features is set as n × 1920, and size of SSI features is set as n × 640 (triple relationship).
Finally, the size of last three hidden layers is 2560 × 1024, 1024 × 512 and 512 × 1, respectively.
4.4. Commonality between NISDL Method I and II

During network training, the same parameters of NISDL method I and II are shown as follows.
Based on data of 1.44 million images, the ratio of training set and test set is 12:4, the number of
validation set is 500. The epoch is 8, which means that the training set was trained 8 times. The input
data batch is 32. When the error of validation set is less than 0.46 ◦ C, the corresponding model (*.h5)
will be saved. Further, when 30,0000 images of training set were trained, the corresponding model
(*.h5) will also be saved. After the generation of the model, the test set images were inputted into
generated model, so that the prediction values of skin temperatures could be obtained.
4.5. Quantitative Comparison

The prediction values of skin temperature are shown in Figure 3. The set of values obtained from
iButton is ground truth. The corresponding error statistics, including mean, median, are shown in
Figure 4, which is a box-whisker plot. The mean values of NISPT, DL, NISDL method I and II in ◦ C are
0.579, 0.359, 0.335 and 0.265, respectively. In addition, the median values of them in ◦ C are 0.343, 0.309,
0.238 and 0.228, respectively. It was shown that deep learning methods (DL, NISDL method I and II)
are all better than the nonlinear model (NIPST) and further that the method with SSI (NISDL method I
and II) is better than the method without SSI (DL).
In this paper, the error distributions are given in Figure 5 and Table 2. The errors of DL, NISDL
method I and II are mainly concentrated in the range of 0 ◦ C and 0.75 ◦ C. NISDL is better than DL,
because two error percentages of NISDL corresponding to [0, 0.25) are 52.25% and 55.62%. In addition,
the error percentages of NISDL corresponding to [0.25, 0.5) and [0.5, 0.75) are less than that of DL.
146
Appl. Sci. 2019, 9, 1375
The error percentage of NIPST is increased from the interval of [0.75, 1), meaning that the performance
of NIPST is worse than DL and NISDL methods I and II.

L%XWWRQJURXQGWUXWK L%XWWRQJURXQGWUXWK
1,367EDVHOLQH, 1,367EDVHOLQH,
'/EDVHOLQH,, '/EDVHOLQH,,
6NLQWHPSHUDWXUH°&
6NLQWHPSHUDWXUH°&
1,6'/PHWKRG, 1,6'/PHWKRG,
1,6'/PHWKRG,, 1,6'/PHWKRG,,

7LPHPLQ 7LPHPLQ
D E

L%XWWRQJURXQGWUXWK L%XWWRQJURXQGWUXWK
1,367EDVHOLQH, 1,367EDVHOLQH,
'/EDVHOLQH,, '/EDVHOLQH,,

6NLQWHPSHUDWXUH°&
6NLQWHPSHUDWXUH°&
1,6'/PHWKRG, 1,6'/PHWKRG,
1,6'/PHWKRG,, 1,6'/PHWKRG,,

7LPHPLQ 7LPHPLQ
F G
Figure 3. Skin temperatures comparison between ground truth, baseline and NISDL (Images of
12 subjects (0.96 million images) were training set; images of 4 subjects (0.32 million images) were test
set. Sub-figures (a), (b), (c) and (d) are the comparison results of 4 test set (data of 4 subjects)).

$EVROXWHHUURUV°&

1,367EDVHOLQH, '/EDVHOLQH,, 1,6'/PHWKRG, 1,6'/PHWKRG,,

1RQLQYDVLYHPHDVXULQJPHWKRGIRUVNLQWHPSHUDWXUH
Figure 4. Error statistics (box-whisker plot) comparison between baseline and NISDL (a). NIPST was
published in references [21]. (b). Skin images were trained directly by DenseNet201 to obtain a model
and predict skin temperature, and SSI was not involved. (c). NISDL method I and NISDL method II all
belong to NISDL. The main difference is that SSI values and skin images are combined at different times.
147
Appl. Sci. 2019, 9, 1375
Table 2. Absolute error distribution.
NIPST DL NISDL Method I NISDL Method II

Absolute Error (°C) ć
(Baseline, %) (Baseline, %) (Figure 1, %) (Figure 2, %)
[0, 0.25) 35.39 37.64 52.25 55.62
[0.25, 0.5) 24.16 35.96 28.09 30.34
[0.5, 0.75) 5.06 19.10 10.11 11.80
[0.75, 1.0) 11.24 5.62 3.37 2.250
[>1.0) 24.16 1.69 6.18 0
5. Discussion
5.1. Situation of Overcoming Challenges

NISDL has overcome the three challenges mentioned in Section 1 to some extent. A kind
of subtleness magnification technology, which is Euler Video Magnification (EVM), was used for
magnifying the skin texture variation, so that the challenges ‘(1)’ given in Section 1 can be overcome.
For overcoming challenges ‘(2)’ which is inter-individual difference, the skin sensitivity index (SSI)
was proposed and SSI is related with skin saturation. Figures 4 and 5 shows that the performance of
NISDL algorithm with SSI is better than that of DL without SSI. In practical application, skin images
will be captured in real-time (30 frames/s), the skin temperature variation can always be obtained.
Therefore, challenges ‘(3)’ proposed in Section 1 can be overcome. Furthermore, piecewise stationary
time series analysis was adopted in this paper for overcoming challenges ‘(3)’. Considering operability,
the breakpoint interval of piecewise stationary signal is set to 5 s, supposing, e.g., that the skin
temperature has a constant value during 5 s.
5.2. The Deep Learning Framework

In this paper, NISDL method I and II all belong to the deep learning method. In addition,
DL generated for algorithm comparison is also a deep learning method. The error distribution
comparison is shown in Figures 4 and 5 and the performance of NISDL II is encouraged. From the
perspective of deep learning, the main reason for this is that big data is adopted.

1,367EDVHOLQH,
'/EDVHOLQH,,
1,6'/PHWKRG,
1,6'/PHWKRG,,

$EVROXWHHUURUGLVWULEXWLRQ

> > > > >!
$EVROXWHHUURULQWHUYDOV°&
Figure 5. Error distribution comparison between baseline and NISDL (a). NIPST was published in [21].
(b). Skin images were used in training directly by DenseNet201 to obtain a model and predict skin
temperature, and SSI was not involved. (c). NISDL method I and NISDL method II all belong to NISDL.
The main difference is that SSI values and skin images are combined at different times.).
148
Appl. Sci. 2019, 9, 1375
5.3. The Proposed SSI

Although NISDL and DL are all better than NIPST, there are still a big gap between NISDL and
DL. Figures 4 and 5 and Table 2 also show that NISDL is better than DL. The main reason is that SSI
is used in NISDL. The NISDL method II is also shown to be better than NISDL method I. Further,
when SSI features are extracted and concatenated with ROI images features in the second half of
network, the performance will be better.
5.4. Reasons of Designing Two Frameworks for NISDL

Some researchers may ask, why do we design NISDL methods I and II together? The main reason
is that we want to extensively confirm the effectiveness of SSI. In this paper, SSI participates in network
training from different locations, and the results are all good. When the SSI values are removed from
network (DL), the corresponding performance decreased significantly. Based on this, we can know
that SSI is helpful for predicting skin temperature through deep learning networks.
5.5. Practical Application

Some researchers may argue that the method proposed in this paper still cannot be applied in
practice right now. In fact, a method is always being gradually improved. For example, the NISDL
proposed in this paper is better than NIPST which was proposed in 2017. In addition, when more
diverse data is captured and used for model training, the performance of NISDL will be better.
Some researchers may argue that the infrared sensor can also be used for measuring skin
temperature, so why do we use a vision-based method? In fact, the study [19,20] focused on thermal
comfort measurement with an infrared sensor. However, the measurement accuracy is limited.
Beyond this, there are other drawbacks in the infrared based measuring method: (1) Distance.
The infrared sensor should be placed close to occupant. (2) Cost. The infrared sensor with high
accuracy is expensive and the accuracy of an infrared sensor with low cost is also low. (3) Information
is limited. From the perspective of the human sensory system, the infrared sensor is ‘touch’ and
vision-based method is ‘sight’. The data captured by vision-based methods is much more than that
collected by infrared sensor. e.g., human poses can be captured by vision sensor rather than infrared
sensor for analyzing human thermal comfort. Based on these three drawbacks, the infrared-based
method is difficult to widely apply in practice.
Furthermore, some other researchers may say that vison-based contactless measuring method
may have concerns related to personal privacy. In fact, there are at least two options to protect personal
privacy issues: (1) Switch button. Based on this switch button, any customer can choose to accept or
reject the implementation of real-time personal service of thermal comfort. 92) Information selection.
In future practical applications, only the information about human thermal comfort be processed and
saved, while other information will be discarded in real-time. (3) Data protection. In order to avoid
data protection issues, processed data related to thermal comfort can be directly transferred to the
HVAC system instead of being saved. Therefore, from the perspective of human-centeredness, the
NISDL algorithm proposed in this paper is helpful.
In future practical applications, the processing capacity of HVAC system should be considered.
The HVAC system is usually equipped with computer server. The framework proposed in this paper
can be embedded on the computer server directly and GPU is required in computer server. Based on
occupant number in the building, it is necessary to prepare one or more GPUs. Further, the best ratio
between GPU number and occupant number should be validated and tuned in practical application
and will not be considered in this paper.
5.6. Exceptions
While human physiology and human thermal comfort remains a complex issue, the possibility to
measure temperature distributions on the body’s surface can provide valuable indicators. Therefore,
149
Appl. Sci. 2019, 9, 1375
in this paper, we just focus on the prediction of skin temperature. However, some exceptions still
should be mentioned. (1) There are some exceptions to the close relationship between skin temperature
and thermal comfort. e.g., while sweating occurs, as the reason of sweat evaporates and heat absorption,
the skin surface temperature will drop. However, the human perception could be hot. (2) There are
some special cases between thermal sensation and thermal comfort. e.g., Occupant sometimes has a
warm sensation, but he or she is very comfortable.
5.7. Others
Some potential limitations should be noted. (1) In this paper, all subjects are Asian females.
Therefore, maybe we only can say that the NISDL is applicable to Asian women at the present
time. Further, more data validation is required. (2) The subject acclimatization time is 10 min
in this paper. The result of a longer time, e.g., more than 30 min, will be better than that of
10 min. When the acclimatization time is 30 min, the subjects are more likely to reach a stable starting
state. (3) Relative constant parameters were set in an experiment chamber. This means that only one
thermal condition was tested. If we handle the physiological experiment in different indoor parameters
(e.g., indoor temperature), more valuable data and conclusions can be obtained.
6. Conclusions
In this paper, a kind of contactless measuring method based on skin sensitivity index for thermal
comfort (NISDL) is proposed. For validating the effectiveness of SSI, two different deep learning
frameworks with SSI were designed. A total of 1.44 million images were used for algorithm validation.
The conclusions can be summarized as follows.
(1) SSI is a good and high weight parameter in contactless measurement of skin temperature based
on a deep learning network.
(2) The location of SSI participation in NISDL network training has little impact on measuring the
performance of skin temperature. Of course, if the SSI features are extracted firstly, and then
merged with the features of ROI images, the corresponding effect is slightly better.
(3) The NISDL method proposed in this paper can be used for measuring thermal comfort and more
diverse data can help it to improve the measuring accuracy.
In practical application, the inter-difference is very large. How to define and calculate suitable SSI
will affect the measuring results. Further, more diverse data comparison is required to improve the
algorithm robustness. These areas will be our research directions in the near future.
Author Contributions: Conceptualization, X.C. and B.Y.; Formal analysis, X.C. and B.Y.; Methodology, X.C., B.Y.,
T.O. and H.L.; Software, X.C. and L.L.; Validation, B.Y., K.T. and E.I.; Writing—original draft, X.C.; Writing—review
& editing, B.Y. and A.H.
Funding: This study was supported by the International Postdoctoral Fellowship Program from China
Postdoctoral Council (No. 20160022), the National Natural Science Foundation of China (No. 61401236), the
Jiangsu Postdoctoral Science Foundation (No. 1601039B), the Key Research Project of Jiangsu Science and
Technology Department (No. BE2016001-3).
Acknowledgments: The authors thank William T. Freeman (MIT) for providing his MATLAB code of Euler Video
Magnification (EVM).
References
1. U.S. Energy Information Administration. International Energy Outlook 2017; IEO2017 Report; U.S. Energy
Information Administration (EIA): Washington, DC, USA, 2017.
2. U.S. Energy Information Administration. Energy Implications of China’s Transition toward Consumption-Led
Growth; IEO2018 Report; U.S. Energy Information Administration (EIA): Washington, DC, USA, 2018.
150
Appl. Sci. 2019, 9, 1375
3. Zagreus, L.; Huizenga, C.; Arens, E.; Lehrer, D. Listening to the occupants: A web-based indoor
environmental quality survey. Indoor Air 2004, 14, 65–74. [CrossRef]
4. Zhao, Q.; Zhao, Y.; Wang, F.; Wang, J.; Jiang, Y.; Zhang, F. A data-driven method to describe the personalized
dynamic thermal comfort in ordinary office environment: From model to application. Build. Environ. 2014,
72, 309–318. [CrossRef]
5. Ghahramani, A.; Tang, C.; Becerik-Gerber, B. An online learning approach for quantifying personalized
thermal comfort via adaptive stochastic modeling. Build. Environ. 2015, 92, 86–96. [CrossRef]
6. Liu, W.; Lian, Z.; Zhao, B. A neural network evaluation model for individual thermal comfort. Energ. Build.
2007, 39, 1115–1122. [CrossRef]
7. Afroz, Z.; Urmee, T.; Shafiullah, G.M.; Higgins, G. Real-time prediction model for indoor temperature in a
commercial building. Appl. Energ. 2018, 231, 29–53. [CrossRef]
8. Wang, D.; Zhang, H.; Arens, E.; Huizenga, C. Observations of upper-extremity skin temperature and
corresponding overall-body thermal sensations and comfort. Build. Environ. 2007, 42, 3933–3943. [CrossRef]
9. Nakayama, K.; Suzuki, T.; Kameyama, K. Estimation of thermal sensation using human peripheral skin
temperature. In Proceedings of the IEEE International Conference on Systems, Man and Cybernetics
(SMC2009), San Antonio, TX, USA, 11–14 October 2009; pp. 2872–2877.
10. Liu, W.; Lian, Z.; Deng, Q.; Liu, Y. Evaluation of calculation methods of mean skin temperature for use in
thermal comfort study. Build. Environ. 2011, 46, 478–488. [CrossRef]
11. Takada, S.; Matsumoto, S.; Matsushita, T. Prediction of whole-body thermal sensation in the non-steady state
based on skin temperature. Build. Environ. 2013, 68, 123–133. [CrossRef]
12. Sim, S.Y.; Koh, M.J.; Joo, K.M.; Noh, S.; Park, S.; Kim, Y.H.; Park, K.S. Estimation of thermal sensation based
on wrist skin temperatures. Sensors 2016, 16, 420. [CrossRef] [PubMed]
13. Wu, Z.; Li, N.; Cui, H.; Peng, J.; Chen, H.; Liu, P. Using upper extremity skin temperatures to assess thermal
comfort in office buildings in Changsha, China. Int. J. Environ. Res. Pu 2017, 14, 1092–1109.
14. Chaudhuri, T.; Zhai, D.; Soh, Y.C.; Li, H.; Xie, L. Thermal comfort prediction using normalized skin
temperature in a uniform built environment. Energ. Build. 2018, 159, 426–440. [CrossRef]
15. Yao, Y.; Lian, Z.; Liu, W.; Jiang, C.; Liu, Y.; Lu, H. Heart rate variation and electroencephalograph - the
potential physiological factors for thermal comfort study. Indoor Air 2009, 19, 93–101. [CrossRef] [PubMed]
16. Dai, C.; Zhang, H.; Arens, E.; Lian, Z. Machine learning approaches to predict thermal demands using skin
temperatures: Steady-state conditions. Build. Environ. 2017, 114, 1–10. [CrossRef]
17. Chaudhuri, T.; Soh, Y.C.; Li, H.; Xie, L. Machine learning based prediction of thermal comfort in buildings of
equatorial Singapore. In Proceedings of the IEEE International Conference on Smart Grid and Smart Cities,
Singapore, 23–26 July 2017; pp. 72–77.
18. Kim, J.; Zhou, Y.; Schiavon, S.; Raftery, P.; Brager, G. Personal comfort models: Predicting individuals’
thermal preference using occupant heating and cooling behavior and machine learning. Build. Environ. 2018,
129, 96–106. [CrossRef]
19. Ghahramani, A.; Castro, G.; Becerik-Gerber, B.; Yu, X. Infrared thermography of human face for monitoring
thermoregulation performance and estimating personal thermal comfort. Build. Environ. 2016, 109, 1–11.
[CrossRef]
20. Ghahramani, A.; Castro, G.; Karvigh, S.A.; Becerik-Gerber, B. Towards unsupervised learning of thermal
comfort using infrared thermography. Appl. Energ. 2018, 211, 41–49. [CrossRef]
21. Cheng, X.; Yang, B.; Olofsson, T.; Li, H.; Liu, G. A pilot study of online non-invasive measuring technology
based on video magnification to determine skin temperature. Build. Environ. 2017, 121, 1–10. [CrossRef]
22. Meier, A.; Dyer, W.; Graham, C. Using human gestures to control a building’s heating and cooling System.
In Proceedings of the International Conference on Energy-Efficient Domestic Appliances and Lighting
(EEDAL 2017), Irvine, CA, USA, 13–15 September 2017.
23. Dziedzic, J.W.; Yan, D.; Novakovic, V. Real time measurement of dynamic metabolic factor (D-MET).
In Proceedings of the 9th International Cold Climate Conference Sustainable new and renovated buildings
in cold climates (Cold Climate HVAC 2018), Kiruna, Sweden, 12–15 March 2018.
24. Peng, B.; Hsieh, S. Data-driven thermal comfort prediction with support vector machine. In Proceedings
of the ASME 2017 12th International Manufacturing Science and Engineering Conference (MSEC),
Los Angeles, CA, USA, 4–8 June 2017; pp. 1–8.
151
Appl. Sci. 2019, 9, 1375
25. Wang, Z.; de Dear, R.; Luo, M.; Lin, B.; He, Y.; Ghahramani, A.; Zhu, Y. Individual difference in thermal
comfort: A literature review. Build. Environ. 2018, 138, 181–193. [CrossRef]
26. Peng, Y.; Rysanek, A.; Nagy, Z.; Schlüter, A. Using machine learning techniques for
occupancy-prediction-based cooling control in office buildings. Appl. Energ. 2018, 211, 1343–1358.
[CrossRef]
27. Fanger, P.O. Thermal Comfort: Analysis and Application in Environmental Engineering; Danish Technical Press:
Copenhagen, Denmark, 1970.
28. Farhan, A.A.; Pattipati, K.; Wang, B.; Luh, P. Predicting individual thermal comfort using machine learning
algorithms. In Proceedings of the IEEE International Conference on Automation Science and Engineering
(CASE), Gothenburg, Sweden, 24–28 August 2015; pp. 708–713.
29. Megri, A.; Naqa, I. Prediction of the thermal comfort indices using improved support vector machine
classifiers and nonlinear kernel functions. Indoor Built Environ. 2016, 25, 6–16. [CrossRef]
30. Li, W.; Zhang, J.; Zhao, T.; Wang, J.; Liang, R. Experimental study of human thermal sensation estimation
model in built environment based on the Takagi-Sugeno fuzzy model. Build. Simul. 2018. [CrossRef]
31. Cosma, A.C.; Simha, R. Machine learning method for real-time non-invasive prediction of individual thermal
preference in transient conditions. Build. Environ. 2019, 148, 372–383. [CrossRef]
32. LeCun, Y.; Bengio, Y.; Hinton, G. Deep learning. Nature 2015, 521, 436–444. [CrossRef] [PubMed]
33. Huang, G.; Liu, Z.; Maaten, L.V.D. Densely connected convolutional networks. In Proceedings of the IEEE
Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI, USA, 21–26 July 2017.
34. Wu, H.; Rubinstein, M.; Shih, E.; Guttag, J.; Durand, F.; Freeman, W. Eulerian video magnification for
revealing subtle changes in the world. Acm T Graph. 2012, 31, 1–8. [CrossRef]
35. Wadhwa, N.; Wu, H.; Davis, A.; Rubinstein, M.; Shih, E.; Mysore, G.; Chen, J.; Buyukozturk, O.; Guttag, J.;
Freeman, W.; et al. Eulerian video magnification and analysis. Commun. Acm 2017, 60, 87–95. [CrossRef]
36. Cheng, X.; Li, B.; Chen, Q. Online structural breaks estimation for non-stationary time series models.
China Commun. 2011, 7, 95–104.
152
applied
sciences
Article
Keeping Doors Closed as One Reason for Fatigue in
Teenagers—A Case Study
Anna Mainka * and Elwira Zajusz-Zubek
Department of Air Protection, Silesian University of Technology, 22B Konarskiego St., 44-100 Gliwice, Poland
Received: 9 July 2019; Accepted: 22 August 2019; Published: 28 August 2019
Featured Application: Presented results are only a case study and the authors are not able to relate
them to the population of children and teenagers in Poland; however, it is worth underlining the
positive effects of opening doors during sleep, since the CO2 concentration decreases 55−64%
without a reduction in thermal comfort. This simple action applied by parents can decrease the
contribution of low indoor environment quality (IEQIV ) in children’s and teenager’s bedrooms
to approximately 1% during the night.
Abstract: (1) Background: Healthy teenagers are often sleepy. This can be explained by their
physiology and behavioral changes; however, the influence of CO2 concentration above 1000 ppm
should not be neglected with respect to sleep dissatisfaction. (2) Methods: CO2 concentrations
were measured in two similar bedrooms occupied by girls aged 9 and 13 years old. The scheme of
measurements included random opening and closing of the bedroom doors for the night. Additionally,
the girls evaluated their sleep satisfaction in a post-sleep questionnaire. (3) Results: During the
night, the CO2 concentration varied from 402 to 3320 ppm in the teenager’s bedroom and from
458 to 2176 ppm in the child’s bedroom. When the bedroom doors were open, inadequate indoor
air quality (IEQIII and IEQIV categories) was observed in both the teenager’s and child’s bedroom
during 11% and 25% of the night, respectively; however, closing the doors increased the contribution
of moderate (IEQIII ) and low (IEQIV ) categories of air to 79% and 86%, respectively. The girls were
dissatisfied only when the bedroom door was closed. The satisfied category of sleep was selected
only by the younger girl. (4) Conclusions: Opening the bedroom door during the night can decrease
the CO2 concentration 55–64% without reducing thermal comfort.
Keywords: teenagers; children; bedroom; IEQ; CO2
1. Introduction
There is a generally accepted opinion in society that teenagers are sleepy. This notion indicates
that sleepiness is a negative issue; however, this particular group of young people has several reasons
for being tired. From a physiological perspective, teenagers undergo the processes of puberty and rapid
increases in height [1], and, from a psychological aspect, they have problems with low self-esteem,
seeking their role in society, and getting easily involved in personal conflicts at school or within the
family [1,2]. The duration of sleep varies among individuals, with an average teenager attending
school requiring at least 8 h of sleep per night; however, the maturational changes combined with a
cell phone or computer use within the hour prior to trying to fall asleep lead to shorter sleeping hours
or sleep deprivation [3]. Sleep issues have drawn worldwide attention in recent years. Sleep problems
threaten health and quality of life for up to 45% of the world’s population, and 35% of individuals feel
that they do not get enough sleep, which has negative effects on both physical and mental health [4].
Rafihi-Ferreira et al. [5] referenced several studies, in which the association between sleep quality

Appl. Sci. 2019, 9, 3533
and behavioral problems was investigated, demonstrated a relationship between externalizing and
internalizing problems and sleep problems in children. This association becomes even more serious
when considering that childhood sleep problems may persist and that they constitute a greater risk for
other behavioral problems. In comparison with psychological state, body condition, and circadian
rhythm, indoor environment quality (IEQ) including thermal conditions and indoor air quality (IAQ)
can improve sleep through environmental control. Matricciani et al. [6] in the meta-review underlined
that sleep duration is associated with adiposity in children, while sleep quality, timing and variability
appear important for children’s health, but further research is needed. Children and teenagers are
more adversely affected by indoor air pollution than adults since they breathe a greater volume of air
relative to their body weight, which may lead to a greater burden of pollutants on their bodies [7,8].
The quality of the indoor environment not only affects health and comfort, but it may also impair
learning ability. Poor IEQ, as a result of inadequate thermal conditions, and IAQ can cause symptoms
such as being too hot or too cold as well as feeling restless or sleepy [7].
A powerful remedial to improve IAQ is to increase the outdoor air supply rate; however, in
recent years, effort has been focussed on decreasing energy consumption. Within houses, the building
airtightness is maximized and the ventilation rates are minimized, leading to a general deterioration in
indoor air quality. Consequently, by focussing on optimal indoor thermal comfort, air infiltration and
dilution of indoor air pollutants are lowered.
Generally, in passive stack ventilation systems, to create optimal air quality conditions for the
maintenance of health and comfort of the occupants, windows should be open slightly. In moderate
climate zones such as Poland, this is the best solution from spring to fall seasons; however, during
winter, unsealed windows decrease thermal comfort. The Upper Silesia region, in comparison with
other Polish regions and European Union countries, is characterized by relatively high particulate
matter concentrations. Although the last three decades of economic changes have forced the greatest
drop in Polish industrial air pollution by closure of the old steelworks, cookeries, coal mines, and
coking plants, emissions from small-scale combustion utilities, such as domestic boilers, together with
re-suspension processes from urban surfaces and road traffic, have become particularly dangerous.
Investigations performed during the winter season confirm that the hazard of domestic sources
originates from the low quality of fuels (coal, biomass, culm, or even refuse) used for heating [9,10].
According to the dominant role of coal combustion in the region, residents are even advised not to
open windows during winter, particularly at low wind speed, since it decreases IAQ.
IAQ is a multi-disciplinary phenomenon and is determined by many pathways in which chemical,
biological, and physical contaminants eventually become a portion of the total indoor environmental
composition. There is spatial and temporal heterogeneity of these contaminants, and the determination
of exposure is difficult due to the diversity of time that occupants spend within the space [11,12].
Among indoor air pollutants, carbon dioxide (CO2 ) is considered a useful and easily measurable
indicator of the ventilation and air quality in indoor environments. Although this pollutant itself does
not cause serious health issues at lower concentrations and, for short durations, higher concentrations
can indicate a lower ventilation level and possible air contamination with other pollutants [13].
In residential non-smoking areas, with the exception of kitchens, occupants are a major source of
CO2 through exhalation. Mean concentrations of CO2 during cooking times are significantly higher
(p < 0.01) than those during non-cooking times, with no significant differences among fuel types [14].
While the CO2 generation rate per person varies as a function of age, activity, and diet, on average,
children aged 6–11 years old produce 448 mg/min/p and adults generate 763 mg/min/p [15]. Persily
and de Jonge in [16] presented CO2 generation rates according to age and gender. During sleep, the
level of physical activity (MET) is 1.0; thus, the calculated metabolic rate in children aged 6–11 years
old is 295 mg/min/p and 271 mg/min/p for boys and girls, respectively, while, for 11–16-year-old
males and females, it is 401 mg/min/p and 342 mg/min/p, respectively. Teenagers generate more CO2
according to their higher body weight, and they demand more privacy, therefore keeping their door
154
Appl. Sci. 2019, 9, 3533
closed all day and night. The metabolic activity itself influences air quality in bedrooms by reducing
the concentration of oxygen and increasing the level of CO2 .
For CO2 , the exposure limits have been derived exclusively on the basis of health considerations.
The Exposure Guidelines for Residential Indoor Air Quality [17] suggest that CO2 concentrations above
1800 mg/m3 (1000 ppm) are indicative of an inadequate supply of fresh air, although complaints have
been documented at concentrations as low as 1100 mg/m3 (600 ppm) [17]. Assumptions regarding these
levels are in accordance with ASHRAE standards, which previously (62-1989) considered 1000 ppm as
the highest acceptable concentration for a minimum sanitary requirement and 8 l/p/s as the minimum
ventilation rate. More recent standards (ASHRAE 62-1999, 62-2001, and 62-2004) recommend that
the indoor-outdoor differential concentration should not exceed 700 ppm [15]. Generally, a sufficient
margin to protect against undesirable changes in the acid-base balance and subsequent adaptive
changes such as the release of calcium from bones is a level of 6300 mg/m3 (3500 ppm) [17]. However, a
higher maximum exposure concentration is recommended for direct physiological effects of exposure
to CO2 as opposed to subjective symptoms. Subjective symptoms such as fatigue, headaches, and
an increased perception of warmth and unpleasant odors have been associated with CO2 levels of
900–5800 mg/m3 (500–3200 ppm) [17].
Polish legal acts do not specify permissible concentrations of CO2 in ambient air or in rooms
intended for the permanent residence of individuals, i.e., apartments and houses. The regulations
of the Ministry of Family, Work, and Social Policy define the highest permissible concentrations and
intensities of agents harmful to health in the work environment [18]. The highest permissible CO2
concentration is 9000 mg/m3 . Thus far, the Polish Committee for Standardisation followed general
guidelines concerning the quality of air inside non-residential buildings (PN-EN 13779 standard [19]),
but, in May 2019, the Polish Committee for Standardisation accepted EN-16798-1 (in English) [20] as
a national standard and in CEN/TR 16798-2 Technical Report [21] including recommended criteria
for the CO2 calculation for demand-controlled ventilation in occupied living rooms and bedrooms.
Design ΔCO2 concentration for bedrooms (ppm above outdoors) are within the limits of ≤380 ppm
(IEQI ), 380–550 ppm (IEQII ), 550–950 ppm (IEQIII ), and ≥950 ppm (IEQIV ) corresponding to high (I),
medium (II), moderate (III), and low (IV) indoor environment quality, respectively.
IEQ evaluation based on CO2 concentration and possible effects on occupants comfort has been
examined [11,22–27]. Polish research on CO2 levels and possible occupant fatigue includes a few
publications [9,28–31]; however, no research regarding IEQ in the bedrooms of Polish children and
teenagers have been performed. Such studies can help to better define how behavioral patterns
influence the possible exposure levels of air pollutants and occupant fatigue. The present paper
presents the results of the measurement of CO2 levels and the corresponding IAQ linked to the possible
tiredness and lethargy of a teenager as compared with a child.

The concentration of CO2 was continuously measured in two bedrooms: one occupied by a
teenage girl aged 13 and a second by a younger girl aged 9. The choice of girls instead of boys was
intentional since, among children and teenagers, more males than females exhibit excessive nightly
use of computer games and consequent sleep disturbances [1]. Moreover, Karjalainen [32] suggested
that female occupants should primarily be used as subjects when examining indoor thermal comfort
requirements, since, if women are satisfied, then it is highly probable that men are also satisfied.
Participation in the present study was voluntary. Prior to starting the measurements, the participants’
parents gave informed consent.
The home of the girls was located in the suburbs of the western part of the Upper Silesia region.
The distance from the urban area was approximately 10 km (Figure 1). The house was built in 2008
and was located in the third row of buildings, 100 m from the local street. It was a detached house
(Figure 1) with an attic and no basement. The house area was 98 m2 with a glass terrace of 12 m2 . The
measurements were performed in the South and North bedrooms, which each had an area of approx.
155
Appl. Sci. 2019, 9, 3533
12 m2 and were occupied by the older and younger girl, respectively; detailed information has been
added in Supplementary Materials Table S1.
Figure 1. Location, south view and ground floor plan of the house.
The sampling strategy for CO2 included indoor CO2 concentration, measurement, and
questionnaires. Precautions necessary to avoid measurements in air directly exhaled by building
occupants were performed as described in [33]. One automatic portable monitor (model 77535, Az
Instruments International Ltd., Hong Kong, China) connected to a PC with RS232 software installed
was used in the present study. According to Mahyuddin et al. [34], one CO2 sensor in a room with
<100 m2 floor area had significant p-value relationships, and the breathing zone within the occupied
space was considered to be between 1.0 and 1.2 m in the middle of a zone as a representative location.
The monitor was equipped with a non-dispersive infrared sensor. The precision of measurements
ranged between +0 and +10,000 ppm CO2 : ±100 ppm CO2 or ±3% at a concentration below 100 ppm.
The monitor also displayed and recorded in real-time the measurements of air temperature and relative
humidity in the bedroom, allowing for logged data to be downloaded for analysis. The selected
sampling interval was 60 s.
To estimate the parameters influencing CO2 concentrations in the bedrooms during the
measurements, parents were asked to note whether each daughter left the room for a longer period
during the night. They provided a diary in which it was specified whether the windows and doors were
closed, ajar, or fully open during the monitored days and nights. Additionally, the girls were asked to
evaluate their sleep satisfaction as categorized into five levels [4]: very dissatisfied (1), dissatisfied (2),
moderate (3), satisfied (4), and very satisfied (5). However, for the analysis, three levels were used:
very dissatisfied and dissatisfied, defined as dissatisfied; moderately satisfied, defined as moderate;
and satisfied and very satisfied, defined as satisfied. The evaluations of sleep quality were included in
the post-sleep questionnaire. In total, 137 questionnaires were acquired, with 121 being valid.
During the study, the family was asked to maintain their regular routine regarding the opening of
doors and windows. It should be underlined that, during daylight hours, the bedroom windows were
fully open only twice for cleaning and ajar for a maximum of 1 h less than 10 times. The measurements
were performed between 16 September 2018 and 2 March 2019. For the final analysis, 102 nights from
9:00 p.m. to 7:00 a.m. (51 nights per room) were included.
156
Appl. Sci. 2019, 9, 3533
3. Results
As mentioned above, proper ventilation controls IAQ. Among the techniques adopted to evaluate
ventilation rates, the one based on the measurement and analysis of the indoor CO2 concentration
and trends is the most common approach; however, it could be improved by the integration of an
electronic nose for odor detection [35]. Acceptable ventilation conditions can be easily achieved in
mechanically ventilated rooms, but it is not equally simple to maintain CO2 and odor levels under
control in indoor environments that are naturally ventilated [35]. Natural ventilation is typical for
detached buildings in Poland, where the ventilation and the IAQ are controlled only by means of
air infiltration through cracks and openings. The IAQ gets worse during the winter season when
the desire for thermal comfort and acceptable IAQ are in conflict. For the characterization of IAQ by
measuring CO2 concentration, several indicators and criteria can be used [25]. We selected the average
CO2 concentration and a time fraction over a limit of ΔCO2 values spent during the night according to
the PN-EN 16798 standards [20,21].
Figure 2 shows the field measurements in the teenager’s and child’s bedrooms, including the CO2
concentration range, when the doors were open and closed.
Figure 2. CO2 concentration inside the teenager’s and child’s bedrooms during the night.
During the night, the short-term (60 s average) concentrations varied from 402 to 3320 ppm in the
teenager’s bedroom and from 458 to 2176 ppm in the child’s bedroom. Thus, the highest concentrations
exceeded the limits characteristic of subjective symptoms such as fatigue, headaches, and an increased
perception of warmth and unpleasant odors [14]. If we compare the data separately by day of the week
(Table 1), it is clear that, despite similar average concentrations, the maximum concentrations were
significantly higher in the teenager’s bedroom.
Table 1. Average, minimum, and maximum CO2 concentrations according to the day of the week.
CO2 Concentrations, ppm

Parameters
Teenager’s Bedroom Child’s Bedroom
Average Minimum Maximum Average Minimum Maximum
Monday 1127 402 3102 1828 1828 1828
Tuesday 998 597 2336 1114 554 1855
Wednesday 1088 452 2897 1144 458 1855
Thursday 1199 411 2695 1118 688 1778
Friday 1195 479 3320 1076 563 1757
Saturday 3061 3061 3061 1262 599 2176
Sunday 1191 405 3188 1993 1993 1993
157
Appl. Sci. 2019, 9, 3533
Distinguishing between open and closed doors in the teenager’s and child’s bedrooms, the average
nightly concentrations were not significantly different (p = 0.76; 1077 and 1103 ppm, respectively);
however, there was a statistically significant difference in CO2 concentration between bedrooms with
open and closed doors (Figure 3). In the teenager’s bedroom, opening the doors during the night
decreased the average CO2 concentration by 55%, and in the child’s bedroom, the observed decrease
was 64%.
Figure 3. Mean temperature, relative air humidity, and CO2 concentration inside the teenager’s and
child’s bedrooms during the night.
Moreover, room temperature and humidity are important factors that may influence perceived
environment quality. They both affect the thermal balance of the human body through effects on the
skin and respiratory organs. Changes in air temperature trigger a sympathetic reflex via the skin
that strengthens with lower air temperature. High temperatures and humidity require the human
body to respond by increasing heat loss through the skin surface via blood circulation. The relation
between indoor temperature and humidity depends on the season and indoor conditions [36]. In
our study in the younger girl bedroom, the temperature and humidity correlation coefficient were
−0.10 and −0.08 (p < 0.05) with opened or closed doors, respectively. In the older girl bedroom, the
temperature and humidity were weakly positively correlated in the case of opened doors 0.26 and
negatively correlated −0.33 in the case of closed doors. The mean room temperature differed between
the teenager’s and child’s bedrooms significantly (p < 0.01). The mean temperature in the teenager’s
bedroom (21.1 ± 1.11 ◦ C) was lower than that in the child’s bedroom (22.05 ± 0.59 ◦ C); however, the
influence of opening and closing the doors to each bedroom during the night was not statistically
significant (p > 0.05). The mean nightly temperature in the teenager’s bedroom was 21.07 ± 1.39 ◦ C
and 21.01 ± 0.76 ◦ C with open and closed doors, respectively, while, in the child’s bedroom, the mean
nightly temperature was 22.06 ± 0.62 ◦ C and 22.03 ± 0.57 ◦ C, respectively. In the case of humidity,
a significant difference (p < 0.01) between the teenager’s and child’s bedrooms was also observed
(46.46 ± 5.79% and 33.39 ± 3.41%, respectively). In the teenager’s bedroom, the humidity did not differ
between nights with respect to open and closed doors (46.04 ± 5.97% and 46.89 ± 5.68%, respectively),
while, in child’s bedroom, this difference was statistically significant (p < 0.01; 31.83 ± 3.20% and
34.88 ± 2.95%, respectively).
Based on WHO regulations [37] and the PN-EN 16798-1 and CEN/TR 16798-2 standards [20,21],
the increase in CO2 concentration in relation to that of outdoor air (ΔCO2 ) was measured in both
bedrooms. During the night, the general IEQ in the bedrooms included 32% in the high category (I),
28% in the medium category (II), 19% in the moderate category (III), and 21% in the low category (IV)
in the older girl’s bedroom and 14%, 29%, 32%, and 24%, respectively, in the younger girl’s bedroom.
This indicates that, for 10 h of rest at night, the teenager spent approximately 3 h in proper air quality,
while the child spent only approximately 1.5 h. The indoor concentrations of CO2 showed a higher
158
Appl. Sci. 2019, 9, 3533
contribution of inadequate air quality in the III and IV categories, corresponding to moderate and
low IEQ, in the younger girl’s bedroom. When the bedroom doors were open, inadequate IEQ was
observed in the teenager’s and child’s bedrooms for 11% and 25% of the night, respectively (Figure 4);
however, closing the doors increased the contribution of the moderate (III) and low (IV) categories to
79% and 86%, respectively.
Figure 4. Categories of the indoor environment inside the teenager’s and child’s bedrooms during
the night.
Despite the fact that the teenage girl spent an extra 1.5 h in the high and medium quality
environment (IEQI and IEQII ), this did not correspond to sleep satisfaction. In the older girl’s opinion,
her sleep during study was never satisfying. During 51 nights of measurement, the teenage girl
evaluated her sleep as moderate (3) for 36 nights, 13 nights as dissatisfied (2), and only two nights
as very dissatisfied (1). The younger girl made a more positive assessment, evaluating seven nights
as satisfied (4) and one night as very satisfied (5), eight nights categorized as dissatisfied (2), while
one night as very dissatisfied (1), and 34 nights as moderate (3). Following the small contribution of
very dissatisfied (1) and very satisfied (5) categories, we gathered categories (1) and (2) into one group
defined as low sleep satisfaction, in addition to categories (4) and (5) in one group defined as high
sleep satisfaction, as presented in Figure 5.
Figure 5. Low, moderate, and high sleep satisfaction of the teenager and child corresponding to
categories: Very dissatisfied and dissatisfied (1) + (2); moderate (3), as well as satisfied and very satisfied
(4) + (5).
159
Appl. Sci. 2019, 9, 3533
Deeper analysis reveals that the girls were dissatisfied only when the bedroom door was closed.
It is interesting that the average CO2 concentration during the nights evaluated as dissatisfied had a
similar average value of 1691 ppm and 1623 ppm in the older and younger girl’s bedroom, respectively.
Moreover, the average concentration ranges were not statistically different (1070 ppm and 1079 ppm to
2308 ppm and 1878 ppm) in the teenager’s and child’s bedroom, respectively. The moderate category
was selected during the nights with open as well as closed doors. The average concentration of CO2
during nights defined as moderate with open doors was 766 ppm and 903 ppm in the teenager’s and
the child’s bedroom, respectively, while, during nights with closed doors, the CO2 concentration was
968 ppm and 1185 ppm, respectively. The satisfied category of sleep was selected only by the younger
girl during nights with open doors, corresponding to an average CO2 concentration of 769 ppm.
4. Discussion
To ensure that the CO2 concentration is at a lower level, it is necessary to supply an adequate
amount of fresh air instead of used air. Bekö et al. [38] summarised that the most influential variables
on the air change rate (ACR) are room volume, number of individuals sleeping in the bedroom, average
window and door opening habits, location of the measured room (ground or higher floor), and year of
building construction. The authors underlined that more door opening results in a higher air change
rate. Our study was performed in two similar bedrooms occupied by girls aged 13 and 9 years old.
Due to a low IAQ outside the building, the window opening was very limited; thus, we focussed
on door opening and compared it with subjective evaluation of sleep satisfaction. Research by Bekö
et al. [38] revealed that, when the bedroom doors are ajar or fully open instead of closed, the ACR
increases from 0.48 to 0.55 (14%) and 0.71 (48%). Although we did not calculate the ACH, the observed
decrease in CO2 concentration while keeping the bedroom doors open during the night was 55% and
64% in the teenager’s and the child’s bedroom, respectively.
Generally, adults appear to be additional immune to the consequences of sleep deprivation,
whereas children, teenagers and particularly young children tend to be a vulnerable subpopulation
as they spend the majority of their time indoors at home while their respiratory and other systems
are under development [8]. When an individual does not get enough sleep to feel awake and alert,
they begin to experience symptoms of sleep deprivation such as yawning, irritability, fatigue, inability
to concentrate, moodiness, forgetfulness, lack of motivation, depression, and poor perspective on
life [3]. In our study, the average concentration during the night was similar in both bedrooms, at
1077 ppm (teenager) and 1103 ppm (child), which is comparable with the average CO2 concentration
in unoccupied bedrooms (999 ppm and 1236 ppm in South and North bedrooms, respectively) as
reported by Bouvier et al. [25]. In occupied bedrooms, the authors in [25] reported higher (1585 ppm
and 1760 ppm in South and North bedrooms, respectively) than our average concentration of CO2 ,
which was 766 ppm and 860 ppm with open doors and 1402 ppm and 1337 ppm with closed doors, in
the teenager’s and the child’s bedroom, respectively (Figure 3). However, it is similar to the average
CO2 concentration of 716 ppm measured in 13 bedrooms located in Athens’ residences [39].
During the night with the doors closed, in the teenager’s and the child’s bedroom, 48% and 43%
of CO2 concentration measurements were higher than 950 ppm above outdoor CO2 concentration
(398 ppm), which corresponds to category IV–low indoor environment quality (IEQIV ), and, when
open, it was < 1% in both bedrooms. The PN-EN 16798-1 standard [20] underlines that a lower level
will not provide any health risk but may decrease comfort. At the same time, the categories are related
to the levels of expectations. A normal level would be medium, but, for occupants with special needs
(children, elderly, persons with disabilities, etc.), a higher level might be selected. During the research
with opened doors, a high IEQ prevailed during 51% of the night in the teenager’s bedroom, and
during 26% of the night in the younger girl’s bedroom. In comparison, the authors in [25] reported
that time spent by an occupant in a CO2 concentration over 1500 ppm in a mechanically ventilated
detached house was approximately 30% of an entire day; thus, the IEQ in researched bedrooms could
be satisfying with the door open; however, only the younger girl evaluated sleep satisfaction as high
160
Appl. Sci. 2019, 9, 3533
(16%). The subjective assessment of sleep satisfaction points to a lower IEQ in the case of the older girl’s
bedroom, which can be supported by a greater CO2 concentration range (Figure 2) in the teenager’s
bedroom. Nevertheless, the sharing of categories of environment calculated based on the PN-EN
16798-1 standard [20] does not explain the lower sleep satisfaction in the older girl’s assessment. The
reason can be seen in Table 1, which presents average CO2 concentrations and ranges depending on
the day of the week. In the teenager’s bedroom, the maximum CO2 concentrations were >3000 ppm
on four days of the week (Mondays, Fridays, Saturdays, and Sundays), while, in the younger girl’s
bedroom, the maximum CO2 concentration exceeded 2000 ppm only on Saturday nights. These results
indicate the important role of short-term high concentrations in sleep satisfaction assessment.
Increasing the effectiveness of the building ventilation system is to decrease the number of
exceedances of CO2 concentration in naturally ventilated buildings. A cheap and simple solution is
the application of ventilation grills to the window frames [40]. We would recommend simple passive
grilles, where the airflow is due to the pressure drop between the indoor and the outdoor environment
in addition to the typical pressure drop of the grille itself. Humidity-sensitive grills might not be very
effective in this case because of lower humidity values (29.7–50.4%) in the bedrooms during the heating
season (Supplementary Materials Table S1).
5. Conclusions
We investigated the variability of CO2 concentration in naturally ventilated bedrooms occupied
by a teenage girl and a female child according to their sleep satisfaction. Despite the fact that the
average CO2 concentration was 1077 ppm in the teenager’s bedroom and 1103 ppm in the child’s
bedroom, the concentration ranges were much wider, from 402 to 3320 ppm and 458 to 2176 ppm,
respectively. The average concentration during the night with the doors closed was 1402 ppm and
1336 ppm, respectively; however, the simple action of door opening decreased the CO2 concentration
by 55% and 64% in the older and younger girl’s bedroom, respectively.
In the studied period of heating season (from 16 September 2018 to 2 March 2019) during the
night (9:00 p.m. to 7:00 a.m.), the time spent by the teenage and younger girls at a concentration over
1348 ppm (IV—low IEQ category) was generally 21% and 24%, respectively. With the doors open, the
low IEQ contributed to <1% of the time in both bedrooms; however, keeping the doors closed increased
the contribution of low IEQ to 48% and 46% in the teenager’s and the girl’s bedroom, respectively.
This highlights the strong influence of simple door opening during children’s sleep.
The evaluation of sleep satisfaction highlights that the teenage girl was more dissatisfied in the
mornings than was the younger girl, indicating the important role of short-term CO2 concentrations
>3000 ppm in sleep satisfaction assessment.
There are several limitations to this work. This study was limited by small sample size and was
confined to one house. The suburban site where the house was located may be different from the
general population of children and teenagers in Poland with respect to socioeconomic status, education,
and other risk factors for adverse health outcomes, all of which may affect the level of IEQ of the home.
The relationships presented here cannot be directly linked to other groups of children, and may not be
applicable to other regions due to community differences and the air quality in the area. In Poland,
during winter, suburban and rural areas according to individual coal heating systems tend to be more
polluted than urban areas connected to collective heating systems, which influences the air quality
in the area. Finally, these results apply to residential indoor exposure and may not apply to indoor
exposure experienced in settings such as multi-family buildings. Of two bedroom IEQ measures
during nights, we found that the parent’s approval of children’s and teenager’s privacy by keeping the
doors closed increased airtightness in the bedrooms, favoring low ventilation rates and poor indoor
environment quality (IEQ). Caution should be taken regarding the fact that the measurement results
may not be able to be interpreted as a human exposure level under regularly ventilated conditions
since the data were obtained after the windows were closed. However, it could be useful for predicting
exposure of CO2 in children and teenagers during the night, since it is not only in Poland where
161
Appl. Sci. 2019, 9, 3533
occupants tend to close the windows and doors while sleeping. Examination of these relationships in
other buildings, over a longer period of time, as well as involving wilder groups of participants, could
help to explain what parameters affect indoor environment and sleep quality, which has effects on both
physical and mental health.
Supplementary Materials: The following are available online at http://www.mdpi.com/2076-3417/9/17/3533/s1,

Table S1: General characteristic of studied bedrooms as well as the distribution of nightly average indoor
temperature and humidity in each studied bedroom depending on door opening.
Author Contributions: Conceptualization, A.M.; Methodology, A.M.; Investigation, A.M.; Data curation, A.M.;
Writing—original draft preparation, A.M.; Writing—review and editing, E.Z.-Z.; Visualization, A.M.; Project
administration, E.Z.-Z.
Funding: This research was funded by the Faculty of Energy and Environmental Engineering, Silesian University
of Technology (statutory research).
Acknowledgments: We would like to thank our colleague Walter Mucha for drawing the scheme of the building.
References
1. Landtblom, A.M.; Engström, M. The sleepy teenager-diagnostic challenges. Front. Neurol. 2014, 5, 1–5.
[CrossRef]
2. Herrmann, J.; Koeppen, K.; Kessels, U. Do girls take school too seriously? Investigating gender differences in
school burnout from a self-worth perspective. Learn. Individ. Differ. 2019, 69, 150–161. [CrossRef]
3. Preety, R.; Devi, R.G.; Priya, A.J. Sleep deprivation and cell phone usage among teenagers. Drug Invent.
Today 2018, 10, 2073–2075.
4. Zhang, N.; Cao, B.; Zhu, Y. Indoor environment and sleep quality: A research based on online survey and
field study. Build. Environ. 2018, 137, 198–207. [CrossRef]
5. Rafihi-Ferreira, R.E.; Laura, M.; Pires, N.; Ferreira, E.; Silvares, M. Behavioral intervention for sleep problems
in childhood: A Brazilian randomized controlled trial. Psicol. Reflexão Crítica 2019, 32, 1–13. [CrossRef]
6. Matricciani, L.; Paquet, C.; Galland, B.; Short, M.; Olds, T. Children’s sleep and health: A meta-review. Sleep
Med. Rev. 2019, 46, 136–150. [CrossRef]
7. Hall, R.; Hardin, T.; Ellis, R. School indoor air quality best management practices manual. 2003. Available
online: https://www.doh.wa.gov/portals/1/Documents/Pubs/333-044.pdf (accessed on 26 August 2019).
8. Stamatelopoulou, A.; Saraga, D.; Asimakopoulos, D.; Vasilakos, C.; Maggos, T. The link between residential
air quality and children’s health. Fresenius Environ. Bull. 2017, 26, 162–176.
9. Mainka, A.; Zajusz-Zubek, E. Indoor air quality in urban and rural preschools in Upper Silesia, Poland:
Particulate patter and carbon dioxide. Int. J. Environ. Res. Public Health 2015, 12, 7697–7711. [CrossRef]
10. Mainka, A.; Zubek, E.Z.; Kaczmarek, K. PM10 composition in urban and rural nursery schools in Upper
Silesia, Poland: A trace elements analysis. Int. J. Environ. Pollut. 2017, 61, 98. [CrossRef]
11. Tham, K.W. Indoor air quality and its effects on humans—A review of challenges and developments in the
last 30 years. Energy Build. 2016, 130, 637–650. [CrossRef]
12. Monn, C. Exposure assessment of air pollutants: A review on spatial heterogeneity and
indoor/outdoor/personal exposure to suspended particulate matter, nitrogen dioxide and ozone. Atmos.
Environ. 2001, 35, 1–32. [CrossRef]
13. Pathirana, S.M.; Sheranie, M.D.M.; Halwatura, R.U. Indoor thermal comfort and carbon dioxide concentration:
A comparative study of air conditioned and naturally ventilated houses in Sri Lanka. In Proceedings of the
Moratuwa Engineering Research Conference (MERCon), Moratuwa, Sri Lanka, 29–31 May 2017; pp. 401–406.
14. Khalequzzaman, M.; Kamijima, M.; Sakai, K.; Chowdhury, N.A.; Hamajima, N.; Nakajima, T. Indoor air
pollution and its impact on children under five years old in Bangladesh. Indoor Air 2007, 17, 297–304.
[CrossRef]
15. Santamouris, M.; Synnefa, A.; Asssimakopoulos, M.; Livada, I.; Pavlou, K.; Papaglastra, M.; Gaitani, N.;
Kolokotsa, D.; Assimakopoulos, V. Experimental investigation of the air flow and indoor carbon dioxide
concentration in classrooms with intermittent natural ventilation. Energy Build. 2008, 40, 1833–1843.
[CrossRef]
162
Appl. Sci. 2019, 9, 3533
16. Persily, A.; de Jonge, L. Carbon dioxide generation rates for building occupants. Indoor Air 2017, 27, 868–879.
[CrossRef]
17. Government of Canada: Health Canada Exposure Guidelines for Residential Indoor Air Quality. Available
online: http://publications.gc.ca/collections/Collection/H46-2-90-156E.pdf (accessed on 26 August 2019).
18. Regulation of the Minister of Family, Labor and Social Policy Regarding the Highest Permissible
Concentrations and Intensities of Harmful Factors in the Work Environment. Available online: http:
//www.ilo.org/dyn/natlex/natlex4.detail?p_lang=en&p_isn=99664 (accessed on 26 August 2019).
19. PN-EN13779. Ventilation for Non-Residential Buildings. Performance Requirements for Ventilation and
Room-Conditioning Systems. Available online: http://www.cres.gr/greenbuilding/PDF/prend/set4/WI_25_
Pre-FV_version_prEN_13779_Ventilation_for_non-resitential_buildings.pdf (accessed on 26 August 2019).
20. PN-EN 16798-1: 2019. Indoor environmental input parameters for design and assessment of energy
performance of buildings addressing indoor air quality, thermal environment, lighting and acoustics.
Available online: http://sklep.pkn.pl/pn-en-16798-1-2019-06e.html (accessed on 27 August 2019).
21. CEN/TR 16798-2. Technical Report: Energy Performance of Buildings-Ventilation for Buildings-Part 2:
Interpretation of the Requirements in EN 16798-1-Indoor Environmental Input Parameters for Design and
Assessment of Energy Performance of Buildings Addressing Indoor. Available online: https://epb.center/
documents/centr-16798-2/ (accessed on 26 August 2019).
22. Bekö, G.; Lund, T.; Nors, F.; Toftum, J.; Clausen, G. Ventilation rates in the bedrooms of 500 Danish children.
23. Satish, U.; Mendell, M.J.; Shekhar, K.; Hotchi, T.; Sullivan, D. Is CO2 an indoor pollutant? Direct effects of
low-to-moderate CO2 concentrations on human decision-making performance. Environ. Health Perspect.
2012, 120, 1671–1678. [CrossRef]
24. Norbäck, D.; Nordström, K.; Zhao, Z.; Nordstro, K.; Norba, D. Carbon dioxide (CO2 ) demand-controlled
ventilation in university computer classrooms and possible effects on headache, fatigue and perceived indoor
environment: An intervention study. Int. Arch. Occup. Environ. Health 2013, 86, 199–209. [CrossRef]
25. Bouvier, J.L.; Bontemps, S.; Mora, L. Uncertainty and sensitivity analyses applied to a dynamic simulation of
the carbon dioxide concentration in a detached house. Int. J. Energy Environ. Eng. 2019, 10, 47–65. [CrossRef]
26. Belmonte, J.F.; Barbosa, R.; Almeida, M.G. CO2 concentrations in a multifamily building in Porto, Portugal:
Occupants’ exposure and differential performance of mechanical ventilation control strategies. J. Build. Eng.
2019, 23, 114–126. [CrossRef]
27. Bakó-Biró, Z.; Clements-Croome, D.J.; Kochhar, N.; Awbi, H.B.; Williams, M.J. Ventilation rates in schools
and pupils’ performance. Build. Environ. 2012, 48, 215–223. [CrossRef]
28. Telejko, M.; Zander-Świercz, E. Attempt to improve indoor air quality in kindergartens. Procedia Eng. 2016,
161, 1704–1709. [CrossRef]
29. Zender-Swiercz, E.; Telejko, M. Indoor air quality in kindergartens in Poland. In Proceedings of the IOP
Conference Series: Materials Science and Engineering, 3rd World Multidisciplinary Civil Engineering,
Architecture, Urban Planning Symposium (WMCAUS 2018), Prague, Czech Republic, 18–22 June 2018;
Volume 471.
30. Krawczyk, D.A.; Rodero, A.; Gładyszewska-Fiedoruk, K.; Gajewski, A. CO2 concentration in naturally
ventilated classrooms located in different climates—Measurements and simulations. Energy Build. 2016, 129,
31. Mainka, A.; Zajusz-Zubek, E.; Kozielska, B.; Bragoszewska,
˛ E. Investigation of air pollutants in rural nursery
school–a case study. In Proceedings of the E3S Web of Conferences, X-th Scientific Conference Air Protection
in Theory and Practice, Zakopane, Poland, 18–21 October 2017; Volume 28, pp. 1–8.
32. Karjalainen, S. Thermal comfort and gender: A literature review. Indoor Air 2012, 22, 96–109. [CrossRef]
[PubMed]
33. Seppanen, O.A.; Fisk, W.J.; Mendell, M.J. Association of ventilation rates and CO2 concentrations with health
and other responses. Indoor Air 1999, 9, 226–252. [CrossRef] [PubMed]
34. Mahyuddin, N.; Awbi, H.B. A Review of CO2 measurement procedures in ventilation research. Int. J. Vent.
2012, 10, 353–370. [CrossRef]
35. Eusebio, L.; Derudi, M.; Capelli, L.; Nano, G.; Sironi, S. Assessment of the indoor odour impact in a naturally
ventilated room. Sensors 2017, 17, 778. [CrossRef] [PubMed]
163
Appl. Sci. 2019, 9, 3533
36. Nguyen, J.L.; Schwartz, J.; Dockery, D.W. The relationship between indoor and outdoor temperature,
apparent temperature, relative humidity, and absolute humidity. Indoor Air 2014, 24, 103–112. [CrossRef]
37. WHO. WHO Air Quality Guidelines for Europe; WHO: Copenhagen, Denmark, 2000.
38. Bekö, G.; Toftum, J.; Clausen, G. Modeling ventilation rates in bedrooms based on building characteristics
and occupant behavior. Build. Environ. 2011, 46, 2230–2237. [CrossRef]
39. Stamatelopoulou, A.; Asimakopoulos, D.N.; Maggos, T. Effects of PM, TVOCs and comfort parameters on
indoor air quality of residences with young children. Build. Environ. 2019, 150, 233–244. [CrossRef]
40. D’Ambrosio Alfano, F.R.; Ficco, G.; Palella, B.I.; Riccio, G.; Ranesi, A. An experimental investigation on the
air permeability of passive ventilation grilles. Energy Procedia 2015, 78, 2869–2874. [CrossRef]
164
applied
sciences
Article
Cooking/Window Opening and Associated Increases
of Indoor PM2.5 and NO2 Concentrations of
Children’s Houses in Kaohsiung, Taiwan
Yu-Chuan Yen 1 , Chun-Yuh Yang 1 , Kristina Dawn Mena 2 , Yu-Ting Cheng 1 and
Pei-Shih Chen 1,3,4,5, *
1 Department of Public Health, College of Health Science, Kaohsiung Medical University, Kaohsiung City 807,
Taiwan; [email protected] (Y.-C.Y.); [email protected] (C.-Y.Y.); [email protected] (Y.-T.C.)
2 Epidemiology, Human Genetics, and Environmental Sciences, School of Public Health, University of Texas
Health Science Center at Houston, Houston, TX 77046, USA; [email protected]
3 Institute of Environmental Engineering, College of Engineering, National Sun Yat-Sen University,
Kaohsiung City 807, Taiwan
4 Department of Medical Research, Kaohsiung Medical University Hospital, Kaohsiung City 807, Taiwan
5 Research Center for Environmental Medicine, Kaohsiung Medical University, Kaohsiung City 807, Taiwan
* Correspondence: [email protected]; Tel.: +886-7-3110811
Received: 29 August 2019; Accepted: 2 October 2019; Published: 14 October 2019
Abstract: High concentrations of air pollutants and increased morbidity and mortality rates are
found in industrial areas, especially for the susceptible group, children; however, most studies use
atmospheric dispersion modeling to estimate household air pollutants. Therefore, the aim of this
study was to assess the indoor air quality, e.g., CO, CO2 , NO2 , SO2 , O3 , particulate matter with
aerodynamic diameter less than 2.5 μm (PM2.5 ), and their influence factors in children’s homes in
an industrial city. Children in the “general school”, “traffic school”, and “industrial school” were
randomly and proportionally selected. Air pollutants were sampled for 24 h in the living rooms and
on the balcony of their houses and questionnaires of time–microenvironment–activity-diary were
recorded. The indoor CO concentration of the traffic area was significantly higher than that of the
industrial area and the general area. In regard to the effects of window opening, household NO2 and
PM2.5 concentrations during window opening periods were significantly higher than of the reference
periods. For the influence of cooking, indoor CO2 , NO2 , and PM2.5 levels during the cooking periods
were significantly higher than that of the reference periods. The indoor air quality of children in
industrial cities were affected by residential areas and household activities.
Keywords: indoor air quality; children’s house; industrial city; window opening; cooking
1. Introduction
According to the Environmental White Paper of Taiwan Environmental Protection Agency (Taiwan
EPA), the annual average concentrations of ambient CO, NO2 , SO2 , and O3 in 2008 were 0.47 ppm,
16.90 ppb, 4.35 ppb, and 29.09 ppb, respectively. The Kaohsiung–Pingtung area was the worse
polluted area in Taiwan and accounted for 5.93% of station-days of the Pollutant Standards Index
(PSI) > 100. Especially, Kaohsiung is a heavy industrial city. In industrial areas, high concentrations
of air pollutants and increased morbidity and mortality rates are found, depending on the types
of industrial activities and exposure concentrations in residential areas [1,2]. Children are more
susceptible to the health effects of air pollution than adults due to not having full development of their
pulmonary metabolic capacity [3]. Long-term exposure of air pollution may affect children’s lung
development [4]. Previously, most of the studies revealed that ambient pollution such as particulate
matter with aerodynamic diameter less than 10 and 2.5 μm (PM10 and PM2.5 ), sulfur dioxide (SO2 ),

Appl. Sci. 2019, 9, 4306
nitrogen dioxide (NO2 ), volatile organic compounds (VOCs), etc. in industrial areas may increase
the risk of respiratory symptoms, and attacks of asthma in children [1,5,6]. Therefore, indoor air
quality of children’s homes may be very important to children’s health, especially in industrial cities,
since children spend most of their time at home [7].
Indoor air quality may be affected by indoor human activities such as cooking, smoking, cleaning,
etc. and the infiltration of outdoor pollutants produced from the traffic or industrial sources [8–11].
For example, SO2 , NOx , PM2.5 , and carbon monoxide (CO), the major conventional air pollutant in steel
plants, oil refineries, and vehicular exhaust emissions [12–14], may enter a house through cracks and
windows [15,16]. In addition, if the indoor air is not well ventilated, the air pollutants may accumulate
in the indoor environment, and then seriously affects the health of the inhabitants [17].
Previously, atmospheric dispersion modeling was used to estimate the household concentrations
of indoor air pollutants in industrial areas [1,18,19]. Only a few studies actually measured individual
exposure [20] and household concentrations [21–23], and these studies only focused on PM mass
concentrations, elemental composition, and VOCs concentrations. However, other air pollutants e.g.,
CO, carbon dioxide (CO2 ), NO2 , SO2 , and ozone (O3 ) in households in industrial cities also need to be
considered. Therefore, the main aim of this study was to assess the indoor air quality including CO, CO2 ,
NO2 , SO2 , O3 , and PM2.5 , temperature and relative humidity, and their influence factors (e.g., window
opening and cooking) in children’s homes in an industrial city—Kaohsiung City. To our knowledge,
this is the first study to assess the indoor air quality including CO, CO2 , NO2 , SO2 , and O3 in children’s
homes in an industrial city. In addition, the second aim was to evaluate potential determinants of
indoor air pollutants levels of occupants’ activities, including cooking and window opening, etc. It is
also the first study to reveal the differences of air pollutants between cooking periods/window opening
periods and reference periods through a time–microenvironment–activity-diary via a questionnaire in
one-hour time segments.
2.1. Study Area

Kaohsiung City (22◦ 38 N, 120◦ 17 E), located in southern Taiwan and with the population density
of 9962.6/km2 in 2010, is the largest industrialized harbor city in Taiwan with intense traffic and heavy
industries including the largest steel plant (the China Steel Corporation, which also ranked the 19th
steel mill in the world in 2005), the largest oil refinery (the CPC Corporation), the largest international
shipbuilding (it ranked 6th in the world in 2005) in Taiwan, and many petrochemical industries.
2.2. Study Design

In April 2010, we selected three elementary schools in Kaohsiung City. One elementary school
had a general air quality monitoring station of Taiwan EPA on the roof of the 4th floor, so we called
this school a “general school”. Another elementary school was 0.33 km from Taiwan EPA’s traffic
air quality monitoring station and was regarded as a "traffic school". The “industrial school” was an
elementary school located near the Xiaogang Industrial Zone in Kaohsiung City and about 0.30 km
from Taiwan EPA’s air monitoring station. The study population was limited to children who attended
these schools. The number of students in the “general school”, “traffic school”, and “industrial school”
were 1669, 987, and 960, respectively. After obtaining the assented of the child and the permission of
the parents, we recorded the subjects who agreed to home visits for environmental sampling. Children
were randomly and proportionally selected from each school to participate in this study. Finally,
the home visits of 32, 16, and 12 participants in the “general school”, “traffic school”, and “industrial
school”, respectively, were completed between April 2010 and October 2010.
166
Appl. Sci. 2019, 9, 4306
2.3. Air Sampling

Indoor air pollutants including CO, CO2 , NO2 , SO2 , O3 , PM2.5 , temperature, and relative humidity
were measured by real-time monitoring equipment for 24 h in the living rooms. We also measured
the atmospheric CO, CO2 , NO2 , SO2 , O3 , and PM2.5 on the balcony as outdoor concentrations.
All instruments were placed on the bench at a height of approximately 1 m above the ground. The PM
was measured by a real-time optical scattering instrument (DUSTTRAK™ DRX Aerosol Monitor
Models 8533, TSI Incorporated, Shoreview, MN, USA) and the measurements were taken every 1 s by
the flow rate of 3.0 L/min with detectable concentration from 0.001 to 150 mg/m3 . The CO, CO2 , NO2 ,
SO2 , indoor temperature, and relative humidity were also recorded (KD-airboxx, KD Engineering,
Blaine, WA, USA) every 15 s with the measuring range of 0 to 500 ppm, 0 to 10,000 ppm, 0 to 20 ppm,
0 to 20 ppm, 0 to 50 ◦ C, and 5% to 95%, respectively. The accuracy of CO, CO2 , NO2 , and SO2 were
±3% of reading or 2 ppm (whichever was greater), ±5% of reading or 60 ppm (whichever was greater),
0.25 ppm, and 0.25 ppm, respectively. The resolution of CO, CO2 , NO2 , and SO2 were 0.1 ppm, 1 ppm,
0.01 ppm, and 0.01 ppm, respectively. In terms of O3 , it was detected by a real-time monitoring (Model
202 Ozone monitorTM , 2B Technologies Inc, Boulder, CO, USA) every 5 min with the measuring range
of 0 to 250 ppm.
All real-time monitors were manufacturer-calibrated for the study in the beginning of this study
and every six months. Before every field sampling, the DUSTTRAK™ DRX Aerosol Monitor Models
8533 was calibrated using emery oil aerosol and nominally adjusted to the respirable mass of standard
ISO 12103-1, A1 test dust, (Arizona Dust); and the KD-airboxx, the Model 202 Ozone monitorTM were
calibrated using zero gas and span gas. In addition, the zero calibrators of instruments were carried
out, and the flow rate of sampling pump also was adjusted by Gilian Gilibrator-2NIOSH Primary
Standard Air Flow Calibrator (Sensidyne, St. Petersburg, FL, USA) before every household sampling.
2.4. Household Characteristics

In addition, household characteristics including the number of occupants, air-conditioning
use, smoking, incense burning, etc. were also recorded in the questionnaires. In addition
to household characteristics, data on potential determinants of indoor air pollutants levels of
occupants’ activities, including cooking and window opening, etc. were obtained through a
time–microenvironment–activity-diary via a questionnaire in one-hour time segments. We also
actually evaluated the effects of window opening and cooking on indoor air pollutants. The window
opening periods were defined from a time–microenvironment–activity-diary and two one-hour
periods before and after window opening periods were defined as the reference periods. In terms
of cooking, the cooking periods were the periods recorded by participants as cooking from a
time–microenvironment–activity-diary and the reference periods were defined as the one-hour periods
before the cooking periods.
2.5. Ethics
This study was approved by the Institutional Review Board of the Kaohsiung Medical University
Chung-Ho Memorial Hospital (the protocol number was KMU-IRB-990045). Informed written consent
was obtained from each child (the phonetic version of the consent form that the children read and
signed) and their legal guardians.

Statistical analyses in this study were performed using SAS version 9.3 (SAS Institute of Taiwan
Ltd, Taipei, Taiwan). Descriptive statistics were used to describe the 24-hour of average of exposure data
(indoor/outdoor air pollutant concentrations, temperature, and relative humidity). The concentrations
of air pollutants were not normally distributed (data not shown), therefore we analyzed our data by
nonparametric statistics, also known as distribution-free statistics. A paired Student’s t-test was used to
167
Appl. Sci. 2019, 9, 4306
assess the difference in the average concentration of air pollutants between indoor and outdoor, between
window opening periods and reference periods, and between cooking periods and reference periods.
With the objective of evaluating significant differences among the areas (general, traffic, and industry)
for all air pollutants variables, data were analyzed using one-way analysis of variance (ANOVA) with
Scheffe multiple comparison test. The generalized estimating equations (GEE) is a general statistical
method in a longitudinal study with small samples for adjusting time interference, in which each time
point is an independent event. Finally, the relationships between the 24-hour average concentrations of
indoor air pollutants (dependent variable) and household characteristics (independent variable) were
analyzed using GEE, adjusting for other household characteristics, and time interference. A p-value of
less than 0.05 was considered significant.
3. Results
Table 1 shows the descriptive statistics of 24-h average indoor and outdoor air pollutants,
temperature, and relative humidity in 60 houses. When indoor air pollutants were paired with
outdoors within the same home, we found that the 24-hour average concentrations of indoor CO, CO2 ,
and NO2 were significantly higher than the 24-hour average of outdoors concentrations, whereas,
outdoor O3 and PM2.5 concentrations were significantly higher than indoor concentrations (all p < 0.01).
The average distance between homes of subjects and their school were 0.86 km, 0.94 km, and 1.46 km
in general, traffic, and industrial areas, respectively, as well as, the average distance between homes of
subjects and the nearest air monitoring station were 1.07 km, 0.97 km, and 1.46 km in general, traffic,
and industrial areas, respectively.
Table 1. Descriptive statistics of 24-h average indoor and outdoor air pollutants, temperature,
and relative humidity in 60 houses.
Standard
Mean Median Minimum Maximum p-Value #
Deviation
indoor 3.47 0.83 4.29 0.00 12.27
CO (ppm)
outdoor 0.60 0.38 0.55 0.00 1.98 0.004 ‡
indoor 655.43 479.55 321.60 413.82 1320.00
CO2 (ppm)
outdoor 322.22 319.92 17.23 285.83 353.90 <0.001 ‡
indoor 185.30 177.97 41.52 127.28 251.41
NO2 (ppb)
outdoor 107.54 118.22 36.83 39.90 149.80 0.008 ‡
indoor 0.00 0.00 0.00 0.00 0.00
SO2 (ppm) 0.193
outdoor 0.01 0.00 0.02 0.00 0.06
indoor 11.04 8.50 8.93 1.06 32.29
O3 (ppb)
outdoor 13.46 9.20 12.34 0.24 45.50 0.006 ‡
indoor 60.00 40.00 50.00 10.00 210.00
PM2.5 (μg/m3 )
outdoor 110.00 90.00 90.00 30.00 410.00 0.001 ‡
Temperature (◦ C) indoor 31.00 31.00 1.76 26.00 34.00 -
Relative humidity (%) indoor 72.00 72.00 4.98 62.00 84.00 -
# Paired Student’s t-test, ‡ p < 0.01.
In comparison with household air pollutants of three areas, Table S1 shows descriptive statistics
of 24-h average concentration of indoor air pollutants in the houses of traffic, industry, and general
areas. Figure 1 shows the 24-h average concentration of indoor air pollutants (A) CO, (B) CO2 , (C) NO2 ,
and (D) O3 in the houses of traffic, industry, and general areas. We found the 24-hour average
concentration of indoor CO concentration of the traffic area was significantly higher than that of the
industrial area, and the general area with all p < 0.01 (Figure 1, Table S1). In addition, the 24-hour
average concentration of indoor CO2 level of the general area was significantly lower than that of the
traffic area, and industrial area (all p < 0.01) (Figure 1, Table S1). Finally, both the 24-hour average
concentration of household NO2 and O3 concentrations of the industrial area were significantly lower
than that of the traffic area, and general area (all p < 0.01) (Figure 1, Table S1). Moreover, there was no
168
Appl. Sci. 2019, 9, 4306
statistical significant difference of the 24-hour average concentration of indoor SO2 and PM2.5 between
the three areas (Table S1).
Figure 1. The 24-h average concentration of indoor air pollutants (A) CO, (B) CO2 , (C) NO2 , and (D) O3
in the houses of traffic, industry, and general areas. * Scheffe test p < 0.01.
Table 2 shows the percentage of household characteristics including window opening, residents
>4 people, cooking, etc. in traffic, industry, and general areas. We found compared with traffic area
and industrial area, the general area had a higher percentage of window opening, cooking, and air-
conditioning use; moreover, a lower percentage of residents > 4 people, smoker, incense burning,
mosquito coil burning, and essential oil using.
Table 2. The percentage (%) of household characteristics in traffic, industry, and general areas.
Area
Traffic Industry General
Window opening 68.75 66.67 87.50
Occupants (>4 people) 40.40 57.01 34.41
Cooking 87.50 83.33 95.83
Air-conditioning use 62.50 83.33 95.83
Making tea 31.25 30.00 0
Smoker 63.64 40.00 26.09
Incense burning 72.73 50.00 29.17
Mosquito coil burning 37.50 22.22 12.50
Essential oil using 31.25 33.33 25.00
The following Table 3 shows the ratios of air pollutants during window opening periods to the
reference periods and the differences in air pollutants between window opening periods and reference
periods. The median ratios of pollutants during window opening periods to the reference periods
for NO2 and PM2.5 were 1.56 and 1.13, respectively with the maximum values up to 5.23 and 1.85
respectively (Table 3). The NO2 and PM2.5 levels during window opening periods were significantly
higher than that of the reference periods, and the maximum increased values were 53.25 ppb and
44 μg/m3 , respectively. Table 4 shows the ratios of air pollutants during cooking periods to reference
169
Appl. Sci. 2019, 9, 4306
periods and the differences in air pollutants between window opening periods and reference periods.
The median ratios of pollutants during cooking periods to the reference periods for CO, CO2 , NO2 ,
and PM2.5 were 0.93, 1.06, 1.11, and 1.09, respectively. The concentrations of CO2 , NO2 , and PM2.5
during the cooking periods were significantly higher than those of reference periods with increased
concentrations of 26.17 ppm, 5.40 ppb, and 5 μg/m3 , respectively. However, the CO level during cooking
periods was significantly lower than that of the reference periods with the decreased concentration of
0.25 ppm.
Table 3. The ratios of air pollutants during window opening periods to reference periods and the
differences in air pollutants between window opening periods and reference periods.
Ratios (Window Opening Differences (Window Opening Periods

Periods/Reference Periods § ) − Reference Periods § ) p-Value #
Median S.D. Min. Max. Median S.D. Min. Max.
CO (ppm) 0.98 1.34 0.57 4.44 0.00 1.31 −2.42 3.67 0.53
CO2 (ppm) 1.05 0.18 0.73 1.43 29 128 −141 296 0.21
NO2 (ppb) 1.56 1.30 0.94 5.23 18.71 16.05 −9.40 53.25 <0.01 ‡
SO2 (ppm) 0.00 0.92 0.00 3.27 0.00 0.02 0.00 0.08 0.21
O3 (ppb) 1.18 0.59 0.56 2.19 0.81 4.44 −11.91 10.05 0.52
PM2.5 (μg/m3 ) 1.13 0.31 0.69 1.85 7 16.20 −6 44 0.04 †
# Paired Student’s t-test, † p < 0.05, ‡ p < 0.01. § Reference periods were two one-hour periods before and after
window opening periods.
Table 4. The ratios of air pollutants during cooking periods to reference periods and the differences in
air pollutants between during cooking periods and reference periods.
Ratios (Cooking Periods/Reference Differences (Cooking Periods −

Periods § ) Reference Periods § ) p-Value #
Median S.D. Min. Max. Median S.D. Min. Max.
CO (ppm) 0.93 0.22 0.46 1.51 −0.25 0.84 −3.53 0.61 <0.01 ‡
CO2 (ppm) 1.06 0.14 0.85 1.58 26.17 90.21 −111.67 342.5 <0.01 ‡
NO2 (ppb) 1.11 0.98 0.51 5.43 5.40 29.71 −71.17 101.75 <0.01 ‡
O3 (ppb) 1.08 0.69 0.46 4.36 0.27 8.89 −35.14 17.08 0.94
PM2.5 (μg/m3 ) 1.09 0.30 0.60 2.56 5 14 −45 56 0.04 †
# Paired Student’s t-test, † p < 0.05, ‡ p < 0.01. § Reference periods were the one-hour period before cooking periods.
Table 5 shows the association between air pollutants concentrations (24-h average concentration
of air pollutants in each house as dependent variable), and household characteristics by using the
generalized estimating equations model. This study revealed that CO concentrations were positively
associated with the number of occupants, cleaning, smoking, incense burning, mosquito coil burning,
and negatively correlated to cooking with a statistical significance. Indoor CO2 concentrations
were positively associated with the number of occupants, air-conditioning use, smoking, incense
burning, and negatively correlated to mosquito coil burning with a statistical significance. In addition,
significantly higher NO2 levels were found in the homes with smokers than homes without smokers.
There were significantly positive associations between indoor SO2 concentrations and smoking and
incense burning. In terms of O3 , indoor O3 concentrations were positively associated with the window
opening and negatively correlated to the number of occupants, incense burning, and essential oil use
with a statistical significance. For PM2.5 , it was positively associated with cleaning and incense burning
with a statistical significance.
170
Appl. Sci. 2019, 9, 4306
Table 5. Association between air pollutants concentrations (24-h average concentration of air pollutants
in each house as dependent variable), and household characteristics: generalized estimating equations.
CO (ppm) CO2 (ppm) NO2 (ppb) SO2 (ppm) O3 (ppb) PM2.5 (μg/m3 )
Window opening
0.32 84.84 −0.61 0.44 24.34 ‡ −0.021
(Yes vs. No)
Occupants 0.52 ‡ 51.62 ‡ 3.02 −0.008 −3.49 ‡ 0.004
Cleaning
4.73 † −317.49 1.39 0.43 −6.24 0.047 †
(Yes vs. No)
Cooking
−3.89 † 228.02 −28.01 −0.21 1.79 0.065
(Yes vs. No)
Fan using
1.42 −32.97 10.58 −0.0003 −2.07 0.002
(Yes vs. No)
Air- conditioning use ‡
−1.22 246.99 87.87 0.25 21.59 0.008
(Yes vs. No)
Making tea
37.04 - −0.45 −0.13 14.21 −0.050
(Yes vs. No)
Smoking † ‡ ‡ ‡
17.21 1988.44 547.36 2.98 1.69 0.173
(Yes vs. No)
Incense burning † † ‡ ‡
18.21 2927.87 193.11 3.66 −108.9 0.416 ‡
(Yes vs. No)
Mosquito coil burning
41.55 ‡ −892.64 † 673.52 2.67 2.29 -
(Yes vs. No)
Essential oil use ‡
12.76 269.25 74.89 −0.66 −89.29 −0.022
(Yes vs. No)
† ‡
Generalized estimating equations (GEE) p < 0.05, p < 0.01.
4. Discussion
Our results showed that the outdoor concentrations of O3 and PM2.5 were significantly higher
than indoor concentrations. The Kaohsiung City is a city with intense traffic and heavy industries,
and previous studies believed SO2 , NOx , PM2.5 , and CO were the major conventional air pollutant in
steel plants, oil refineries, and vehicular exhaust emissions [12–14,24]. In addition, outdoor O3 might
be formed by the photochemical reaction of nitrogen oxides absorbing sunlight, and VOCs [25,26].
According to the PSI database from 2010 to 2012 of Taiwan EPA, only O3 and total suspended particulate
(TSP) would exceed the standard [27]. This may be the reason why outdoor PM2.5 and O3 concentrations
were higher than indoor concentrations. In our study, outdoor median PM2.5 levels (90 μg/m3 ) were
higher than both the National Ambient Air Quality Standards of UAS and Taiwan EPA with the
24-hour standard for PM2.5 of 35 μg/m3 . In addition, the median value of indoor PM2.5 concentrations
(40 μg/m3 ) was also higher than the criteria of indoor air quality (IAQ) standards of Taiwan EPA
(35 μg/m3 /24 h). In our study, indoor CO, CO2 , and NO2 levels were significantly higher than outdoor
levels. The number of occupants and human activities such as cooking, smoking, etc. might be the
factors affecting indoor pollutants whereas liquefied petroleum gas (LPG), not electric stoves, was the
main cooking way in Kaohsiung City [28,29]. In addition, most of the houses were just by the roads
and very close to the mobile sources in Kaohsiung City, which was thought of as a traffic-intensive
city with the number of cars and motorcycles of approximately 430,000 and 1,230,000, respectively,
in 2010 [30]. Thus, the main combustion products of vehicular engines such as CO, NOX , etc. entering
the houses through cracks and windows might be the reason why the indoor concentrations of CO,
CO2 , and NO2 were higher than outdoor concentrations [15,16].
In comparison with traffic, industrial, and general areas, the highest household CO concentration
was found in the traffic area among the three areas. According to the previous study, the greatest
source of CO (more than 90%) in cities was motor vehicles [24]. The high traffic flow in the traffic
area might be the reason for the observation. For CO2 , our study indicated that the lowest household
CO2 level was in the general area among the three areas. The main source of CO2 was from human
respiration [24,31]. The number of residents might be one possible reason since the number of residents
> 4 people in the traffic area, the industrial area, and general area were 40.40%, 57.01%, and 34.41%,
respectively. We also found both household NO2 and O3 concentrations of the industrial area were
171
Appl. Sci. 2019, 9, 4306
lowest among the three areas, which was not consistent with the observations of previous studies
that ambient NO2 was related to industrial activities [24], and outdoor O3 might be formed by the
photochemical reaction of nitrogen oxides absorbing sunlight, and VOCs [25,26]. We believed these
may be related to Taiwan EPA’ s policies and efforts to control air pollution from stationary sources
after that the “Stationary Pollution Source Air Pollutant Emissions Standards” was passed in 1992,
and the “Air Pollution Control Act Enforcement Rules” was also implemented in 2003.
In regard to the effects on the window opening, our study displayed that household NO2 and
PM2.5 concentrations during window opening periods were significantly higher than that during
reference periods. NOX and PM were related to traffic emissions [24,32], and most of the houses in
Taiwan were adjacent to roads, so window opening might increase indoor NO2 and PM2.5 . For the
influence of cooking, there were many simulated experiments exploring the air pollutant emissions of
cooking-related fuel combustion [29,33–36], and they demonstrated that CO, CO2 , NOX , and PM2.5
would be emitted by the process of the experiments. Although CO also was produced by cooking,
it was revealed that combustion of high-grade fuels (such as natural gas, and LPG which contained
propane, butane, etc.), the main fuel-burning stoves use in Taiwan households usually produce much
less CO than combustion of low-grade fuels [29,33]. In the previous study, Delp et al. revealed the
residential cooking exhaust hoods could not completely capture the pollutants and their efficiency
was highly variable [37]. Our results showed that indoor CO2 , NO2 , and PM2.5 levels during cooking
periods were significantly higher than during reference periods, but the indoor CO level during cooking
periods was lower than during reference periods, possibly indicating that the emission rate of CO2 ,
NO2 , and PM2.5 might be higher than the capture rate of the exhaust hood and the emission rate of CO
might be lower than the pollutants capture rate of the exhaust hood.
In terms of influence factors, we found there were significantly positive correlations between the
number of occupants and CO and CO2 concentrations. Our study was consistent with the observations of
the previous study that CO2 was produced by human respiration [24,31]. In addition to the combustion,
the indoor CO also was related to the status of residents; the previous studies revealed either a smoking
person or person with inflammatory diseases exhaled higher CO levels than control group [38,39].
We also found smoking was significantly positively associated with household CO in our study.
According to previous studies, smoking, incense burning, and mosquito coil burning were significantly
positively associated with CO, CO2 , SO2 , NOX , and PM [40–42], and these results were consistent with
our observation. The cleaning behavior would increase indoor PM2.5 and CO levels; it was consistent
with the previous study that indoor PM2.5 and PM5 levels could be elevated by the cleaning behavior of
dry dust, and vacuuming [43]. In addition, commercial cleansers and disinfectants contain VOCs [44],
and El Fadel et al. found VOCs concentration was positively correlated with CO concentration [45].
We also revealed that air-conditioning use was positively associated with indoor CO2 concentrations
with a statistical significance, which was consistent with a previous observation that CO2 levels were
higher in mechanically ventilated buildings than in naturally ventilated buildings [46]. There was a
significantly negative association between essential oil use and O3 concentration. The commercially
available essential oils contain many VOCs (e.g., D-limonene, α- pinene, etc.) [47], in addition, a study
displayed that indoor VOCs level had increased significantly after burning essential oils [48]. O3 was
one of the indoor oxidants [49,50], and Waring et al. demonstrated that 68% of all O3 reactions were
with D-limonene, and 26% of all O3 reactions are with α-pinene [50]. This might be the reason why
the essential oil use could decrease the O3 level. Finally, by questionnaire, it was found that window
opening was significantly correlated with increased O3 concentration, which was not consistent with
the results from the time–microenvironment–activity-diary that only NO2 and PM2.5 levels during the
window opening periods were significantly higher than that of reference periods. We believed O3 was
a major component of photochemical pollution, so it is more relevant to outdoor sources than indoor
sources. Thus, compared with the households which closed the windows, the households which
opened the windows had a significantly higher 24-hour average concentration of O3 . When comparing
the window opening periods with the reference periods (two one-hour periods before and after window
172
Appl. Sci. 2019, 9, 4306
opening periods), there was no significant variation in atmospheric O3 concentration in a short time
(within three hours). For PM2.5 and NOX levels, there was no significant difference between households
which closed and opened the windows, the possible reason might be that PM2.5 and NOX could come
from both indoor (cooking) and outdoor (traffic) sources.
5. Conclusions
This study explored the concentration of indoor air pollutants in different areas including traffic,
industrial, and general areas within an industrial city. Moreover, this study also revealed household
NO2 and PM2.5 concentrations during window opening periods were significantly higher than that of
the reference periods with increased concentrations of 18.71 ppb, and 7 μg/m3 , respectively. For the
influence of cooking, indoor CO2 , NO2 , and PM2.5 levels during the cooking periods were significantly
higher than that of the reference periods with increased concentrations of 26.17 ppm, 5.40 ppb,
and 5 μg/m3 , respectively.
Supplementary Materials: The following are available online at http://www.mdpi.com/2076-3417/9/20/4306/s1,

Table S1: Descriptive statistics of 24-h average concentration of indoor air pollutants in the houses of traffic,
industry, and general areas.
Author Contributions: Conceptualization, C.-Y.Y.; formal analysis, Y.-C.Y.; writing—original draft preparation,
Y.-T.C. and Y.-T.C.; writing—review and editing, P.-S.C. and K.D.M.
Funding: This research was funded by “Wang Jhan-Yang Public Trust Fund”, grant number 108-002-6.
References
1. Idavain, J.; Julge, K.; Rebane, T.; Lang, A.; Orru, H. Respiratory symptoms, asthma and levels of fractional
exhaled nitric oxide in schoolchildren in the industrial areas of Estonia. Sci. Total. Environ. 2019, 650, 65–72.
[CrossRef] [PubMed]
2. Martuzzi, M.; Pasetto, R.; Martin-Olmedo, P. Industrially Contaminated Sites and Health. J. Environ. Public
Health 2014, 2014, 198574. [CrossRef] [PubMed]
3. Kurt, O.K.; Zhang, J.; Pinkerton, K.E. Pulmonary health effects of air pollution. Curr. Opin. Pulm. Med. 2016,
4. Rice, M.B.; Rifas-Shiman, S.L.; Litonjua, A.A.; Oken, E.; Gillman, M.W.; Kloog, I.; Luttmann-Gibson, H.;
Zanobetti, A.; Coull, B.A.; Schwartz, J.; et al. Lifetime Exposure to Ambient Pollution and Lung Function in
Children. Am. J. Respir. Crit. Care Med. 2016, 193, 881–888. [CrossRef] [PubMed]
5. Deng, Q.; Lu, C.; Norbäck, D.; Bornehag, C.G.; Zhang, Y.; Liu, W.; Yuan, H.; Sundell, J. Early life exposure to
ambient air pollution and childhood asthma in China. Environ. Res. 2015, 143, 83–92. [CrossRef] [PubMed]
6. Marchetti, P.; Marcon, A.; Pesce, G.; Paolo, G.; Guarda, L.; Pironi, V.; Fracasso, M.E.; Ricci, P.; de Marco, R.
Children living near chipboard and wood industries are at an increased risk of hospitalization for respiratory
diseases: A prospective study. Int. J. Hyg. Environ. Health 2014, 217, 95–101. [CrossRef] [PubMed]
7. Silvers, A.; Florence, B.T.; Rourke, D.L.; Lorimor, R.J. How Children Spend Their Time: A Sample Survey for
Use in Exposure and Risk Assessments. Risk Anal. 1994, 14, 931–944. [CrossRef]
8. Cannistraro, G.; Cannistraro, A.; Cannistraro, M.; Galvagno, A. Analysis of air pollution in the urban center
of four cities sicilian. Int. J. Heat Technol. 2016, 34, S219–S225.
9. Cannistraro, M.; Cannistraro, G.; Chao, J.; Ponterio, L. New Technique Monitoring and Transmission
Environmental Data with Mobile. Instrum. Meas. Metrol. 2018, 18, 549–562. [CrossRef]
10. Cannistraro, M.; Chao, J.; Ponterio, L. Experimental Study of Air Pollution in the Urban Centre of the City of
Messina. Model. Meas. Control C 2018, 79, 133–139. [CrossRef]
11. Leung, D.Y.C. Outdoor-indoor air pollution in urban environment: Challenges and opportunity. Front.
Environ. Sci. 2015. [CrossRef]
12. Brand, A.; McLean, K.E.; Henderson, S.B.; Fournier, M.; Liu, L.; Kosatsky, T.; Smargiassi, A. Respiratory
hospital admissions in young children living near metal smelters, pulp mills and oil refineries in two
Canadian provinces. Environ. Int. 2016, 94, 24–32. [CrossRef] [PubMed]
173
Appl. Sci. 2019, 9, 4306
13. Tian, L.; Zeng, Q.; Dong, W.; Guo, Q.; Wu, Z.; Pan, X.; Li, G.; Liu, Y. Addressing the source contribution of
PM2.5on mortality: An evaluation study of its impacts on excess mortality in China. Environ. Res. Lett. 2017,
12, 104016. [CrossRef]
14. Topacoglu, H.; Katsakoglou, S.; Ipekci, A. Effect of exhaust emissions on carbon monoxide levels in employees
working at indoor car wash facilities. Hippokratia 2014, 18, 37–39. [PubMed]
15. Srithawirat, T.; Latif, M.T.; Sulaiman, F.R. Indoor PM10 and Its Heavy Metal Composition at a Roadside Residential
Environment; Atmosfera: Phitsanulok, Thailand, 2016; Volume 29, pp. 311–322.
16. United States Environmental Protection Agency. Available online: https://www.epa.gov/report-environment/
indoor-air-quality (accessed on 20 August 2019).
17. World Health Organization. Household (Indoor) Air Pollution. 2014. Available online: https://www.who.int/
indoorair/en/ (accessed on 20 August 2019).
18. Han, I.; Guo, Y.; Afshar, M.; Symanski, E.; Stock, T.H. Comparison of trace elements in size-fractionated
particles in two communities with contrasting socioeconomic status in Houston, TX. Environ. Monit. Assess.
2017, 189, 67. [CrossRef] [PubMed]
19. Hart, J.E.; Garshick, E.; Dockery, D.W.; Smith, T.J.; Ryan, L.; Laden, F. Long-term ambient multipollutant
exposures and mortality. Am. J. Respir. Crit. Care Med. 2011, 183, 73–78. [CrossRef] [PubMed]
20. Milà, C.; Salmon, M.; Sanchez, M.; Ambrós, A.; Bhogadi, S.; Sreekanth, V.; Nieuwenhuijsen, M.; Kinra, S.;
Marshall, J.D.; Tonne, C. When, Where, and What? Characterizing Personal PM2.5 Exposure in Periurban
India by Integrating GPS, Wearable Camera, and Ambient and Personal Monitoring Data. Environ. Sci.
Technol. 2018, 52, 13481–13490. [CrossRef]
21. Pekey, B.; Bozkurt, Z.B.; Pekey, H.; Doğan, G.; Zararsız, A.; Efe, N.; Tuncel, G.; Zararsiz, A. Indoor/outdoor
concentrations and elemental composition of PM10/PM2.5 in urban/industrial areas of Kocaeli City, Turkey.
Indoor Air 2010, 20, 112–125. [CrossRef]
22. Pekey, H.; Arslanba, D. The Relationship Between Indoor, Outdoor and Personal VOC Concentrations in
Homes, Offices and Schools in the Metropolitan Region of Kocaeli, Turkey. Water Air Soil Pollut. 2008, 191,
23. Liu, W.; Zhang, J.; Zhang, L.; Turpin, B.; Weisel, C.; Morandi, M.; Stock, T.; Colome, S.; Korn, L. Estimating
contributions of indoor and outdoor sources to indoor carbonyl concentrations in three urban areas of the
United States. Atmos. Environ. 2006, 40, 2202–2214. [CrossRef]
24. Najjar, Y. Gaseous Pollutants Formation and Their Harmful Effects on Health and Environment. Innov.
Energy Policies 2011, 1. [CrossRef]
25. Du, X.; Liu, J. Relationship between outdoor and indoor ozone pollution concentration. Trans. Tianjin Univ.
2009, 15, 330–335. [CrossRef]
26. Jianhui, B.; Singh, N.; Chauhan, S.; Singh, K.; Saud, T.; Saxena, M.; Soni, D.; Mandal, T.K.; Bassin, J.K.;
Gupta, P.K. Study on surface O3 chemistry and photochemistry by UV energy conservation. Atmos. Pollut.
Res. 2010, 1, 118–127. [CrossRef]
27. Taiwan Environmental Protection Agency. Available online: https://taqm.epa.gov.tw/taqm/tw/
PSIOver100MonthlyReport.aspx (accessed on 20 August 2019).
28. Persily, A.; De Jonge, L. Carbon dioxide generation rates for building occupants. Indoor Air 2017, 27, 868–879.
[CrossRef] [PubMed]
29. Penney, D.; Benignus, V.; Kephalopoulos, S.; Kotzias, D.; Kleinman, M.; Verrier, A. WHO Guidelines for Indoor
Air Quality: Selected Pollutants; WHO Regional Office for Europe: Denmark, 2010.
30. Taiwan Environmental Protection Agency. Available online: https://erdb.epa.gov.tw/DataRepository/
Statistics/StatSceAreapop.aspx (accessed on 20 August 2019).
31. Satish, U.; Mendell, M.J.; Shekhar, K.; Hotchi, T.; Sullivan, D.; Streufert, S.; Fisk, W.J. Is CO2 an Indoor
Pollutant? Direct Effects of Low-to-Moderate CO2 Concentrations on Human Decision-Making Performance.
Environ. Health Perspect. 2012, 120, 1671–1677. [CrossRef]
32. Chao, H.-R.; Hsu, J.-W.; Ku, H.-Y.; Wang, S.-L.; Huang, H.-B.; Liou, S.-H.; Tsou, T.-C. Inflammatory Response
and PM2.5 Exposure of Urban Traffic Conductors. Aerosol. Air Qual. Res. 2018, 18, 2633–2642. [CrossRef]
33. Bilsback, K.R.; Dahlke, J.; Fedak, K.M.; Good, N.; Hecobian, A.; Herckes, P.; L’Orange, C.; Mehaffy, J.;
Sullivan, A.; Tryner, J.; et al. A Laboratory Assessment of 120 Air Pollutant Emissions from Biomass and
Fossil Fuel Cookstoves. Environ. Sci. Technol. 2019, 53, 7114–7125. [CrossRef]
174
Appl. Sci. 2019, 9, 4306
34. Shen, G.; Hays, M.D.; Smith, K.R.; Williams, C.; Faircloth, J.W.; Jetter, J.J. Evaluating the Performance of
Household Liquefied Petroleum Gas Cookstoves. Environ. Sci. Technol. 2018, 52, 904–915. [CrossRef]
35. Shen, G.; Gaddam, C.K.; Ebersviller, S.M.; Wal, R.L.V.; Williams, C.; Faircloth, J.W.; Jetter, J.J.; Hays, M.D.
A Laboratory Comparison of Emission Factors, Number Size Distributions, and Morphology of Ultrafine
Particles from 11 Different Household Cookstove-Fuel Systems. Environ. Sci. Technol. 2017, 51, 6522–6532.
[CrossRef]
36. Smith, K.R.; Uma, R.; Kishore, V.V.N.; Zhang, J.; Joshi, V.; Khalil, M.A.K. Greenhouse Implications of
Household Stoves: An Analysis for India. Annu. Rev. Energy Environ. 2000, 25, 741–763. [CrossRef]
37. Delp, W.W.; Singer, B.C. Performance Assessment of U.S. Residential Cooking Exhaust Hoods. Environ. Sci.
Technol. 2012, 46, 6167–6173. [CrossRef] [PubMed]
38. Deveci, S.E.; Deveci, F.; Açik, Y.; Ozan, A.T. The measurement of exhaled carbon monoxide in healthy
smokers and non-smokers. Respir Med. 2004, 98, 551–556. [CrossRef] [PubMed]
39. Ryter, S.W.; Choi, A.M. Carbon monoxide in exhaled breath testing and therapeutics. J. Breath Res. 2013, 7,
017111. [CrossRef] [PubMed]
40. Alberts, W.M. Indoor air pollution: NO, NO2 , CO, and CO2 . J. Allergy Clin. Immunol. 1994, 94, 289–295.
[CrossRef] [PubMed]
41. Lee, S.C.; Guo, H.; Kwok, N.H. Emissions of air pollutants from burning of incense by using large
environmental chamber. In Proceedings of the Indoor Air 2002: 9th International Conference on Indoor Air
Quality and Climate, Monterey, CA, USA, 30 June–5 July 2002; Volume 1.
42. Lee, S.C.; Wang, B. Characteristics of emissions of air pollutants from mosquito coils and candles burning in
a large environmental chamber. Atmos. Environ. 2006, 40, 2128–2138. [CrossRef]
43. Ferro, A.R.; Kopperud, R.J.; Hildemann, L.M. Source Strengths for Indoor Human Activities that Resuspend
Particulate Matter. Environ. Sci. Technol. 2004, 38, 1759–1764. [CrossRef]
44. United States Environmental Protection Agency. Available online: https://www.epa.gov/indoor-air-quality-
iaq/volatile-organic-compounds-impact-indoor-air-quality (accessed on 20 August 2019).
45. El Fadel, M.; Alameddine, I.; Kazopoulo, M.; Hamdan, M.; Nasrallah, R. Carbon Monoxide and Volatile
Organic Compounds as Indicators of Indoor Air Quality in Underground Parking Facilities. Indoor Built
46. Sribanurekha, V.; Wijerathne, S.N.; Wijepala, L.H.S.; Jayasinghe, C. Effect of Different Ventilation Conditions
on Indoor CO2 Levels. In Proceedings of the International Conference on Disaster Resilience, At Kandalama,
Sri Lanaka, 19–21 July 2011.
47. Nematollahi, N.; Kolev, S.D.; Steinemann, A. Volatile chemical emissions from essential oils. Air Qual. Atmos.
Heal. 2018, 11, 949–954. [CrossRef]
48. Chao, C.J.; Wu, P.C.; Chang, H.Y.; Su, H.J. The effects of evaporating essential oils on indoor air quality.
In Proceedings of the Indoor Air 2005, 10th International Conference on Indoor Air Quality and Climate,
Beijing, China, 4–9 September 2005.
49. Young, C.J.; Zhou, S.; Siegel, J.A.; Kahan, T.F. Illuminating the Dark Side of Indoor Oxidants. Environ. Sci.
Process. Impacts 2019, 21, 1229–1239. [CrossRef]
50. Waring, M.S.; Wells, J.R. Volatile organic compound conversion by ozone, hydroxyl radicals, and nitrate
radicals in residential indoor air: Magnitudes and impacts of oxidant sources. Atmos. Environ. 2015, 106,
175
applied
sciences
Article
Evaluation of Performance of Inexpensive Laser
Based PM2.5 Sensor Monitors for Typical Indoor and
Outdoor Hotspots of South Korea
Sungroul Kim *, Sujung Park and Jeongeun Lee
Department of Environmental Health Sciences, Soonchunhyang University, Asan 31538, Korea;
[email protected] (S.P.); [email protected] (J.L.)
Received: 27 March 2019; Accepted: 25 April 2019; Published: 12 May 2019
Featured Application: In consideration of relatively stable outcomes with the application of a

correction factor for relative humidity, recently introduced inexpensive real-time monitors (IRMs),
ESCORTAIR (ESCORT, Seoul, Korea) or PurpleAir (PA) (PurpleAir U.S.A.), our study supports
their usage in PM2.5 monitoring for various urban hotspots.
Abstract: Inexpensive (<$300) real-time particulate matter monitors (IRMs), using laser as a light
source, have been introduced for use with a Wi-Fi function enabling networking with a smartphone.
However, the information of measurement error of these inexpensive but convenient IRMs are still
limited. Using ESCORTAIR (ESCORT, Seoul, Korea) and PurpleAir (PA) (PurpleAir U.S.A.), we
evaluated the performance of these two devices compared with the U.S. Environmental Protection
Agency (EPA) Federal Equivalent Monitoring (FEM) devices, that is, GRIMM180 (GRIMM Aerosol,
Germany) for the indoor measurement of pork panfrying or secondhand tobacco smoking (SHS)
and Beta-ray attenuation monitor (BAM) (MetOne, Grants Pass, OR) for outdoor measurement at
the national particulate matter (PM2.5 ) monitoring site near an urban traffic hotspot in Daejeon,
South Korea, respectively. The PM2.5 concentrations measured by ESCORTAIR and PA were strongly
correlated to FEM (r = 0.97 and 0.97 from indoor pan frying; 0.92 and 0.86 from indoor SHS; 0.85
and 0.88 from outdoor urban traffic hotspot). The two IRMs showed that PM2.5 mass concentrations
were increased with increased outdoor relative humidity (RH) levels. However, after applying
correction factors for RH, the Median (Interquartile range) of difference compared to FEM was (14.5
(6.1~23.5) %) for PA and 16.3 (8.5–28.0) % for ESCORTAIR, supporting their usage in the home or
near urban hotspots.
Keywords: PM2.5 ; sensor; correction; pan frying; secondhand smoke; urban traffic
1. Introduction
A large volume of previous epidemiological studies relied on the use of ground-based fixed
national monitoring stations [1,2]. However, recently, inexpensive (<$300) particulate matter (PM)
monitors (PM) have been introduced for home usage in South Korea. These devices can provide PM
distribution patterns at high temporal and spatial resolution [3–5] which is a substantial improvement
on establishing a pollution monitoring networking system as well as environmental epidemiologic
study [6], as compared to traditional approaches that relied on relatively small number of ground-based
fixed national air monitoring stations or mobile sampling techniques.
Most of these low-cost devices are classified into two groups, that is, optical particle counters (OPCs)
or photometers. OPCs use the light scattered from individual particle to estimate the concentration
of particles in different size ranges [7,8]. These data, along with assumptions of the particle shape

Appl. Sci. 2019, 9, 1947
and density, can be converted to estimate mass concentrations that compare favorably with reference
instruments [7,8]. However, it has been reported that there may be bias when aerosol type or size is
unknown [7,8].
Photometers use a light source to illuminate sensing zones that contain many particles at one
time [9,10] and obtained that the mass concentration of aerosol scales linearly with the amount of light
scattered by an assembly of particles captured at a discrete angle from the incident light [11]. The
light scattered by the assembly of particles is measured by a photodetector at an angle specific to the
photometer model, often 90◦ from the incident light [12]. The intensity of scattered light is directly
proportional to gravimetrically measured mass concentration, although the relationship is dependent
on the light scattering characteristics, density and size distribution of the particle [12].
The cost of research grade light scattering instruments (approximately, $10,000 or higher) limits
their use to conduct studies at high temporal and/or spatial resolution. In Korea, a laser-based
inexpensive OPC based IRM, for example, ESCORTAIR, was recently introduced. However, the
reliability of ESCORTAIR have been unknown. Therefore, it may be necessary to test this new device
before it is applied in a high spatial-temporal resolution exposure assessment study. As a proper
calibration protocol providing a correction factor can have a dramatic impact on precision, accuracy
and bias of a real-time monitor, researchers evaluated the OPC or photometers for use in the laboratory,
outdoors or in the home in other countries [3,9,13–16]. A recent article reported that “one size fits all”
approach to obtain PM2.5 mass concentrations by OPC result in relatively high uncertainty in complex
exposure situations. Although OPC Therefore, corresponding conversion curve approach may be most
valuable when a relatively high contrast is expected in exposure levels for example, daytime home
with indoor combustion sources, BBQ or secondhand smoke versus night time or day time outside
with heavy traffic volume versus night time [16]. To our knowledge, no one has rigorously evaluated
the performance of IRMs, operated with OPC or photometer, in Korea with comparison of the U.S.
Federal Equivalent Method (US FEM) [17].
In this study, we evaluated the performance of inexpensive (less than $300) real-time PM monitors
(IRMs), with high cost (about $2000–$10,000) and cross-comparisons between them and research grade
PM monitors (RGMs). We used US FEMs as reference instruments (approximately $20,000 or higher)
and provided a final error of mass concentration (PM2.5 ) measurement after applying correction factors
in this study.
2.1. PM2.5 Real-Time Monitors

A laser-based light-scattering PM2.5 sensor monitor (ESCORTAIR, ESCORT, Seoul, Korea) (weight
<300 g, volume <510 cm3 ) consisted of an optical particle counting (OPC) PM sensor (INNOSIPLE1),
CO2 sensor, temperature relative humidity sensor, data transfer networking module and light-emitting
diode (LED) display screen (Figure 1). In the ESCORTAIR, the sensing volume is illuminated with
a laser and airborne particles are counted and processed one at a time. There were various IRMs
commercially available in South Korea. In this study, however, we chose ESCORTAIR as they allowed
us to directly transfer data to our data server using its Wi-Fi function.
For comparison purposes, another inexpensive photometer type of PM monitor (PA, PurpleAir,
Draper, UT, USA) (https://www.purpleair.com/), mounting two Plantower sensors in a monitor, was
used. PA is recommended its usage by AQ-SPEC (Air quality sensor performance evaluation center,
South Coast Air Quality Management District, CA, USA) or US EPA (Environmental Protection Agency,
NC, USA) as an IRM. The inlet system of these IRMs did not have an impactor or a cyclone unlike that
of the RGM.
The performance of the two IRMs (one OPC, that is, ESCORTAIR and one photometer, that is, PA)
costed less than $300 were simultaneously compared with those of high-cost devices ($10,000 or so),
178
Appl. Sci. 2019, 9, 1947
that is, research-grade laser photometers including PDR-1500 (Thermo Scientific, Waltham, MA, USA)
and SIDEPAK AM510 (TSI, Inc., Shoreview, MN, USA).
These research-grade monitors have a cyclone inlet or an impact inlet for the measurement of the
respirable fraction of airborne particulate matters in different environments and can provide real-time
data. The SIDEPAK is a portable battery-operated personal aerosol monitor with an impact inlet and
light-scattering laser photometer that provides real-time aerosol mass concentration. The PDR-1500
(Personal DataRAM 1500) is a nephelometric monitor with a cyclone inlet for the measurement of
the respirable fraction of the airborne particulate matters. The PDR-1500 can simultaneously collect
particles on a 37 mm filter for the gravimetric analysis by passing through the sensing zone.
(a)
(b)
Figure 1. Layout of ESCORTAIR installed INNOSOPLE-1 PM sensor for measuring particulate matter
(PM) concentrations.
179
Appl. Sci. 2019, 9, 1947
2.2. Federal Equivalent Method

As mentioned above, to conduct a comparative measurement of IRMs, we used the U.S.
Environmental Protection Agency (EPA) Federal Equivalent Method (FEM), that is, GRIMM 180
(GRIMM Aerosol, Technik Ainring GmbH & Co. KG, Ainring, Germany) for indoor testing and
BAM-1022 (MetOne, Grants Pass, OR, USA) for outdoor field tests [17].
The Grimm Technologies, Inc. Model EDM 180 PM2.5 Monitor is a light scattering OPC monitor
operated for 24 h at a volumetric flow rate of 1.2 L/min, configured with a Nafion® - type air sample
dryer. BAM-1022, a beta-ray attenuation mass monitor has a PM2.5 particle size separator. Using BAM,
we obtained 24 1-h average measurements at the national PM2.5 monitoring supersite operated by
National Institute of Environmental Research, Daejeon, Korea.
2.3. Flow Rate Inspection

Before each indoor and outdoor experiment, RGMs, that is, the PDR-1500 and SIDEPAK were
zeroed with an in-line high efficiency particulate air (HEPA) filter and the flow of each device (1.52 L/min,
1.7 L/min and 0.5 L/min) was checked by a mass flowmeter (TSI, Inc., Shoreview, MN, USA). The
temperature and relative humidity data were also downloaded from the devices if the device is
equipped with sensors for these data. Flow rate of GRIMM, one of FEMs, was also checked with a
similar way. The values from BAM, the other FEM, were used as reference, since it is located at the
KOREA national PM2.5 monitoring site (Daejeon) and operated with high QAQC programs [18,19] to
report hourly outdoor PM2.5 data.
2.4. Experimental Setting

We collected PM2.5 concentrations by performing both an indoor exposure test on March 2018
and outdoor PM2.5 monitoring at the national supersite located in Daejeon, Korea from June to July
2018 including rainy days. We used the 2 sets of each IRM or RGM for indoor and outdoor testing
(serial numbers of devices: Table 1).
2.4.1. Indoor Test

The indoor test included scenarios of frying pork in a pan and exposure to secondhand smoke
(SHS). Indoor pan-frying tests were conducted at inside of an empty laboratory (4 m × 10 m × 3.5 m,
W D H), according to the protocol described in detail in our previous article [20]. In brief, this
experiment was carried out over a 2 h measurement period per trial including first 9 min simulating
the barbequing of pork belly (100 g). Standard operating protocol: a portion of pork belly (100 g) was
pan-fried for 9 min: 3 min on Side A, 3 min on Side B; then 1.5 min on Side A again and finally 1.5 min
on Side B again. When we fried pork belly, after the first 9 min, we opened a window (0.5 m × 0.8 m)
to allow the ventilating air to naturally reduce the PM2.5 concentration.
Measurement of the PM2.5 levels from the exposure to secondhand smoke was conducted with a
lighted cigarette burned. We opened the same window after 30 min during our secondhand smoke
exposure level test. Then, we collected concentration data over next two hours.
During our indoor test, we maintained a minimum distance of 20 cm among these devices, at
least 50 cm from the emission sources and 1 m above the floor. To minimize the effect of additional
source contribution to our PM2.5 measurement results, we reported our PM2.5 results after subtracting
the field background PM2.5 concentrations measured at the baseline. We collected with the frying pan
test and the secondhand exposure test three times with GRIMM 180 on separate days. The indoor test
data from each device (80-s interval for PA, 60 s for the remaining devices) were calculated to the 5 min
average level to be compared with the outcomes from GRIMM. The final number of data for the 2 h
panfrying test was approximately 50 (2 sets of each device × 12 data point/h × 2.0 h) and that for the
2.5 h secondhand smoke exposure test was about 60 (2 sets of each device × 12 data point/h × 2.5 h) for
PM2.5 , as well as temperature and relative humidity.
180
Table 1. Summary of PM2.5 measurement range of concentration, measurement interval, weight and Wi-Fi availability reported by manufacturers and unit cost in
South Korea by 30 June 2018.
Sampling
Device Measurement
Sensor Type b Pump Flow Precision c Log Interval c Unit Price ($) Weight (g) Wi-Fi
Classification a Range
Rate
97% over the
Appl. Sci. 2019, 9, 1947
GRIMM (EDM180) 1
0~3000,000 whole
(GRIMM Aerosol, Germany) FEM OPC 1.2 L/min, 5 s to 1 h 19,000 20,000 No
particles/Liter measuring
S/N #: 11R15047
range
Exceeds
BAM-1020 2 .
Beta ray US-EPA Class
(MetOne, OR) FEM 0~1000 mg/m3 16.7 L/min 1 min to 1 h 23,750 24,500 No
Attenuation III PM2.5 FEM
S/N #: N11181
standards
ESCORTAIR 3
(ESCORT, Seoul, South Korea) ±10%@100~500
IRM OPC 1000 μg/m3 NA 30 s 300 400 Yes
S/N #: 6a:c6:3a:c7:83:bf μg/m3
6a:c6:3a:c7:88:b1
PA 4 (PurpleAir, CA, USA)
0~500 μg/m3 as ±10%@100~500
S/N #: A0:20:A6:A:AD:1B. IRM Photometer NA 80 s 300 450 Yes
effective range μg/m3
A0:20:A6:B:83:32
PDR-1500 5
±2% of reading
(Thermo Scientific, MA, USA) Adjustable
181
RGM Photometer 0.001~400 mg/m3 or ±0.005 1 s to 1 h 9000 1200 No
S/N #: CM17422007, 0 to 3.5 L/min
mg/m3
CM17422017
SIDEPAK 6 ±0.001 mg/m3
Adjustable
(TSI, MN, USA) RGM Photometer 0.001~100 mg/m3 over 24 h as 1 s to 60 s 6000 460 No
0 to 1.8 L/min
S/N #: 11104037, 11008055 zero stability
a Federal equivalent method (FEM), Research Grade Monitor (RGM) and inexpensive real-time monitor (IRM); b Optical particle count (OPC); c Information
from manufacture. 1. GRIMM: https://www.grimm-aerosol.com/fileadmin/files/grimm-aerosol/General_Downloads/The_Catalog_2018_web.pdf. 2. BAM: https://

metone.com/wp-content/uploads/2017/08/bam-1020-9803_touch_screen_manual_rev_k.pdf. 3. ESCORTAIR: This study. 4. PA: https://www.purpleair.com/sensors. 5.
PDR: https://www.newstarenvironmental.com/air-toxic-monitors/personal-dataram-pdr1500-aerosol-monitor.html?_vsrefdom=adwords&gclid=CjwKCAiA2fjjBRAjEiwAuewS_
cYiCFBPpx0d0bTaRKtmxe-1Kt22JVs352pQKq9e63XyqT_pIbAZsRoCn9UQAvD_BwE. 6. SIDEPAK https://www.tsi.com/getmedia/84b5be22-c339-49bc-ab97-e3c4baee16c1/SidePak%
20AM520_US_5001737_Web_1.
Appl. Sci. 2019, 9, 1947
2.4.2. Outdoor Test

Using the same configuration of monitoring devices, we also measured the ambient PM2.5
concentration at the outdoor Roof-top of one of the national PM2.5 Supersites located in Daejeon, South
Korea, operating BAM (MetOne, Grants Pass, OR, USA). Main body of BAM was installed at inside of
an experiment laboratory of the Supersite while the inlet of BAM was located at the Roof-top. The
outdoor temperature and RH during were measured by the sensors in ESCORTAIR, PA and PDR-1500
and the values were crosschecked. Measurements from IRMs were collected every minute, except
for PA, which provided a response every 80 s. To compare the hourly concentration values provided
by the Supersite, we calculated the hourly mean values using measured values acquired at hourly
intervals from each IRM device. Final sample size for the outdoor data was 240 (2 sets of each device ×
24 data points/day × 5 days) for PM2.5 , as well as temperature and relative humidity.

The Spearman correlation tests were used to evaluate the associations among the outcomes
of devices measured at indoor or outdoor environments, considering that variables were not
normally distributed.
Using outdoor measurement data, we evaluate the association of device response with various
relative humidity level. We also evaluated the associations of the daily mean concentrations from
IRMs with those obtained with FEM methods using multivariate linear regression models. We used
the values obtained with the FEM as dependent variables and those obtained with real-time devices
as independent variables to obtain a correction factor. The hourly mean temperature and relative
humidity data for each sampling date, which was previously compared with the nearest national
meteorological monitoring sites, were used to adjust the effects of relative humidity on the association
between IRMs and FEM outcomes. The corresponding slopes and coefficients of determinant (R2) for
each RT monitor were also provided. The final measurement error (%) was calculated based on the
U.S. EPA performance evaluation program for PM instruments;
Measurementi − FEMi
Di f f erencei = (1)
FEMi
EPA specifies that the percent bias goal for acceptable measurement uncertainty should be within
±20% [17]. Here, we reported median (IQR) value of the differences per device because the distribution
of differences was not normal. We also provided mean of the differences just to compare with value in
the guideline [21]. All analyses were conducted with SAS (Version 9.4) and R software (Version 2.15.3,
R Development Core Team).
3. Results
3.1. PM2.5 Concentration

In our indoor test, the median (IQR) PM2.5. concentration over the pan-frying test was 86.8 μg/m3
(17.8–254.4 μg/m3 ) by the real-time ESCORTAIR, 104.9 μg/m3 (43.9–228.2 μg/m3 ) and 236.2 μg/m3
(49.3–648.7 μg/m3 ) by the real-time PA or PDR-1500 devices and 153.2 μg/m3 (46.2–409.7 μg/m3 ) by
using GRIMM (Table 2). The median (IQR) concentrations for secondhand smoke (SHS) test were 20.9
(17.4– 156.6) μg/m3 , 31.2 (14.4–194.3) μg/m3 and 28.4 (12.8–314.0) μg/m3 for ESCORTAIR, PA and PDR,
respectively, whereas GRIMM provided 23.5 (15.9–107.1) μg/m3 .
Simultaneous outdoor PM2.5 monitoring results are also provided in Table 2. The median (IQR) of
the hourly average values of ESCORTAIR and PA were 13.7 (7.3~21.2) and 19.7 (9.3~35.8) μg/m3 , which
were an overestimation of the values obtained by BAM, one of U.S. EPA FEMs. During indoor testing,
the median temperature and RH were approximately 20~22 ◦ C and 37%. During outdoor testing, the
median values (IQR) were 30.7 (25.6~40.7) ◦ C and 56.4 (34.8~71.4) %, respectively (Table 2).
182
Appl. Sci. 2019, 9, 1947
Table 2. PM2.5 mass concentration (μg/m3 ) (median, IQR) measured by real-time sensor devices
and the federal equivalent method as well as the temperature and relative humidity throughout the
sampling period.
Indoor—Pan-Frying Indoor—SHS Outdoor—Urban Traffic Hotspot

(n = 50) (n = 60) (n = 240)
GRIMM 153.2 (46.2–409.7) 23.5 (15.9–107.1) NA
BAM NA NA 9.0 (4.0–22.0)
ESCORTAIR 86.8 (17.8–254.4) 20.9 (17.4–156.6) 13.7 (7.3–21.2)
PA 104.9 (43.9–228.2) 31.2 (14.4–194.3) 19.7 (9.3–35.8)
PDR-1500 236.2 (49.3–648.7) 28.4 (12.8–314.0) 13.8 (6.8–34.8)
SIDEPAK 261.3 (71.5–800.0 50.0 (21.0–652.0) 29.1 (15.6–59.9)
Temp. (◦ C) 21.7 (21.1–21.7) 20.1 (19.7–20.5) 30.7 (25.6–40.7)
RH (%) 37.0 (35.0–39.0) 37.0 (35.0–38.0) 56.4 (34.8–71.4)
3.2. Correlations among Devices and the Fem

High-level correlations were obtained between the IRMs (ESCORT, PA) and FEM. The Spearman
correlation coefficients were 0.97 (P = 0.0001) or 0.92 (P = 0.0001) between ESCORTAIR and GRIMM for
indoor pan-frying or the SHS test, respectively. The correlation coefficients between PA and GRIMM
were similar (0.97 (P = 0.0001) or 0.86 (P = 0.0001)) for the indoor test (Table 3). Our outdoor test also
showed high correlation between ESCORTAIR and BAM (0.84 (P = 0.0001)). The correlation coefficient
for PA was 0.88 (P = 0.0001).
Similarly, the measurements by RGMs (PDR-1500, SIDEPAK) and FEM were highly correlated:
0.97–0.98 for the Pan-frying test, 0.88~0.96 for the SHS test and 0.84–0.91 for the outdoor test. Between
IRMs, that is, ESCORTAIR and PA, the association of measurements were strong to each other (r = 0.93:
Indoor pan frying; 0.85: Indoor-SHS; 0.93: Outdoor-urban traffic). In addition, they showed similar
correlation patterns to PDR-1500 or SIDEPAK (Table 3).
Table 3. Scatter plots and Spearman correlation coefficients between real-time PM2.5 monitoring devices
and FEM.
Indoor-Pan-Frying Indoor-SHS Outdoor Urban Traffic Hotspot

FEM E PA P S FEM E PA P S FEM E PA P S
FEM 1 1 1
E 0.97 1 0.92 1 0.85 1
PA 0.97 0.93 1 0.86 0.85 1 0.88 0.93 1
PDR 0.98 0.95 0.99 1 0.96 0.93 0.94 1 0.84 0.93 0.99 1
S 0.98 0.99 0.96 0.98 1 0.88 0.86 0.88 0.93 1 0.91 0.91 0.99 0.99 1
FEM: Federal Equivalent Method: In this study Indoor: GRIMM Optical particle courting, OPC), Outdoor: BAM
(Beta ray attenuation monitor). E: ESCORTAIR, PA: PurpleAir, PDR: PDR-1500, S: SIDEPAK.
3.3. Effects of Ambient Humidity for Outdoor Measurement

In this study, as seen in Supplementary Figure S1, we observed that the PM2.5 concentration of
IRMs were significantly increased with the increase in relative humidity level. The slope obtained from
a simple regression line of FEM on IMRs, that is, (BAM = Slope * ESCORTAIR + intercept) at a relative
humidity above 80%, was smaller than the slope at < 20.0%, 20.1~40.0%, 40.1~60.0%, 60.1~80.0%. The
degree of decreasing trend was larger with IRMs, compared to two RGMs (Figure 2).
183
Appl. Sci. 2019, 9, 1947

(6&257$,5

5 5 5 5
5 !

3$

5
5 5 5 5 !

6ORSH&RQILGHQFH,QWHUYDO
3'5

5 5 5 5
5 !

6,'(3$.

5 5 5 5
5 !
5HODWLYH+XPLGLW\>@&RHIRI'HWHUPLQDWLRQ
Figure 2. Change of the slope with 95% confidence interval obtained from the regression line of BAM
(Beta ray attenuation monitor) on IRMs (Inexpensive real-time particulate matter monitors) or RGMs
(research grade monitors) by the degree of relative humidity.
3.4. Correction Factor

Since we found the effect of RH on the measurements of ESCORTAIR or PA, we developed our
own correction models for those IRM devices by performing stepwise linear regression. The correction
factors were obtained from the regression models (1.11 for ESCORTAIR and 1.92 for PA, p < 0.01), (1.00
for ESCORTAIR and 0.87 for PA, p < 0.01) and (1.15 for ESCORTAIR and 0.70 for PA, p < 0.01) for the
measurement of PM2.5 resulting from indoor pan-frying, SHS or the urban traffic hotspot of South
Korea, respectively.
After adjusting for temperature and relative humidity, the results were unchanged for the indoor
tests (1.10 for ESCORTAIR and 1.90 for PA, p < 0.01), (0.97 for ESCORTAIR and 0.81 for PA, p < 0.01)
(Table 3). For the outdoor measurement, the change of slopes (0.72 for ESCORTAIR and 0.77 for PA,
p < 0.01) was relatively small but the R2 values were changed (Table 4).
184
Appl. Sci. 2019, 9, 1947
Table 4. Slopes obtained from stepwise linear calibration models with adjusted R2 (Dependent variable:
US EPA FEM, Independent variable: IRM).
Indoor—Pan-Frying Indoor—SHS Outdoor—Urban Traffic Hotspot

Single Multivariate * Single Multivariate * Single Multivariate *
β R2 β R2 β R2 β R2 β R2 β R2
ESCORTAIR 1.11 0.98 1.10 0.98 1.00 0.92 0.97 0.92 1.15 0.70 1.14 0.81
PA 1.92 0.94 1.90 0.94 0.87 0.89 0.81 0.90 0.70 0.83 0.71 0.87
PDR-1500 0.33 0.98 0.33 0.98 0.54 0.91 0.49 0.92 0.33 0.72 0.36 0.80
SIDEPAK 1 0.34 0.98 0.32 0.99 0.28 0.90 0.31 0.92 0.35 0.84 0.36 0.89
* results obtained after adjusting for temperature and relative humidity.
3.5. Bias after Application of Correction Factors

We then determined the extent to which the original PM2.5 data measured by ESCORTAIR were
improved after applying the correction factor obtained from the model by performing comparative
analyses using the outcomes of multivariate regression models (Table 4). Using the corrected data, the
final coefficient of determination (R2) between FEM (y) and ESCORTAIR (x) was 0.81. The coefficient
for PA was 0.87. We found the difference (median (IQR)) with the calibrated data, compared to FEM, to
be 16.3. (8.5~28.0)% for ESCORTAIR and 14.5 (6.1 to 23.5)% for PA for outdoor environments (Figure 3).
The bias (mean of the difference) was 13.1% for ESCORTAIR and 7.8% for PA for outdoor. The bias for
indoor data was at least similar or lower than the bias level obtained from the outdoor test.
(6&257$,50HGDLQ,45 a

3$0HGDLQ,45 a

'LIIHUHQFH >0HDVXUHPHQW)(0)(0@

3'50HGDLQ,45 a

6,'(3$.0HGDLQ,45 a

7LPH+RXU
Figure 3. Distribution of difference of outdoor measurements against the FEM; BAM for outdoors.
185
Appl. Sci. 2019, 9, 1947
4. Discussion
In this study, we compared PM2.5 concentrations measured with IRMs and RGMs to those
measured with U.S. EPA FEMs. A relationship between the real-time PM2.5 concentration and
FEM were acceptable (R = 0.97 to 0.98: Indoor pan frying; 0.86 to 0.96: Indoor-SHS; 0.84 to 0.91:
Outdoor-urban traffic, respectively).
This study conducted indoor testing with common PM2.5 sources of Korea, that is, frying pork in a
pan or smoking. The indoor test was conducted in an indoor laboratory in which the room temperature
was maintained at 20~22 ◦ C and the relative humidity at 37% because we assumed that most low-cost
PM sensing devices would be established inside susceptible populations’ homes where the indoor
temperature and humidity level would be relatively stable, compared to the outdoor environment.
In addition, in this study, we extended our comparison test by performing them outside with
a federal equivalent method. We discovered that, on days with a high level of outdoor RH (80% or
higher), our IRMs overestimated the PM2.5 level. Thus, finally, we got the correction model providing
a correction factor for ESCORTAIR to adjust for the effect of temperature and relative humidity as such
a process has been conducted in a regulatory monitoring site (South Coast Air Quality Monitoring
District, SCAQMD) for the field use of these kinds of inexpensive sensing devices including PA [22,23].
Several previous studies provided a correction factor for SIDEPAK monitors: 0.77 in Northern
California, U.S.A. (ambient air), 0.43 or 0.52 in Italy (ambient air in urban or rural areas) and 0.42
in Italy (indoor-outdoor mixed environment) [24–26], which were comparable to our results (0.36).
Our correction factor (0.36) for PDR-1500 was smaller than results reported by Wang et al. (2016) [27]
(0.71 compared to PDR-1500 using its own filtering method) but very similar to the values obtained
by Ramachandran et al. (2000) (0.33) and Wallace et al. (2011) (0.38), who conducted their studies on
atmospheric environments [28,29].
A lack of quantitative information on the speciation of particles, traffic volume, type of vehicle or
the particular sampling time or season limits further exploration of the basis for the differences in the
correction factors between these studies and ours. Nevertheless, our correction model for ESCORTAIR,
with consideration of the RH level, was derived in a similar way to that with which we obtained the
factor for SIDEPAK or PDR-1500. Thus, we consider no significant systematic errors to have been
involved in the calculation process. A good linear relationship has been obtained between the PM2.5
mass concentration of FEM and the responses of low-cost PM sensors as reported previously in other
country [8,9,25,30]. We demonstrated the urban hotspot specific correction factor for a light-scattering
sensor in Korean urban environment of interest to enable our findings and methodology to be extended
and replicated by researchers who are interested in the utility of low-cost sensing device, such as
ESCORTAIR, in South Korea.
OPCs are reportedly good at estimating mass if they have numerous bins, such as GRIMM [31].
However, estimating the mass concentration from a limited number of bins may be subject to a
measurement error during the conversion process with the factory-provided internal conversion
algorithm. Therefore, we used an additional correction factor for ESCORTAIR after comparison to
GRIMM for proper usage under Korean circumstances.
It is well established that the response of monitors based on light scattering varies with aerosol size
distribution, composition and optical properties and need a proper calibration process [13,32,33]. No
single calibration model (or correction model) can enable accurate performance for all particle sources
in microenvironments. This challenge applies to both research and consumer monitors. Although
gravimetric measurements may be used to determine a source- or environment-specific calibration
for a research study, the approach is not practical for routine monitoring in homes. A key objective
of continuous monitoring—to activate controls—can be achieved if the monitor reliably and clearly
responds to sources that account for the majority of particles in the home even if responses are
not quantitative.
We conducted this study by assuming that indoor PM2.5 emission sources exist, that is, from
frying pork in a pan and smoking, with consideration of Korean life style [20] and relatively high
186
Appl. Sci. 2019, 9, 1947
smoking prevalence [34]. We determined that our linear model obtained from the indoor test for each
single-aerosol type showed excellent performance (R2 = 0.98 or 0.92 for ESCORTAIR; 0.94 or 0.90 for
PA), compared to FEM responses. However, the responses of ESCORTAIR as well as PA were relatively
less precise but good (R2 = 0.81, 0.87) for monitoring in urban traffic hotspots suggesting that IRMs
need a site-specific calibration with a reference method before they are used [13,32,33].
The shape of the response curve can be related to the type of OEM sensors integrated with the
monitor and relative humidity. Because we did not have information about the internal conversion
factors (count to mass concentration for OPC type or light intensity to mass concentration for
photometer), using GRIMM or BAM, known as US FEM, we tried to obtain our own correction factors
for usage of ESCORTAIR or PA in urban indoor or outdoor settings.
The limitations of our study should be noted. First, the sample size in this study was relatively
small. However, in our tests, we determined a PM2.5 range of 10 to 3000 μg/m3 including both indoor
and outdoor measurement. This range ensures that the concentration distributions would not be
systematically biased. Our PM2.5 concentrations might not be representative of each sampling season
or area as a result of spatial-temporal variations. Additional experiments are needed to understand
the stability of our correction factor in different seasons and/or in other locations, that is, in industrial
or rural areas of Korea. Extending the sampling periods for each season and location would ensure
that our results are more representative. Furthermore, in future studies, measurements of the wind
direction and speed are expected to provide improved correction factors between the IRM and FEM
methods. As we mentioned in the method section, we checked flow rate for our RGMs or FEMs prior to
our experiment. For IRMs, we have considered measuring flow rate but due to its open wide inlet and
very low flow rate, we could not connect it to our mass flowmeter properly. Instead, especially, before
our outdoor test, we operated 5 ESCORTAIRs and 5 PAs simultaneously and checked measurement
errors between devices. Then, we selected 2 of them which provide best outcomes, compared to FEM.
A preparation of QC/QA test program for massive products of IRMs are recommended. And for
ESCORTAIR, like PA, application of weather proof design is suggested. In addition, future studies
may be necessarily conducted to obtain site-specific correction factors including at coal power plants
or in rural areas.
In this study, the performances of the IRM with RGM and that of the FEM operated with a high
QCQA program were compared with one-hour monitoring intervals at national PM2.5 monitoring
site. This makes this study unique compared to previous studies, which were mostly conducted with
one-day interval gravimetric methods at ordinary sampling sites.
Despite the growing public interest in reducing personal exposure levels to PM2.5 in Korea, IRM
monitoring still faces challenges in terms of providing real-time concentration information. Although
the number of national PM2.5 monitoring sites in Korea is increasing, additional IRMs in hotspots
or communities are required because they can detect continuous spatial—temporal variations and
identify nearby exposure sources on a real-time basis in a micro-environment of hotspots.
This study found that the measurement of PM2.5 concentrations with recently developed
laser-based IRM under- or over-estimates PM2.5 concentrations obtained from FEM while its bias could
be approximately 11 to 16% even at urban outdoor hotspots with traffic sources with high relative
humidity levels. Therefore, the application of a correction factor is strongly suggested for inexpensive
laser-based monitoring devices.
5. Conclusions
Our study determined that on days with a high level of outdoor RH (80% or higher), our IRMs
overestimated the outdoor PM2.5 level and showed the necessity of a correction factor for IRMs to
adjust for the effect of temperature and relative humidity. PM2.5 concentrations measured with IRMs
need to be subjected to quality control and quality assurance evaluation before these monitors are used
for the quantification of PM2.5 levels in urban indoor or outdoor atmospheric environments in Korea.
In consideration of relatively stable outcomes with the application of correction factors for recently
187
Appl. Sci. 2019, 9, 1947
developed new IRMs—ESCORTAIR or PA—our study supports their usage in networking monitoring
for various urban hotspots.
Supplementary Materials: The following are available online at http://www.mdpi.com/2076-3417/9/9/1947/s1.

Figure S1. Scatter plots of PM2.5 concentrations of IRMs with those of FEM, i.e., BAM, by the level of relative
humidity. The data points refer to each paired hourly mean value (n = 240, except ESCORT which had 120 points
due to malfunction of a device with records of high internal temperature values during our outdoor test).
Author Contributions: S.K. designed this study, wrote manuscript and conducted interpretation of the quantitative
aspects of data analysis. S.P. performed modeling simulation and J.L. provided editorial efforts. S.K. supervised
the whole study.
Funding: This research was funded by Environmental Health Research Center Project (2016001360002) by Korea
Environmental Industry & Technology Institute, Ministry of Environment, South Korea
Acknowledgments: The authors deeply appreciate the technical comments for sensor evaluation from. Andrea
Clements and Timothy Buckley at the national exposure research laboratory, U.S. EPA. The authors also appreciate
the assistance of Sungmin Jung and the staff (Minhee Lee, Taekyung Hwang, Jieun Jeong) at the Korea PM2.5
supersite of Daejeon, South Korea for their support with data collection and the help of Juhee Kim in data screening.
The content is solely the responsibility of the authors and does not necessarily represent the official views of the
US EPA or Korea NIER (National Institute for Environmental Research).
References
1. Orellano, P.; Quaranta, N.; Reynoso, J.; Balbi, B.; Vasquez, J. Effect of outdoor air pollution on asthma
exacerbations in children and adults: Systematic review and multilevel meta analysis. PLoS ONE 2017, 12,
e0174050. [CrossRef]
2. Prieto Parra, L.; Yohannessen, K.; Brea, C.; Vidal, D.; Ubilla, C.A.; Ruiz Rudolph, P. Air pollution, PM2.5
composition, source factors, and respiratory symptoms in asthmatic and nonasthmatic children in Santiago,
Chile. Environ. Int. 2017, 101, 190–200. [CrossRef] [PubMed]
3. Gillooly, S.E.; Zhou, Y.; Vallarino, J.; Chu, M.T.; Michanowicz, D.R.; Levy, J.I.; Adamkiewicz, G. Development
of an in~home, real-time air pollutant sensor platform and implications for community use. Environ. Pollut.
2019, 244, 440–450. [CrossRef] [PubMed]
4. Genikomsakis, K.N.; Galatoulas, N.; Dallas, P.I.; Ibarra, L.M.C.; Margaritis, D.; Ioakimidis, C.S. Development
and On Field Testing of Low Cost Portable System for Monitoring PM2.5 Concentrations. Sensors 2018, 18,
1033. [CrossRef] [PubMed]
5. Shao, W.; Zhang, H.; Zhou, H. Fine Particle Sensor Based on Multi Angle Light Scattering and Data Fusion.
Sensors 2017, 17, 1033. [CrossRef] [PubMed]
6. Aneja, V.P.; Pillai, P.R.; Isherwood, A.; Morgan, P.; Aneja, S.P. Particulate matter pollution in the coal
producing regions of the Appalachian Mountains: Integrated ground based measurements and satellite
analysis. J. Air Waste Manag. Assoc. 2017, 67, 421–430. [CrossRef] [PubMed]
7. Burkart, J.; Steiner, G.; Reischl, G.; Moshammer, H.; Neuberger, M.; Hitzenberger, R. Characterizing
the performance of two optical particle counters (Grimm OPC1.108 and OPC1.109) under urban aerosol
conditions. J. Aerosol. Sci. 2010, 41, 953–962. [CrossRef] [PubMed]
8. Sousan, S.; Koehler, K.; Hallett, L.; Peters, T.M. Evaluation of the Alphasense Optical Particle Counter (OPC
N2) and the Grimm Portable Aerosol Spectrometer (PAS 1.108). Aerosol. Sci. Technol. 2016, 50, 1352–1365.
[CrossRef]
9. Dacunto, P.J.; Klepeis, N.E.; Cheng, K.C.; Acevedo Bolton, V.; Jiang, R.T.; Repace, J.L.; Ott, W.R.;
Hildemann, L.M. Determining PM2.5 calibration curves for a low cost particle monitor: Common indoor
residential aerosols. Environ. Sci. Process. Impacts 2015, 17, 1959–1966. [CrossRef] [PubMed]
10. Lanki, T.; Alm, S.; Ruuskanen, J.; Janssen, N.A.; Jantunen, M.; Pekkanen, J. Photometrically measured
continuous personal PM(2.5) exposure: Levels and correlation to a gravimetric method. J. Expo. Anal.
Environ. Epidemiol. 2002, 12, 172–178. [CrossRef] [PubMed]
11. O’Shaughnessy, P.T.; Slagley, J.M. Photometer response determination based on aerosol physical characteristics.
AIHA J. 2002, 63, 578–585. [CrossRef]
188
Appl. Sci. 2019, 9, 1947
12. Baron, P.A. Aerosol Photometers for Respirable Dust Measurements. In NIOSH Manual of Analytical
Methods; 1998. Available online: https://www.cdc.gov/niosh/docs/2003--154/pdfs/chapter~{}g.pdf (accessed
on 25 March 2019).
13. Kelly, K.E.; Whitaker, J.; Petty, A.; Widmer, C.; Dybwad, A.; Sleeth, D.; Martin, R.; Butterfield, A. Ambient and
laboratory evaluation of a low cost particulate matter sensor. Environ. Pollut. 2017, 221, 491–500. [CrossRef]
[PubMed]
14. Semple, S.; Ibrahim, A.E.; Apsley, A.; Steiner, M.; Turner, S. Using a new, low cost air quality sensor to
quantify second hand smoke (SHS) levels in homes. TOB Control. 2015, 24, 153–158. [CrossRef] [PubMed]
15. Zhou, X.; Aurell, J.; Mitchell, W.; Tabor, D.; Gullett, B. A small, lightweight multipollutant sensor system
for ground mobile and aerial emission sampling from open area sources. Atmos. Environ. 2017, 154, 31–41.
[CrossRef] [PubMed]
16. Franken, R.; Maggos, T.; Stamatelopoulou, A.; Loh, M.; Kuijpers, E.; Bartzis, J.; Steinle, S.; Cherrie, J.W.;
Pronk, A. Comparison of methods for converting Dylos particle number concentrations to PM2.5 mass
concentrations. Indoor Air 2019, 29, 450–459. [CrossRef]
17. Williams, R.; Vasu Kilaru, E.; Snyder, A.; Kaufman, T.; Dye, A.; Rutter, A.; Russell, A.; Hafner, H. Air Sensor
Guidebook; EPA/600/R-14/159 (NTIS PB2015-100610); U.S. Environmental Protection Agency: Washington,
DC, USA, 2014.
18. National Institute of Environmental Research (NIER). 2015 Annual Report of Intensive air Quality Monitoring
Station, NIER GP2016–160; NIER: Incheon, Korea, 2016.
19. Yu, G.; Park, S.; Ghim, Y.; Shin, H.; Lim, C.; Ban, S.; Yu, J.; Kang, H.; Seo, Y.; Kang, K.; et al. Difference
in Chemical Composition of PM2.5 and Investigation of its Causing Factors between 2013 and 2015 in Air
Pollution Intensive Monitoring Stations. J. Korean Soc. Atmos. Environ. 2018, 34, 16–37. [CrossRef]
20. Lee, S.; Yu, S.; Kim, S. Evaluation of Potential Average Daily Doses (ADDs) of PM2.5 for Homemakers
Conducting Pan Frying Inside Ordinary Homes under Four Ventilation Conditions. Int J. Environ. Res. Public
Health 2017, 14, 78. [CrossRef]
21. Rosner, B. Hypothesis Testing, Fundamentals of Biostatistics; Duxbury: Parific Grove, CA, USA, 2000; Chapter 7.
22. Kim, B.M.; Teffera, S.; Zeldin, M.D. Characterization of PM2.5 and PM10 in the South Coast Air Basin of
southern California: Part 1 Spatial variations. J. Air Waste Manag. Assoc. 2000, 50, 2034–2044. [CrossRef]
23. Kim, B.M.; Teffera, S.; Zeldin, M.D. Characterization of PM2.5 and PM10 in the South Coast Air Basin of
southern California: Part. 2 Temporal variations. J. Air Waste Manag. Assoc. 2000, 50, 2045–2059. [CrossRef]
[PubMed]
24. Borgini, A.; Tittarelli, A.; Ricci, C.; Bertoldi, M.; De Saeger, E.; Crosignani, P. Personal exposure to PM2.5
among high school students in Milan and background measurements: The EuroLifeNet study. Atmos.
25. Jiang, R.T.; Acevedo Bolton, V.; Cheng, K.C.; Klepeis, N.E.; Ott, W.R.; Hildemann, L.M. Determination of
response of real-time SidePak AM510 monitor to secondhand smoke, other common indoor aerosols, and
outdoor aerosol. J. Environ. Monit. 2011, 13, 1695–1702. [CrossRef]
26. Karagulian, F.; Belis, C.A.; Lagler, F.; Barbiere, M.; Gerboles, M. Evaluation of a portable nephelometer
against the Tapered Element Oscillating Microbalance method for monitoring PM(2.5). J. Environ. Monit
2012, 14, 2145–2153. [CrossRef] [PubMed]
27. Wang, Z.; Calderon, L.; Patton, A.P.; Sorensen Allacci, M.; Senick, J.; Wener, R.; Andrews, C.J.; Mainelis, G.
Comparison of real-time instruments and gravimetric method when measuring particulate matter in a
residential building. J. Air Waste Manag. Assoc. 2016, 66, 1109–1120. [CrossRef]
28. Ramachandran, G.; Adgate, J.L.; Hill, N.; Sexton, K.; Pratt, G.C.; Bock, D. Comparison of short term variations
(15 min averages) in outdoor and indoor PM2.5 concentrations. J. Air Waste Manag. Assoc. 2000, 50, 1157–1166.
[CrossRef]
29. Wallace, L.A.; Wheeler, A.J.; Kearney, J.; Van Ryswyk, K.; You, H.; Kulka, R.H.; Rasmussen, P.E.; Brook, J.R.;
Xu, X. Validation of continuous particle monitors for personal, indoor, and outdoor exposures. J. Expo. Sci
Environ. Epidemiol. 2011, 21, 49–64. [CrossRef] [PubMed]
30. Shi, J.; Chen, F.; Cai, Y.; Fan, S.; Cai, J.; Chen, R.; Kan, H.; Lu, Y.; Zhao, Z. Validation of a light scattering PM2.5
sensor monitor based on the long term gravimetric measurements in field tests. PLoS ONE 2017, 12, e0185700.
[CrossRef]
189
Appl. Sci. 2019, 9, 1947
31. Peters, T.M.; Ott, D.; O’Shaughnessy, P.T. Comparison of the Grimm 1.108 and 1.109 portable aerosol
spectrometer to the TSI 3321 aerodynamic particle sizer for dry particles. Ann. Occup. Hyg. 2006, 50, 843–850.
[PubMed]
32. Njalsson, T.; Novosselov, I. Design and Optimization of a Compact Low Cost Optical Particle Sizer. J. Aerosol.
Sci. 2018, 119, 1–12. [CrossRef] [PubMed]
33. Northcross, A.L.; Edwards, R.J.; Johnson, M.A.; Wang, Z.M.; Zhu, K.; Allen, T.; Smith, K.R. A low cost particle
counter as a realtime fine particle mass monitor. Environ. Sci. Process. Impacts 2013, 15, 433–439. [CrossRef]
34. Kim, S.; Jung, A. Optimum cutoff value of urinary cotinine distinguishing South Korean adult smokers from
nonsmokers using data from the KNHANES (2008–2010). Nicotine Tob Res. 2013, 15, 1608–1616. [CrossRef]
190
applied
sciences
Article
Real-Time Monitoring of Indoor Air Quality with
Internet of Things-Based E-Nose
Mehmet Taştan * and Hayrettin Gökozan
Department of Electronic, Turgutlu Vocational School, Manisa Celal Bayar University, 45400 Manisa, Turkey
* Correspondence: [email protected]; Tel.: +90-236-313-55-02
Received: 22 June 2019; Accepted: 15 August 2019; Published: 20 August 2019
Abstract: Today, air pollution is the biggest environmental health problem in the world. Air pollution
leads to adverse effects on human health, climate and ecosystems. Air is contaminated by toxic
gases released by industry, vehicle emissions and the increased concentration of harmful gases and
particulate matter in the atmosphere. Air pollution can cause many serious health problems such as
respiratory, cardiovascular and skin diseases in humans. Nowadays, where air pollution has become
the largest environmental health risk, the interest in monitoring air quality is increasing. Recently,
mobile technologies, especially the Internet of Things, data and machine learning technologies have a
positive impact on the way we manage our health. With the production of IoT-based portable air
quality measuring devices and their widespread use, people can monitor the air quality in their living
areas instantly. In this study, e-nose, a real-time mobile air quality monitoring system with various
air parameters such as CO2 , CO, PM10 , NO2 temperature and humidity, is proposed. The proposed
e-nose is produced with an open source, low cost, easy installation and do-it-yourself approach.
The air quality data measured by the GP2Y1010AU, MH-Z14, MICS-4514 and DHT22 sensor array
can be monitored via the 32-bit ESP32 Wi-Fi controller and the mobile interface developed by the
Blynk IoT platform, and the received data are recorded in a cloud server. Following evaluation of
results obtained from the indoor measurements, it was shown that a decrease of indoor air quality
was influenced by the number of people in the house and natural emissions due to activities such as
sleeping, cleaning and cooking. However, it is observed that even daily manual natural ventilation
has a significant improving effect on air quality.
Keywords: internet of things; e-nose; indoor air quality; smart home; ESP32
1. Introduction
In recent years, air pollution has been a major environmental problem and a global concern that
has exceeded recommended national limits. Air pollution has negative effects on human health and
ecosystems, as well as affecting the world’s climate [1]. Air pollution can be classified as internal or
external air pollution, depending on where the activities take place [2]. Outdoor air pollution occurs
in an open environment, covering the entire atmosphere. Fossil fuels used to meet the energy needs
of factories, industries and vehicles are the main activities contributing to agricultural and mining
outdoor air pollution. The external air pollutants are mainly composed of nitrogen oxides (NOx),
nitrogen dioxide (NO2 ), sulfur dioxide (SO2 ), ozone (O3 ), carbon monoxide (CO), hydrocarbons and
particulate matter (PM) of different particle sizes. Indoor air pollution, found in offices, hospitals,
schools, libraries, entertainment areas, gymnasiums, public transport vehicles, etc., is classified as the
pollution of the air of indoor areas [3]. Major indoor air pollutants include NOx, SO2 , O3 , CO, carbon
dioxide (CO2 ), volatile and semi-volatile organic compounds VOCs, PM, radon and microorganisms.
The air quality index includes an internationally adopted parameter for assessing air quality.
Ground-level ozone reflects five pollution standards including PM, carbonic oxide, SO2 and
NO2 [4]. Indoor air quality (IAQ) is very important for many people who spend most of their lives

Appl. Sci. 2019, 9, 3435
in closed spaces, such as the elderly, disabled, infants and patients [5]. Indoor air pollution occurs
due to household activities and products used in these activities. Home cleaning products and paint
materials emit toxic chemicals into the air and cause air pollution. Exposure to air pollution, especially
indoor air pollution, is one of the largest environmental health risk factors and is directly related to
millions of premature deaths worldwide each year [6]. Air pollution comes second on the list of deaths
caused by non-contagious health reasons [7].
In the United States, indoor and outdoor air quality is regulated by the Environmental Protection
Agency (EPA). The EPA has shown that indoor pollutant levels may be up to 100 times higher than
the level of external pollutants and poor air quality is one of the five most important environmental
risks threatening public health [8]. PM, a complex mixture of solid and liquid particles of organic and
inorganic substances suspended in the air, is considered the most common pollutant. Air pollution from
PM (PM2.5 and PM10 ) poses a significant health threat to people living in cities. The health-damaging
PM, which can penetrate deep into the lungs, contributes to the risk of developing cardiovascular
diseases as well as lung cancer [9]. In addition to health problems affecting the low IAQ nervous
system, its effects are associated with other long-term diseases [10]. Simple interventions, such as
natural ventilation provided by homeowners, have significant positive effects on IAQ. However,
consideration of the household’s thermal comfort and climate conditions in natural ventilation is
crucial [11]. Furthermore, power losses due to manually regulated natural ventilation should be
considered [12]. Therefore, indoor real-time monitoring of IAQ is very important to detect unhealthy
conditions [13].
Today, cities face interesting challenges and problems to meet socio-economic development
and quality of life goals, and the concept of “smart cities” responds to these challenges. The smart
city is directly related to a strategy to reduce the problems caused by rapid urbanization [14].
The IoT is a network where physical objects are linked to each other or to larger systems. This
network collects large amounts of data from different devices we use in our daily life and converts
it into utilizable information [15]. The rapidly emerging IoT concept supports many different areas
and applications including health, education, agriculture, industry and environmental monitoring.
Real-time environmental monitoring and analysis is an important area of research for IoT [16].
IoT provides remarkable features to smart cities which improve environmental quality control,
innovative real-life solutions and services. The smart home is an indispensable element of smart cities.
In the future, smart homes will be fully managed through smartphone applications and will include
IoT wearable devices [17] that are supported by microsensors. To date, several have been conducted in
the literature with the aim of establishing real-time monitoring solutions for air quality analysis.
Numerous IoT-based air quality monitoring systems using micro-sensors for data collection
have been proposed, including open source technologies for data processing and transmission [18].
According to an IAQ analysis [19], it has been reported that indoor PM2.5 pollution mainly comes
from outdoor sources. In another study, three typical activities such as a having fireplace in the house,
cooking with kitchen appliances and toasting were investigated, and indoor pollutant levels such as
PM2.5 , PM10 and VOC were found to be above the limit values [20]. Monitoring of the PM concentration
for 63 participants between the ages of 18–65 years in the Perth metropolitan area indicates that the PM
concentration increases the number of heartbeats by 4–6 beats per minute (bpm) and has a significant
effect on systolic blood pressure (SBP) [21]. An IoT-based IAQ measurement system was created using
the WEMOS D1 mini microcontroller and the PMS5003 PM sensor. This system allows the household
individual to intervene for ambient assisted living (AAL) [22].
In this study, a low-cost, portable, IoT-based and real-time monitoring system that can measure
ambient air quality with a range of sensors is proposed. For the proposed system, an Android
interface has been designed primarily by the Blynk platform. Then, with the embedded architecture
ESP32 module, the controller unit of the system is created. Using this IoT controller with an internal
Wi-Fi module, all measured air quality data are displayed in the mobile app and these data are
stored on a cloud server. The mobile interface provides users with numerical and graphical data on
192
Appl. Sci. 2019, 9, 3435
contaminating gas concentrations, temperature and humidity. The proposed e-nose measurement
system sends a notification to the users via the mobile application if any gas concentration levels reach
health-threatening values. In this way, households can take measures to reduce gas concentrations
when necessary. The data can be viewed graphically through the mobile interface which allows users to
observe the effects of activities such as sleep, cleaning and cooking on the gas concentration. In addition,
the IoT-based e-nose air quality measurement system has a low-cost (about $100), easy-to-install and
open-source feature produced by a DIY approach.
2. Related Work
Due to the fast advancement in IoT and sensor techniques, interest in air quality measurements
continues to grow. In [23] an environmental monitoring system was propose that included the sensors
Raspberry Pi and MICS-4514, which enabled the assessment both indoors and outdoors at a university
campus of multiple environmental parameters such as temperature, humidity, light, noise level, CO,
NO2 . A proposed system for detecting environmental contamination [24] used various environmental
detectors to identify temperature, humidity, ambient light, gas sensors and PM. All data measured
by the system using the sensors STM32f4xx and Sharp GP2Y1010 for PM detection, and TGS5342
for CO, are stored in the internal storage and on an Internet server over the Wi-Fi network. Using
the ATmega328AVR controller, DHT22 temperature sensor, Sharp GP2Y1010AU0F dust sensor and
UVM-30A UV sensors, an Integrated Environmental Monitoring System (IEMS) [25] was proposed
to detect the microenvironment. In another study [26], the Nano Environmental Monitoring System
(nEMoS) was proposed using an IoT-based indoor environmental quality (IEQ) assessment system
created using an Arduino Uno module and low-cost sensors such as DHT22. To determine the
quality of packaged products, an IoT-based measuring system was intended, including low-cost
pressure, temperature, humidity, gas sensors (BME680, DHT22 and MQ5), Arduino and XBee wireless
module [27]. In the study, in which iAQ, an air quality surveillance system based on IoT architecture,
was proposed for AAL [28], the wireless sensor nodes (WSN), Gateway and Android user interface
provided environmental data such as temperature, humidity and CO2 to users through the mobile
application. In the IoT-based system [29] developed for real-time IAQ surveillance, multiple gases
such as CO2 , NO2 , ethanol, methane and propane were detected using ESP8266 as the controller and
sensors MICS-6814 and MICS-6814 as the detection unit. This system, based on open source technology,
has a mobile phone application that transmits real-time notifications to users. The solution is based on
the IoT concept and is entirely wireless in a study that presented an IAQ surveillance solution that
can measure temperature, humidity, PM10 , CO2 and light intensity in real-time [30]. The Arduino
UNO emitter utilizes the ESP8266 controller and the “ThingSpeak” open-source IoT platform to record
wireless data to enable wireless Internet access. In this study, a real-time surveillance system based on
IoT architecture for PM surveillance was presented [31]. The iDust system was created using open
source technologies and low-cost sensors. System measurement data, consisting of a WEMOS D1 Wi-Fi
controller and PMS 5003 dust sensor, are transferred to users via an IoT-based implementation.

Low air quality poses a significant health threat for individuals who spend most of their
time indoors. Some pollutants such as tobacco smoke, CO, NO2 , formaldehyde, asbestos fibres,
microorganisms and allergens are known to be closely related to health problems. Temperature and
humidity monitoring are part of everyday life, but in the vast majority of buildings, real-time air quality
monitoring is not performed. In this study, air quality measurement system with an IoT-based e-nose
has been proposed for real-time, low-cost and easy-to-install air quality monitoring. With the proposed
e-nose system the ambient temperature and humidity values are measured in real-time in addition to
the polluting gases such as CO2 , CO, PM10 and NO2 . Information on the presence of these monitored
gases in excessive quantities is transmitted to users via the mobile application as a notification. This is
193
Appl. Sci. 2019, 9, 3435
a completely wireless solution developed using the ESP32 module, which integrates the IEEE 802.11
b/g/n network protocol into the IoT architecture.
The architecture of the proposed e-nose system is given in Figure 1. A real-time air quality
monitoring system provides information about the concentration of pollutants in the environment.
It provides precise and detailed information about the air quality of the living environment and helps
to plan interventions that lead to improved air quality. The e-nose air quality monitoring system in
Figure 1 consists of two parts: The first part is the detection and communication unit consisting of an
ESP32 microcontroller-based sensor array with built-in Wi-Fi; and the second is the Android/iOS-based
mobile user interface.
Figure 1. The proposed e-nose system architecture.
The ESP32 module with a built-in Wi-Fi module is used in the e-nose system created for
monitoring air quality. The low-cost and high-performance 32-bit controller is frequently preferred in
IoT applications. An ESP32 has a dual-core structure and has many internal modules such as Wi-Fi,
Bluetooth, RF, IR, CAN, Ethernet module, temperature sensor, hall effect sensor and touch sensor
needed for smart home applications. In the ESP32 module structure, the Harvard Tensilica Xtensa
LX6 32-bit Dual Core features a processor capable of operating at up to 240 MHz. The detection
unit includes the sensors GP2Y1010AU, MH-Z14, MICS-4514 and DHT22, which measure air quality
parameters such as CO2 , CO, PM10 , NO2 , temperature and humidity.
The GP2Y1010AU is a dust sensor with an analogue output system. An infrared light-emitting
diode (IRED) and a phototransistor are arranged across from each other. The IR beams reflected
from dust entering the air chamber of the sensor is detected by the phototransistor and generates a
corresponding voltage [32].
The MH-Z14A CO2 sensor module uses the non-dispersive infrared (NDIR) principle. It measures
between 0–5000 ppm, 5 ppm resolution with an accuracy of ±50 ppm. The sensor module sends the
CO2 concentration in three different output modes: Serial output (RS-232), analogue output and pulse
width modulation (PWM) [33].
The MICS-4514 is mainly used for measuring emissions from automobile exhausts but is also
used for measuring concentrations of gases such as NO2 , CO and hydrocarbons. The sensor has a
built-in heating element and a micro-sensing diaphragm on the upper side. The MICS-4514 includes
two sensor chips with independent heaters and delicate layers. One sensor chip detects oxidizing
gases (OX) and the other sensor detects reducing gases (RED) [34].
The DHT22 consists of two parts: A thermistor temperature sensor and a capacitive humidity
sensor. The DHT22 is an advanced sensor unit that provides a calibrated digital signal output. It is
equipped with an 8-bit microcontroller and has a short response time. It has a relative error of ±0.5 ◦ C
in temperature measurement and ±2% rH in humidity measurement [35]. The electronic features of
the sensors used in the IoT-based e-nose system are given in Table 1.
194
Appl. Sci. 2019, 9, 3435
Table 1. Electronic features of the sensors in the IoT-based e-nose measuring system.
ID Equipment Name Types Electronic Features

Detection range 0–10000 ppm; operating voltage: 4–6 V;
MH-Z14
1 CO2 gas sensor accuracy: ± 50 ppm ±5%; resolution: 5 ppm;
[36,37]
output Voltage: 0.4–2 V; operating temperature: 0–50 ◦ C
Detection range 1–1000 ppm (CO); 0.05–5 ppm (NO2 );
MICS-4514
2 NO2 , CO gas sensor operating voltage: 4.9–5.1 V; operating temperature:
[38–40]
−35–85 ◦ C; heating current: 58 mA
Operating voltage: 5 V; output voltage: 0.9 (no dust)–3.4 V;
GP2Y1010AU
3 Dust sensor operating current: max 20 mA; operating temperature:
[41–43]
−10 to 65 ◦ C, accuracy ±15%
Temperature range: −40 ◦ C to 80 ◦ C; humidity range: 0%
to 100%; operating voltage: 3.5–5.5 V;
Temperature and DHT22
4 operating current: 60 uA; output: serial; resolution: 0.1 ◦ C
Humidity sensor [44–46]
and ±1 rH%; accuracy: ±0.5 ◦ C and ±1 rH%;
resolution: Temperature and humidity are 16-bit.
The images of the IoT-based e-nose system used in air quality measurements and formed
from different sensors are displayed in Figure 2. The sensors used in the system are mounted in a
17 × 12 × 8 cm size sealed box per the measurement specifications. In addition to the sensors used
for measurement, the box includes a printed circuit board, a fan for airflow to the dust sensor, and a
power supply for the energy of the system.
Figure 2. Images of the IoT-based e-nose system.
The measurement system can work with a 12 V DC adapter, and additionally has a mobile use
feature when paired with a power bank that can be connected to the 5 V DC USB port. This feature
provides a great advantage for short-term measurements. When supplied from a 5 V power supply, the
measuring device draws a current of 160 mA. With a 5000 mAh power bank, it has a measuring time of
approximately 30 h. Using a wireless internet connection, it provides the opportunity to measure air
quality from many common living areas such as parks, gardens, highways, industrial areas, sports
fields, public transportation, cafes, restaurants, schools, hospitals.
The climate parameters and gas concentrations are a median of 12 measurements taken at 5 s
intervals. This minimizes the effect of erroneous measurements brought about by faulty sensors.
Users are notified by exceeding the designated gas concentration and climate parameter thresholds.
Five-minute averages are calculated and delivered to users as notifications, preventing false warnings
that may mislead users. In this e-nose system which measures air quality, no memory element is used
for data recording. The received data are recorded directly to the Blynk cloud server via the mobile
interface. The data are sent to the registered e-mail address of the user when requested. The status of
the internet connection is checked before each data transmission. When not connected, data packets
that cannot be temporarily transmitted are stored to ensure data integrity, and then, when the internet
connection is re-established, these packets are sent to the cloud server using past time tags. In this way,
data loss can be prevented by providing continuity in the data flow.
195
Appl. Sci. 2019, 9, 3435
The front panel images of the developed Android-based mobile user interface are given in Figure 3.
Control over mobile devices in IoT applications has become very common. There are many free options
available for Android and iOS devices. Blynk is one of these applications and it is an IoT platform
developed for iOS and Android applications that enables management of different controllers such
as Raspberry Pi, ESP8266, ESP32, chipKIT, Intel, LeMarker, Onion Omega, SparkFun and STM32.
Using the Blynk cloud server service, digital data such as temperature, humidity, current, voltage
measurements and control systems are stored and can be easily accessed at any time. The Blynk
graphical components (widgets) allow real-time clock and calendar (RTCC) features to be used.
Figure 3. Mobile user interface developed for the e-nose system.
Table 2 shows the cost table of the IoT-based e-nose measurement system. The total cost of the
software and hardware components required to install the system is approximately $100. Compared
to most commercially accessible non-IoT-based home air quality meters, this cost is very economical
and convenient.
Table 2. Cost table for IoT-based e-nose measurement system.
Component Cost ($)

ESP-32 Controller 5
GP2Y1010AU Sensor 5
MH-Z14 Sensor 25
MICS-4514 Sensor 20
DHT22 Sensor 3
5V Fan 3
5V Power Sup. 2
PCB 3
Plastic Box 5
Cable, Socket 5
Power Supply/Bank 15
Arduino IDE free
Blynk IoT Platform 5
Total Cost 100 $

For air quality analysis with IoT-based e-nose, data were collected in a residence for 4 days
at one-minute intervals. The housing area where air quality measurements were taken is 150 m2 .
196
Appl. Sci. 2019, 9, 3435
The room in which the measurements were made is 25 m2 and the measurements were taken from a
height of 1.5 m. The five-person house is heated by a central heating system. Although there were
changes in the number of inhabitants during the day, there were always at least two people in the
household. There was no air cleaning system in the house and the ventilation of the environment was
done by manually opening the windows.
The doors in the household were not closed during the measurement and thus the air was
homogeneous throughout the house. Ventilation was carried out once a day for a duration of one hour.
Ventilation started at 11:30 on the first day and at 07:30 on the other days.
Figure 4 displays 4-day measurement graphs showing the relationships between humidity–CO2 ,
CO–CO2 , CO2 –NO2 , CO–NO2 , PM10 –CO and PM10 –NO2 .
Figure 4. Relationship between gases during four-day measurement, (a) Humidity- CO2 , (b) CO-CO2 ,
(c) CO2 - NO2 , (d) CO-NO2 , (e) PM10 -CO, (f) CO-NO2 .
Figure 5 displays PM10 , CO2 , CO and NO2 measurements over 4 days. When the graph is examined,
it shows that the indoor PM10 , CO, NO2 and CO2 PM and gas concentrations vary depending on the
number of household inhabitants and individual activities. It is seen that natural ventilation in the
morning hours caused a decrease in CO, NO2 and CO2 gas concentrations. In addition, an increase in
PM10 is observed in the morning as a result of increased activity in the household. These increased
rates at the PM10 level are 27.1%, 32.2%, 14.4% and 12.5%, respectively.
Figure 6 displays the daily changes of temperature, humidity, PM10 , CO, CO2 and NO2 values.
One-day time, period 1 (00:00–07:30), period 2 (07:30–16:30) and period 3 (16:30–00:00) are divided
into three parts; period 1 sleeping of household inhabitants, period 2 during which the daily activities
(cooking, cleaning, etc.) of the individuals in the household are performed, and the period 3 of
eating and resting of the household inhabitants. The PM10 value seen in Figure 6a is in the order of
28.18 μg m−3 , 29.93 μg m−3 and 28.40 μg m−3 for periods 1-3, respectively. It can be seen that PM10
has the lowest value in period 1 when the household is asleep. The CO2 values in the periods are as
follows: period 1, 565 μg m−3 ; period 2, 317 μg m−3 ; and period 3 has a mean value of 297 μg m−3 .
The value of CO2 reaches the highest value in period 1 when the household inhabitants are asleep.
197
Appl. Sci. 2019, 9, 3435
Figure 5. Daily changes of PM and gases, (a) PM10 , (b) CO2 , (c) CO, (d) NO2 .
Figure 6. Hourly changes of climate parameters and gases values of day 3, (a) PM10 - CO2 , (b) NO2 -CO,
(c) temperature, (d) humidity.
When Figure 6b is examined, it is seen that NO2 concentration has average values of 53.15 μg m−3
for period 1, 41.59 μg m−3 for period 2 and 42.22 μg m−3 for period 3. It is seen that the NO2
concentration value reached the highest value in the period 1, similar to CO2 . The CO concentration
has an average of 5.48 mg m−3 in the period 1, 1.38 mg m−3 in the period 2 and 1.09 mg m−3 in the
period 3. The CO value, and also CO2 and NO2 values, are observed to be highest in the period 1 while
the household is asleep. Also, natural ventilation by opening the windows in the morning results in a
significant reduction of CO, CO2 and NO2 concentrations.
As a result of natural ventilation, the indoor concentration of CO is seen to fall from 6.51 mg m−3
to 0.4 mg m−3 , NO2 concentration from 54.7 μg m−3 to 38.6 μg m−3 and CO2 concentration from
608 μg m−3 to 282 μg m−3 .
As a result of one hour of natural ventilation, the increased gas concentrations in the closed and
unventilated environment resulted in a significant decrease of 93.8% in CO, 29.4% in NO2 and 53.6%
in CO2 .
Figure 7 shows the minimum, maximum and average values of daily gas concentrations. When
the values given in Figure 7 are examined, it is seen that the largest variance is observed in the CO
concentration. The CO values increased by nine times on day 1 and 16 times on day 3. Similarly, CO2
concentration increased by 1.36 times on day 2. NO2 , with a lower change, increased by half on day 4
and PM10 concentration increased by half on day 2.
198
Appl. Sci. 2019, 9, 3435
Figure 7. Minimum, maximum and average values of daily gas concentrations and PM, (a) PM10 ,
(b) CO2 , (c) CO, (d) NO2
The daily average concentration values of indoor CO gas are as follows: 3.21, 2.75, 2.65 and
3.40 mg m−3 , daily concentrations of NO2 , 47.9, 49.2, 45.6 and 42.8 μg m−3 , CO2 , with mean values of
500, 426, 394 and 424 μg m−3 and PM10 30.2, 31.1, 28.8 and 30.1 μg m−3 .
The correlation relationships between air quality dates have been estimated by the Konstanz
Information Miner (KNIME) “Linear Correlation” algorithms. Figure 8 shows the linear correlation
workflow of KNIME.
Figure 8. Workflow of Linear Correlation in KNIME application.
Tables 3–5 demonstrate the period 3 linear correlation matrices on 22.03.2019. Table 3 demonstrates
the period 1 linear correlation matrix. There was a positive correlation between CO–CO2 , NO2 –humidity
and NO2 –temperature at values of 0.943, 0.871, 0.755, respectively, during the period when the
household was asleep. In addition, there were 0.979, 0.911 positive and 0.941 negative correlations
between time–CO, time–CO2 and time–NO2 , respectively. There was a strong negative correlation
between climate parameters and values of CO and CO2 and a strong positive correlation between
climate parameters and values of NO2 . On the other hand, the correlation between PM10 and gas
concentrations and climate parameters were minuscule. The findings indicate that in period 1 there
was an elevated positive and negative correlation between gas concentrations and climatic parameters,
which represents sleep hours.
199
Appl. Sci. 2019, 9, 3435
Table 3. Linear correlation matrix for period 1.
Time CO NO2 CO2 Temp Humidity PM10 1

Time 1 0.979 −0.941 0.911 −0.866 −0.944 −0.251
CO 1 −0.919 0.943 −0.907 −0.969 −0.254
NO2 1 −0.841 0.755 0.871 0.245 0
CO2 1 −0.886 −0.918 −0.213
Temp. 1 0.941 0.248
Humidity 1 0.258
PM10 1 −1
Table 4 demonstrates the period 2 linear correlation matrix. There was a positive correlation
between CO–CO2 . CO–NO2 and NO2 –CO2 at values of 0.966, 0.881, 0.861, respectively, and a negative
correlation of 0.915 for temperature–humidity during this period, in which households are undertaking
domestic activities (cooking, cleaning, washing, etc.). There were strong negative correlations between
time, CO, CO2 , humidity and NO2 at −0.844, −0.752, −0.735 and −0.669 during this period. In addition,
this period’s correlations between PM10 value and gas levels and climate parameters were very low.

Time 1 −0.844 −0.669 −0.752 0.640 −0.735 −0.242
CO 1 0.881 0.966 −0.266 0.415 0.014
NO2 1 0.861 −0.098 0.254 −0.092 0
CO2 1 −0.186 0.314 −0.020
Temp. 1 −0.915 −0.327
Humidity 1 0.317
PM10 1 −1
The linear correlation matrix provided in Table 5 belongs to period 3, in which families carry out
activities such as eating and resting. There was a positive correlation between CO–NO2 and CO–CO2
at 0.833, 0.671, respectively, during this period. Furthermore, there was also a positive correlation
between time–CO, time–NO2 , time–CO2 and time–PM10 at 0.968, 0.846, 0.622 and 0.407, respectively.

Time 1 0.968 0.846 0.622 0.234 0.304 0.407
CO 1 0.833 0.671 0.213 0.296 0.379
NO2 1 0.474 0.213 0.328 0.347 0
CO2 1 0.051 0.001 0.192
Temp. 1 0.629 0.230
Humidity 1 0.169
PM10 1 −1
We can conclude that the cost of the proposed e-nose measurement system is very low compared
to commercial products sold on the market. Moreover, the majority of domestic air quality meters
available for commercial use measure only a restricted amount of air quality and climate parameters.
In this system, CO, CO2 , NO2 , PM10 , and six parameters of temperature and humidity can be
evaluated. Another significant advantage is that our measuring system is IoT-based and provides
users with real-time data transfer through its mobile interface. The major drawback of the proposed
e-nose system is that no comparison and calibration is conducted with any conventional measuring
system. The measuring precision of the system is equivalent to the measuring precision indicated by
manufacturers. To rectify this, in our future studies the measuring system will be calibrated first, then
the sensor node created and longer-term readings made. Thus, the change in IAQ through individual
activities will be disclosed more obviously.
200
Appl. Sci. 2019, 9, 3435
5. Conclusions
Although many people spend most of their lives indoors, they have very limited information
about the air quality in their environment. IAQ, which plays an especially important role in the health
of children and the elderly, should be measured and necessary ventilation measures should be taken.
In recent years, individual air quality measuring devices have been produced owing to the developing
information communication technologies, Wi-Fi based microcontrollers and low-cost sensors.
In this study, an IoT-based personalized air quality measurement and monitoring system was
proposed by using air quality sensors and DIY approach. The developed e-nose measurement
system made data measurements at one-minute intervals and recorded these data to the cloud server.
The measurement data can be monitored instantaneously via the Blynk mobile interface, and if the limit
values are exceeded, the application sends a notification to the user to take the necessary measures.
According to the four-day measurement results, the following inferences have been obtained:
• IAQ is directly related to the number of people in the household and the activities carried out in
the household.
• Activities such as cooking, sleeping, cleaning have a significant effect on CO, NO2 and CO2
gas concentrations.
• CO, NO2 , CO2 gas concentration values of the house reach the highest level during the sleep
period (period 1). During this period, PM10 concentration has the lowest value.
• PM10 concentration reaches the highest level in period 2, when daily routine tasks such as cleaning
and house arrangement are carried out.
• When the daily maximum and minimum values were taken into consideration, there was a 16-fold
daily maximum change on day 3 in CO concentration, with minimum values of 0.4 mg m−3 and
maximum 6.51 mg m−3 .
• The largest change in PM10 concentration was achieved with a minimum of 27.6 μg m−3 and a
maximum of 40.7 μg m m−3 on day 2, resulting in a 47% change.
• Although CO and CO had the lowest concentration in period 3, NO2 had the lowest concentration
in period 2.
• A rapid decrease in CO, NO2 , CO2 gas concentrations was observed from the moment of natural
ventilation, while a concurrent increase in PM10 was observed.
• In period 1, the highest positive correlations occurred between time–CO and CO–CO2 ; the lowest
negative correlations occurred between CO–humidity and time–humidity.
• In period 2, the largest positive correlations occurred between CO–CO2 and CO–NO2 , whereas
the smallest negative correlations occurred between temp–humidity and time–CO.
• The highest positive correlations between time–CO, time–NO2 and CO–NO2 occurred in period
3. There was no negative correlation between the weather parameters and gas concentrations
during this period.
IAQ has been observed to change depending on the daily activities of the inhabitants. Low air
quality is undoubtedly an important parameter that directly affects our health. Even simple measures
such as opening only windows to reduce the concentration of harmful gases in the environment can
significantly improve IAQ. The numerical data obtained show that the e-nose system is a feature that
can contribute to a healthier living environment. In future studies, it is planned to model the effects of
individual activities on IAQ by using the e-nose air quality measurement system used in this study.
Author Contributions: Conceptualization, M.T. and H.G.; methodology, H.G.; software, M.T.; validation, M.T.;
formal analysis, M.T.; investigation, M.T. and H.G.; visualization, M.T. and H.G.; resources, H.G.; data curation,
M.T.; writing—original draft preparation, M.T.; review and editing, H.G.
201
Appl. Sci. 2019, 9, 3435
References
1. Sharma, A.; Mitra, A.; Sharma, S.; Roy, S. Estimation of Air Quality Index from Seasonal Trends Using Deep
Neural Network. In Proceedings of the International Conference on Artificial Neural Networks, Island of
Rhodes, Greece, 5–7 October 2018; pp. 511–521.
2. Leung, D.Y. Outdoor-indoor air pollution in urban environment: Challenges and opportunity. Front. Environ. Sci.
2015, 2, 69. [CrossRef]
3. De Gennaro, G.; Dambruoso, P.R.; Loiotile, A.D.; Di Gilio, A.; Giungato, P.; Tutino, M.; Porcelli, F. Indoor air
quality in schools. Environ. Chem. Lett. 2014, 12, 467–482. [CrossRef]
4. Chen, M.; Yang, J.; Hu, L.; Hossain, M.S.; Muhammad, G. Urban healthcare big data system based on crowd
sourced and cloud-based air quality indicators. IEEE Commun. Mag. 2018, 56, 14–20. [CrossRef]
5. Marques, G. Ambient Assisted Living and Internet of Things. In Harnessing the Internet of Everything (IoE) for
Accelerated Innovation Opportunities; IGI Global: Hershey, PA, USA, 2019; pp. 100–115.
6. Wei, W.; Ramalho, O.; Mandin, C. Indoor air quality requirements in green building certifications.
7. Neira, M.; Prüss-Ustün, A.; Mudu, P. Reduce air pollution to beat NCDs: From recognition to action. Lancet
2018, 392, 1178–1179. [CrossRef]
8. Seguel, J.M.; Merrill, R.; Seguel, D.; Campagna, A.C. Indoor air quality. Am. J. Lifestyle Med. 2017, 11, 284–295.
[CrossRef] [PubMed]
9. Kampa, M.; Castanas, E. Human health effects of air pollution. Environ. Pollut. 2008, 151, 362–367. [CrossRef]
10. Annesi-Maesano, I.; Baiz, N.; Banerjee, S.; Rudnai, P.; Rive, S.; Sinphonie Group. Indoor air quality and
sources in schools and related health effects. J. Toxicol. Environ. Health Part B 2013, 16, 491–550. [CrossRef]
11. Tong, Z.; Chen, Y.; Malkawi, A.; Liu, Z.; Freeman, R.B. Energy saving potential of natural ventilation in
China: The impact of ambient air pollution. Appl. Energy 2016, 179, 660–668. [CrossRef]
12. Chen, Y.; Tong, Z.; Wu, W.; Samuelson, H.; Malkawi, A.; Norford, L. Achieving natural ventilation potential
in practice: Control schemes and levels of automation. Appl. Energy 2019, 235, 1141–1152. [CrossRef]
13. Jones, A.P. Indoor air quality and health. Atmos. Environ. 1999, 33, 4535–4564. [CrossRef]
14. Gökozan, H.; Taştan, M.; Sarı, A. Smart Cities and Management Strategies; 2017 Socio-Economic Strategies;
Lambert Academic Publishing: Riga, Latvia, 2017; pp. 115–123, ISBN 978-3-330-06982-4.
15. Gubbi, J.; Buyya, R.; Marusic, S.; Palaniswami, M. Internet of Things (IoT): A vision, architectural elements,
and future directions. Future Gener. Comput. Syst. 2013, 29, 1645–1660. [CrossRef]
16. Rathore, P.; Rao, A.S.; Rajasegarar, S.; Vanz, E.; Gubbi, J.; Palaniswami, M. Real-time urban microclimate
analysis using internet of things. IEEE Internet Things J. 2018, 5, 500–511. [CrossRef]
17. Taştan, M. IoT Based Wearable Smart Health Monitoring System. Celal Bayar Univ. J. Sci. 2018, 14, 343–350.
[CrossRef]
18. Pitarma, R.; Marques, G.; Caetano, F. Monitoring indoor air quality to improve occupational health. In New
Advances in Information Systems and Technologies; Springer: Cham, Switzerland, 2016; pp. 13–21.
19. Yang, Y.; Liu, L.; Xu, C.; Li, N.; Liu, Z.; Wang, Q.; Xu, D. Source apportionment and influencing factor analysis
of residential indoor PM2. 5 in Beijing. Int. J. Environ. Res. Public Health 2018, 15, 686. [CrossRef] [PubMed]
20. Canha, N.; Lage, J.; Galinha, C.; Coentro, S.; Alves, C.; Almeida, S. Impact of Biomass Home Heating,
Cooking Styles, and Bread Toasting on the Indoor Air Quality at Portuguese Dwellings: A Case Study.
Atmosphere 2018, 9, 214. [CrossRef]
21. Rumchev, K.; Soares, M.; Zhao, Y.; Reid, C.; Huxley, R. The association between indoor air quality and adult
blood pressure levels in a high-income setting. Int. J. Environ. Res. Public Health 2018, 15, 2026. [CrossRef]
22. Wang, S.; Bolling, K.; Mao, W.; Reichstadt, J.; Jeste, D.; Kim, H.C.; Nebeker, C. Technology to Support Aging
in Place: Older Adults’ Perspectives. Healthc. Multidiscip. Digit. Publ. Inst. 2019, 7, 60. [CrossRef]
23. Alvarez-Campana, M.; López, G.; Vázquez, E.; Villagrá, V.; Berrocal, J. Smart CEI moncloa: An iot-based
platform for people flow and environmental monitoring on a Smart University Campus. Sensors 2017,
17, 2856. [CrossRef]
24. Cho, H. An Air Quality and Event Detection System with Life Logging for Monitoring Household
Environments. In Smart Sensors at the IoT Frontier; Springer: Cham, Switzerland, 2017; pp. 251–270.
25. Wong, M.; Yip, T.; Mok, E. Development of a personal integrated environmental monitoring system. Sensors
2014, 14, 22065–22081. [CrossRef]
202
Appl. Sci. 2019, 9, 3435
26. Salamone, F.; Danza, L.; Meroni, I.; Pollastro, M. A Low-Cost Environmental Monitoring System: How to
Prevent Systematic Errors in the Design Phase through the Combined Use of Additive Manufacturing and
Thermographic Techniques. Sensors 2017, 17, 828. [CrossRef] [PubMed]
27. Popa, A.; Hnatiuc, M.; Paun, M.; Geman, O.; Hemanth, D.J.; Dorcea, D.; Ghita, S. An Intelligent IoT-Based
Food Quality Monitoring Approach Using Low-Cost Sensors. Symmetry 2019, 11, 374. [CrossRef]
28. Marques, G.; Pitarma, R. An indoor monitoring system for ambient assisted living based on internet of things
architecture. Int. J. Environ. Res. Public Health 2016, 13, 1152. [CrossRef] [PubMed]
29. Marques, G.; Pitarma, R. A cost-effective air quality supervision solution for enhanced living environments
through the internet of things. Electronics 2019, 8, 170. [CrossRef]
30. Marques, G.; Pitarma, R. Monitoring health factors in indoor living environments using internet of things.
In Proceedings of the World Conference on Information Systems and Technologies, Madeira, Portugal, 11–13
April 2017; pp. 785–794.
31. Marques, G.; Roque Ferreira, C.; Pitarma, R. A system based on the Internet of Things for real-time particle
monitoring in buildings. Int. J. Environ. Res. Public Health 2018, 15, 821. [CrossRef] [PubMed]
32. GP2Y1010AU0F: Compact Optical Dust Sensor. Available online: http://www.socle-tech.com/doc/IC%
20Channel%20Product/sharp%20products.pdf (accessed on 11 June 2019).
33. Zhengzhou Winsen Electronics Technology CO. Ltd. MH-Z14 CO2 Module Datasheet. 2013. Available
online: http://www.futurlec.com/Datasheet/Sensor/MH-Z14.pdf (accessed on 17 June 2019).
34. MICS-4514 Datasheet. Available online: http://files.manylabs.org/datasheets/MICS-4514.pdf (accessed on 11
June 2019).
35. Aosong Electronics CO. Ltd. DHT22 Temperature and Humidity Sensor Datasheet. Available online:
https://www.sparkfun.com/datasheets/Sensors/Temperature/DHT22.pdf (accessed on 13 June 2019).
36. Hong, C.S.; Ghani, A.S.A.; Khairuddin, I.M. Development of an Electronic Kit for detecting asthma in
Human Respiratory System. In Proceedings of the IOP Conference Series: Materials Science and Engineering,
Yogyakarta, Indonesia, 7–8 December 2017; IOP Publishing: Bristol, UK, 2018; Volume 319, p. 012040.
37. Bimaridi, A.; Putra, K.D.; Djunaedy, E.; Kirom, M.R. Assasment of Outside Air Supply for Split AC
system–Part A: Affordable Instrument. Procedia Eng. 2017, 170, 248–254. [CrossRef]
38. Suárez, J.I.; Arroyo, P.; Lozano, J.; Herrero, J.L.; Padilla, M. Bluetooth gas sensing module combined with
smartphones for air quality monitoring. Chemosphere 2018, 205, 618–626. [CrossRef]
39. McKercher, G.R.; Vanos, J.K. Low-cost mobile air pollution monitoring in urban environments: A pilot study
in Lubbock, Texas. Environ. Technol. 2018, 39, 1505–1514. [CrossRef]
40. Nguyen, T.N.T.; Ha, D.V.; Do, T.N.N.; Nguyen, V.H.; Ngo, X.T.; Phan, V.H.; Bui, Q.H. Air pollution monitoring
network using low-cost sensors, a case study in Hanoi, Vietnam. In Proceedings of the IOP Conference
Series: Earth and Environmental Science, Pan Pacific Hanoi, Vietnam, 23–24 January 2019; IOP Publishing:
Bristol, UK, 2019; Volume 266, p. 012017.
41. Liu, H.Y.; Schneider, P.; Haugen, R.; Vogt, M. Performance assessment of a low-cost PM2.5 sensor for a near
four-month period in Oslo, Norway. Atmosphere 2019, 10, 41. [CrossRef]
42. Budde, M.; El Masri, R.; Riedel, T.; Beigl, M. Enabling low-cost particulate matter measurement for
participatory sensing scenarios. In Proceedings of the 12th International Conference on Mobile and
Ubiquitous Multimedia, Lulea, Sweden, 2–5 December 2013; p. 19.
43. Carminati, M.; Ferrari, G.; Sampietro, M. Emerging miniaturized technologies for airborne particulate matter
pervasive monitoring. Measurement 2017, 101, 250–256. [CrossRef]
44. Muangprathub, J.; Boonnam, N.; Kajornkasirat, S.; Lekbangpong, N.; Wanichsombat, A.; Nillaor, P. IoT and
agriculture data analysis for smart farm. Comput. Electron. Agric. 2019, 156, 467–474. [CrossRef]
45. Xie, J.; Gao, P.; Wang, W.; Xu, X.; Hu, G. Design of Wireless Sensor Network Bidirectional Nodes for Intelligent
Monitoring System of Micro-irrigation in Litchi Orchards. IFAC-PapersOnLine 2018, 51, 449–454.
46. Wijaya, D.R.; Sarno, R.; Zulaika, E.; Sabila, S.I. Development of mobile electronic nose for beef quality
monitoring. Procedia Comput. Sci. 2017, 124, 728–735. [CrossRef]
203
applied
sciences
Article
An Internet of Things-Based Environmental Quality
Management System to Supervise the Indoor
Laboratory Conditions
Gonçalo Marques and Rui Pitarma *
Unit for Inland Development, Polytechnic Institute of Guarda, Avenida Doutor Francisco Sá Carneiro N◦ 50,
6300-559 Guarda, Portugal; [email protected]
Received: 19 December 2018; Accepted: 24 January 2019; Published: 28 January 2019
Abstract: Indoor air quality (IAQ) is not only a determinant of occupational health but also influences
all indoor human behaviours. In most university establishments, laboratories are also used as
classrooms. On one hand, indoor environment quality (IEQ) conditions supervision in laboratories
is relevant for experimental activities. On the other hand, it is also crucial to provide a healthy
and productive workplace for learning activities. The proliferation of cost-effective sensors and
microcontrollers along with the Internet of Things (IoT) architectures enhancements, enables the
development of automatic solutions to supervise the Laboratory Environmental Conditions (LEC).
This paper aims to present a real-time IEQ-laboratory data collection system-based IoT architecture
named iAQ Plus (iAQ+). The iAQ+ incorporates an integrated Web management system along
with a smartphone application to provide a historical analysis of the LEC. The iAQ+ collects IAQ
index, temperature, relative humidity and barometric pressure. The results obtained are promising,
representing a meaningful contribution for IEQ supervision solutions based on IoT. iAQ+ supports
push notifications to alert people in a timely way for enhanced living environments and occupational
health, as well as a work mode feature, so the user can configure setpoints for laboratory mode
and schoolroom mode. Using the iAQ+, it is possible to provide an integrated management of data
information of the spatio-temporal variations of LEC parameters which are particularly significant
not only for enhanced living environments but also for laboratory experiments.
Keywords: IAQ; enhanced living environments; IEQ; IoT; smart cities; LEC
1. Introduction
Indoor environment quality (IEQ) in buildings incorporates indoor air quality (IAQ), acoustics,
thermal comfort, and lighting [1]. It is well known that poor IEQ has a negative effect on occupational
health, particularly on children and old people.
Poor IAQ is a significant problem which affects particularly the underprivileged people in the
world, as they remain most exposed, presenting itself as a critical problem for global health such as
tobacco use or the problem of sexually transmitted diseases [2]. The Environmental Protection Agency
(EPA) is responsible for indoor and outdoor air quality supervision in the United States. The EPA
assessment recognises that IAQ pollutants concentration can be up to 100 times greater when compared
with the outdoor pollutants concentration levels and established poor air quality, which are in the top
five environmental risks to global health [3].
Indoor living environments include several types of spaces and workplaces such as offices, hospitals,
public service centres, schools, libraries, leisure spaces and also the cabins of vehicles [4]. In particular,
schools are an essential place to monitor. Typically, the large number of occupants, the time spent indoors,

Appl. Sci. 2019, 9, 438
and the higher density of occupants justifies the need to develop automatic supervision systems to provide
a healthful and productive workplace for students, teachers and school staff [5].
Buildings are responsible for 40% of the global energy consumption and 30% of the CO2 emissions.
Thermal comfort assurance is related to a substantial percentage of the referred energy consumption.
The adoption of personalised conditioning systems is apparently a reliable approach to improve user
acceptability with the environment thermal conditions since thermal comfort is a complicated subject
with several interrelated aspects that need to be understood [6].
Currently, IEQ in buildings is based on random sampling. However, these procedures are only
providing information relating to a specific sampling and are devoid of details of spatio-temporal
variations, which are particularly relevant in laboratory tests such as thermography experiments.
The fundamental concept of the Internet of Things (IoT) is the pervasive presence of several types
of devices with communication and collaboration skills between them to reach a mutual purpose [7].
The IoT will not only improve everyday life behaviours and activities in different areas such as
smart homes, assisted living environments and smart-health, but also in present innovative data and
computational resources to develop novel software solutions [8]. One of the fields where the IoT plays
a significant role is IAQ monitoring [9]. The rapid increasing of Information and Communication
Technologies (ICTs) and IoT proliferation offer significant opportunities in the development of
healthcare information systems. However, challenges still exist in achieving safety, security and
privacy for healthcare applications [10,11].
Ambient Assisted Living (AAL) is a research area which studies the conceptualisation of an
ecosystem of diverse kinds of sensors, computers, mobile gadgets, wireless networks and software
development for enhanced personal healthcare supervision and telemedicine solutions [12]. At 2050,
a significant proliferation of diseases will lead to high healthcare costs, lack of caregivers, dependency
and notable social impact, because 20% of the world population will be aged 60 or over [13].
Most people (87%) prefer to stay in their houses and pay a significant charge for nursing care [14].
However, there are numerous difficulties in AAL solutions conceptualisation and development
such as data structure, user-interface, human–computer communication, ergonomics, usability and
accessibility [15]. Furthermore, the AAL solutions adopted by the aged population is influenced by
social and ethical obstacles. Therefore, privacy and confidentiality are fundamental to AAL solutions’
development. Actually, technology should not substitute human care but must be seen as a useful
supplement. IoT and AAL research can be conducted side by side in order to present multiple
advantages to various healthcare activities such as not only the identification, authentication and
tracking of objects and patients but also in automated data acquisition, consulting and storage [16].
The “smart city” approach has recently been proposed as an important method to include current
urban production circumstances in a mutual framework especially to focus the ICTs influence in the
last 20 years on city performance [17]. The smart city concept is associated with a recent approach
to decrease the obstacles caused by the growth of urban population and rapid urbanisation [18].
The interoperability of heterogeneous technologies and devices is a significant challenge in smart cities;
IoT can offer the interoperability needed to create unified urban-scale ICT [19].
IoT incorporates notable features to develop innovative real-life solutions and services for smart
cities [20], particularly for environmental quality supervision. The smart home is an indispensable
element in smart cities [21]. In the future, smart homes should incorporate IoT wearable gadgets
managed by smartphone applications and powered by miniaturised built-in sensors [22]. Smart homes
follow the AAL paradigm, which enables access to health-care services for patients and medical staff as
well as other AAL applications [23]. The use of the paradigm of cognitive dynamic systems associated
with IoT architectures in intelligent houses is proposed by Reference [24]. According to this study,
the smart home can benefit from several building blocks of the cognitive dynamic system (CDS) along
with the incorporation of IoT.
This paper aims to present an environmental quality solution based on IoT to supervise
Laboratory Environmental Conditions (LEC) named iAQ+. This low-cost wireless solution for
206
Appl. Sci. 2019, 9, 438
IEQ supervision incorporates mobile computing technologies for data consulting, easy installation,
significant notifications for enhanced living conditions, and laboratory activities.
2. Related Work
Several studies have shown the IAQ importance in schoolrooms. The IAQ of public primary
Portuguese schools was analysed in 73 primary classrooms in Porto [25]. This study indicates
that IAQ deficiencies can persist in classrooms with pollutant sources and defective ventilation.
Therefore, pollutant source regulation procedures are the most effective methods for the prevention of
unfavourable health impacts on children in scholar institutions. The effectiveness of manual airing
strategies on the IAQ of naturally ventilated Italian schools was presented by Reference [26]. A study
conducted in New York State primary and secondary schools organised by 501 teachers concluded
that many classroom characteristics are probably associated with bad IAQ and over 40% of the
teachers described a minimum of one health symptom associated with the building construction [27].
A review of the association between IAQ and its consequences on respiratory wellbeing in Malaysian
students was proposed by Reference [28]. Despite the relatively small-scale epidemiologic evidence,
Malaysian research proposes effective and relatively consistent evidence among IAQ and children’s
respiratory health. Another study that characterises the levels of several indoor air contaminants at
scholar institutions in Hong Kong, correlates the calculated concentrations with proper standards,
and recommend methods to decrease the exposure of students to unwanted pollutants was proposed
by Reference [29].
On the one hand, in most university establishments, laboratories have a large number of polluting
sources. On the other side, these spaces are very often used as classrooms. Therefore, university
laboratories need to be monitored for two purposes: as a classroom (IAQ parameters, including thermal
comfort) and/or to ensure different conditions for sampling and performing experiments with reliable
quality and data. Most people consider thermal comfort with higher importance when compared with
visual and acoustic comfort and IAQ [30]. The comfort temperature might be as low as 17 ◦ C and
as high as 30 ◦ C. Thermal comfort is influenced by several factors such as air temperature, radiant
temperature, air velocity, humidity, clothing insulation and metabolic heat. The first four factors can
be measured and the last two are personalised factors [31]. Although, for laboratory experiments,
the recommendation is 23 ◦ C (±5 ◦ C) and <70 % RH for temperature and relative humidity respectively.
A study on the thermal comfort in a Portuguese school was presented by Reference [32].
Multiple studies on environmental quality supervision are accessible in the literature. This section
presents several prominent low-cost solutions which include not only open-source but also mobile
computing technologies.
A wireless sensor network (WSN) low-cost solution for proper greenhouse pepper cultivation
which incorporates proper supervision methods such as remote administration for drip irrigation and
equipment control was proposed by Reference [33].
A real-time WSN architecture for environmental supervision which provides acoustic levels,
temperature, relative humidity and particulate matter concentration data collection for smart cities
was proposed by Reference [34].
A WSN approach for temperature distribution supervision in large-scale indoor environments
was proposed by Reference [35]. This methodology aims to enhance the quality of the data transmitted
by wireless signals, classify the temperature distribution model and improve the allocation of supplied
air measure flow levels to various supply air terminals which fulfill the area taking into account the
temperature distribution model.
Numerous IoT solutions for IAQ supervision which merge open-source technologies not only for
processing and data transmission and sensors for data collection but also to provide data consulting
from distinct areas at the same time using Web and smartphone applications in real time are proposed
by References [36–43].
207
Appl. Sci. 2019, 9, 438
The iAQ+ solution proposes a valuable instrument for enhanced living environments in smart
cities. The advantages for well-being, comfort and productivity of healthy IAQ levels can be enhanced
by reducing the pollutant concentration when the ventilation is still unchanged [44]. Consequently,
the authors propose a wireless system which incorporates an ESP8266 module which performs the
IEEE 802.11 b/g/n networking protocol. The ESP8266 support built-in Wi-Fi technology, therefore,
this module is used not only as a processing unit but also for data transmission. For data consulting,
this solution uses a smartphone application developed using SWIFT language for the iOS operating
system (Figure 1).
Taking into account the IAQ influence on health, the development of a low-cost, open-source
supervision system is a trending idea. Therefore, several monitoring solutions have been created [45–52].
Regarding the quality and important contribution of the referred solutions, a review of these studies is
presented in Table 1.
Table 1. A summary comparison of real-time IAQ monitoring studies.

Data Easy
Authors MCU Sensors Architecture Low Cost Open-Source Connectivity Notifications
Access Installation
P. Srivatsa and A. Raspberry √ √
CO2 WSN/IoT Wi-Fi Web × ×
Pandhare [45] Pi
F. Salamone et al. Arduino √ √
CO2 WSN ZigBee × × ×
[46] UNO
CO, CO2 , PM, √
S. Bhattacharya
Waspmote Temperature, WSN × ZigBee Desktop × ×
et al. [47]
Relative Humidity
Temperature,
F. Salamone et al. Arduino Relative Humidity, √ √ ZigBee /
IoT Mobile × ×
[48] UNO CO2, Ligth, Air BLE
velocity
Temperature, √ √
Wang, S.K
Arduino Relative Humidity, WSN ZigBee Desktop × ×
et al. [49]
CO2
Jen-Hao Liu TI CO, Temperature, √ √
WSN ZigBee × × ×
et al. [50] MSP430 Relative Humidity
CO, Temperature, √
J.Kang and K. TI
Relative Humidity, IoT × ZigBee × × ×
Hwang [51] MSP430
VOC, PM
CO2 , CO, SO2,
Benammar M. et Raspberry NO2 , O3 , Cl2 , CO, √
IoT/WSN × ZigBee/Ethernet × × ×
al [52] Pi Temperature,
Relative Humidity
√
MCU: microcontroller; : apply; ×: not apply.
Figure 1. iAQ+ architecture.
Proposed iAQ+ Management System

Considering the importance of developing a cost-effective system, only the Bosch BME680
sensor was chosen. This sensor was selected since it is an integrated environmental sensor produced
particularly for portable applications and wearables where dimension and low power consumption
208
Appl. Sci. 2019, 9, 438
are essential requirements. This sensor provides temperature, relative humidity, barometric pressure
and qualitative air quality, the main parameters that should be supervised for both school and
laboratory scenarios. The proposed system is composed of a hardware prototype for environmental
data acquisition and a Web portal and mobile application compatible for data access (Figure 1).
The iAQ+ is based on open-source technologies and is a totality Wi-Fi solution with various benefits
when related to current solutions, such as its modularity, scalability, low-cost, real-time notifications
and easy installation. The data collected is stored in a SQL Server database management system and a
smartphone application or a Web portal can be used for data consulting. The iAQ+ incorporates the
ESP8266 microcontroller with built-in Wi-Fi communication technology and has been tested in the
context of infrared thermography experiments.
Taking into consideration the SQL Server native .NET framework integration feature, this database
management system was chosen. Furthermore, the SQL Server incorporates several advantages such
as data-recovery features, professional management utilities and built-in data compression and data
encryption features. The SQL Server supports easy administration and audit functionalities for
enhanced data management.
3. Laboratory Environmental Conditions

LEC supervision is critical from a quality perspective to provide a consistent and regulated state
of control for equipment and samples for proper lab activities. LEC is influenced by the building
envelope, thermal loads, obstructions and buildings climatisation system, which can produce air flow
patterns of hot and cold spots. In some cases, these hot and cold spots can produce environmental
conditions variations during experiment activities which lead to unwanted effects on samples as well
as manipulation of the testing equipment’s samples. Therefore, the LEC must be monitored in real-time
not only to assure proper regulation and maintenance of indoor conditions but also to correlate the
collected samples with these conditions.
On the one hand, the laboratory activities must be monitored and stored to ensure that they are
stable when the test is conducted and at the data collection moment as they influence the quality
of the results. LEC has a significant influence on test results and on the precision and reliability of
testing records that remain concerned with environmental parameters. On the other hand, monitored
results must be saved in the laboratory information management system that should provide the
following requisites: 1) must incorporate authentication and access control for data safety; 2) must
incorporate data recovery and data adulteration protection methods; 3) must be manipulated in an
environment that satisfies the provider or laboratory specs or, in the situation of non-computerized
methods, provides conditions which preserve the precision of manual reporting and transcription;
4) must be managed to guarantee the information integrity; and 5) must implement reporting system
crashes support to allow prompt corrective operations.
In general, temperature, humidity and barometric pressure are assumed as the main conditions to
be monitored for enhanced laboratory environments. The recommendation for LEC is 23 ◦ C (±5 ◦ C)
and <70 % RH for temperature and relative humidity respectively. For temperature data collection,
the measurements must be made with calibrated sensors that must be placed far from the equipment
under analysis to anticipate every heating consequence which can lead to inaccurate ambient samples.
Regarding the humidity measurements, the data collected must be done at the same altitude from
the ground as the equipment under test and, preferably, in a similar place if possible. The barometric
levels can be consulted from local airports. However, for critical laboratory experiences, barometer
sensors are required and must be placed in the laboratory to detect indoor building environmental
limitations and barometric fluctuations.
IAQ has a significant influence on LEC [53–56]. On the one hand, IAQ should be supervised
in order to provide a healthy and productive workplace for the researchers. On the other hand,
IAQ should be monitored to minimise the impact the laboratory experiments’ samples. For instance,
209
Appl. Sci. 2019, 9, 438
IQ is extremely significant in the clinical embryology laboratory activities and is recognised as a

significant parameter which influences in vitro fertilisation (IVF) success levels [57].

Taking into account, not only the necessity to keep a healthful and productive workplace for
the students, teachers and the school staff, but also to provide a consistent and regulated state
of control for laboratory activities, the iAQ+ solution has been created by the authors. The iAQ+
solution provides a low-cost and reliable method for IAQ supervision which incorporates easy
configuration and easy installation features. This system is a low-cost, reliable system that can
be easily configured and installed by the average user without supporting the cost of an installation
done by certified professionals.
The main objective is to provide proper supervision of the LEC, such as air temperature,
humidity and barometric pressure and a qualitative air quality index. Therefore, the authors
selected a cost-effective BOSCH BME680 sensor, a four-in-one multi-functional microelectromechanical
system (MEMS) environmental sensor which integrates a Volatile Organic Compounds (VOC) sensor,
temperature sensor, humidity sensor and barometer. The DFRobot Gravity BME680 environmental
sensor was used as it provides a Gravity I2C connector, is plug & play and is easy to connect.
The iAQ+ is based on a microcontroller with built-in Wi-Fi compatibility, a FireBeetle ESP8266
(DFRobot). Figure 2 presents the hardware prototype developed by the authors. In this section,
the hardware and software development materials and methods will be discussed.
Figure 2. iAQ+ prototype.
A brief introduction of each component used is shown below.
• FireBeetle ESP8266 is 32-bit Tensilica L106 microcontroller module which supports IEEE802.11
b/g/n WiFi (2.4 GHz~2.5 GHz). This module support one 10-bit analogue input, 10 digital inputs
which incorporate multiple interfaces such as SPI, I2C, IR, and I2S. The clock speed is 80MHz and
can reach a maximum 160MHz; in addition, it includes a 50KB SRAM and 16MB flash memory.
It supports a low-power-consumption mode of 46uA and the operating voltage is 3.3 V.
• DFRobot Gravity BME680 is an I2C environmental VOC sensor, temperature sensor, humidity
sensor and barometer. It supports an input voltage of 3.3–5.0 V; the operating current consumption
is 5 mA without air quality sensing and 25 mA with air quality features. This sensor module
size is 30 × 22 mm / 1.18 × 0.87 inches. The temperature range is from -40 ◦ C to +85 ◦ C with
a precision of ±1.0 ◦ C (0–65 ◦ C). The humidity range is from 0 to 100% r.H with a precision of
a ±3% r.H. (20–80% r.H., 25 ◦ C). The atmospheric pressure measurement range is from 300 to
1100 hPa with a precision of ±0.6 hPa (300–1100 hPa, 0–65 ◦ C).
• DFRobot Buzzer Module is a buzzer module that supports an input voltage of 3.3–5.0 V.
210
Appl. Sci. 2019, 9, 438
• 5V Green LED—a 5 V green LED is used to notify the end-user of a good IEQ conditions.
• 5V Red LED—a 5 V red LED is used to notify the end-user of poor IEQ conditions.
The BME680 sensor calculates the sum of VOCs in the surrounding air to provide qualitative air
quality data. This sensor incorporates a background auto-calibration feature in order to provide reliable
IAQ qualitative data. This process regards the recent measurement records to guarantee that IAQ index
~25 matches to typical good air and IAQ index ~250 states for typical polluted air. The sensor output
resistance value varies according to VOCs concentrations, the higher the concentration of reducing
VOCs, the lower the resistance and vice versa. The IAQ qualitative range is from 0 to 500 (the larger,
the worse). Table 2 illustrates the IAQ qualitative index meaning of the selected sensor.
Table 2. BME680 qualitative IAQ index meaning.
IAQ index Air Quality

0–50 Good air quality
51–100 Normal air quality
101–150 Little poor air quality
151–300 Poor air quality
201–300 Bad air quality
301–500 Very bad air quality
The mobile application, designated as iAQ+Mobile, was created using SWIFT programming
language in XCode IDE (Integrated Development Environment), and the minimum requirement is
the iOS 12. The iAQ+Mobile incorporates significant highlights as it allows not only real-time data
access of the last temperature, humidity, air pressure and qualitative IAQ index information but also
to receive real-time warnings to notify the occupants when the IEQ has severe deficiencies (Figure 3).
The smartphone application also allows viewing the collected data in chart form.
Figure 3. iAQ+Mobile.
The iAQ+ was developed by the authors to be a centralised supervision solution. Therefore, the
Web application, iAQ+Web, was built in ASP.NET C# using the Visual Studio IDE. Using the iAQ+Web,
the build manager can consult the IEQ data in real time. This Web portal saves the IEQ conditions
history for future analysis (Figure 4).
211
Appl. Sci. 2019, 9, 438

Figure 4. iAQ+Web.
The iAQ+Web allows historical data export for enhanced reporting and analysis. The Web portal
supports hourly, daily and monthly charts and tables of the monitored data for enhanced building
management and audits.
To support the two main activities in the university laboratories, the laboratory experiments
and teaching activities, the iAQ+ software (Web and smartphone) supports two functional modes,
the laboratory mode and the schoolroom mode. The iAQ+ software allows the user to configure
setpoints for both modes and the laboratory manager can change the status mode using the Web or the
smartphone app. The smartphone app is an agile and effective way to change the status. However,
the software allows the end-user to schedule the working modes (Figure 5).
Figure 5. iAQ+Mobile Settings.
212
Appl. Sci. 2019, 9, 438

In Portugal, the majority of indoor environments have natural ventilation. The nature of dwelling,
construction, heating and ventilation assumes an important influence on the air permeability variations.
It is expected that over 66% of business/services constructions which support natural ventilation
mechanisms are notably airtight, and the other 33% have a tendency to be leakier.
For testing purposes, one laboratory of a Portuguese university were on-site supervised using one
iAQ+ prototype. Figure 6a presents the experiment conducted by the authors of LEC supervision of
thermography activities. As in most laboratories, the supervised space incorporates natural ventilation
and does not have dedicated ventilation slots on the facades. The indoor air is reheated and recirculated
through a couple of standard air–water fan-coils of the heating system, and the air exchange is
performed through infiltrations and by opening windows manually. The outdoor air is employed to
afford ventilation, to reduce the temperature or when the occupants detect the severe or disturbing
odour; therefore, the IAQ is frequently deficient.
(a) (b)
Figure 6. Laboratory layout: (a) experiments research area (e.g. thermography experiments supported
by iAQ+); (b) classroom learning space (teaching support).
The iAQ+ was powered by the power grid using a 230V-5V AC-DC 2A power supply. The LEC
exposure data were collected in thermography experiments which showed that indoor conditions can
be different from the recommended values and can affect the integrity of the collected data.
The tests conducted show the system capability not only to keep a healthful and productive
workplace for the students, teachers and the school staff but also to provide a consistent and regulated
state of control for laboratory activities. Figure 6b represents the laboratory space used as a schoolroom.
Figure 7 presents a sample of the graphics of the results achieved in the tests conducted. It should
be noted that the graphs displayed the results obtained in the monitored rooms with induced
simulations using tobacco smoke.
On one hand, monitoring environmental conditions and maintaining laboratory temperature and
humidity requirements is extremely important for high-quality experimental activities. On the other
hand, supervising IEQ is significant in creating a healthy and productive workplace for learning and
teaching. Using the smartphone application, the user can carry all the monitored data in his pocket.
This solution not only supports alerts and setpoints configuration for real-time notifications via
e-mail, SMS and push notifications but also provides an integrated dashboard for the monitored
real-time collected data. Quite apart from this, the iAQ+ incorporates a built-in visual and audio alarm,
two LED’s and a buzzer respectively. When the configurated setpoints are met, the correspondent
LED is triggered, green to inform the good IEQ status and red to notify the poor IEQ status along with
the buzzer activation. In the last case, the buzzer will ring for 30 seconds with the aim to notify the
occupants to act in real-time to enhance the indoor conditions.
213
Appl. Sci. 2019, 9, 438
Figure 7. Results of IAQ index, air temperature and relative humidity obtained in the experiments
conducted in a real laboratory environment: (a) IAQ index; (b) relative humidity (%); (c) air temperature (◦ C).
The data collected by the system is analysed before being inserted into the database. If the data
exceeds the parameterised limits, the user is notified and an email or SMS is triggered (Figure 8).
The SMS notification is performed using Twilio, a cloud communication platform, as a service.
This service makes it possible to programmatically send an SMS and other communication functions
using its Web service APIs. The push notifications are performed using the Firebase Cloud Messaging
that is a cross-platform solution for messages and notifications for Android, iOS and Web applications.
The real-time alerts promote behaviour changes. In fact, these messages alert the user to act in
real-time to perform actions to increase IAQ. On the other hand, with this real-time feature, the building
manager can understand when the recurrent unhealthy cases are detected and implement new
adjustments to prevent them. Consecutively, iAQ+ provides the requirements to act in real-time
for enhanced living environments and laboratory conditions. The alert indicators architecture is shown
in Figure 9.
214
Appl. Sci. 2019, 9, 438
(a) (b)
Figure 8. Examples of SMS (a) and E-mail (b) notifications.

Figure 9. Alerts indicators architecture.
Firstly, iAQ+ can help in the maintenance of LEC by providing an integrated management system
that supports daily records of historical behaviours and variations as well as remote notifications when
the configurated setpoints are met. Secondly, this system can not only be used by the building manager
to detect unhealthy situations in real-time but also as a decision-making tool to address behavioural
changes to promote IEQ for enhanced productive environments.
From a quality point of view, supervising the laboratory environment is crucial to achieving a
continuous LEC in the laboratory activities. Irregular maintenance of LEC can lead to unwanted effects
215
Appl. Sci. 2019, 9, 438
to samples as well as to laboratory equipment’s ability to produce stable results. The iAQ+ allows the
user to store the historical data of the LEC; therefore, it is possible to check the integrity of the collected
data and perform data analysis and comparison taking into account the indoor conditions.
Through charts visualisation of IEQ conditions, the application provides a better acknowledgement
of the supervised data when compared to the numeric table format. Consequently, the proposed solution
is a significant decision-making tool to plan interventions in order to promote a healthful and productive
living environment but also for IEQ analysis.
The iAQ+ incorporates a wireless connection interface for Internet access along with an easy Wi-Fi
network configuration. The iAQ+ should be connected to a Wi-Fi hotspot for data transmission and
to store the network credentials on the flash memory after successful connection. At system start-up,
the iAQ+ searches for a stored Wi-Fi network to connect to. However, if there no saved Wi-Fi networks
available, the system will turn to hotspot mode and create a Wi-Fi network with an SSID “iAQ+”.
At this stage, the user must connect to the referred hotspot in order to configure the credentials of the
Wi-Fi network to which the system will be connected (Figure 10).
Figure 10. Wi-Fi configuration.
In this way, the iAQ+ can be easily installed by the end-user which not only follows the original
paradigm of IoT solutions but also contributes to the low-cost aim of the presented solution.
Compared to several similar solutions proposed and described in Table 1, the iAQ+ supports push
notifications to alert people in a timely way for enhanced living environments and occupational health
by providing ventilation, deactivation of pollutant equipment and by activating air purification systems.
When the parameters meet the setpoints for the correspondent working mode, the user is alerted
to ensure a productive environment for teaching or correct laboratory environmental parameters.
The iAQ+ offers a work mode feature; the user can configure setpoints for laboratory mode and
schoolroom mode. Using the iAQ+Mobile, the user can easily change the work mode in real-time.
On one hand, compared to other systems proposed by References [46,47,49,50] based on WSN,
the iAQ+ provides several advantages in scalability and installation in indoor living environments
as it is only necessary to configure the Wi-Fi Internet connection and it is not required to configure
the sensor nodes and coordinators. On the other hand, compared to the other systems proposed
by References [46,50–52] which do not use mobile computing, the iAQ+ provides smartphone and
Web compatibility for data access. The solution proposed by Reference [48] supports mobile access.
216
Appl. Sci. 2019, 9, 438
However, the proposed mobile app allows only to access the last collected data and does not support
historical data access. The iAQ+Mobile supports graphical presentation for an easy overview of the
IEQ data by the end-user.
The IoT design facilitates the scalability of the proposed solution offering flexibility and
expandability as the area can be monitored using only one iAQ+ unity, but other unities can be
added if needed.
Enhancements to the system hardware and software are scheduled in order to incorporate
other IEQ parameters such as noise supervision and/or experiment variables. So, as future work
improvements and adaptions for specific laboratory tests, such as thermographic experiments applied
to wood and trees are planned. The proposed solution has multiple benefits in installation and
configuration as the iAQ+ incorporates wireless communication technology for data transmission,
but also provides compatibility with both the standard domestic homes and smart homes.
Apache Kafka is an open-source stream-processing software platform for distributed
high-throughput systems. This platform provides low-latency, built-in partitioning, replication and
inherent fault-tolerance for real-time data handling. Apache Spark is a framework for large-scale data
processing which can be used with Kafka to stream the data to solve the response time and system
throughput problems. The SQL Server response time and system throughput tests were not performed.
However, these evaluations are planned by the authors as long as the study of the incorporation of
open-source platforms such as Apache Kafka and Apache Spark to face response time and system
throughput difficulties.
It is imperative to enhance the IEQ conditions, and the authors consider that the first step is
to implement real-time supervision to identify its variation and to plan interventions for enhanced
living environments.
Managerial Implication
Regarding managerial implications, the results provide insights for enhanced living environments
and laboratory experiments. In most university establishments, laboratories are also used as classrooms.
Therefore, iAQ+ is relevant to provide a healthful and productive living environment but also to
support experimental activities. LEC has a direct impact on test results and on the accuracy and
consistency of test data that are affected by environmental conditions. Thus, the proposed solution
offers an integrated management system for historical data analysis in order to guarantee the integrity
of test results. The two functional modes of the iAQ+ solution, the laboratory mode and the schoolroom
mode allow the user to configure setpoints for both modes. This will allow switching the functional
mode in an agile way, contributing to increasing the quality of the laboratory experiments and teaching
activities. The iAQ+ easy installation feature follows the original paradigm of IoT solutions and
contributes to the low cost of the proposed solution. Most importantly, the real-time alerts promote
behaviour changes to support the occupants to act promptly in order to provide a healthy and
productive workplace for learning and a regulated environment for laboratory activities. The proposed
solution provides a mobile application and a Web portal for IAQ data analytics which facilitates data
access and analysis.
6. Conclusions
This paper presented an IoT real-time supervision system architecture named iAQ+ composed by
a hardware prototype for data acquisition along with Web and mobile compatibility for data access.
The proposed system provides temperature, relative humidity, barometric pressure and qualitative air
quality supervision for both school and laboratory scenarios.
With the proliferation of IoT technologies, there is great potential to create automatic IEQ
supervision solutions for enhanced living environments and occupational health.
The results achieved are promising, indicating an important contribution to LEC supervision
solutions based on IoT. Using iAQ+, the monitored data can be particularly valuable to analyse and
217
Appl. Sci. 2019, 9, 438
store the laboratory activities’ conditions to ensure that they are stable in the course of the experiments
conducted, as they influence the quality of the results.
Compared to existing systems, the iAQ+ supports push notifications to alert people in a timely
way for enhanced living environments and occupational health; in addition, by supporting a work
mode feature, the user can configure setpoints for laboratory mode and schoolroom mode. This system
offers flexibility and expandability as the user can start with only one iAQ+ unity and add more
unities if needed. However, quality assurance (QA) and quality control (QC) testing of the proposed
study and the SQL Server response time and system throughput analysis was not done by the authors.
In the future, the QA/QC should be implemented for enhanced product quality traceability and
these response time and systems throughput evaluations will be conducted. The authors have also
planned software and hardware improvements to adapt the system to specific laboratory tests, such as
thermographic experiments applied to wood and trees. In spite of the influence of indoor environments
in daily human activities, systems like this will contribute to ensuring a productive environment for
teaching and proper LEC.
Author Contributions: G.M. and R.P. designed the study, developed the methodology, performed the analysis,
and wrote the manuscript.
Funding: This research is framed in the project “TreeM—Advanced Monitoring & Maintenance of Trees” N. º
023831, 02/SAICT/2016, co-financed by CENTRO 2020 and FCT, Portugal 2020 and structural funds UE-FEDER.
Acknowledgments: The financial support from the Research Unit for Inland Development of the Polytechnic
Institute of Guarda is acknowledged.
References
1. Vilcekova, S.; Meciarova, L.; Burdova, E.K.; Katunska, J.; Kosicanova, D.; Doroudiani, S. Indoor
environmental quality of classrooms and occupants’ comfort in a special education school in Slovak Republic.
2. Bruce, N.; Perez-Padilla, R.; Albalak, R. Indoor air pollution in developing countries: A major environmental
and public health challenge. Bull. World Health Organ. 2000, 78, 1078–1092. [PubMed]
3. Seguel, J.M.; Merrill, R.; Seguel, D.; Campagna, A.C. Indoor Air Quality. Am. J. Lifestyle Med. 2016, 11, 284–295.
[CrossRef] [PubMed]
4. De Gennaro, G.; Dambruoso, P.R.; Loiotile, A.D.; Di Gilio, A.; Giungato, P.; Tutino, M.; Marzocca, A.;
Mazzone, A.; Palmisani, J.; Porcelli, F. Indoor air quality in schools. Environ. Chem. Lett. 2014, 12, 467–482.
[CrossRef]
5. Madureira, J.; Paciência, I.; Rufo, J.; Ramos, E.; Barros, H.; Teixeira, J.P.; de Oliveira Fernandes, E. Indoor air
quality in schools and its relationship with children’s respiratory symptoms. Atmos. Environ. 2015, 118, 145–156.
[CrossRef]
6. Rupp, R.F.; Vásquez, N.G.; Lamberts, R. A review of human thermal comfort in the built environment.
Energy Build. 2015, 105, 178–205. [CrossRef]
7. Giusto, D.; Iera, A.; Morabito, G.; Atzori, L. (Eds.) The Internet of Things: 20th Tyrrhenian Workshop on Digital
Communications; Springer: New York, NY, USA, 2010; ISBN 978-1-4419-1673-0.
8. Gubbi, J.; Buyya, R.; Marusic, S.; Palaniswami, M. Internet of Things (IoT): A vision, architectural elements,
and future directions. Future Gener. Comput. Syst. 2013, 29, 1645–1660. [CrossRef]
9. Ibaseta, D.; Molleda, J.; Díez, F.; Granda, J.C. Indoor Air Quality Monitoring Sensor for the Web of Things.
Proceedings 2018, 2, 1466. [CrossRef]
10. Yin, Y.; Zeng, Y.; Chen, X.; Fan, Y. The internet of things in healthcare: An overview. J. Ind. Inf. Integr. 2016, 1, 3–13.
[CrossRef]
11. Bhatt, Y.; Bhatt, C. Internet of Things in HealthCare. In Internet of Things and Big Data Technologies for
Next Generation Healthcare; Bhatt, C., Dey, N., Ashour, A.S., Eds.; Springer International Publishing: Cham,
Switzerland, 2017; Volume 23, pp. 13–33. ISBN 978-3-319-49735-8.
218
Appl. Sci. 2019, 9, 438
12. Marques, G.; Pitarma, R. An indoor monitoring system for ambient assisted living based on internet of
things architecture. Int. J. Environ. Res. Public Health 2016, 13, 1152. [CrossRef]
13. United Nations. World Population Ageing, 1950–2050; Department of Economic and Social Affairs, Population
Division, United Nations: New York, NY, USA, 2002; pp. 11–13. ISBN 92-1-051092-5.
14. Centers for Disease Control and Prevention. The state of aging and health in America 2007. N. A. on an
Aging Society. 2007. Available online: http://https://www.cdc.gov/aging/pdf/saha_2007.pdf (accessed on
1 December 2018).
15. Koleva, P.; Tonchev, K.; Balabanov, G.; Manolova, A.; Poulkov, V. Challenges in designing and implementation
of an effective Ambient Assisted Living system. In Proceedings of the 2015 12th International Conference
on Telecommunication in Modern Satellite, Cable and Broadcasting Services (TELSIKS), Niš, Serbia,
14–17 October 2015; pp. 305–308.
16. Atzori, L.; Iera, A.; Morabito, G. The Internet of Things: A survey. Comput. Netw. 2010, 54, 2787–2805.
[CrossRef]
17. Caragliu, A.; Del Bo, C.; Nijkamp, P. Smart Cities in Europe. J. Urban Technol. 2011, 18, 65–82. [CrossRef]
18. Chourabi, H.; Nam, T.; Walker, S.; Gil-Garcia, J.R.; Mellouli, S.; Nahon, K.; Pardo, T.A.; Scholl, H.J.
Understanding Smart Cities: An Integrative Framework. In Proceedings of the 2012 45th Hawaii
International Conference on System Sciences, Maui, HI, USA, 4–7 January 2012; pp. 2289–2297.
19. Zanella, A.; Bui, N.; Castellani, A.; Vangelista, L.; Zorzi, M. Internet of Things for Smart Cities. IEEE Internet
Things J. 2014, 1, 22–32. [CrossRef]
20. Hernández-Muñoz, J.M.; Vercher, J.B.; Muñoz, L.; Galache, J.A.; Presser, M.; Hernández Gómez, L.A.;
Pettersson, J. Smart Cities at the Forefront of the Future Internet. In The Future Internet; Domingue, J.,
Galis, A., Gavras, A., Zahariadis, T., Lambert, D., Cleary, F., Daras, P., Krco, S., Müller, H., Li, M.-S.,
et al., Eds.; Springer Berlin Heidelberg: Berlin/Heidelberg, Germany, 2011; Volume 6656, pp. 447–462,
ISBN 978-3-642-20897-3.
21. Skouby, K.E.; Lynggaard, P. Smart home and smart city solutions enabled by 5G, IoT, AAI and CoT services.
In Proceedings of the 2014 International Conference on Contemporary Computing and Informatics (IC3I),
Mysore, India, 14–17 December 2014; pp. 874–878.
22. Alaa, M.; Zaidan, A.A.; Zaidan, B.B.; Talal, M.; Kiah, M.L.M. A review of smart home applications based on
Internet of Things. J. Netw. Comput. Appl. 2017, 97, 48–65. [CrossRef]
23. Theoharidou, M.; Tsalis, N.; Gritzalis, D. Smart Home Solutions: Privacy Issues. In Handbook of Smart Homes,
Health Care and Well-Being; Van Hoof, J., Demiris, G., Wouters, E.J.M., Eds.; Springer International Publishing:
Cham, Switzerland, 2017; pp. 67–81. ISBN 978-3-319-01582-8.
24. Feng, S.; Setoodeh, P.; Haykin, S. Smart Home: Cognitive Interactive People-Centric Internet of Things. IEEE
Commun. Mag. 2017, 55, 34–39. [CrossRef]
25. Madureira, J.; Paciência, I.; Pereira, C.; Teixeira, J.P.; Fernandes, E.d.O. Indoor air quality in Portuguese
schools: Levels and sources of pollutants. Indoor Air 2016, 26, 526–537. [CrossRef] [PubMed]
26. Stabile, L.; Dell’Isola, M.; Russi, A.; Massimo, A.; Buonanno, G. The effect of natural ventilation strategy on
indoor air quality in schools. Sci. Total Environ. 2017, 595, 894–902. [CrossRef] [PubMed]
27. Kielb, C.; Lin, S.; Muscatiello, N.; Hord, W.; Rogers-Harrington, J.; Healy, J. Building-related health symptoms
and classroom indoor air quality: A survey of school teachers in New York State. Indoor Air 2015, 25, 371–380.
[CrossRef] [PubMed]
28. Choo, C.P.; Jalaludin, J. An overview of indoor air quality and its impact on respiratory health among
Malaysian school-aged children. Rev. Environ. Health 2015, 30, 9–18. [CrossRef]
29. Lee, S.; Chang, M. Indoor and outdoor air quality investigation at schools in Hong Kong. Chemosphere
2000, 41, 109–113. [CrossRef]
30. Yang, L.; Yan, H.; Lam, J.C. Thermal comfort and building energy consumption implications—A review.
Appl. Energy 2014, 115, 164–173. [CrossRef]
31. Havenith, G.; Holmér, I.; Parsons, K. Personal factors in thermal comfort assessment: Clothing properties
and metabolic heat production. Energy Build. 2002, 34, 581–591. [CrossRef]
32. Pereira, L.D.; Cardoso, E.; da Silva, M.G. Indoor air quality audit and evaluation on thermal comfort in
a school in Portugal. Indoor Built Environ. 2015, 24, 256–268. [CrossRef]
33. Srbinovska, M.; Gavrovski, C.; Dimcev, V.; Krkoleva, A.; Borozan, V. Environmental parameters monitoring
in precision agriculture using wireless sensor networks. J. Clean. Prod. 2015, 88, 297–307. [CrossRef]
219
Appl. Sci. 2019, 9, 438
34. Sanchez-Rosario, F.; Sanchez-Rodriguez, D.; Alonso-Hernandez, J.B.; Travieso-Gonzalez, C.M.;

Alonso-Gonzalez, I.; Ley-Bosch, C.; Ramirez-Casanas, C.; Quintana-Suarez, M.A. A low consumption
real time environmental monitoring system for smart cities based on ZigBee wireless sensor network.
In Proceedings of the 2015 International Wireless Communications and Mobile Computing Conference
(IWCMC), Dubrovnik, Croatia, 24–28 August 2015; pp. 702–707.
35. Zhou, P.; Huang, G.; Zhang, L.; Tsang, K.-F. Wireless sensor network based monitoring system for a large-scale
indoor space: Data process and supply air allocation optimization. Energy Build. 2015, 103, 365–374. [CrossRef]
36. Marques, G.; Pitarma, R. IAQ Evaluation Using an IoT CO2 Monitoring System for Enhanced Living
Environments. In Trends and Advances in Information Systems and Technologies; Rocha, Á., Adeli, H., Reis, L.P.,
Costanzo, S., Eds.; Springer International Publishing: Cham, Switzerland, 2018; Volume 746, pp. 1169–1177.
ISBN 978-3-319-77711-5.
37. Pitarma, R.; Marques, G.; Ferreira, B.R. Monitoring Indoor Air Quality for Enhanced Occupational Health.
J. Med Syst. 2017, 41, 23. [CrossRef]
38. Pitarma, R.; Marques, G.; Caetano, F. Monitoring Indoor Air Quality to Improve Occupational Health. In New
Advances in Information Systems and Technologies; Rocha, Á., Correia, A.M., Adeli, H., Reis, L.P., Mendonça
Teixeira, M., Eds.; Springer International Publishing: Cham, Switzerland, 2016; Volume 445, pp. 13–21. ISBN
978-3-319-31306-1.
39. Marques, G.; Pitarma, R. Smartwatch-Based Application for Enhanced Healthy Lifestyle in Indoor
Environments. In Computational Intelligence in Information Systems; Omar, S., Haji Suhaili, W.S.,
Phon-Amnuaisuk, S., Eds.; Springer International Publishing: Cham, Switzerland, 2019; Volume 888,
pp. 168–177. ISBN 978-3-030-03301-9.
40. Marques, G.; Pitarma, R. Using IoT and Social Networks for Enhanced Healthy Practices in Buildings.
In Information Systems and Technologies to Support Learning; Rocha, Á., Serrhini, M., Eds.; Springer International
Publishing: Cham, Switzerland, 2019; Volume 111, pp. 424–432. ISBN 978-3-030-03576-1.
41. Marques, G.; Pitarma, R. Monitoring Health Factors in Indoor Living Environments Using Internet of Things.
In Recent Advances in Information Systems and Technologies; Rocha, Á., Correia, A.M., Adeli, H., Reis, L.P.,
Costanzo, S., Eds.; Springer International Publishing: Cham, Switzerland, 2017; Volume 570, pp. 785–794.
ISBN 978-3-319-56537-8.
42. Marques, G.; Roque Ferreira, C.; Pitarma, R. A System Based on the Internet of Things for Real-Time Particle
Monitoring in Buildings. Int. J. Environ. Res. Public Health 2018, 15, 821. [CrossRef]
43. Akkaya, K.; Guvenc, I.; Aygun, R.; Pala, N.; Kadri, A. IoT-based occupancy monitoring techniques for
energy-efficient smart buildings. In Proceedings of the 2015 IEEE Wireless Communications and Networking
Conference Workshops (WCNCW), New Orleans, LA, USA, 9–12 March 2015; pp. 58–63.
44. Wargocki, P.; Wyon, D.P.; Sundell, J.; Clausen, G.; Fanger, P.O. The Effects of Outdoor Air Supply Rate in
an Office on Perceived Air Quality, Sick Building Syndrome (SBS) Symptoms and Productivity. Indoor Air
2000, 10, 222–236. [CrossRef]
45. Srivatsa, P.; Pandhare, A. Indoor Air Quality: IoT Solution. In Proceedings of the National Conference
NCPCI, 19 March 2016; Volume 2016, p. 19.
46. Salamone, F.; Belussi, L.; Danza, L.; Galanos, T.; Ghellere, M.; Meroni, I. Design and Development of a
Nearable Wireless System to Control Indoor Air Quality and Indoor Lighting Quality. Sensors 2017, 17, 1021.
[CrossRef]
47. Bhattacharya, S.; Sridevi, S.; Pitchiah, R. Indoor air quality monitoring using wireless sensor network.
In Proceedings of the 2012 Sixth International Conference on Sensing Technology (ICST), Kolkata, India,
18–21 December 2012; pp. 422–427.
48. Salamone, F.; Belussi, L.; Danza, L.; Ghellere, M.; Meroni, I. Design and Development of nEMoS, an All-in-One,
Low-Cost, Web-Connected and 3D-Printed Device for Environmental Analysis. Sensors 2015, 15, 13012–13027.
[CrossRef]
49. Wang, S.K.; Chew, S.P.; Jusoh, M.T.; Khairunissa, A.; Leong, K.Y.; Azid, A.A. WSN based indoor air quality
monitoring in classrooms. AIP Conf. Proc. 2017, 1808, 020063.
50. Liu, J.; Chen, Y.; Lin, T.; Lai, D.; Wen, T.; Sun, C.; Juang, J.; Jiang, J.-A. Developed urban air quality monitoring
system based on wireless sensor networks. In Proceedings of the 2011 Fifth International Conference on
Sensing Technology, Palmerston North, New Zealand, 28 November–1 December 2011; pp. 549–554.
220
Appl. Sci. 2019, 9, 438
51. Kang, J.; Hwang, K.-I. A Comprehensive Real-Time Indoor Air-Quality Level Indicator. Sustainability 2016, 8, 881.
[CrossRef]
52. Benammar, M.; Abdaoui, A.; Ahmad, S.; Touati, F.; Kadri, A. A Modular IoT Platform for Real-Time Indoor
Air Quality Monitoring. Sensors 2018, 18, 581. [CrossRef] [PubMed]
53. Klein, R.C.; King, C.; Kosior, A. Laboratory air quality and room ventilation rates: An update. J. Chem. Health
Saf. 2011, 18, 21–24. [CrossRef]
54. Stuart, R.; Sweet, E.; Batchelder, A. Assessing general ventilation effectiveness in the laboratory. J. Chem.
Health Saf. 2015, 22, 2–7. [CrossRef]
55. Yau, Y.H.; Chew, B.T.; Saifullah, A.Z.A. Studies on the indoor air quality of Pharmaceutical Laboratories in
Malaysia. Int. J. Sustain. Built Environ. 2012, 1, 110–124. [CrossRef]
56. Ugranli, T.; Gungormus, E.; Sofuoglu, A.; Sofuoglu, S.C. Indoor Air Quality in Chemical Laboratories.
In Comprehensive Analytical Chemistry; Elsevier: Amsterdam, The Netherlands, 2016; Volume 73, pp. 859–878.
ISBN 978-0-444-63605-8.
57. Morbeck, D.E. Air quality in the assisted reproduction laboratory: A mini-review. J. Assist. Reprod. Genet.
2015, 32, 1019–1024. [CrossRef]
221
applied
sciences
Article
A Promising Technological Approach to Improve
Indoor Air Quality
Thomas Maggos 1, *, Vassilios Binas 2 , Vasileios Siaperas 3 , Antypas Terzopoulos 3 ,
Panagiotis Panagopoulos 1 and George Kiriakidis 2
1 Environmental Research Laboratory, NCSR “Demokritos”, 15310 Ag. Paraskevi, Athens, Greece;
[email protected]
2 Institute of Electronic Structure and Laser, Foundation for Research and Technology, 70013 Heraclion, Crete,
Greece; [email protected] (V.B.); [email protected] (G.K.)
3 691 Industrial Base Factory, 19011 Avlonas, Greece; [email protected] (V.S.); [email protected] (A.T.)
Received: 18 September 2019; Accepted: 5 November 2019; Published: 12 November 2019
Abstract: Indoor Air quality (IAQ) in private or public environments is progressively recognized as a
critical issue for human health. For that purpose the poor IAQ needs to be mitigated and immediate
drastic measures must be taken. In environmental science and especially in advanced oxidation
processes and technologies (AOPs-AOTs), photocatalysis has gained considerable interest among
scientists as a tool for IAQ improvement. In the current study an innovative paint material was
developed which exhibits intense photocatalytic activity under direct and diffused visible light for
the degradation of air pollutants, suitable for indoor use. A laboratory and a real scale study were
performed using the above innovative photo-paint. The lab test was performed in a special design
photo-reactor while the real scale in a military’s medical building. Nitrogen Oxide (NO) and Toluene
concentration was monitored between “reference” rooms (without photo paint) and “green” rooms
(with photo-paint) in order to estimate the photocatalytic efficiency of the photo-paint to degrade the
above pollutants. Results of the study showed a decrease up to 60% and 16% for NO and toluene
respectively under lab scale tests while an improvement of air quality up to 19% and 5% under real
world conditions was achieved.
Keywords: IAQ improvement; photo-paint; NO; Toluene degradation
1. Introduction
Indoor air quality (IAQ) is an important determinant of human health, comfort and productivity.
For that purpose, high quality indoor air is desirable. Indoor air pollution can be addressed through
the two approaches of prevention and removal. The latter includes the use of air cleaning technologies
especially in buildings where ventilation rates are being reduced in order to save energy. Indicative
air cleaning technologies which have been developed during the recent years are: filtration and
adsorption, electrostatic air purification, air filtration and gas adsorption filtration [1–3], ozonation [4–7],
non-thermal plasma [8,9] and photocatalytic oxidation (PCO) [10–13]. PCO is a general air cleaning
technology, which is able to degrade Volatile Organic Compounds (VOCs), such as aromatics, alkanes,
odor compounds etc. Air cleaning photocatalytic technology is based on the principle that radiation
of suitable wave-lengths can be absorbed by semiconductors, which leads to the creation of reactive
oxygen species (ROS) that can degrade air pollutants. TiO2 is the most commonly-used semiconductor
in PCO research. However, over the last years, scientists have combined TiO2 with other materials
such as activated carbon and zeolite hybrid catalysts in order to enhance the PCO degradation of air
pollutants [14–16]. Furthermore, numerous TiO2 photocatalysts with different morphological designs
have been developed: nanoparticles, nanotubes, hollow fibers and mesoporous. To this end, the need to

Appl. Sci. 2019, 9, 4837
evaluate the photocatalytic performance of the above materials in a common methodological approach
has been raised and extensive research efforts have been devoted to it [17,18].
However, in most of these studies, only a single compound was tested, using a photocatalytic
reactor and a methodological approach, which have been developed by the same lab that produced
the material. It is well known that indoor air contains numerous contaminants; thus, tests of only
one compound may be misleading. Furthermore, many studies proved the generation of by-products
during the photocatalytic processes, such as formaldehyde, ozone, benzaldehyde, acetaldehyde etc.
It is obvious that in some cases, by-products could be more harmful than the target pollutant [19–22].
For that purpose, and although photocatalytic technology is promising, the synthesis route of a
photocatlytic material should be designed carefully in order to avoid contaminants which could
lead to the formation and emission in the gas phase of intermediate products as they can be more
hazardous than the target pollutant. The latter is clearly demonstrated during the evaluation of
two photocatalytic air-purification in a mock-up air cabin. Although two symptoms, dizziness and
claustrophobia, were reported to decrease when either one of the photocatalytic air-purification devices
was operated, intermediates (acetaldehyde and formaldehyde) were detected as a result of ethanol
photodegradation [23,24].
The present research addresses the indoor air purification study using a photocatalytic paint,
which was tested under both laboratory and real world indoor conditions (application in building
walls). A modified TiO2 was chosen as photocatalyst in order to be activated by visible light which
is the dominant spectrum in indoor environments. Traditional TiO2 is activated only under UV
light. More specifically a Mn-doped TiO2 photocatalyst (powder) was used in the synthetic route of a
photocatalytic paint production. The powder named TCM-1 [25] has already been successfully tested
for the oxidation of air pollutants under indoor-like illumination conditions and when mixed either with
calcareous or cementitious base matrices demonstrated a unique ability to efficiently degrade volatile
organic compounds (VOCs) such as BTX (Benzene, Toluene, Xylene), formaldehyde and nitrogen
oxides (NOx) [26–28]. Additionally, it has been proven effective on eliminating bacteria such as E. coli
and Klebsiella pneumoniae and phages such as MS2 [29–32]. Furthermore, no dangerous by-products
were produced during the degradation process. In a previous work, in order to demonstrate the
photocatalytic effectiveness of TCM-1, the powder has been incorporated in/on different construction
matrices such as glass, plywood, wood, ceramic and concrete substances, which were, subsequently,
tested indicating very promising results [33].
In the current study TCM-1 incorporated into paint production process in order to produce paint
with the ability to photocatalytically improve IAQ by degraded indoor air pollutants. NO and toluene
were chosen as the target pollutants since they are typical indoor air pollutants that can be emitted from
various indoor sources, such as cooking, tobacco smoke, furniture, building materials and fireplaces,
as well as from outdoor sources, e.g., traffic, domestic heating. These indoor pollutants can have
significant health impacts and for that purpose their elimination from the indoor environment should
be appropriately addressed. In order to simulate the real indoor air conditions, visible light was used
as the source for the PCO reaction in the current study. The latter is different from what we usually find
in the literature. More specifically, UV light is the most commonly used parameter in the experimental
procedures, which usually applied to test the efficiency of a photo-material.
The experimental study of the current work was carried out using the European Committee for
Standardization (CEN) Technical Specification (TS) 16980:2016 as a basic reference to perform the
lab scale experiments, while the real scale experiments were based on the comparison between the
concentration level of a pollutant in a reference and the “green” room, respectively. The latter was
performed at the military medical center of the Cadets Training Camp in Heraklion Crete.
224
Appl. Sci. 2019, 9, 4837
2.1. Materials
Chemicals for the preparation of the photocatalytic powder such as Titanium (IV) oxysulfate
hydrate (TiOSO4·xH2O), manganese (II) acetate tetrahydrate Mn (CH3COO)2) and ammonium
hydroxide (25% NH4OH) purchased from Aldrich were applied.
Chemicals for the preparation of the photo-paint were purchased from VINAVIL EGYPT
(New Cairo, Egypt), DOW chemicals (Midland, Michigan, MI, United States), BYK (Wesel, Germany),
DuPont (Midland, Michigan, MI, United States) and Dionyssos Marbles (Penteli, Greece) and are
readily available raw materials mainly used at the coating industry.
2.2. Methods
2.2.1. Preparation of Photo-Material (powder)

In the present study, an optimum powder (0.1% Mn-doped) named TCM-1 was used for the
preparation of photocatalytic paint. TCM-1 was synthesized by a co-precipitation method with
0.1% of manganese. TCM-1 was precipitated at pH ∼ 7 from aqueous solution of titanium (IV)
oxysulfate hydrate and manganese by the addition of ammonia. After aging the suspension overnight,
the precipitate was filtered and dried under air at 373 K. The residue was crushed to a fine powder and
calcined in a furnace at 973 K for 3 h. More details in the synthesis procedure and the characterization
are given in previous work along with the preparation details [25].
2.2.2. Preparation of Photo-Paint

TCM-1 powder (0.1% Mn-doped) was added to a specially formulated architectural coating.
The paint was consisted 10% w/w of the TCM-1 powder partly replacing the Titanium Oxide (normally
used as a white pigment). The raw materials for the preparation of the photo-paint are commonly
used in the coating industry. A polyvinyl acetate copolymer binder was used and the Pigment Volume
Content (PVC) was adjusted, but kept below its critical value in order for the paint to have a high
quality, matte finish.
The production process of the photo-paint was consisted of three discreet phases. At phase 1,
the mill-base (the minimum amount of liquid that can wet the solid particles of the paint, added at
phase 2) was produced. Water was used as solvent with additives, such as sodium hexametaphosphate
(dispersion agent), propylene glycol (antifreezing agent), silica modified surfactant (defoaming
agent), cellulose (rheology modifier), isothiazolinone based mixture (biocide agent) and a mixture
of Alkanol-amines (pH adjusting agent). At phase 2, we applied high shear forces to the mill-base
by using a disperser equipped with a blade dissolver disc and gradually added the solid phase of
the paint [Calcium carbonate powder (20 μm), Titanium Oxide (0.2–0.5 μm), Talc Powder (6 μm) and
TCM-1 powder]. The goal of this phase is to maximize dispersion by eliminating the presence of
any agglomerate in the final paint. At phase 3, we added the Polyvinyl Acetate copolymer binder
emulsion and a coalescent solvent mixture, necessary for the final product to have a smooth surface
(filmer). The dispersion was achieved by using low shearing forces and the pigment dispersion was
stabilized in order to prevent the formation of uncontrolled flocculates by using non-ionic urethane
rheology modifiers.
2.3. Characterization
Powder X-ray diffraction patterns were obtained by a Rigaku D/MAX-2000H rotating anode
diffractometer (CuKα radiation) equipped with a secondary pyrolytic graphite monochromator
operating at 40 kV and 80 mA over the 2θ collection range of 20–80◦ (scan rate was 0.05◦ s−1 ). The grain
225
Appl. Sci. 2019, 9, 4837
size (nm) of TCM-1 was calculated from the line broadening of the X-ray diffraction peak according to
the following Scherrer formula:
D = kλ/βcosθ (1)
where k is the Scherrer contact (~ 0.9), λ is the wavelength of the X-ray radiation (1.54 Ǻ for CuKα),
β is the full width at half maximum (FWHM) of the diffraction peak measured at 2θ, and θ is the Bragg
angle. The morphology and elemental analysis were performed using scanning electron microscopy
(SEM) and energy dispersive X-ray spectroscopy (EDX), on a JSM-6390LV microscope (Jeol, Tokyo,
Japan).
The photo-paint was evaluated through typical quality control measurements. Viscosity was
measured using a Brookfield KU-2 viscometer (Brookfield Engineering Laboratories, INC. Middleboro,
MA, USA) at a temperature of 25 ◦ C with the help of a laboratory water bath and the measurements
were conducted according to ASTM D 562-10 (2018) standard test method.. Density was measured
using a 100 ml pyknometer (density cup) by Elcometer (Manchester, United Kingdom) and an analytical
scale, and the measurements were conducted according to ISO 2811 -1:2016 standard. Fineness of Grind
was evaluated using a Hegman gauge and the evaluation was conducted according to ASTM D 1210-05
(2014) standard test method. Lastly, pH was measured using a Hanna HI 83141V pH meter (Hanna
Instruments, Greece) and the measurements were conducted according to ISO 787-7:2009 standard.
2.4. Photocatalytic Evaluation_Lab-Scale

The experimental methodology that was applied and the required scientific equipment were based
on CEN Technical Specification (TS) 16980:2016. The photocatalytic effect of the optimized material
studied in a continuous flux photocatalytic reactor (Figure 1), which consists of a) a gas transfer and
mixing unit in order to adjust the concentration and humidity levels; b) the photocatalytic reactor main
body made of special plastic so as radiation intensity and wavelength of the radiation is not affected;
(c) the sample irradiation system (OMICRON FS LED Rodgau-Dudenhofen, Germany) consisting of
an LED device connected via software to a computer in order to achieve the optimum efficiency and
accuracy in measuring the radiation; d) a NOx and VOCs analyzer installed on line with the reactor for
continuous monitoring of the pollutant concentration.
Furthermore, in order to ensure optimum mixing of atmospheric pollutants in the reactor, a fan
is installed inside the chamber, while its intensity is adjusted externally to ensure the stability of the
experimental conditions throughout the experiments. The NO concentration is set to (0.50 ± 0.05)
ppmv, while the relative humidity to 40 ± 5%. The illumination provided an average irradiance to the
test specimen surface within the range of wavelengths that are mostly adsorbed by the photocatalyst,
equal to (10.0 ± 5%) W/m2 .
Figure 1. Photocatalytic reactor.
226
Appl. Sci. 2019, 9, 4837
2.5. Photocatalytic Evaluation_Real-Scale

A building of the Hellenic Army in Crete was used to test the de-polluting efficiency of the
photo-paint under real scale application. In general, the photocatalytic efficiency of a material while
applied in a real scale environment and more specifically in indoor building environment could be
estimated through two approaches:
a) A first approach could be the installation of air quality and environmental monitoring
systems (passive inorganic (NO) and organic (BTX) samplers, temperature and humidity recorders)
in the buildings prior to the application of the photo-paint. To that end, a reliable record on the
concentration levels of air pollutants in the case study building (without the photo paint) for at least
12 months should be obtained. Accordingly, the photo-paint should be applied in the buildings and
environmental parameters should be monitored through passive sampling techniques for another
12 months. The potential changes in the indoor air quality due to the photocatalytic action of the
photo-paint will be recorded and quantified to illustrate the capability of the photo-paint to improve
IAQ. A restriction of this approach could be the variations in outdoor air quality and meteorological
conditions during the different sampling periods. In order to eliminate the effect of the above restrictions,
the experiments has to take place the same season (e.g., winter) and the outdoor concentration should
be considered on the final results.
b) A second approach to evaluate the efficiency of the photo-paints is to estimate the IAQ differences
between “reference” rooms (without photo paint) and “green” rooms (with photo-paint), which are
located on the same level (in a raw) and where the same activities take place. The current approach
overcomes the restrictions of the previous one and for that purpose was used in the current study.
More specifically a room of 120 m2 in the ground floor was paint with the innovative photocatalytic
paint (“green room”) and compared with a same size and usage room (“reference room”) located
very close to the “green” one. The outcome of this approach was compared with the outcome of the
lab tests in order to estimate the differences of the photocatalytic performance of a material when
studied in a control experimental reactor and under real world environmental conditions. Passive
samplers for NO and Toluene were applied. More specifically, four passive samplers/pollutants were
installed in each of the rooms for 30 days and then analyzed in the lab using the well-established
Saltzmann spectrophotometric method for NO, while the Toluene samplers were and desorbed by
carbon disulphide and analyszed by gas chromatography (GC/FID).
3. Results
3.1. Physical and Chemical Properties

Figure 2 shows the X-Ray Difraction (XRD) pattern of TCM-1(dopant concentration 0.1 wt%),
calcinated at 700 °C for 3 h 1. The peaks at 2θ values of 25.3◦ , 37.6◦ , and 48.2◦ correspond to the (101),
(004) and (200) planes, respectively, and they are all anatase signature peaks. The grain size for the
TCM-1was 38.69 nm and it was determined from the full width at half maximum (FWHM) of the
(101) anatase peak according to the Scherrer’s formula. In our previous work, it was proven that the
doping and the role of Mn in TiO2. in the case of 0.1 wt% manganese shows only 2+ oxidation state in
comparison with high concentrations [34].
227
Appl. Sci. 2019, 9, 4837
Figure 2. Powder XRD patterns of TCM-1.
Figure 3 shows the UV-Vis absorption as a function of wavelength for TCM-1 and exhibited an
absorption edge in the visible light range (400–800 nm). The band gap energy was 2.87 eV for Mn
doped catalyst.
Figure 3. UV-vis absorption of TCM-1and energy gaps calculated from Kubenka–Munk plots.
Figure 4a shows the morphology of the photocatalyst were investigated with SEM, without
specific morphology, while the spherical shape particles of all the samples demonstrated some degree
of agglomeration and the diameter ranged from 0.1 to 40 μm. Figure 4b shows the characteristic peaks
of Mn and Ti atoms.
228
Appl. Sci. 2019, 9, 4837
D
E
Figure 4. (a,b) SEM image of TCM-1 and EDX analysis.
As far as the photo-paint is concerned, the main physicochemical properties are listed in Table 1.
The measured values are typical for an indoor emulsion architectural paint, and although TCM-1
was used at 10% w/w, TCM-1 behaved as expected from a Titanium Oxide white pigment. A paint
formulated with the process described in Section 2.2.2 and with the use of Ti Pure 902+ Titanium Oxide
white pigment instead of TCM-1 has identical physicochemical properties.
Table 1. Physicochemical Properties of photo-paint.
Property Value Test Method

Viscosity at 25◦ C (KU) 100–110 ASTM D 562
Density at 25◦ C (kg/l) 1.55 ISO 2811
Fineness and Dispersion <40 μm ASTM D1210
pH at 25◦ C 8.5–9 ISO 787-9
PVC (%) 66 Calculated
Usage rate for a 50 μm dry film thickness (m2 /kg) 5.95 Calculated
3.2. Lab and Real-Scale Photocatalytic Performance

In order to estimate the background, experiments in the absence of the photocatlatyic paint were
performed. The photocatalytic experiments determine the total pollutant degradation involving both
UV photolysis and photocatalysis on the photo-paint. The net photocatalytic effect is calculated by the
subtraction of the background contribution from the photocatalytic experiments. More specifically,
blank tests were carried out by polluting the reactor with NO and Toluene in the absence of the
229
Appl. Sci. 2019, 9, 4837
photocatalyst without and with irradiation, respectively; then, the same experiments were carried out
in the presence of the photocatalysts.
Figure 5a,b present the elimination of NO and toluene under the irradiation of the photo-paint
from Vis-light. It is obvious that just after the irradiation of the sample (Time 30), a sharp decrease in
NO and a smaller but significant decrease in toluene concentration is observed, which demonstrate the
immediate response of the photocatalytic system and provide the photo-efficiency of the paint.
Adsorption of NO onto the chamber’s wall area and photolysis are the main sinks of NO during
blank tests. Calculations have shown that both these mechanisms did not contributed to the total NO
and toluene removal during the photocatalytic experiments.
The photocatalytic activity was evaluated by the calculation photocatalytic yield (% n, Equation (2))
and photodegradation rate (r, Equation (3)). The corresponding equations were used for toluene.
The results in Table 2 showed the possibility of developing a very promising and highly active to air
pollutants photocalatylic paint (Table 2).
OUT,light
CIN
NO
− CNO
NO =
% ηtotal × 100 (2)
CIN
NO
where:
CIN
NO
: the concentration of NO at reactor inlet
OUT,light
CNO : the concentration of NO at reactor outlet under stable conditions with irradiation (lamp on)
The rate of photocatalytic yield of the material is calculated by the formula below (Equation (3))
and expressed in μgm−2 s−1 :
⎛ ⎞
613F ⎜⎜⎜⎜ ηNO ηdark ⎟⎟
total
photo NO ⎟⎟
rNO = ⎜ − ⎟⎟ (3)
S ⎜⎝ 1 − ηtotal (1 − ηNO ) ⎠
dark
NO
where
F: the gas flow (m3 h−1 )

S: The area of the test surface (m2 )
The photodegradation rate provides a more accurate measure of the photocatalytic activity of
the material in comparison with the % photocatalytic decomposition. The latter is attributed to the
fact that r is taking into consideration the initial concentration of the pollutant, the sample’s area and
the irradiation time. It is expressed as μg of converted NO/toluene per m2 of material per second of
irradiation. The current parameter was calculated only for the lab tests, where the pollutant flow rate
in the chamber was known.
ηdark
NO
The losses in the system are minimal, and as a consequence the fraction is zero.
(1−ηdark
NO
)
Table 2. Photocatalytic parameters for lab and real scale tests.
Parameter Lab Scale Real Scale

NO Toluene NO Toluene
%η 59.08 16.7 18.8 5.26
rphoto (μg/m2 s) 3.89 1.05 - -
230
Appl. Sci. 2019, 9, 4837
(a)
(b)
Figure 5. (a) Photocatalytic yields (% η) concentration profile versus time for the photo-oxidation of
NO in the presence of the catalyst. (b) Photocatalytic yields (% η) concentration profile versus time for
the photo-oxidation of Toluene in the presence of the catalyst.
The variations on NO and Toluene concentrations during the experimental procedure of

photocatalysis are presented in Figure 5a and Figure 5b, respectively. At first, the pollutant was
introduced into the photo-reactor and the system remained in equilibrium for 30 min. Then, irradiation
was followed and a sharp reduction in NO and significant but less intense than NO elimination in
toluene concentration is observed, which stabilized after 3 h and remained stable for approximately 2 h
where the irradiation stopped. The results from the laboratory tests showed that for both pollutants,
the photocatalytic paint gave very promising results. It is worth to note that in the case of toluene,
the degradation efficiency is significantly lower than NO. However, a reduction of VOCs even at that
level could have significant effects in the improvement of IAQ. Furthermore, the lower degradation
efficiency in VOCs compare with NO is in line with the results of various studies [35–38]. The latter
could be attributed to the low adsorption capacity of toluene molecules at the catalyst surface.
The extremely hydroxylated surface of the paint material due to physically or chemically bound water,
could constrain its interaction with the active radicals through surface diffusion [39,40].
3.3. Real-Scale Photocatalytic Performance

As far as the real scale application is concerned, the deppolution efficiency was calculated by the
absolute difference in the concentration levels of NO and toluene, which was measured in the two
231
Appl. Sci. 2019, 9, 4837
rooms of the military building: the “green” and the “reference” rooms. Results are shown in Table 2.
It is observed that real scale tests showed significant lower values of NO and toluene degradation
than in the lab-scale application. The latter was also observed in other studies [41–45]. The control
environment of the laboratory versus the more complex and polyparametric environment of a real
application could be the main reason that led to significant differences in the values that characterize
the photocatalytic efficiency of the paint. However, they provide an indication of the photocatalytic
efficiency of the paints to degrade pollutants in situ and a basis for photo-paint applications in order to
improve IAQ.
4. Discussion
The effect of photocatalytic paint in the improvement of IAQ has been studied in both laboratory
and real scale tests. It is interesting to note that NO could be effectively converted by the photocatalytic
paint, while toluene showed lower photocatalytic removal in the same photocatalytic process (lab
and real scale). A possible explanation of the different photocatalytic performance of NO and toluene
under the same application could be explained as follows: the dominant oxidants in a photocatalytic
reaction are hydroxyl radicals (OH• ) and hydroperoxyl radicals (HO2• ), which are generated from the
redox reactions of positive holes (h+ ) and electrons (e− ) with O2 , H2 O or OH− . The kinetic coefficients
of gas phase reactions for HOx radicals and targeted gas (NO and toluene in this case) have only minor
differences [46,47]. However, NO is better adsorbed by the alkaline constituents of the paint due to its
acidic property. This significantly increases their reaction potential with HOx radicals. On the other
hand, the adsorption capacity of toluene is lower than NO due to the extremely hydroxylated surface
of the paint, which constrain its interaction with the active radicals. Toluene is absorbed more easily in
a less hydrophilic TiO2 surface than a more hydroxylated one [41]. It is obvious that different gaseous
pollutants present significant variations in their photocatalytic activity due to the difference in their
diffusion in the paint matrix. Efficient adsorption of the pollutant molecules at the catalyst surface
could promote the photocatalytic reaction.
Comparing lab versus real scale experiments results, it is observed that laboratory tests showed
significant higher values of both pollutants (NO and toluene) degradation than in real scale application.
The translation from the laboratory results to “real” site efficiency is difficult because of the great number
of parameters involved such as traffic and environmental parameters (temperature, light intensity,
relative humidity wind speed). For that purpose, the more complex and polyparametric environment
of a real application lead to significant differences in the values, which characterize the photocatalytic
efficiency of the paint. In any case, precaution has to be taken with the interpretation of data obtained
from the real scale experiments since these results are limited over time. However, they provide an
indicative picture of the efficiency of the photocatalytic paints to eliminate air pollutants under real
world conditions and a basis on which to improve their photocatalytic capacity for future applications.
The need for large scale applications is more imperative nowadays, as many photo-materials have
indicated very promising results during lab scale tests, but their capacity under real world conditions
has not been proven. The latter is critical and proven by the outcome of the current work, which showed
significant differences between the photo-efficiency of the same material under two different scales.
Additionally, the demonstration of the effectiveness of photocatalytic materials on site should also
include negative effects, such as the formation and emission of by-products during the photocatalytic
reactions, as well as the durability of the photo-paint mechanical properties. The latter is of high
importance, as in most of the cases, TiO2 oxidizes not only the air pollutants on the photo-material
surface, but also their organic and inorganic components. For that purpose, special attention is given in
the formulation of such materials in order to avoid it. As it is not in the scope of the current manuscript
to study the durability of the photo-paint mechanical properties, measurements to characterize the
mechanical durability of the paint were not performed. However, it is worth mentioning that almost
two years after the photo-paint application, cracks or other surface damages have not been observed.
Nevertheless, beyond the visual observation, the latter should be certified by analytical measurements,
232
Appl. Sci. 2019, 9, 4837
which should be the topic of future study. Furthermore, better results could be obtained by using
mathematical models to simulate the photocatalytic processes, validating the model using the outcomes
of the lab measurements and then implement various parameters in order to assess the real life
effects [48].
5. Conclusions
This article addresses the effect of a photocatalytic paint on the elimination of air pollutants,
and more specifically NO and toluene, for application in indoor environments. The following
conclusions can be drawn:
• The physicochemical properties (including the mechanical parameters) of the photo-paint does
not seem to be affected from the introduction of the photocatalytic powder (TCM-1) in the
synthesis route.
• The photocatalytic efficiency of the paint on NO removal was significantly higher than toluene.
The potential of a pollutant removal depends on the intrinsic properties of gas and the chemical
nature of the paint in which the TiO2 particles are embedded. However, the removal rate of
toluene was very promising for the improvement of IAQ while using the studied photo-paint.
• Lab tests showed better photocatalytic properties of the paint than the results from the real scale
application due to the great number of parameters involved in the case of real scale application.
• There is a need for large scale applications to demonstrate the effectiveness of photocatalytic
materials on site, including any negative effects of the application, such as the emission of
by-products (e.g., carbonyl compounds, O3 etc.).
Author Contributions: Conceptualization, T.M., V.B. and G.K.; Data curation, T.M. and P.P.; Investigation, G.K.;
Methodology, T.M., V.B., V.S. and A.T.; Project administration, V.B.; Writing—original draft, T.M.; Writing—review
& editing, T.M., V.B., V.S., A.T., P.P. and G.K.
Funding: This research was funded by the project “Intelligent technologies for enhancing Quality of Life through
the provision of health monitoring at home and support for everyday life activities” (MIS 5002464), funded by the
Operational Programme “Competitiveness, Entrepreneurship and Innovation” (NSRF 2014-2020) and co-financed
by Greece and the European Union (European Regional Development Fund).
References
1. Ren, H.; Koshy, P.; Chen, W.-F.; Qi, S. Charles Christopher Sorrell Photocatalytic materials and technologies
for air purification. J. Hazard. Mater. 2017, 325, 340–366. [CrossRef] [PubMed]
2. Zhang, Y.; Moa, J.; Li, Y.; Sundell, J.; Wargocki, P.; Zhang, J.; Little John, C.; Corsi, R.; Deng, Q.; Leung, M.H.
Can commonly-used fan-driven air cleaning technologies improve indoor air quality? A literature review.
Atmos. Environ. 2011, 45, 4329–4343. [CrossRef]
3. Bolashikov, Z.D.; Melikov, A.K. Methods for air cleaning and protection of building occupants from airborne
pathogens. Build. Environ. 2009, 44, 1378–1385. [CrossRef]
4. Mamaghani, A.H.; Haghighat, F.; Lee, C.-S. Photocatalytic oxidation technology for indoor environment air
purification: The state-of-the-art. Appl. Catal. B Environ. 2017, 203, 247–269. [CrossRef]
5. Fan, Z.; Lioy, P.; Weschler, C.; Fiedler, N.; Kipen, H.; Zhang, J. Ozone-initiatedreactions with mixtures of
volatile organic compounds under simulatedindoor conditions. Environ. Sci. Technol. 2003, 37, 1811–1821.
[CrossRef] [PubMed]
6. Boeniger, M.F. Use of ozone generating devices to improve indoor airquality. Am. Ind. Hyg. Assoc. J. 1995,
7. Zhong, L.; Haghighat, F. Ozonation air purification technology in HVACapplications. ASHRAE Trans. 2014,
120 (Pt 1), 8.
8. Bahri, M.; Haghighat, F.; Rohani, S.; Kazemian, H. Impact of design parameterson the performance of
non-thermal plasma air purification system. Chem. Eng. J. 2016, 302, 204–212. [CrossRef]
233
Appl. Sci. 2019, 9, 4837
9. Bahri, M.; Haghighat, F. Plasma-based indoor air cleaning technologies: Thestate of the art-review. CLEAN-Soil
Air Water 2014, 42, 1667–1680. [CrossRef]
10. Zhong, L.; Haghighat, F.; Lee, C.-S. Ultraviolet photocatalytic oxidation for indoor environment applications:
Experimental validation of the model. Build. Environ. 2013, 62, 155–166. [CrossRef]
11. Farhanian, D.; Haghighat, F. Photocatalytic oxidation air cleaner: Identification and quantification of
by-products. Build. Environ. 2014, 72, 34–43. [CrossRef]
12. Mo, J.; Zhang, Y.; Xu, Q. Effect of water vapor on the by-products anddecomposition rate of ppb-level toluene
by photocatalytic oxidation. Appl. Catal. B Environ. 2013, 132–133, 212–218. [CrossRef]
13. Debono, O.; Thévenet, F.; Gravejat, P.; Héquet, V.; Raillard, C.; Le, L.; Locoge, N. Gas phase photocatalytic
oxidation of decane at ppb levels: Removal kinetics, reaction intermediates and carbon mass balance.
J. Photochem. Photobiol. A Chem. 2013, 258, 17–29. [CrossRef]
14. Ao, C.H.; Lee, S.C.; Mak, C.L.; Chan, L.Y. Photodegradation of volatile organic compounds (VOCs) and NO
for indoor air purification using TiO2: Promotion versus inhibition effect of NO. Appl. Catal. B Environ. 2003,
42, 119–129. [CrossRef]
15. Ao, C.H.; Lee, S.C.; Zou, S.C.; Mak, C.L. Inhibition effect of SO2 on NOx and VOCs during the
photodegradation of synchronous indoor air pollutants at parts perbillion (ppb) level by TiO2. Appl.
Catal. B Environ. 2004, 49, 187–193. [CrossRef]
16. Selishchev, D.S.; Kolinko, P.A.; Kozlov, D.V. Adsorbent as an essential participant in photocatalytic processes
of water and air purification: Computer simulation study. Appl. Catal. A Gen. 2010, 377, 140–149. [CrossRef]
17. Verbruggen, S.W. TiO2 photocatalysis for the degradation of pollutants ingas phase: From morphological
design to plasmonic enhancement. J. Photochem. Photobiol. C Photochem. 2015, 24, 64–82. [CrossRef]
18. Ismail, A.A.; Bahnemann, D.W. Mesoporous titania photocatalysts: Preparation, characterization and reaction
mechanisms. J. Mater. Chem. 2011, 21, 11686–11707. [CrossRef]
19. Hodgson, A.T.; Destaillats, H.; Sullivan, D.P.; Fisk, W.J. Performance of ultraviolet photocatalytic oxidation
for indoor air cleaning applications. Indoor Air 2007, 17, 305–316. [CrossRef] [PubMed]
20. Muggli, D.S.; McCue, J.T.; Falconer, J.L. Mechanism of the photocatalytic oxidation of ethanol on TiO2.
J. Catal. 1998, 173, 470–483. [CrossRef]
21. Mo, J.H.; Zhang, Y.P.; Xu, Q.J.; Zhu, Y.F.; Lamson, J.J.; Zhao, R.Y. Determination and risk assessment of by-products
resulting from photocatalytic oxidation of toluene. Appl. Catal. B Environ. 2009, 89, 570–576. [CrossRef]
22. Kovalevskiy, N.S.; Lyulyukina, M.N.; Selishcheva, D.S.; Kozlova, D.V. Analysis of air photocatalytic
purification using a total hazard index: Effect of the composite TiO2/zeolite photocatalyst. J. Hazard. Mater.
2018, 358, 302–309. [CrossRef] [PubMed]
23. Denny, F.; Permma, E.; Scott, J.; Wang, J.; Pui, D.Y.H.; Amal, R. Integrated Photocatalytic Filtration Array for
Indoor Air Quality Control. Environ. Sci. Technol. 2010, 44, 5558–5563. [CrossRef] [PubMed]
24. Yu, H.; Zhang, K.; Rossi, C. Experimental Study of the Photocatalytic Degradation of Formaldehyde in Indoor
Air using a Nano-particulate Titanium Dioxide Photocatalyst Indoor. Built Environ. 2007, 16, 529–537. [CrossRef]
25. Binas, V.D.; Sambani, K.; Maggos, T.; Katsanaki, A.; Kiriakidis, G. Synthesis and photocatalytic activity of
Mn-doped TiO2 nanostructured powders under UV and visible light. Appl. Catal. B Environ. 2012, 113–114,
79–86. [CrossRef]
26. Cacho, C.; Geiss, O.; Barrero-Moreno, J.; Binas, V.D.; Kiriakidis, G.; Botalico, L.; Kotzias, D. Studies on
photo-induced NO removal by Mn-doped TiO2 under indoor-like illumination conditions. J. Photochem.
Photobiol. A Chem. 2011, 222, 304–306. [CrossRef]
27. Karafas, E.S.; Romanias, M.N.; Stefanopoulos, V.; Binas, V.; Zachopoulos, A.; Kiriakidis GPapagiannakopoulos, P.
Effect of metal doped and co-doped TiO2 photocatalysts oriented to degrade indoor/outdoor pollutants for air
quality improvement. A kinetic and product study using acetaldehyde as probe molecule. J. Photochem. Photobiol.
A-Chem. 2019, 371, 255–263. [CrossRef]
28. Binas, V.; Stefanopoulos, V.; Kiriakidis, G.; Papagiannakopoulos, P. Photocatalytic oxidation of gaseous
benzene, toluene and xylene under UV and visible irradiation over Mn-doped TiO2 nanoparticles. J. Mater.
2019, 5, 56–65. [CrossRef]
29. Venieri, D.; Tournas, F.; Gounaki, I.; Binas, V.; Zachopoulos, A.; Kiriakidis, G.; Mantzavinos, D. Inactivation
of Staphylococcus aureus in water by means of solar photocatalysis using metal doped TiO2 semiconductors.
J. Chem. Technol. Biotechnol. 2017, 92, 43–51. [CrossRef]
234
Appl. Sci. 2019, 9, 4837
30. Venieri, D.; Gounaki, I.; Bikouvaraki, M.; Binas, V.; Zachopoulos, A.; Kiriakidis, G.; Mantzavinos, D. Solar
photocatalysis as disinfection technique: Inactivation of Klebsiella pneumoniae in sewage and investigation
of changes in antibiotic resistance profile. J. Environ. Manag. 2017, 195, 140–147. [CrossRef] [PubMed]
31. Venieri, D.; Gounaki, I.; Binas, V.; Zachopoulos, A.; Kiriakidis, G.; Mantzavinos, D. Inactivation of MS2
coliphage in sewage by solar photocatalysis using metal-doped TiO2 . Appl. Catal. B Environ. 2015, 178,
54–64. [CrossRef]
32. Venieri, D.; Fraggedaki, A.; Kostadima, M.; Chatzisymeon, E.; Binas, V.; Zachopoulos, A.; Kiriakidis, G.;
Mantzavinos, D. Solar light and metal-doped TiO2 to eliminate water-transmitted bacterial pathogens: Photocatalyst
characterization and disinfection performance. Appl. Catal. B Environ. 2014, 154–155, 93–101. [CrossRef]
33. Binas, V.; Papadaki, D.; Maggos, T.; Katsanaki, A.; Kiriakidis, G. Study of innovative photocatalytic cement
based coatings: The effect of supporting materials. Constr. Build. Mater. 2018, 168, 923–930. [CrossRef]
34. Binas, V.; Venieri, D.; Kotzias, D.; Kiriakidis, G. Modified TiO2 based photocatalysts for improved air and
health quality. J. Mater. 2017, 3, 3–16.
35. Ramirez, A.M.; Demeestere, K.; De Belie, N.; Mäntylä, T.; Levänen, E. Titanium dioxide coated cementitious
materials for air purifying purposes: Preparation, characterization and toluene removal potential. Build.
36. Tsoukleris, D.S.; Maggos, T.; Vassilakos, C.; Falaras, P. Photocatalytic degradation of volatile organics on
TiO2 embedded glass spherules. Catal. Today 2007, 129, 96–101. [CrossRef]
37. Kannangara, Y.Y.; Wijesena, R.; Rajapakse, R.G.M.; Nalin de Silva, K.M. Heterogeneous photocatalytic
degradation of toluene in static environment employing thin films of nitrogen-doped nano-titanium dioxide.
Int. Nano Lett. 2018, 8, 31–39. [CrossRef]
38. Pei, C.C.; Leung, W.W.-F. Photocatalytic oxidation of nitrogen monoxide and o-xylene by TiO2 /ZnO/Bi2 O3
nanofibers: Optimization, kinetic modeling and mechanisms. Appl. Catal. B Environ. 2015, 174–175, 515–525.
[CrossRef]
39. Chen, J.; Kou, S.-c.; Poon, C.-s. Photocatalytic cement-based materials: Comparison of nitrogen oxides and
toluene removal potentials and evaluation of self-cleaning performance. Build. Environ. 2011, 46, 1827–1833.
[CrossRef]
40. Ardizzone, S.; Bianchi, C.L.; Cappelletti, G.; Naldoni, A.; Pirola, C. Photocatalytic degradation of toluene in
the gas phase: Relationship between surface species and catalyst features. Environ. Sci. Technol. 2008, 42,
6671–6676. [CrossRef] [PubMed]
41. Boonen, E.; Beeldens, A. Photocatalytic roads: From lab tests to real scale applications. Eur. Transp. Res. Rev.
2013, 5, 79–89. [CrossRef]
42. Beeldens, A. Air purification by pavement blocks: Final results of the research at the BRRC. In Proceedings
of the Transport Research Arena Europe–TRA, Ljubljana, Slovenia, 21–24 April 2008.
43. Gignoux, L.; Christory, J.P.; Petit, J.F. Concrete roadways and air quality–Assessment of trials in Vanves in
the heart of the Paris region. In Proceedings of the 12th International Symposium on Concrete Roads, Sevilla,
Spain, 13–15 October 2010.
44. Maggos, T.; Plassais, A.; Bartzis, J.G.; Vasilakos, C.; Moussiopoulos, N.; Bonafous, L. Photocatalytic
degradation of NOx in a pilot street canyon configuration using TiO2 -mortar panels’. Environ. Monit. Assess.
2007, 136, 35–44. [CrossRef] [PubMed]
45. Maggos, T.; Bartzis, J.G.; Liakou, M. Gobin Photocatalytic degradation of NOx gases using TiO2 -containing
paint: A real scale study. J. Hazard. Mater. 2007, 146, 668–673. [CrossRef] [PubMed]
46. Atkinson, R.; Baulch, D.L.; Cox, R.A.; Crowley, J.N.; Hampson, R.F.; Hynes, R.G.; Jenkin, M.E.; Rossi, M.J.;
Troe, J. Evaluated kinetic and photochemical data for atmospheric chemistry: Volume I-gas phase reactions
of Ox , HOx , NOx and SOx species. Atmos. Chem. Phys. 2004, 4, 1461–1738. [CrossRef]
47. Finlayson-Pitts, B.J.; Pitts, J.N. Chemistry of the Upper and Lower Atmosphere; Academic Press: San Diego, CA,
USA, 2000.
48. Hunger, M.; Hüsken, G.; Brouwers, H.J.H. Photocatalytic degradation of air pollutants–from modeling to
large scale application. Cem. Concr. Res. 2010, 40, 313–320. [CrossRef]
235
applied
sciences
Article
Study of Passive Adjustment Performance of Tubular
Space in Subway Station Building Complexes
Junjie Li 1 , Shuai Lu 2, *, Qingguo Wang 3 , Shuo Tian 1 and Yichun Jin 1
1 School of Architecture and Design, Beijing Jiaotong University, Beijing 100044, China;
[email protected] (J.L.); [email protected] (S.T.); [email protected] (Y.J.)
2 School of Architecture and Urban Planning, Shenzhen University, Shenzhen 518060, China
3 China Design and Research Group, Beijing 100042, China; [email protected]
* Correspondence: [email protected]; Tel.: +86-10-18500234716; Fax: +86-10-62785691
Received: 11 January 2019; Accepted: 20 February 2019; Published: 26 February 2019
Abstract: The stereo integration of subway transportation with urban functions has promoted the
transformation of urban space via extensive two-dimensional plans to intensive three-dimensional
development. As sustainable development aspect, it has posed new challenges for the design of
architectural space to be better environmental quality and low energy consumption. Therefore,
subway station building complexes with high-performance designs should be a primary focus.
Tubular space is a very common spatial form in subway station building complexes; it is an important
space carrier for transmitting airflow and natural light. As such, it embodies the advantages of
effectively utilizing natural resources, improving the indoor thermal and light environments, refining
the air quality, and reducing energy consumption. This research took tubular space, which has a
passive regulation function in subway station building complexes as its research object. It firstly
established a scientific and logical method for verifying the value of tubular space by searching causal
relationships among the parameterized building space information factors, occupancy satisfaction
elements, physical environment comfort aspects, and climate conditions. Secondly, based on the
actual field investigation, a database of physical environment performance data and users’ subjective
satisfaction information was collected. Through the fieldwork results and analysis, the research
thirdly concluded that the potential passive utilization of tubular space in subway station building
complexes can be divided into two aspects: improvement in comfort level itself and utilization of
climate between natural or artificial. Finally, three typical integrated design method for tubular spaces
exhibiting high levels of performance and low amounts of energy consumption in subway station
building complexes was put forward. This interdisciplinary research provides a design basis for
subway station building complexes seeking to achieve high levels of performance and low amounts
of energy consumption.
Keywords: passive space design; tubular space; physical building environment; fieldwork test;
subway station building complex
Highlights:
(1) Construction of a multi-criteria analysis framework to analyze the passive adjustment
performance of tubular space in subway station building complexes;
(2) Establishment of a database of physical environment performance and occupants’ subjective
satisfaction, based on actual field investigations;
(3) Development of an integrated design idea for tubular spaces in subway station building
complexes that displays a high level of performance and low amount of energy consumption as the
target orientation; and
(4) Proposal of three typical design concepts for compound tubular space.

Appl. Sci. 2019, 9, 834
1. Introduction
1.1. Research Background

With the rapid evolution of urban construction, the Transit-Oriented Development (TOD) mode
has gradually formed a new organizational model for urban public spaces [1–3]. With the expansion
of city subway, subway station building complexes have also entered a period of reinvention [4,5].
The stereo integration of subway transportation with urban functions has promoted the transformation
of urban spaces from extensive two-dimensional plans, to intensive three-dimensional development [6].
Due to China’s rapidly-advancing urbanization, the demand for sustainable development is becoming
more and more urgent [7,8], and the issues of improving occupant comfort and reducing environmental
load must be optimized [9,10]. The significant flow rate of people mainly in pass-through mode has led
to lower environmental quality in above- and underground spaces at the junctions in subway station,
and this may directly affect occupant comfort [11] and health [12]. In addition, large and complex
public buildings tend to occupy a significant proportion of a city’s energy consumption, threatening
the sustainable development of human living environments [13,14].
1.2. Passive Design of Tubular Space in Subway Station Building Complexes

Passive design, which affects the sustainability of architecture from the prototype stage onward,
is an important aspect of green building design [15]. Passive building design does not rely on active
system equipment, but it does depend on a strong capacity for climate adaptability and self-adjustment,
which creates a harmonious indoor coexistence of people and the outdoor environment [16,17]. Passive
architecture describes buildings that are designed to cope with climate factors by providing enduring
and natural comfortable indoor conditions [18,19]. The term “passive” conveys the idea of self-defense
or self-protection of users in architectural design, with respect to the local natural environment [20,21].
A quality passive design avoids the possibility of high levels of energy consumption, saving up to 50%
over traditional methods [22]. Therefore, the architectural prototype generally determines the degree
of sustainability of the building.
Tubular space includes horizontal and vertical corridors in buildings, usually in slender shape,
such as ventilation shafts, patios and lighting tubes, and tunnel corridors for connection. Tubular
space occupies an important proportion of buildings in subway station building complexes, and
its passive regulation has not been deeply investigated [23]. Subway station building complexes
are affected by the characteristics of the mode of space utilization, wherein it is very common to
use tubular space in ground-level and underground spaces (as shown in Figure 1), including patios
and lighting tubes to improve natural lighting efficiency. Other uses include ventilation shafts for
improving the indoor thermal environment and air quality, a station’s traffic tubes for connecting
ground-level and underground stations, and tunnels for traffic transmission. Tubular space can be
seen as a “communications device” that transmits people, mass, and energy to different spaces [24].
This space type is a passive adjustment strategy located between the external and interior environments
of the building. It uses natural energy sources (such as wind, solar energy, and rainwater) and the
natural environment to regulate microclimates and improve the indoor atmosphere. In subway
station building complexes, tubular space has the potential to play an important role in passive
adjustment performance, especially with regards to natural lighting and ventilation, passive cooling,
etc., to optimize comfort and user satisfaction with the indoor space, and greatly reduce the energy
consumption of the building’s operating phase [25].
238
Appl. Sci. 2019, 9, 834
Figure 1. Examples of tubular spaces in subway station building complexes.
In their preliminary analyses of spatial design and climatic contradiction factors, some scholars
have considered the particularity of using underground spaces over ground-level spaces, with respect
to climate [26,27]. From the perspective of the physical environment of underground spaces, the degree
of thermal and light comfort are of great importance. The comfort level of subway transit spaces
and vehicle interiors has been verified by the coupling of actual tests with digital simulations [28].
After six years of actual tests of underground civil air defense space, Yong Li argued for a suitable
acceptable thermal temperature range for underground areas [29]. In recent years, scholars have paid
more attention to occupant health and placed a greater emphasis on ventilation and air quality, by
conducting typological studies, and control and defense research on pollution and particulate matter
(PM). Min Jeong Kim and others have proposed ventilation systems that can improve the platform
PM10 levels and reduce ventilation energy, as compared to manual systems [30]. Practices based on
this theory can be found as early as in ancient Rome, where ancient architects used tubular space to
create more comfortable living environments. For instance, the underground corridor is a very good
example of a kind of air cooling system in use at this time [31]. In terms of modern urban architecture,
Hikarie Shibuya, a Japanese transportation complex designed as an integration concept that considers
the passive utilization relationship between subway station buildings and above-ground structures,
solved the problem of subway station lighting and ventilation [32]. In the retrofitting of the Les Halles
area of Paris, the utilization of tubular space was adopted in underground spaces to provide natural
lighting [33].
1.3. Objective of this Study

This study addressed a variety of forms of tubular space in city subway station building complexes,
screening those spaces for passive adjustment potential in order to study the effect on the comfort
level of the indoor environment and overall energy consumption. Through a fieldwork evaluation
of building performance and occupant satisfaction in actual built projects, this research conducted a
quantitative performance evaluation of passive architectural design strategies for tubular spaces.
Through an analysis of the current situation and excavation of the spatial potential, this work
determined the passive adjustment performance effects for subway station building complexes in
terms of sustainable development, therefore providing a basis for improving architectural design
methods to show higher levels of performance and lower amounts of energy consumption. This
research pursued the following three objectives: (1) provide a basis for design by analyzing the types of
passive function and specific variables for the technical strategies employed by subway station building
complexes; (2) test the passive adjustment effects of subway station building complexes and improve
the authenticity and objectivity of the designs via actual and effective environmental monitoring
evaluation; and (3) explore typical further-optimized strategy models for the compound tubular space
systems of subway station building complexes, and provide guidance for design optimization.
2. Methodology
This research was based on the dual perspectives of architecture and the built environment.
With regards to architectural design, this work produced a space prototype and deconstructed the
239
Appl. Sci. 2019, 9, 834
factors affecting the passive adjustment performance of architecture. According to a comparison of the
factors that influence the quality of indoor buildings and actual built environments, a comprehensive
evaluation was made of the passive adjustment performance of tubular spaces in subway station
building complexes [34]. First, the factors that affect the passive adjustment performance were
analyzed. Then, according to the analytic factors and taking the urban Beijing subway station building
complexes as an example, a long-term physical environment test was carried out. The subject was
a subway station building complex with a typical amount of tubular space. This research focused
on the physical environment, as tube as the passive function of potential space and its influence on
surrounding functional areas. It included actual measurement results such as the thermal conditions,
air ventilation, lighting environments, indoor air quality, occupant satisfaction and comfort, and other
subjective feedback. Through this objective investigation of the physical environment and subjective
feedback of the occupants’ degree of comfort, the problems with objective space were able to be
studied and analyzed, and the potential for spatial optimization put forth from the perspective of
passive adjustment performance. Finally, the database established through this research assisted in
highlighting the design goals for tubular space in subway station building complexes. A model for
three typical kinds of composite tubular spaces was constructed with the goal of achieving high levels
of performance and low amounts of energy consumption. Therefore, this research method was divided
into the following four steps. (as shown in Figure 2)
Figure 2. Research methodology route.
2.1. Stage One: Factor Analysis of the Effect of Passive Design on Tubular Space
Passive design belongs to the category of sustainable development in architecture, and is an
essential part of addressing three factors: the environment, society, and the economy [35]. Tubular
space is a typical passive spatial design strategy that coordinates contradictions among these three
factors and architecture, leading building construction in a more positive direction. Architecture
is a carrier of the climate and its human occupants; building space and outdoor climate conditions
can be seen as the reasons for indoor physical environments and occupant satisfaction [36]. This
research is based on an AHP (Analytic Hierarchy Process) methodology which decomposes complex
issues into several group factors, and compares those factors to one another to determine their relative
importance [37,38]. It adopted the method of factor quantification analysis to classify climate conditions,
building spaces, physical environmental comfort, and occupancy satisfaction. This logical framework
was established through measurements and simulations; the influences therein were determined by
a correlation analysis, based on the acquired data. Therefore, the passive design of tubular space
was divided into four factor groups: spatial parameters, climate parameters, the degree of physical
environmental comfort, and occupancy satisfaction. (as shown in Figure 3)
240
Appl. Sci. 2019, 9, 834
6SDWLDOLQIRUPDWLRQSDUDPHWHUVDIIHFWLQJWKH
SHUIRUPDQFHRIZHOOVSDFH
Q DVV
WLR R
DQD FLDWLR
RFLD LV
DVV DO\V 6SDWLDO3DUDPHWHU O\V Q
LV
DQ
3K\VLFDOHQYLURQPHQW
SHUIRUPDQFHPRQLWRULQJ
LQFOXGLQJWKHUPDO
2FFXSDQWVVXEMHFWLYH 3DVVLYHGHVLJQRIZHOO 3K\VLFDO(QYLURQPHQW
2FFXSDQF\6DWLVIDFWLRQ HQYLURQPHQWZLQGVSHHG
VDWLVIDFWLRQYRWLQJ VSDFH &RPIRUW OLJKWLQJHQYLURQPHQWDQG
DFRXVWLFVHQYLURQPHQW
LQVLGHZHOOVSDFH
Q
DVV WLR
R &OLPDWH3DUDPHWHU RFLD LV
DQD FLDWLR DVV DO\V
O\V Q
LV DQ
&OLPDWHSDUDPHWHUVVXFKDVVRODUUDGLDWLRQ
WHPSHUDWXUHZLQGVSHHGUDLQIDOODQGVRRQ

Figure 3. Analysis of the factors influential on tubular space in subway station building complexes.
According to authors pervious research in passive space design [36,39], the building space can be
recognized as four categories including “shape”, “mass”, “quantity”, and “connection”. This research
also adopts this research framework that the factor groups were further quantified and subdivided.
The spatial parameters were again divided into four sub-factors: geometric dimensions, interface
properties, and internal and external related categories (as shown in Table 1). Researchers have made
useful contributions to indoor environment comfort research [40]. For example, Professor Fanger
highlighted six elements that influence the comfort level of an indoor thermal environment (the mean
radiation temperature, air temperature, relative humidity, air speed, clothing insulation, and metabolic
rate) to form the Predicted Mean Vote (PMV) model [41]. LEED [42] in USGBC and China ESGB [43]
lists lighting environment evaluation factors (including natural light, artificial light, view, and control)
and offers clear standard indicators. As the tubular space can have a significant wind effect that affects
the indoor air flow, while most tubular space exists underground, and the outdoor noise environment
has little influence on the interior, therefore, outdoor climate parameters can be divided into four
additional sub-factors: thermal environment, lighting environment, air quality, and wind speed.
Corresponding to the outdoor climatic conditions, the indoor physical environment also includes four
sub-factors: thermal environment, lighting environment, air quality, and air velocity. There are many
factors involved in occupancy satisfaction. The method of semantic differential (SD) analysis [44] was
used to divide the influencing factors into eight aspects: thermal conditions, humidity, light, air quality,
air velocity, ease of use, cleanliness and maintenance, and overall environmental satisfaction (as shown
in Table 1).
241
Appl. Sci. 2019, 9, 834
Table 1. Factor analysis of the passive adjustment effect of tubular space.
Parameter Acquisition
Factor Group Factors Parameter Unit
Method
geometric dimensions (L:W:H) distance measurement m
material thermal
interface property (U-value) W/(m2 ·K)
performance calculation
Spatial parameter
internal related categories (visitor flow rate) statistics N/h
external related categories (outdoor,
judgment N/a
platform, commercial, none)
thermal environment ◦ C, %
(temperature, relative humidity)
Climate parameter lighting (illuminance) Lux
air quality (PM2.5, PM10, CO2 ) μg/m3 , ppm
wind speed measurement m/s
thermal environment ◦ C, %
(temperature, relative humidity)
Degree of comfort
with the physical lighting (illuminance) Lux
environment air quality (PM2.5, PM10, HCHO, CO2 ) μg/m3 , ppm
air velocity, wind temperature m/s, ◦ C
thermal comfort
humidity
air quality
Occupancy lighting Vote score

occupant survey
satisfaction ventilation [−3~3]
ease of use
cleanliness and maintenance
overall environmental quality satisfaction
2.2. Stage Two: Field Survey

Corresponding with the public factors affecting the passive function of tubular space,
the actual field investigation involved spatial drawings, monitoring the indoor and outdoor physical
environments, and determining occupants’ subjective levels of satisfaction. The relationship between
buildings and people, especially in terms of the healthiness of the indoor environment, has a significant
influence on human survival and sustainable development. Since the normal operating phase tends
to be from 6:00 a.m. to 10:00 p.m., the opening time of building complexes (offices or businesses) is
usually included within that time frame. Therefore, the survey had a clear research plan regarding
a day cycle time, from 6:00 a.m. to 10:00 p.m., including two rush hours where there was peak
human flow. The physical environment test called for the selection of a typical space, such as tube
entrance, middle tubular space, connection point between a subway and complex building, station
hall, or subway platform where the long-term physical environment could be monitored. The data
were collected every five minutes. The physical quantities included nine parameters: temperature,
humidity, illuminance, CO2 concentration, PM2.5, PM10, HCHO, air velocity, and wind temperature.
The test contents including long-term consecutive days of outdoor temperatures which measurement
interval was 5 min, temperature measurements for each (selected) test point which measurement from
6:00 a.m. to 10:00 p.m. for each typical day; measurement interval was 5 min.
In addition, since piston wind can affect the interior tubular space and connection points of urban
complexes during subway operation, and the aerodynamic forces of piston wind may be usable as a
source of renewable energy [45], the fieldwork test also included instantaneous wind speed changes at
242
Appl. Sci. 2019, 9, 834
the subway platform level. The observation frequency was 3 s, with a train cycle of arrival, stay, and
departure (as shown in Table 2).
Table 2. Building the physical environment fieldwork test framework.
Parameter Test
Measurement Items Properties of the instruments
Type Content
outdoor temperature test
temperature ◦C
indoor temperature test for each Portable infrared temperature meter,
Thermal (selected) test point Biaozhi GM700, Range: −50~700 ◦ C,
environment
indoor humidity test for each Resolution: 0.1 ◦ C
humidity %
(selected) test point
outdoor luminance test Portable luminance meter, Reggiani
Lighting lux/daylight
indoor luminance test for each luminance DT-1301, Range: 0~50,000 Lux,
factor %
(selected) test point Resolution: 1 Lux
outdoor CO3 concentration test Portable and self-record CO2 meter,

CO2 ppm TJHY-EZY-1, Range: 0~5000 ppm,
indoor CO2 concentration test concentration Resolution: 1 ppm
for each (selected) test point
outdoor PM2.5/10
IAQ concentration test PM2.5/10
μg/m3
concentration Portable air quality meter, temopt
indoor PM2.5/10 concentration
LKC-1000S+, Range: 0~999 mg/m3 ,
test for each (selected) test point
Resolution: 0.01 mg/m3
indoor HCHO concentration
HCHO g/cm3
test for each (selected) test point
Self-record instrument for wind
indoor air velocity test for each velocity and wind temperature,
air velocity m/s
(selected) test point TJHY-FB-1, Range: 0~10 M/S,
Resolution: 0.01 M/S
Ventilation
Self-record instrument for
indoor air temperature of each ◦C
air temperature environment, TJHY-HCZY-1, Range:
(selected) test point
0~5000 ppm, Resolution: 1 ppm
The object of this investigation was the tubular spaces in a subway station building complexes,
so the function and shape of the spaces were simpler and more explicit than other common building
spaces. The design of the subjective questionnaire focused on the degree of occupancy comfort and the
space’s influence on human health during short stays. In general, the questionnaire included three
categories: satisfaction with the physical environment; space satisfaction votes, such as ease of use,
cleanliness, and maintenance; and overall satisfaction with the environment’s quality.
In order to improve the efficiency of the investigation, spatial drawings, subjective satisfaction
research, and studies of comfort related to the objective physical environment were all conducted.
The characteristics of the subway space in utilization mode were special in that mostly the area was
designed for a rapidly-passing crowd who would be present only for a short stay. The connection
spaces often fluctuated in terms of the physical environment, and users’ moods and physical health
conditions likely varied considerably. Therefore, the method of subjective investigation also needed
to be more diverse than in other similar types of research. The questionnaire was collected mainly
by three means: website-based and on-site surveys, and on-site interviews. The purpose of the
website-based questionnaire was to avoid misunderstandings related to temporally-subjective factors,
and dispel individual elements through the long-term accumulation of memory. For the satisfaction
and self-reported productivity questions, the survey used a 7-point semantic differential scale with
endpoints of “very dissatisfied” and “very satisfied.” For the purposes of comparison, the scale
was assumed to be roughly linear, with ordinal values for each of the points that ranged from
−3 (very dissatisfied) to +3 (very satisfied) and 0 as the neutral midpoint [46]. The on-site interviews
were with specifically-selected occupants who stayed in the space for a long period of time, such as
retail vendors, subway station operators, commercial building security, cleaning and maintenance
staff, etc. The research methods were not rigidly obeyed for prescribed problems and formats, and
243
Appl. Sci. 2019, 9, 834
were altered via conversations to help the researchers understand the respondent’s age, cultural and
economic background, space satisfaction, environmental problems, etc (as shown in Table 3).
Table 3. Occupancy satisfaction voting framework.
Parameter
Test Items Test Content Test Content
Type
thermal
comfort
Physical web-based survey/
humidity 7-point scale [−3,−2,−1,0,1,2,3]
environment vote fieldwork-based survey/
very dissatisfied to very satisfied
satisfaction air quality human perception test
lighting
ventilation
ease of use web-based survey/ 7-point scale [−3,−2,−1,0,1,2,3]
Space fieldwork-based survey/ SD of feelings about space’s
vote
satisfaction cleanliness and
human perception test atmosphere
maintenance
web-based survey/
Overall environmental 7-point scale [−3,−2,−1,0,1,2,3]
vote fieldwork-based survey/
quality satisfaction very dissatisfied to very satisfied
human perception test
2.3. Stage Three: Problem and Analysis of the Spatial Potential

Stage two set up a framework for a comprehensive system that focused on three factors:
architecture, humans, and the environment. Based on the conclusions made during that stage, the
passive strategy factors affecting the research object were classified and recombined to analyze the
functional characteristics of different positions of space that can be found throughout a subway
station building complexes. According to the research by Margarita N. Assimakopoulos regarding
the thermal environment in Greek subways [47], Teresa Moreno and colleagues’ work on airborne
particulate matter in the Barcelona subways, and John Burnett [48] and associates’ investigation and
analysis of the lighting environment in the Hongkong metro in China [49], obvious problems such
as high humidity, low thermal comfort, and poor air quality and lighting environments in subway
halls and platform spaces all emerged as worthy of further research. Many scholars also argued for
energy-saving strategies in subway systems by means of passive ventilation designs for complete
ventilation systems, for instance by developing ventilation systems in subway stations that could
control indoor air pollutants [50,51]. The core of the third stage of this study includes two aspects. First,
through an investigation and analysis of the status quo, the existing environmental problems were
extracted and a design strategy put forward to resolve certain issues. Secondly, through statistical data,
the researchers discovered the potential capacity of passive space, and identified design opportunities
that could improve comfort, health, and energy efficiency.
2.4. Stage Four: Set up New Target Orientation and Space Update
The fourth step of this research put forward the spatial design goal of tubular space in
subway station building complexes, from the perspective of sustainable development. Certain space
assumptions and a particular design procedure for the compound tubular space were promoted to
provide guidance for the optimized design.
Through data and problem analyses and potential excavation, this research searched typical
models of complex tubular space complex systems that would be applicable to subway station building
complexes. The key point was to determine an applicable and feasible space utilization model that
could provide a design basis. The researchers put special emphasis on viable applications for potential
natural resources in subway spaces, such as tunnel and piston wind, pull shafts, lighting, and landscape
tubes, that could be further integrated into the design. Typical composite tubular spaces can be in the
form of a tube tunnel (a solar chimney composite space system), combined wind tunnel (a displacement
244
Appl. Sci. 2019, 9, 834
ventilation complex space system), combined active and passive ground source heat pump (wind
tunnel composite tubular space system), hot air shaft ventilation, or lighting composite space system.
Figure 3 offers an overall view of the study.
3.1. Building Space Information Factors

Based on the above-mentioned factors, a multi-criteria evaluation method for tubular space was
proposed. The survey selected five typical subway station building complexes in Beijing, including
Xizhimen (W1), Haidianhuangzhuang (W2), Guomao (W3), Dawanglu (W4), and Wangfujing Stations
(W5) (as shown in Figure 4). The selected five stations involved six main subway lines: #1, #2, #4, #10,
#13, and #14. All of the surveyed stations were urban subway transit hubs. Xizhimen Station (W1) is
where the #2, #4, and #13 subway lines converge; it is also home to three high-rise commercial office
buildings and the Beijing subway station. Haidianhuangzhuang (W2), Guomao (W3), and Dawanglu
Stations (W4) all are points of convergence for more than two subway lines, and offer connections
with urban complexes. Wangfujing Station (W5) is located in the middle of Beijing, and is connected to
the largest integrated commercial building in Asia. As such, it features a substantial people flow rate;
the location is also of geographical importance (as shown in Table 4).
Figure 4. Five typical subway station building complexes in Beijing.
The test period was selected to be from 28th June 2017 to 20th July 2017, the highest-temperature
time period in Beijing. It is a typical and continuous testing period of 3 weeks. The data excluded
unstable factors which may conduct to instantaneous mutations data such as weather mutations,
active equipment interference, people behavioral interference, misuse of testing instruments, etc,
and used a mean value within the 3 weeks. The purpose of this research was to investigate the
performance of the tubular space in the physical environment under the most unfavorable conditions
in the summer climate.
漏̓漐
245
Appl. Sci. 2019, 9, 834
Table 4. Survey object information.
Number
City Building Building Subway
No. Station Name of Test Type of Test Point
Complex Function Area (m2 ) Line
Points
middle tubular space,
Commercial,
W1 Xizhimen Cade mall 89,000 #2, #4, #13 3 tube entrance,
office
platform layer
Gate City middle tubular space,
W2 Haidianhuangzhuang Commercial 47,000 #4, #10 3
mall tube entrance
Commercial,
W3 Guomao Yintai mall 350,000 #1, #10 4 tube entrance, platform
office
layer, station hall
China
Trade Commercial,
W4 Dawanglu 710,000 #1, #14 4 tube entrance, platform
Center office
layer, station hall
mall
Oriental
W5 Wangfujing Commercial 120,000 #1 4 tube entrance, platform
Plaza mall
layer, station hall
Figure 5. Test station space plan and test points location.
The first step in the fieldwork survey was to acquire the building information. This was performed
in order to obtain the data supporting the tubular spaces. After the fieldwork test, the study first
246
Appl. Sci. 2019, 9, 834
drew out five site plans and geometric scale parameters for the tubular space (as shown in Figure 5).
The researchers then selected three or four test points for each site station; all tests contained 18
measuring points. Each test point had a certain representativeness. The W1 site contained three
test points; tp1 was located in the middle of an above-ground glass corridor between the station
and the complex, and was a middle tubular space.Tp4 and tp11 were located between the sites and
underground associated tubes, which were all classified as middle tubular space. Tp8 was a link
between a subway station and commercial building, which was also middle tubular space. Tp3, tp7,
and tp14 belonged to the first kind of tube entrance space, and connected the commercial complex
building at one end of the tube shaft. Tp2, tp5, tp10, and tp13 belonged to the second kind of tube
entrance space, and connected to the station hall. Tp6, tp9, and tp12 were the test points at the subway
station hall (as shown in Figure 5). Finally, tp15, tp16, tp17, and tp18 were the subway platform test
points for W1, W3, W4, and W5, respectively.
Because the locations of the tubular spaces in each subway station building complexes were
different, the environmental problems varied dramatically. Therefore, this research compared the
physical environment parameters of the 18 measuring points, according to five types: middle
tubular space (four test points), tube entrance (connection to building) (three test points), tube
entrance (connection to subway) (four test points), station hall (four test points), and platform layer
(four test points).
3.2. Field Survey Results and Analysis
3.2.1. Physical test results and analysis

Professor Fanger highlighted six elements that influence the comfort level of an indoor thermal
environment (mean radiation temperature, air temperature, relative humidity, air speed, clothing
insulation, and metabolic rate); all six are necessary to form the Predicted Mean Vote (PMV) model [41].
Tubular space in subway station building complexes is mostly underground, so the influence of
radiation temperature can be ignored and people’s metabolic rates can be set to the same level of
1.5 met [41]. During the test period, the hottest period in summer in Beijing was selected, so clothing
level was chosen as 0.35 clo [41]. Data from this survey were collected according to eight physical
parameters: environment temperature, humidity, illuminance, air velocity, PM2.5, PM10, and the
HCHO and CO2 concentrations at each point. The average result values are shown in Tables 5–9.
According to the current national standards and norms, the typical range of thermal comfort is defined
as between 16~28 ◦ C [43], humidity comfort is between 30~60% [42], illumination should be no higher
than 150 lux [52], and indoor air velocity in winter should be lower than 0.15 m/s and 0.25 m/s in
summer. The concentrations of PM2.5 and PM10 should be lower than 75 μg/m3 and 150 μg/m3 ,
respectively, according to the 24-hour average concentration limits of the two grades listed in the
national standard [53]. The concentration of HCHO and CO2 should be lower than 0.08 mg/m3 [54]
and 1000 ppm [55], respectively.
(1) Comfort analysis
Figures 6–9 are box diagrams of the physical test results for thermal conditions, lighting, IAQ,
and ventilation in all five types of tubular space. The red zones show locations where the occupant
comfort values were beyond the related comfort standard.
As regards the thermal environment, tp1 was a solar corridor space with a higher temperature
than the other three points, which were at the boundary of the comfort zone (the point at which the
human body would no longer enjoy thermal comfort). The tube lengths of tp4 and tp8 were 150 m and
140 m, respectively, and the humidity levels of the two test points significantly exceeded the standard.
The highest reached 84.7%. The temperature in the space was lower than the human comfort level
standard would deem acceptable, and the excessive humidity could easily cause mildew and affect
users’ health. As regards the lighting environment, the illuminance levels of tp4 and tp8 in the middle
tubular space were not sufficient; the values did not reach the national standard requirement and
247
Appl. Sci. 2019, 9, 834
thus could be hiding potential dangers. Due to the large amount of natural light at tp1, the average
illuminance could reach 990 lux without artificial lighting. A better lighting environment would also
improve the quality of the indoor environment. Past research results have shown that occupants
become uncomfortable and can even experience headaches and chest tightness, causing their work
ability to decline, when the CO2 concentration is over 1000 ppm. In the test of CO2 concentration
for the middle tube, the two longer tube sat tp4 and tp8 had lower air quality; the highest CO2
concentration was at tp4, which reached 1931 ppm. This is over two times the standard. The maximum
PM2.5 concentration at tp8 reached 121 g/m3 ; the outdoor concentration was 74.6 g/m3 . Thus, the
disadvantages were greatly exacerbated. Although the air velocity values at tp4 and tp11 were slightly
higher than the standard, moderate ventilation in a thermal environment can lower body surface
temperature, taking away the sweat that collects on the human body’s surface.
To sum up, the main problems as regards comfort in the middle tubular space were its high
humidity, poor light environment, and low air quality.
Table 5. Average values for the physical test results in the middle tubular space.
Test Air
Site Temperature Humidity Illuminance
Point Velocity PM2.5 PM10 HCHO CO2
No. (◦ C) (%) (Lux)
Number (m/s)
W1 Tp1 32.1 47.7 990 0.196 47.9 67.2 0.022 681
W2 Tp4 28.1 70.2 72.0 0.265 45.4 64.3 0.030 1163
W3 Tp8 25.5 76.8 19.0 0.0 66.1 92.0 0.057 1000
W4 Tp11 28.3 57.4 153.6 0.307 52.0 73.1 0.044 892.1
Note: The data in the grey background indicate that it is beyond comfortable zone.
Figure 6. diagram of the physical test results from the middle tubular space.
Figure 7 shows the performance of the physical environment of the tube entrance space approach
to one end of a shopping mall. All of the environment’s temperatures and most of the humidity
index values for the three test points exceeded the comfort zone. The maximum value of tp3 was
32.3 ◦ C, while the mean outdoor temperature was 29 ◦ C. The average illuminance levels of the three
test points in the tube entrance (connection to building) did not reach the national standard; tp3 and
tp7 were especially low. Both of the test points located at the indoor and outdoor junctions had light
248
Appl. Sci. 2019, 9, 834
levels that would require human eyes a significant amount of time to adjust to the darkness. This can
cause sensations of insecurity and discomfort when entering from a strong outdoor light environment.
The concentration of CO2 in these spaces was also high, with maximum values for tp3 and tp7 reaching
1761 ppm and 1690 ppm, respectively.
To sum up, because most of the tube entrance space was connected to a shopping mall and was
close to the outdoors, the thermal environment and humidity levels were low, light problems were
obvious, and concentrations of CO2 were high.
Table 6. Average values for the physical test results from the tube entrance (connection to a building).
Test Air
No. (◦ C) (%) (Lux)
Number (m/s)
W2 Tp3 30.6 62.1 7.5 0.117 54.8 76.3 0.055 981.9
W3 Tp7 30.4 65.2 24.4 0.144 73.5 102.8 0.126 1308.2
W5 Tp14 28.9 52.9 147.9 0.604 32.2 45.0 0.047 928.9
Figure 7. Box diagram of the physical test results from the tube entrance (connection to building).
According to Figure 8, the thermal and humidity environments at the tube entrance space near
the subway were obvious. The temperatures at the four test points all exceeded the comfort level.
The average temperature at tp5, with an elevation depth of 14.4 m, was 2.5 ◦ C lower than that of tp10
and tp13, which were at an elevation depth of 9.2 m. The higher wind speed relieved the high humidity
and CO2 concentrations at the tube entrance space (connection to subway). Generally, the average
humidity was lower than 70% and both the average and maximum values of CO2 were greatly reduced.
However, due to the influence of the subway piston wind, the wind speed presented a sinusoidal
fluctuation and the direction of the wind speed changed periodically. Thus, the wind environment
became the most unfavorable factor with regards to comfort.
To sum up, ventilation was one of the most detrimental comfort factors at the tube entrance space
(connection to subway). It was affected by the piston wind so that cold air (from the air conditioning)
was sometimes sucked out of the subway platform layer and hot air was pumped out of the tubular
space. People who stayed at that location for long periods of time were frequently affected by two
249
Appl. Sci. 2019, 9, 834
kinds of wind that had large temperature and direction differences. They expressed great discomfort
and likely were experiencing threats to their health.
Table 7. Average values for the physical test results from the tube entrance (connection to subway).
Test Air
No. (◦ C) (%) (Lux)
Number (m/s)
W1 Tp2 30.8 50.0 154 0.365 53.5 75.9 0.019 688.4
W2 Tp5 28.6 66.0 32.5 0.859 52.7 73.6 0.019 840.5
W4 Tp10 31.3 68.3 107.3 1.73 107.6 151.1 0.021 920.8
W5 Tp13 31.1 70.0 91.2 0.683 149.2 201.7 0.026 802.4
Figure 8. Box diagram of the physical test results at the tube entrance (connection to subway).
The physical environment test at the station hall focused on Beijing subway line #1, and included
three test points. As can be seen from Figure 9, the environment temperature in the station hall was
significantly higher than the comfort zone and 5 to 6 ◦ C higher than the outdoor temperature at that
time. The humidity in the station hall was also relatively high, but the light environment basically
met the national standard. Due to a large number of people flowing through the middle of the station
hall and the relatively longer tube lengths (130 m, 190 m, and 160 m), the CO2 concentrations at
the three points were correspondingly higher. The peak concentration of CO2 at the tp6 measuring
point reached 1833 ppm. Tp12 was located at a point with a smaller people flow rate, so its CO2
concentration stayed within the standard range. However, due to the impact of the subway piston
wind, there was a high level of PM pollution in the subway tunnel that was brought into the entrance
hall, with a peak concentration of 160 g/m3 . Compared with the outdoor concentration 74.6 g/m3 ,
this was two times the outdoor concentration at that time. Thus, it can be seen that a large number
of people were gathered at the station and hall levels, resulting in a higher concentration of CO2 .
The station hall layer was directly connected to the train platform and had a higher PM concentration
due to the influence of the piston wind from the subway.
In summary, influenced by the space size and flow rate, the physical environment of the hall at all
stations was the worst, which is reflected in the high temperature and humidity, and poor air quality.
250
Appl. Sci. 2019, 9, 834
Table 8. Average values of the physical test results from the station hall.
Test Air
Site Temperature Humidity Illuminance PM PM
Point Velocity HCHO CO2
No. (◦ C) (%) (Lux) 2.5 10
No. (m/s)
W3 Tp6 33.5 62.2 144.1 0.724 68.6 96.0 0.038 1374.4
W4 Tp9 32.6 61.9 152.3 0.446 100.9 139.2 0.023 1153.6
W5 Tp12 31.2 71.4 92.0 0.172 119.6 167.2 0.021 922.5
Figure 9. Box diagram of the physical test results from the station hall.
The physical environment of the platform layer is shown in Figure 10. Since the wind speed at the
platform layer was significantly affected by the movement of the subway vehicles, it will be discussed
in detail in the next chapter. The lighting environment met the national lighting standards. There
were two key problems with comfort: the thermal environment and air quality. The temperature was
generally too high; the highest value was from W4, which reached 34.5 ◦ C. The values were 3.5 ◦ C
higher than the outdoor temperature at that time. For air quality, the most significant problem was the
PM concentration. The PM2.5 data for almost all of the test sites were higher than the human body’s
comfort range; the PM10 concentration was too high at the W4 site as tube, reaching a maximum of
225.7 g/m3 , while the outdoor concentration was 83.2 g/m3 .
To sum up, the temperature and PM values were the key problems with comfort at the
platform layer.
Table 9. Average values for the physical test results from the platform layer.
Test
Site Temperature Humidity Illuminance Air Velocity PM
Point PM2.5 HCHO CO2
No. (◦ C) (%) (Lux) (m/s) 10
No.
W1 tp15 30.2 48 260.3 Instantaneous 85.3 121.7 0.017 974.5
W3 tp16 31.6 64.3 263 wind speed 98.5 138.3 0.019 867
W4 tp17 33.9 58.8 259.5 (as shown in 137.6 192.5 0.02 1005
W5 tp18 33.7 69.6 251.2 Figure 12) 72.7 102.3 0.019 1283
251
Appl. Sci. 2019, 9, 834
Figure 10. Box diagram of the physical test results from the platform layer.
(2) Changes in physical quantities with time and human flow

a: Temperature
The green curve in Figure 11 shows the fluctuation in outdoor temperature. Except for the
aboveground space of tp1, the temperature at other points was not directly affected by solar radiation;
therefore, the temperature curve hardly varied over time. The overall temperature environment at
the middle of the tubular space was the best, followed by the tube entrance space near the shopping
center. The third best was the tube entrance space near the subway measurement point. The thermal
environment of the station hall was the worst, maintaining the highest temperature nearly the entire
day. Changes in the flow rate had little influence on the thermal environment of the tubular space;
the relative position was the decisive factor for the temperature there.
252
Appl. Sci. 2019, 9, 834
(a) (b)
(c) (d)
Figure 11. Comparison curves of hourly temperature data for each point during the monitoring day.
(a) Middle tubular space; (b) Tube entrance (connection to building); (c) Tube entrance (connection to
subway); (d) Station hall.
b: CO2 concentration
In the diurnal variation curve for CO2 concentration, there was almost no consistent fluctuation
pattern for the same type of tubular space. From the overall curvilinear relation in Figure 12, it can be
seen that there were two main trends. The first was the stable value for the whole day; the range of
change was not large, as illustrated by test points tp8, tp13, and tp14. The second was the dramatic
changes during two time periods, 8:00~9:00 a.m. and 4:30~7:30 p.m., which indicated that the
concentration of CO2 was significantly affected by indoor people flow, and the elevated concentration
occurred during morning and evening peak hours. In addition, tp3, tp4, tp5, tp6, tp7, tp9, tp10, and
tp12 exposed a lack of air adjustment capacity when a large number of people gathered together.
253
Appl. Sci. 2019, 9, 834
(a) (b)
(c) (d)
Figure 12. Comparison curves of hourly CO2 concentration data for each point during the monitoring
day. (a) Middle tubular space; (b) Tube entrance (connection to building); (c) Tube entrance (connection
to subway); (d) Station hall.
c: PM2.5/PM10
As shown in Figure 13, the PM2.5 and PM10 data for each test point varied with time. This study
took the middle of the tubular space to be the object of analysis. There were change rules over time for
four test points: tp1, tp4, tp8, and tp11. The survey found that three points—tp1, tp4, and tp8—had a
sudden increase in PM concentration at 11:00 a.m. and 6:00 p.m., almost two times that of other times.
Tp11 was located in the tubular space near the #14 subway line, and was affected by the piston wind
from the subway platform. The variations in PM concentration throughout the day were large, and the
regularity presented was closely related to the subway’s operation time.
254
Appl. Sci. 2019, 9, 834
W1 tp1 W2 tp4
W3 tp8 W4 tp11
Figure 13. Comparison curves of hourly particulate matter 2.5 (PM2.5)/PM10 concentration data for
each point during the monitoring day.
d: Wind speed
The test point height was 1.5 m (tp1, as shown in Figure 14a). One operating cycle for a train
(arrival, stay, and departure) was around three minutes. Due to the piston effect caused by the train’s
operation, the tunnel wind reached remarkable levels. The wind speed when the train was arriving
lasted for 15 s, with an average wind speed of 1.2 m/s. The stage during which the train stayed at
the station lasted for 40~50 s (at the test point it was 48 s), with an average wind speed of 0.84 m/s.
The train’s departure stage lasted longer, about 120 s, with a maximum instantaneous wind speed of
up to 3.6 m/s and an average wind speed of 1.79 m/s. During each three-minute cycle, the maximum
wind speed was 3.6 m/s during the departure stage, while the minimum wind speed was 0 m/s
during the period when the train stayed at the station (as shown Figure 14b). The wind direction was
opposite during the arrival and departure stages. These two wind directions were off set when the
train stayed at the station, presenting a brief state of calm. According to the coupling experiment in
the Beijing subway conducted by Mingliang Ren and others, the maximum wind speed near the train
could reach 7.5 m/s at the test point height of 2 m [56] (tp2, as shown in Figure 14a).
3.2.2. Occupancy satisfaction survey results and analysis

The subjective occupancy survey was comprised of two parts: on-site questionnaires and on-site
interviews. The subjective questionnaire adopted a 7-point scale, where −3 corresponded with
“Very Dissatisfied”, 3 referred to the respondent being “Very Satisfied”, and 0 was neutral (as shown in
Table 10). The subjective questionnaire was issued at each test point (tp1 to tp14), 40 copies each, for a
total of 560 copies. There was a total of 551 valid questionnaires completed, for a recovery rate of 98%.
The test period was selected to be the same as the physical environment test, the highest temperature
255
Appl. Sci. 2019, 9, 834
period for the Beijing area, from 28 June 2017, to 20 July 2017. The purpose of this research was to
investigate the performance of the tubular space in terms of occupancy satisfaction under the most
unfavorable conditions in the summer climate. In addition, the survey also included on-site interviews.
Researchers talked with subway staff, retail traders, security guards, cleaners, and passers-by for a
substantial period of time.
(a) (b)
Figure 14. Wind speed curve for one operating cycle of a train. (a) Test point location; (b) Curve line.
Table 10. Questionnaire for the occupancy satisfaction survey.
Occupancy
Very Dissatisfied Neutral Very Satisfied
Satisfaction
Thermal comfort
Humidity
Air quality
Lighting
Ventilation −3 −2 −1 0 1 2 3

Ease of use
Cleanliness and
maintenance
Overall
environmental
quality satisfaction
Table 11 shows the average data collected from the questionnaire. Corresponding with the analysis
of the physical environment test results, this research also classified the 14 test points according to
their location. There were a total of four types: middle tubular space, tube entrance (connection to
building), tube entrance (connection to subway), and station hall.
Based on a histogram analysis of Figure 15, the researchers found that occupants had higher levels
of satisfaction in the middle tubular space, and most of the collected data were positive. The problems
with higher temperatures and poor lighting were clear in the physical environment test, but the
subjective feelings of the users were not obvious. However, in terms of humidity, both the subjective
questionnaire and the interviews reflected the occupants’ discomfort.
The satisfaction results at the tube entrance (connection to building) fell into two categories.
The first type was in relation to where the tube entrance connected the building to the outdoor space
(tp3 and tp7). All indexes of satisfaction were low, especially in terms of thermal comfort, humidity,
256
Appl. Sci. 2019, 9, 834
air quality, and convenience. The other type related to where the tube entrance connected to the
underground building (tp14), which demonstrated positive advantages for all indicators.
Table 11. Results data from occupancy satisfaction survey.
Occupancy
Satisfaction Tp1 Tp2 Tp3 Tp4 Tp5 Tp6 Tp7 Tp8 Tp9 Tp10 Tp11 Tp12 Tp13 Tp14
Vote
Thermal Comfort −0.23 1.1 −0.9 1.3 −1.1 −1.8 −2.3 2.1 −2.1 −1.2 0.9 −2.2 −2.5 1.9
Humidity −0.25 0.2 −1.5 −1.6 −2.5 −2.2 −2.2 −0.6 −2.4 −1.5 0.8 −2.3 −2.1 1.5
Air Quality −0.15 −0.3 −1.2 0.2 −1.9 −2.6 −1.9 1.1 −1.5 −1.2 1.1 −2.3 −0.5 1.3
Lighting 2 −0.9 0.5 0.3 −0.8 0.2 −0.6 1.2 0.8 0.8 1.2 −0.5 0.2 1.5
Wind Comfort −0.9 1.1 −0.6 0.5 −2.6 −0.5 −0.5 1.4 0.5 −1.6 1.2 −1.1 −2.4 −0.1
Ease of Use −0.15 0.6 −2.1 1.1 0.2 0.9 1.1 0.9 0.6 −0.2 1.1 0.3 0.2 0.2
Cleaning and
0.6 0.5 1.6 0.5 1.8 1.2 −1.6 1.3 0.7 0.2 1.3 0.2 0.3 0.3
Maintenance
Overall
Environmental 0.5 0.6 1.1 0.5 −0.9 0.4 −1.2 1.4 0.5 0.3 1.3 −1.2 −0.9 0.8
Quality
Overall 0.18 0.36 −0.39 0.35 −0.98 −0.55 −1.15 1.10 −0.36 −0.55 1.11 −1.14 −0.96 0.93
Figure 15. Results analysis of the occupancy satisfaction survey.
The most complaints from occupants were from the area near the subway. Both the space
entrance near the subway and the subway platform level received lower occupancy satisfaction scores.
The average thermal comfort of the tube entrance space near the subway was −1.85, the humidity
score was −1.48, and air velocity comfort score was −1.38. Compared with the physical environment
test results, the thermal environment of the tube entrance space near the subway had the greatest
influence on the physiology and psychology of the human body. It was also the weakest point of the
inner tubular space. In the physical environment test, the performance of the station hall was the worst,
which is reflected in its high temperature, humid environment, and poor air quality. The averages
of the three categories from the subjective survey data—thermal environment, humidity, and air
quality—were as follows. Tp6 was −2.2, Tp9was −2, and tp12 was −2.3, which indicates extreme
occupant dissatisfaction.
All of the questionnaire items can be sorted as follows: tp11 (middle tubular space) > tp8 (middle
tubular space) > tp14 (tube entrance type1) > tp2 (tube entrance type2) > tp4 (middle tubular space) >
tp1 (middle tubular space) > tp9 (station hall) > tp3 (tube entrance type1) > tp6 (station hall) = tp10
257
Appl. Sci. 2019, 9, 834
(tube entrance type2) > tp13 (tube entrance type2) > tp5 (tube entrance type2) > tp12 (station hall) >
tp7 (tube entrance type 1).
3.2.3. Satisfaction–Comfort Matrix Results and Analysis

Figure 16 shows the results of the Satisfaction–Comfort Matrix in the overall 14 target test points.
The horizontal axis of the matrix corresponds to the level of each comfort (thermal, lighting, ventilation,
and air quality) with the physical environment inside of the tubular space. The data was calculated
into a comfort percentage obtained from the test results, and divided into six classes from 0 to 100%.
The vertical axis of the matrix corresponds to the subjective analysis of occupancy space satisfaction,
following with the −3~3 score scale outlined in the above research method.
(a) (b)
(c) (d)
Figure 16. Test Point Satisfaction-Comfort Matrix results. (a) Middle tubular space; (b) Tube entrance
(connection to building); (c) Tube entrance (connection to subway); (d) In subway station.
The Satisfaction–Comfort Matrix is divided into four quadrants. The results located in the first
quadrant indicate that the animate space has a positive effect in terms of both satisfaction and comfort.
The results located in the second quadrant mean it has a positive effect in terms of satisfaction but a
negative effect in terms of comfort. The results located in the third quadrant imply neither satisfaction
nor comfort. The results located in the fourth quadrant mean it has a negative effect in terms of
258
Appl. Sci. 2019, 9, 834
satisfaction but a possible positive effect in terms of comfort [39]. Each quadrant’s evaluation is
divided into four grades (as shown in Figure 16).
3.3. Problem Analysis and Potential

According to the investigation of the objective physical environment and subjective comfort
feelings, it can be concluded that the potential for passive utilization of tubular space includes
two aspects: Improvement in comfort and utilization of climate.
3.3.1. Comfort Improvement from the Perspective of Passive Space Design

Tube entrance space plays a role in cohesion and intermediary conversion. However, it was
revealed that in the Beijing subway station building complex in the underground tube entrance
convergence space, the instantaneous wind speed was too high in winter and transition seasons; this
seriously affected user comfort and the long-term health of security personnel. Combining a design
for the complex building’s tube entrance transitional buffer space with setting a reasonable wind
barrier along the narrow tube entrance area, as tube as identifying transfer space in the underground
connected tubular space, would help to avoid the local wind outlet issue (as shown in Figure 17).
(a)
(b)
Figure 17. Passive strategy for relieving excessive wind velocity at the tube entrance. (a) Setting
conversion space in the underground interconnected tubular space; (b) Transition buffer space at the
tube entrance that combines with the complex building design.
The tubular space in the subway station area is always long and narrow. People in these locations
experience high traffic flow, and the space lacks connection with the outdoor natural environment;
this often results in poor air quality. Because it is a pass-through space, it is regularly neglected.
Considering the complex building shape, setting aside areas such as air shafts, light tubes, atriums, and
sunken courtyards would be an effective way of improving the air quality and indoor environment
through space-based strategies.
259
Appl. Sci. 2019, 9, 834
3.3.2. Climate Utilization from the Perspective of Passive Space Design

However, tubular space as a means of connecting the external and internal environments has
the advantage of transmitting natural resources. Effective natural resources can be transferred to the
ground and underground spaces in subway transportation areas to solve comfort-based shortcomings
in tubular spaces. These spaces can improve the quality of the area and reduce possible energy
consumption during operation. In subway station building complexes, using aerodynamic, piston,
and mixed ventilations, and exporting hot air to the outside through ventilation shafts or entrances can
reduce energy consumption from air conditioning equipment, as tube as the area’s levels of heat and
humidity. Research has explored new modes for composite space systems, such as tubes with wind
tunnels and solar chimneys [57], air supply tunnels that displace wind for better ventilation, active
and passive combinations of wind tunnels and ground source heat pumps, and lighting from hot-press
ventilation channels. These are all typical spatial patterns that can improve the utilization efficiency of
passive tubular space. In addition, the shape of the tubular space has the advantage of introducing
natural light. Natural light can improve the comfort level of the light environment and reduce lighting
energy consumption. According to the actual test data, in a summer outdoor temperature of 28 ◦ C,
the air temperature in a space covered by four meters of soil can be maintained at 10 ◦ C. When the
outdoor temperature drops to −5 ◦ C, the air temperature in that same underground area will remain
stable at 10 ◦ C [58]. In subway station building complexes, using geothermal energy with tubular
spaces will not only improve the thermal environment quality and passive space design, but also
transfer comfortable temperatures to other functional areas surrounding the tubular space, improving
comfort and reducing the cost of operation.
3.4. Design Target and Space Conception of Tubular Composite Spaces from the Perspective of
Sustainable Development
3.4.1. Passive Design for High Performance and Low Energy Consumption-Oriented Tubular Spaces
in Subway Station Building Complexes
It was determined that there is often low levels of comfort and user satisfaction problems with
tubular space. To design a new integrated subway station building complex both at home and abroad,
one must continuously attempt to optimize the indoor environment and further coordinate the design
strategy with the setting. Thus, this research studied tubular spaces in subway station building complex
integration designs, forming a passive design method for tubular spaces that would offer high levels
of performance and low amounts of energy consumption. This high level of performance is expressed
in greater comfort and improved human health related to the indoor thermal environment, light and
air quality, and user experience. Low energy consumption comes in the form of an energy-saving role
in the building, such as by providing passive cooling, a fresh air supply, natural light, and wind energy
utilization via the functional and shape advantages of tubular space (as shown in Figure 18).
Figure 18. Type analysis of the passive parameters of high performance and low energy
consumption-oriented tubular space in subway station building complexes.
260
Appl. Sci. 2019, 9, 834
3.4.2. Three Typical Design Concepts for Composite Tubular Space

An integrated design concept was presented in the study; it included two key aspects:
the integration of an independent original design division that considered the organization of a
single building, the underground station, and station space in the subway station building complex
design process; and the integration of the space system with the goals of high performance and low
energy consumption, to optimize the comprehensive performance of the composite space. Based on
these issues, three design concepts for composite spaces are described below. The verification of their
performance will be analyzed in detail in future research.
(a) Tubular complex space system for a wind tunnel and solar chimney (building atrium) for
air filtration
The subway station building complex is a relatively complicated building space. Because of
its large area, functional space coordination, and significant number of users, atriums are usually
employed to organize the functional relationship. Vertically high atrium spaces can serve as a draft for
solar chimneys. Combining them with a horizontal tunnel shaft design will help improve the quality
of the indoor space’s thermal environment and reduce the building’s amount of energy consumption.
In terms of the thermal environment, tunnel air temperature tends to be low, and combining the
low temperature of tunnel air with air conditioning will reduce energy consumption in summer.
Combining wind tunnels with the construction of solar chimneys and using the air dynamics principle
of cold air sinking and hot air rising will benefit passive air circulation in buildings, exhausting
high-temperature air and sucking low-temperature air from the tunnels. In addition, moderate thermal
pressure ventilation in buildings is beneficial to the circulation of air and can reduce surface body
temperature in summer, thus improving thermal comfort. However, due to the wind tunnels the air
quality will remain low. Therefore, the composite tubular space system should be combined with
an air filtration system to improve the air quality inside the tube and air circulation throughout the
subway station building complex (as shown in Figure 19).
Figure 19. Sketch map of a complex tubular space system for wind tunnel and solar chimney
air filtration.
(b) Complex tube path space system for lighting using hot air ventilation
The types of tubular space in a building can be divided into vertical, horizontal, and mixed.
The advantage of a vertical tubular space is that it penetrates space in a vertical direction, which is
beneficial for energy flowing from the top to the bottom, or vice versa. Natural lighting is a scarce
resource in underground space, and thus underground areas consume much more energy for lighting
261
Appl. Sci. 2019, 9, 834
than do general ground-level buildings. In most subway station building complex buildings, atrium
spaces or pass heights are the most common organizational structures. The integrated design idea
connects building complexes and subway station halls to provide natural lighting in tubular space,
forming light tubes with the atrium space. In addition, vertical tubes in atriums can offer useful hot air
ventilation. The low air temperature of tunnel wind can result in a substantial temperature difference
from the higher air temperatures at the tops of atriums, which can lead low-temperature air into the
building. This composite space system can also integrate the composite tubular space system of a
wind tunnel with a solar chimney (atrium) for air filtration, and integrate natural light with the air
filtration system, ventilation, and other functions resulting in a composite function space. This yields
further improvement in the passive space adjustment effect (as shown in Figure 20).
Figure 20. Sketch map of lighting through hot pressing the ventilation in a composite space system.
(c) Piston wind power generation for an auxiliary lighting system

According to past research, if a large number of fans are installed in subway tunnels, the blades
produce air resistance, resulting in resistance to the train operation. This can lead to more energy
consumption. However, if a few small micro-sized wind power generators are laid out in the subway
station, the machines will generate renewable energy for the platform, station hall layer, and artificial
lighting for the complex connection space. This is a useful way of using wind energy in subway
tunnels. According to the survey distributed for this research, the average wind speed at a height of
1.5 m above the platform is 1.5 m/s, and the average wind speed near the upper end of the metro
vehicle can reach 7.5 m/s. Previous research has shown that large, lightweight twist blades are the
best choice for subway tunnels. In micro-sized wind turbines, the starting wind speed for vertical
axis force generator equipment is 1 m/s, and the rated wind speed is 11 m/s. The safe velocity is
45~60 m/s, and the rated power is 200 W, 300 W, or 400 W. If the subway’s daily schedule runs from
6:00 a.m. to 10:00 p.m., the system will be in operation for 16 hours a day. If the 400 W fan is selected,
two units can be installed at the corner of each platform floor (for a total of eight units). This would
result in 51.2 KW·h of electricity sent per day for artificial lighting demands. According to the general
industrial and commercial electricity fee collection standards in Beijing, the annual payback period for
each of these typhoon machines would be two to three years, with an average value of 1 Rmb/degree
(as shown in Table 12).
262
Appl. Sci. 2019, 9, 834
Table 12. Economic benefit calculation for piston wind power generation for an auxiliary
lighting system.
Annual Daily Annual Payback

Rated Impeller Lump Sum Additional Service
Maintenance Power Power Period of
Power Diameter Investment Investment Life
Cost Output Output Investment
200 0.47 2000 1200 10 100 3.2 1100 2.99
300 0.66 2100 1500 10 150 4.8 1750 2.24
400 0.66 2400 2000 10 300 6.4 2300 2.16
1. A one-time investment would be the total cost of the system’s installation.

Note 2. Additional investments would be used to replace two battery charges for 10 years of life.
4. Conclusions
This study focused on the various tubular space forms in subway station building complexes; the
goal was to identify those that would have a moderating effect on passive potential space. The result
of this investigation, research, and analysis was an improvement of the indoor environment in terms
of comfort and energy consumption. This work followed four key pathways.
(1) It attempted to establish a scientific and logical method for verifying the value of tubular
space by establishing causal relationships among the parameterized building space information
factors, occupancy satisfaction elements, physical environment comfort aspects, and climate conditions.
This research adopted an analytical hierarchical process methodology for classifying each quantized
building information factor, and then compared this information to the actual fieldwork physical
environment test and occupant satisfaction vote data, in order to discern the key advantages and
weaknesses in tubular space design in subway station building complexes.
(2) Based on the actual field investigation, a database of physical environment performance data
and users’ subjective satisfaction information was established. The results showed that 59% and 57.4%
time in in middle tubular space was out of thermal and humidity comfort zone, and the uncomfortable
temperature and humidity at the tube entrance measuring point reached 100%. Air quality was the
worst in middle tube space, 65% of the time is unhealthy which exposing significant environmental
problems. By analyzing correlations therein, the comfort and health problems found in different
locations that were related to tubular spaces, as tube as the potential for passive utilization, were able
to be discussed.
(3) According to the investigation of the objective physical environment and subjective comfort
feelings, it can be concluded that the potential for passive utilization of tubular space includes two
aspects: improvement in comfort level itself and utilization of climate between natural or artificial.
Based on the measured data, an integrated design method for tube path spaces exhibiting high levels
of performance and low amounts of energy consumption in subway station building complexes was
put forward.
(4) The research described three typical composite tubular space designs, including wind
tunnels/solar chimneys (atriums), which are composite tubular spaces with air filtration systems;
composite space systems that provide lighting by hot-pressing the ventilation; and piston wind power
generation for use in auxiliary lighting systems. These provide methods and ideas for future research
and design.
Author Contributions: Conceptualization, J.L.; methodology, J.L.; software, S.L.; validation, Q.W.; formal analysis,
J.L.; investigation, J.L., S.T., Y.J.; data curation, J.L.; writing—original draft preparation, J.L.; writing—review and
editing, S.L.; visualization, Q.W.; project administration, J.L.; funding acquisition, J.L.
Funding: This work was supported by the Fundamental Funds for Beijing Natural Science Foundation of China
(Grant No. 8182043), and the National Natural Science Foundation of China (Grant No. 51708019).
Acknowledgments: Bejing Jiaotong University; Beijing MTR Construction Administration Corporation.
263
Appl. Sci. 2019, 9, 834
References
1. Wey, W.; Zhang, H.; Chang, Y. Alternative transit-oriented development evaluation in sustainable built
environment planning. Habitat Int. 2016, 55, 109–123. [CrossRef]
2. Yang, J.; Chen, J.; Le, X.; Zhang, Q. Density-oriented versus development-oriented transit investment:
Decoding metro station location selection in Shenzhen. Transp. Policy 2016, 51, 93–102. [CrossRef]
3. Papa, E.; Bertolini, L. Accessibility and Transit-Oriented Development in European metropolitan areas.
J. Transp. Geogr. 2015, 47, 70–83. [CrossRef]
4. Peng, Y.; Li, Z.; Choi, K. Transit-oriented development in an urban rail transportation corridor. Transp. Res.
Part B Methodol. 2017, 103, 269–290. [CrossRef]
5. Yang, Y.; Zhang, P.; Ni, S. Assessment of the Impacts of Urban Rail Transit on Metropolitan Regions Using
System Dynamics Model. Transp. Res. Procedia 2014, 4, 521–534. [CrossRef]
6. Liu, J. Rail Transit Complex Design Based on TOD; Beijing Jiaotong University: Beijing, China, 2013.
7. Hu, Y.; Zhao, C. Brief Discussion on Architectural Design of Green Subway Station. Build. Sci. 2014, 6,
132–138.
8. Song, Y.; Wang, J.; Zhu, N. Pondering over the Passive Design Strategy for Native Green Buildings of China.
Archit. J. 2013, 7, 94–99.
9. Wang, Z.; Chen, F.; Shi, Z. Prediction on Medium and long term energy consumption of urban rail transit
network in Beijing. China Railw. Sci. 2013, 34, 133–136.
10. Chen, J.; Gao, G.; Wang, X.; Wang, X. Calculation method of whole life-cycle energy consumption for urban
rail transit. J. Traffic Transp. Eng. 2014, 14, 89–97.
11. Gao, G.; Guan, W.; Li, J.; Dong, H.; Zou, X.; Chen, W. Experimental investigation of an active–passive
integration energy absorber for railway vehicles. Thin-Walled Struct. 2017, 117, 89–97. [CrossRef]
12. Marzouk, M.; Abdelaty, A. Monitoring thermal comfort in subways using building information modeling.
Energy Build. 2014, 84, 252–257. [CrossRef]
13. Li, X.; Wang, Z.; Ma, C.; Liu, L.; Liu, X. Energy Consumption Test and Analysis of Large Public Buildings
Based on Gray Box Model. Procedia Eng. 2016, 146–150. [CrossRef]
14. He, J.; Yan, Z.F.; Liu, H.J. On the Current Energy Consumption and Countermeasures of Large Public
Buildings in China. Adv. Mater. Res. 2012, 840–842. [CrossRef]
15. Lam, J.C.; Yang, L.; Liu, J. Development of passive design zones in China using bioclimatic approach. Energy
Convers Manag. 2006, 47, 746–762. [CrossRef]
16. Badescu, V.; Laaser, N.; Crutescu, R.; Crutescu, M.; Dobrovicescu, A.; Tsatsaronis, G. Modeling, validation
and time-dependent simulation of the first large passive building in Romania. Renew. Energy 2011, 36,
17. Sadineni, S.B.; Madala, S.; Boehm, R.F. Passive building energy savings: a review of building envelope
components. Renew. Sustain Energy Rev. 2011, 15, 3617–3631. [CrossRef]
18. Feist, W.; Schnieders, J.; Dorer, V.; Haas, A. Re-inventing air heating: convenient and comfortable within the
frame of the Passive House concept. Energy Build. 2005, 37, 1186–1203. [CrossRef]
19. Zhang, H.; Li, J.; Dong, L.; Chen, H. Integration of sustainability in Net-zero House: Experiences in Solar
Decathlon China. Energy Procedia 2014, 57, 1931–1940. [CrossRef]
20. Yiing, C.F.; Yaacob, N.M.; Hussein, H. Achieving Sustainable Development: Accessibility of Green Buildings
in Malaysia. Procedia Soc. Behav. Sci. 2013, 101, 120–129. [CrossRef]
21. Russell-Smith, S.V.; Lepech, M.D.; Fruchter, R.; Littman, A. Impact of progressive sustainable target value
assessment on building design decisions. Build. Environ. 2015, 85, 52–60. [CrossRef]
22. Olgyay, V. Design with Climate, New ed.; John Wiley & Sons Inc.: Hoboken, NJ, USA, 1992.
23. Li, J. Passive Adjustment Performance of Intermediary Space in Buildings; Tsinghua University: Beijing,
China, 2016.
24. Jiang, Y. The Study on Ventilation Design of Deep Plan Buildings with Vertical Sapce; Shenyang Architecture
University: Shenyang, China, 2011; Volume 12, p. 15.
25. Brown, G.Z.; Dekey, M. Sun, Wind & Light: Architecture Design Strategies; John Wiley & Sons Inc.: Hoboken,
NJ, USA, 2006.
26. Pieter, D.W.; Marinus, V.D.V. Providing computational support for the selection of energy saving building
components. Energy Build. 2004, 36, 749–758.
264
Appl. Sci. 2019, 9, 834
27. Kenji, M.; Keishi, D.; Li, M. Tokyo Tokyu Toyoko line Shibuya Station. Archit. J. 2009, 4, 40–45.
28. Katavoutas, G.; Assimakopoulos, M.N.; Asimakopoulos, D.N. On the determination of the thermal comfort
conditions of a metropolitan city underground railway. Sci. Total Environ. 2016, 566, 877–887. [CrossRef]
[PubMed]
29. Li, Y.; Geng, S.B.; Zhang, X.S.; Zhang, H. Study of thermal comfort in underground construction based on
field measurements and questionnaires in China. Build. Environ. 2017, 116, 45–54. [CrossRef]
30. Kim, M.J.; Richard, D.B.; Kim, J.T.; Yoo, C.K. Indoor air quality control for improving passenger health in
subway platforms using an outdoor air quality dependent ventilation system. Build. Environ. 2015, 92,
31. Chip Sullivan.Garden and Climate; China Building Industry Press: Beijing, China, 2005.
32. Wu, C.; Wang, Z.; Lu, Z. Strategy of urban development from the view of Shibuya Hikarie. Archit. Tech. 2015,
11, 40–47.
33. Zhang, F.; Zhou, X. From the utilization of underound space to the conformity of underground space and
the city: a case study of the two renovations of Les Hasses. Mod. Urban Res. 2014, 29, 29–38.
34. Li, J. Multi-criteria Approach to Impact Evaluation of Passive Adjustment Performance of Intermediary
Space in Buildings. Archit. J. 2016, 2, 50–55.
35. Song, Y.; Li, J.; Wang, J.; Hao, S.; Zhu, N.; Lin, Z. Multi-criteria approach to passive space design in buildings:
Impact of courtyard spaces on public buildings in cold climates. Build. Environ. 2015, 89, 295–307. [CrossRef]
36. Li, B. The Research on Climatic-Active Design Strategy of Building Skin in Hot-Summer and Cold-Winter Zone;
Tsinghua University: Beijing, China, 2004; pp. 127–281.
37. Qiu, J. Evaluation Science: Theory Method Practice; Science Press: Beijing, China, 2010. (In Chinese)
38. Haas, R.; Meixner, O. An Illustrated Guide to the Analytic Hierarchy Process. Available online: https:
//mi.boku.ac.at/ahp/ahptutorial.pdf (accessed on 22 February 2019).
39. Li, J.; Song, Y.; lv, S.; Wang, Q. Impact evaluation of indoor environmental performance of animate space in
buildings. Build. Environ. 2015, 94, 353–370. [CrossRef]
40. McMulla, R. Environmental Science in Building; Macmillan: Basingstoke, UK, 2007.
41. Fanger, P.O. Fundamentals of thermal comfort. Adv. Sol. Energy Technol. 1988, 4, 3056–3061.
42. Leadership in Energy and Environmental Design (LEED). The U.S. Green Building Council (USGBC):
Washington, DC, USA, 2001. Available online: http://www.usgbc.org/ (accessed on 22 February 2019).
43. MOHURD. Assessment Standard for Green Building of China; GB/T50378-2014; China Building Industry Press:
Beijing, China, 2014.
44. Kang, J.; Zhang, M. Semantic differential analysis of the soundscape in urban open public spaces.
45. Sun, Z.; Wei, D.; Xie, J. Research on subway tunnel wind power system. Renew. Energy Resour. 2016, 9,
1333–1341.
46. Hummelgaard, J.; Juhl, P.; Sabjornsson, K.O.; Clausen, G.; Toftum, J.; Langkilde, G. Indoor air quality
and occupant satisfaction in five mechanically and four naturally ventilated open-plan office buildings.
47. Assimakopoulos, M.N.; Katavoutas, G. Thermal comfort conditions at the platforms of the Athens Metro.
Procedia Eng. 2017, 180, 925–931. [CrossRef]
48. Moreno, T.; Reche, C.; Minguillón, M.C.; Capdevila, M.; de Miguel, E.; Querol, X. The effect of ventilation
protocols on airborne particulate matter in subway systems. Sci. Total Environ. 2017, 584, 1317–1323.
[CrossRef] [PubMed]
49. Burnett, J.; Pang, A.Y. Design and performance of pedestrian subway lighting systems. Tunn. Undergr.
Space Technol. 2004, 19, 619–628. [CrossRef]
50. Han, L.; Feng, L.; Yuan, Y. Numerical Analysis of Cooling Effect of Tunnel Ventilation System in Subway
Station. Refrig. Air Cond. 2016, 2, 1–4.
51. Yan, L. The Effect of Piston Wind on Subway Environment and Energy Saving Character; Beijing University of
Technology: Beijing, China, 2015.
52. Standard for Urban Rail Transit Lighting; GB/T16275-2008; Ministry of Housing and Urban-Rural Development
of the People’s Republic of China: Beijing, China, 2008.
53. Indoor Air Quality Standard GB/T18883-2002; Ministry of Housing and Urban-Rural Development of the
People’s Republic of China: Beijing, China, 2002.
265
Appl. Sci. 2019, 9, 834
54. Environemt Air Quality Standard GB3095-2012; Ministry of Housing and Urban-Rural Development of the
People’s Republic of China: Beijing, China, 2012.
55. ASHRAE (American Society of Heating, Refrigerating, and Air Conditioning Engineers). ASHRAE Standard
62.1—2013.Ventilation for Acceptable Indoor Air Quality; ASHRAE: Atlanta, GA, USA, 2013.
56. Ren, M.; Guo, C.; Guo, Q.; Yang, Y.X.; Kang, G.Q.; Luo, H.L. Numerical Analysis and Effectively Using of
Piston-effect in Subway. J. Shanhaijiaotong Univ. 2008, 8, 1376–1380, 1391.
57. Hu, Y. An Analysis of the Transitional Space of Traffic Hubs Case Studies of Traffic Hubs in Japan. Archit. J.
2014, 6, 109–113.
58. Wang, S.; Zhu, X.; Jiang, Y. Measurement and analysis of wall heat flow in thermal environment of Beijing
Subway. J. Undergr. Work. Tunn. 1997, 3, 32–37.
266
applied
sciences
Article
Application of Airborne Microorganism Indexes in
Offices, Gyms, and Libraries
Pietro Grisoli 1, *, Marco Albertoni 1 and Marinella Rodolfi 2
1 Department of Drug Sciences, Laboratory of Microbiology, University of Pavia, 27100 Pavia, Italy;
[email protected]
2 Department of Earth and Environmental Sciences, Mycology Section, University of Pavia, 27100 Pavia, Italy;
marinella.rodolfi@unipv.it
Received: 19 December 2018; Accepted: 9 March 2019; Published: 15 March 2019
Abstract: The determination of microbiological air quality in sporting and working environments
requires the quantification of airborne microbial contamination. The number and types of
microorganisms, detected in a specific site, offer a useful index for air quality valuation.
An assessment of contamination levels was carried out using three evaluation indices for
microbiological pollution: the global index of microbiological contamination per cubic meter
(GIMC/m3 ), the index of mesophilic bacterial contamination (IMC), and the amplification index
(AI). These indices have the advantage of considering several concomitant factors in the formation
of a microbial aerosol. They may also detect the malfunction of an air treatment system due to the
increase of microbes in aeraulic ducts, or inside a building compared to the outdoor environment.
In addition, they highlight the low efficiency of a ventilation system due to the excessive number
of people inside a building or to insufficient air renewal. This study quantified the levels of
microorganisms present in the air in different places such as offices, gyms, and libraries. The air
contamination was always higher in gyms that in the other places. All examined environments are in
Northern Italy.
Keywords: airborne microorganisms; bacteria; fungi; gyms; indoor air quality; libraries; offices
1. Introduction
The evaluation of microbiological air contamination is an important aspect of applied microbiology
and industrial hygiene [1–3]. In Italy, several legislative decrees, in particular the Legislative Decree
81/2008, impose that employers are responsible for the assessment of risks of biological origin arising
from the activities in the workplace; moreover, they must adopt suitable measures so that the air in the
workplace be healthy [4]. Therefore, the realization of measuring methods of airborne contamination
is a matter of prime importance in the most complex procedure of risk evaluation that must involve,
necessarily, different scientific skills [5–8]. Air contamination, both in confined environments and
outdoors, can be caused not only by chemical agents but also by biological sources such as pollen, dust,
mites, insects, pet allergens, bacteria, fungi, and viruses [9]. The monitoring of bioaerosols represents
an important tool both for assessing the risk in the working environment [10] and for the evaluation
of the environmental impact of certain activities that take place outdoors, such as biopurification or
waste disposal [11]. Several investigations have been directed to study the disease called Sick Building
Syndrome (SBS), which is characterized by varied and non-specific symptoms observed in workers
employed in confined environments [12–14]. The World Health Organization has defined SBS as
“an increase in frequency in the occupants of non-industrial buildings, of not specific acute symptoms
(irritation of eyes, nose, throat, headache, fatigue, nausea) that improve when building is left” [15].
The American Conference of Governmental Industrial Hygienists (ACGIH) does not suggest threshold

Appl. Sci. 2019, 9, 1101
limit values (TLV) for the environmental concentration of biological agents, because the existing
information does not allow to establish a scientifically acceptable dose–response relationship [16].
However, the assessment of culturable microbial agents is currently the easiest and most practical
measure to determine changes in the air quality of confined workplaces. The purpose of these
controls may be to set guide values, technologically attainable by adopting containment measures
of contamination established on the basis of repeated samplings in specific locations. With this aim,
this study examined structures destined for different types of work and recreational activity such as
offices, gyms, and libraries, measuring air-diffused contamination to identify and apply indices useful
for the microbiological classification of the air quality [17].
The indoor air quality was evaluated by applying the following indices: the global index
of microbiological contamination per cubic meter (GIMC/m3 ), the index of mesophilic bacterial
contamination (IMC), and the amplification index (AI). The GIMC considers different microbial types
and emphasizes their ability to grow in a wide temperature range. GIMC may be a simple method
to evaluate a potential biological risk in indoor and outdoor environments and to monitor sources
of microbiological contamination. The IMC reveals the presence of obligated mesophilic bacteria,
organisms of probable human origin in the indoor air. This index represents a practical instrument to
underline bacterial growth caused by hypoventilation and overcrowding. The AI is important as it
reveals microbial pollution in indoor ventilation systems. These indices showed their applicability in
other types of environments [18,19].
2.1. Sampling Method

Environmental monitoring was performed in Northern Italy in 10 diverse offices, gyms,
and libraries situated in buildings equipped with a ventilation system able to work in the following
modes: heating, air conditioning, and simple ventilation. The offices are placed in bank edifices, the
gyms in fitness centers, while the libraries belong to university buildings. The samplings were realized
every month for one year, inside the various environments, during normal people activity, and outside
the building. Quantitative data were determined in triplicate by means of an orthogonal impact
Microflow Air Sampler (AQUARIA, Lacchiarella, Italy), kept 1.5 m above ground level, in the center of
the room for offices, in the middle of the weight-training room for gyms, and in the reading rooms
for libraries. Air samplers worked at a fixed speed of 1.5 Ls−1 , collecting a volume of 200 L. For the
different types of environments, the number of persons (n. p.) and the air speed (m/s) were recorded,
so that in three sampling periods, the following average values were obtained: offices (n. p. 4.7; 0.03
m/s), gyms (n. p. 21.3; 0.060 m/s), and libraries (n. p. 16.2; 0.05 m/s).
2.2. Microorganisms Assessed

Bacteria were collected using Tryptone Soya Agar (TSA, Oxoid, Basingstoke, UK), and the cultures
were incubated at 37 ◦ C for 48 h for mesophilic bacteria and at 20 ◦ C for 6 days for psychrophilic
bacteria. Fungi were collected on Sabouraud Dextrose Agar (SAB Oxoid, Basingstoke, UK), and the
cultures were incubated at 20 ◦ C for 6 days. The outdoor air quality used as control was analyzed
following the same criteria. All total microbial counts are indicated as the number of colony-forming
units per cubic meter of air (CFU/m3 ), calculated as an average of three determinations from three
samples collected serially. The assessment of microbial contamination was effected by using bacterial
and fungal counts, and the following indexes were calculated: GIMC per cubic meter (GIMC/m3 ),
which is the sum of the values of the total microbial counts determined for mesophilic bacteria,
psychrophilic bacteria, and fungi in all sampled areas; IMC, derived from the ratio between the values
of CFU per cubic meter measured for mesophilic and psychrophilic bacteria at the same sampling
point; AI, resulting from the ratio between the GIMC/m3 values measured inside the building and
those measured outdoor.
268
Appl. Sci. 2019, 9, 1101
2.3. Data Analysis

The air microbial contamination values were expressed as colony-forming units (CFUs),
and the limit of quantification was 1 CFU/m3 . The number of CFUs for contact plate after
appropriate incubation was corrected using the positive-hole correction table provided by the supplier.
Statistical analysis of the data was performed by comparing the results obtained in the different
environments by means of analysis of variance one-way (post hoc test) values transformed into
logarithms (natural log). The analyses were conducted using Prism 3.0. The significance level
was p < 0.05.
3. Results
Table 1 shows the results of microbiological contamination of 10 different offices located in
buildings with centralized systems for ventilation. The mean values were higher for mesophilic bacteria
during the air conditioning phase compared to heating (p = 0.034) and during simple ventilation
compared to heating (p = 0.023). The contamination of psychrophilic bacteria was greater during
simple ventilation compared to air conditioning (p = 0.016) and to heating (p = 0.032). The mean
values of CFU/m3 for the fungi were not significantly different during the different periods of air
treatment. Although the average contamination was low, there were high maximum values of CFU/m3
for mesophilic bacteria in the periods of air conditioning and simple ventilation and for psychrophilic
bacteria during simple ventilation.
Table 1. Total microbial concentrations measured in the offices during heating, air conditioning, and
simple ventilation.
Mesophilic Bacteria Psychrophilic Bacteria Fungal Count
Functioning Modes N
CFU/m3 CFU/m3 CFU/m3
M ± SD Min–max M ± SD Min–max M ± SD Min–max
Heating 10 191.20 ± 144.53 8–450 176.70 ± 117.21 7–366 136.00 ± 32.80 22–654
Simple ventilation 10 667.60 ± 1602.60 6–5200 761.90 ± 1519.7 10–5000 131.50 ± 72.30 5–190
Air Conditioning 10 1089.70 ± 2392.20 19–7800 495.60 ± 307.20 76–860 119.60 ± 48.84 28–199
The number of samples (N), mean (M), standard deviation (SD), and range of values are indicated for each sampling
period. CFU: colony-forming units.
The calculation of GIMC/m3 confirmed the presence of higher values of contamination during
conditioning (p = 0.01) and simple ventilation (p = 0.01) compared to heating. The average value of
IMC during heating amounted to 7.30 and was higher than the values of the other modes of operation,
corresponding to 0.80 during the simple ventilation (p = 0.04) and 1.60 during air conditioning (Table 2).
Table 2. Global index of microbial contamination per cubic meter (GIMC/m3 ) and index of
mesophilic bacterial contamination (IMC) measured in offices during heating, air conditionining
and simple ventilation.
Functioning Modes N GIMC/m3 IMC

M ± SD Min–max M ± SD Min–max
Heating 10 503.90 ± 282.22 124–1111 7.30 ± 20.03 0.20–64.30
Simple ventilation 10 1561.00 ± 3154.80 21–10,450 0.80 ± 0.40 0.10–1.30
Air Conditioning 10 1704.90 ± 2537.57 229–8720 1.60 ± 2.67 0.02–9.10
The number of samples (N), mean (M), standard deviation (SD), the range of values are indicated for each
sampling period.
The data regarding the microbiological contamination found in 10 different gyms with centralized
ventilation systems are summarized in Table 3. The results show highest mean values for mycetic
contamination during the air conditioning phase compared to simple ventilation (p = 0.04) and heating
(p = 0.01). As regards bacteria, the average contamination values for psychrophilic bacteria were
superior to those for mesophilic bacteria in the three air treatment modalities and were significantly
different during conditioning (p = 0.02).
269
Appl. Sci. 2019, 9, 1101
Table 3. Total microbial concentrations measured in gyms during heating, air conditioning,

Functioning Modes N
M ± SD Min–max M ± SD. Min–max M ± SD Min–max
Heating 10 1096.60 ± 924.40 140–2850 1446.40 ± 1356.82 180–4650 187.50 ± 167.02 20–568
Air Conditioning 10 666.50 ± 477.76 200–1800 1509.00 ± 1354.02 440–4400 2430.90 ± 3678.31 8–10,848
The number of samples (N), mean (M), standard deviation (SD), and range of values are indicated for each
sampling period.
There was a notable increase in fungal count when the central heating was switched off. In fact, the
highest maximum values of CFU/m3 were found for fungi during simple ventilation and conditioning,
in contrast to what observed for bacteria (p = 0.016).
Table 4 shows the results of microbial contamination based on the GIMC/m3 and on the IMC
measured in the gyms. Because of the elevated fungal count observed in this period, the mean of
GIMC/m3 was higher during air conditioning (p = 0.019), with a maximum value of 15,248. The mean
of the IMC values was always <1, which implies that the counts of the psychrophilic bacteria were
almost always higher than the counts of the mesophilic bacteria; however, no significant differences in
the index were observed when central heating was on or off.
Table 4. GIMC/m3 and IMC measured in gyms during heating, air conditioning,

Heating 10 2703.50 ± 2144.84 1160–7090 0.90 ± 0.74 0.12–2.40
Simple ventilation 10 2420.90 ± 1645.70 910–6548 0.50 ± 0.30 0.17–1.20
Air Conditioning 10 4606.40 ± 4428.70 1480–15,248 0.60 ± 0.39 0.20–1.40
sampling period.
Similar sampling and monitoring methods were adopted for the evaluation of the air quality of
libraries. The average levels of contamination detected during air sampling showed small oscillations
with very restricted values of CFU/m3 both for bacterial loads (mesophilic and psychrophilic) and
for fungi. The results of airborne bacteria and fungi recovered from the libraries showed that the
CFU/m3 values were higher during heating than during simple ventilation or air conditioning.
However, no significant differences in the concentrations of all microbial counts were evidenced in the
three functioning modes of the ventilation system (Table 5).
Table 5. Total microbial concentrations in libraries during heating, air conditioning,


Functioning Modes N
Heating 10 171.20 ± 72.43 112–320 301.10 ± 237.26 125–725 123.90 ± 54.52 62–220
Air Conditioning 10 71.80 ± 30.50 30–120 200.60 ± 127.65 97–400 76.60 ± 36.55 35–120
sampling period.
In addition, the transformation of total microbial count values into the indices evidenced a general
reduction of the GIMC/m3 values and of the IMC values in air conditioning compared to heating,
but the statistical analysis showed no significant differences. The IMC index assumed values lower
than 1 and showed that air contamination was mainly due to environmental bacteria. All values were
limited to a small range in every period of the monitoring campaign (Table 6).
270
Appl. Sci. 2019, 9, 1101
Table 6. GIMC/m3 and IMC measured in libraries during heating, air conditioning,

Heating 10 596.20 ± 335.08 300–1250 0.70 ± 0.22 0.40–0.90
Simple ventilation 10 283.50 ± 131.70 153–578 0.50 ± 0.20 0.25–0.70
Air Conditioning 10 349.00 ± 176.43 182–652.50 0.40 ± 0.18 0.20–0.60
sampling period.
In indoor environments, microbiological air contamination can also be described by the

Amplification Index of microbial contamination (AI). To calculate this index, it was necessary to use the
GIMC values measured outside the buildings in the different sampling periods (Table 7). These values
were significantly different in the three sampling points; in particular, the widest variations were
detectable outside gyms in the air conditioning period (p < 0.0001).
Table 7. GIMC/m3 measured outside gyms, libraries, and offices during heating, air conditioning,
and simple ventilation periods.
Heating Simple Ventilation Air Conditioning

Sampling Points N
GIMC/m3 GIMC/m3 GIMC/m3
Gyms- outdoor 10 714.40 ± 294.00 328–1268 703.10 ± 226.40 484–1236 3586.10 ± 2326.42 1325–9240
Libraries outdoor 10 410.65 ± 187.81 266–880 596.55 ± 167.19 359–892 508.40 ± 141,01 364–830
Offices-outdoor 10 421.70 ± 179.62 248–890 619.90 ± 169.30 392–933 703.80 ± 226.20 485–1068
sampling period.
AI is determined by calculating the ratio between the GIMC/m3 values measured inside
a building and those measured outside. This index is greater than 1 when the microbial contamination
inside a building is higher than the outdoor contamination. In most cases, the values of microbial
contamination of indoor air are higher than the outdoor values. In the heating phase, AI showed
a worsening of air contamination especially in gyms compared the other environments considered
(3.7). On the contrary, during the conditioning period, the highest values were recorded in offices
(3.0) (Figure 1).
Figure 1. Amplification Index of microbial contamination (AI) measured in the three types of
environments during heating, air conditioning, and simple ventilation.
4. Discussion
It can be reasonably assumed that the microflora existing in recreational and working indoor
environments has a concentration lower than or equal to the external one detected in the same
271
Appl. Sci. 2019, 9, 1101
location and in the same climatic conditions [20,21]. In buildings equipped with centralized ventilation
systems there is a reduction of microbial contamination compared to the outside when air filtration
and treatment systems are maintained in appropriate conditions [22]. The air inside buildings may
be contaminated by the growth of microorganisms on floors, walls, aeraulic plants or because of
inadequate air changes [23]. The ventilation systems can have very different structural characteristics,
ranging from the simple aspiration of exhaust air to integrated heating, cooling, and humidification.
It is, therefore, clear that the air introduced from these systems can be considered as a matrix whose
microbiological quality can be evaluated and classified, as for any other product that may have effects
on human health. In addition, the activities carried out in a confined environment may be responsible
for the spread of microorganisms.
Dissenting opinions have been expressed on the possibility to monitor microbiological
environmental contamination in workplaces. Information on the dose–response relationship related
to the exposure to microorganisms is still not available today. The ACGIH does not indicate TLV for
biological agents. Furthermore, different analytical methods do not allow to recover and identify all
the microorganisms present in the air. In fact, the percentage of viable, culturable bacteria recovered
with normal sampling systems oscillates in a range of values that varies between 0.1% and 10% of
the total bacteria present in the air [24–26]. The presence of microorganisms should also be sought
in working environments because of their toxigenic potential and the possibility of spreading cell
fragments and volatile organic compounds into the environment. The lack of viable, culturable cells
does not necessarily indicate a healthy environment. Some studies, carried out in diversified working
environments, report environmental contamination values referable to the number of viable cells
belonging to a single class of microorganisms [27–29]. For example, a microbiological classification of
the air quality in non-industrial environments and homes considers the contamination from bacteria
that develop at 20–25 ◦ C. In this classification, for non-industrial environments, a level <100 CFU/m3
corresponds to the category of low contamination, while a value >2000 CFU/m3 corresponds to
the category of very high contamination [30]. It is clear that such assessment is incapable not only
to describe all the factors that determine the accumulation and spread of microorganisms in the
environment but also to identify possible risks for workers.
In this research, GIMC was used in order to include in the same data several categories of
microorganisms. This index considers microorganisms proliferating at different temperatures, such as
mesophilic and psychrophilic bacteria, and fungi capable of adapting to different types of environments.
These characteristics make this quantitative measure of microbiological contamination significant
and allow to evaluate the salubrity of a work environment; in fact, it considers microorganisms
that can develop in a wide range of temperatures, including ambient temperatures, typical of
saprophytic life, and 37 ◦ C, which is the temperature of development of pathogens. This index
is particularly useful because it is able to highlight even anomalous situations in indoor environmental
microbiological contamination: in fact, in the offices we analyzed, the mesophilic bacteria during
the conditioning phase were higher than the psychrophilic ones. This value represents an exception
because it is usually reasonable to assume that psychrophilic bacteria, which might grow well in
air conditioning systems, are more numerous than mesophilic bacteria during the cooling season,
as shown in the results for gyms and libraries buildings and in previous researches [22]. In particular,
GIMC attributes importance to bacteria that can proliferate in a wide range of temperatures. It is
true that the two incubation temperatures (20 ◦ C and 37 ◦ C) do not differentiate the two categories of
bacteria completely. Nevertheless, it is useful to determine the two total counts during environmental
monitoring because they have different significances and allow a more complete evaluation of airborne
bacterial contamination. In fact, the purpose of the index is to provide a measure of biological risk and
to aggregate several “environmental indicators”. IMC is mainly an index of anthropic contamination;
it highlights the share of mesophilic bacteria in the microbial population examined. This index derives
from the ratio between the value of CFU/m3 at 37 ◦ C and that at 20 ◦ C. This value, determined outdoor,
is always very close to 1, whereas in indoor environments, it may be higher, also depending on the
272
Appl. Sci. 2019, 9, 1101
number of people present. In fact, mesophilic bacteria derive from the normal bacterial flora of humans
and can therefore constitute the predominant population in confined environments, as already verified
in other works [9]. Finally, the amplification index is fundamental to detect the accumulation and
proliferation of microorganisms in ventilation systems or in buildings. AI describes global indoor
aerial modification. Generally, there are no relations between indoor and outdoor fungal and bacterial
counts. However, it is important to judge Indoor Air Quality (IAQ) not only by the measurement of
single parameters, but also using the global value of microbial contamination. High AI values may
only be indicative of IAQ deterioration when caused by high total fungal and bacterial counts [10,22].
Moreover, AI is essential to determine the environmental impact of outdoor work activities with
potential spread of pathogenic and non-pathogenic microorganisms; it must be calculated by referring
to the contamination detected in a control point that must provide the background value [18].
AI and the other two indices could provide criteria to evaluate indoor environmental quality more
uniformly [31]. In fact, these indices have different advantages: they consider several factors which
contribute to the development of different types of microorganisms (determination of GIMC), they may
reveal an inadequate functioning of the air conditioning system due to an excessive number of people
in a building or insufficient ventilation (IMC determination), and they may suggest a malfunction of
the air system due to the increase of microbial contamination in the ventilation ducts or in the building
compared to the outside environment (AI measures). The utilization of these indices can enable the
determination of threshold values as a function of the structures (buildings) analyzed. In previous
research in other environments, such as university classrooms, a reference value of GIMC/m3 = 1000
was connected to a correct maintenance of the aeraulic system [19]. In this work, the measured GIMC
values were higher than 1000, with the exception of libraries that presented very low contamination.
In particular, both offices and gyms had the highest values of pollution during the conditioning
phase, corresponding to GMIC/m3 of 1704.90 and 4606.40, respectively. GIMC exceeding 1000 does
not necessarily indicate a health risk; however, it is appropriate to analyze the contamination levels
through the calculation of IMC and AI. During air conditioning in the offices, IMC and AI showed
that mesophilic bacteria underwent a real amplification due to few air changes or the accumulation of
microorganisms in the air system. In fact, as reported by other authors, variations and fluctuations
in indoor humidity and temperature have significant effects on microbial diffusion and growth [32].
For gyms, on the contrary, the GIMC values were higher in summer because the proportion of
mesophilic bacteria decreased passing from the heating phase to the conditioning phase, but the
portion of psychrophilic bacteria and mycetes increased consistently, as evidenced by the IMC.
It is important to note that the rise of the microbial load inside the gyms coincided with an increase
of the microorganisms outdoors, as shown by the calculation of AI. Therefore, the presence of a greater
microbial contamination during summer could depend on the incorrect functioning or poor controls of
the ventilation systems, which resulted in the accumulation of microorganisms from the outside [33].
In conclusion, our research results show that the microbial contamination changes depending
on the environment analyzed and highlight the easy applicability of the proposed indices. GIMC,
IMC, and AI complete the information provided by the single classes of airborne microorganisms and,
together, they allow the analysis of indoor air quality. On the basis of sufficient data, these indices
classify environments in function of their microbiological contamination and identify guide values to be
adopted for routine monitoring and for the implementation of containment and remediation measures.
Author Contributions: Conceptualization, P.G.; methodology, P.G.; formal analysis, P.G., M.R.; investigation, P.G.,
M.A., M.R.; funding acquisition, P.G., M.A.; writing—original draft preparation, P.G. and M.R.
Funding: This work was financially supported by a research fund of the Laboratory of Microbiology, Department
of Drug Sciences, University of Pavia.
273
Appl. Sci. 2019, 9, 1101
References
1. Anderson, A.M.; Weiss, N.; Raine, F.; Salkinoja-Salonen, M.S. Dust-borne bacteria in animal sheds,
schools and children’s day care centers. J. Appl. Microbiol. 1999, 86, 622–634. [CrossRef]
2. Bernasconi, C.; Rodolfi, M.; Picco, A.; Grisoli, P.; Dacarro, C.; Rembges, D. Pyrogenic activity of air to
characterize bioaerosol exposure in public buildings: A pilot study. Lett. Appl. Microbiol. 2010, 50, 571–577.
[CrossRef] [PubMed]
3. Maroni, M.; Seifert, B.; Lindvall, T. Indoor Air Quality: A comprensive reference book. In Air Quality
Monographs; Elsevier: Amsterdam, The Netherlands, 1995; Volume 3, ISBN 9780444816429.
4. Testo unico in materia di tutela della salute e della sicurezza nei luoghi di lavoro. D.Lgs n. 81/2008.
In Gazzetta Ufficiale della Repubblica Italiana; n. 101 (30 Aprile 2008); Ministero della Giustizia,
Ufficio Pubblicazione Leggi e Decreti: Rome, Italy.
5. Brief, R.S.; Bernath, T. Indoor pollution: guidelines for prevention and control of microbiological respiratory
hazards associated with air conditioning and ventilation systems. Appl. Ind. Hyg. 1988, 3, 5–10. [CrossRef]
6. Jensen, P.A.; Schafer, M.P. Sampling and characterization of bioaerosols. In NIOSH Manual of Analytical
Methods Niosh/DPSE; NIOSH Publications: Washington, DC, USA, 1998; Volume 2, pp. 82–112.
7. Švajlenka, J.; Kozlovská, M.; Pošiváková, T. Assessment and biomonitoring indoor environment of buildings.
Int. J. Environ. Health Res. 2017, 27, 427–439. [CrossRef] [PubMed]
8. Burge, P.S. Sick building syndrome. Occup. Environ. Med. 2004, 61, 185–190. [CrossRef] [PubMed]
9. Cabo Verde, S.; Almeida, S.M.; Matos, J.; Guerreiro, D.; Meneses, M.; Faria, T.; Botelho, D.; Santos, M.; Viegas, C.
Microbiological assessment of indoor air quality at different hospital sites. Res. Microb. 2015, 166, 557–563. [CrossRef]
[PubMed]
10. Dacarro, C.; Grisoli, P.; Del Frate, G.; Villani, S.; Grignani, E.; Cottica, D. Micro-organisms and dust exposure
in an Italian grain mill. J. Appl. Microbiol. 2005, 98, 163–171. [CrossRef] [PubMed]
11. Prazmo, Z.; Krysinska-Traczyk, E.; Skorska, C.; Sitkowska, J.; Cholewa, G.; Dutkiewicz, J. Exposure to
bioaerosols in a municipal sewage treatment plant. Ann. Agric. Environ. Med. 2003, 10, 241–248. [PubMed]
12. Mendell, M.J.; Lei-Gomez, Q.; Mirer, A.G.; Seppänen, O.; Brunner, G. Risk factors in heating, ventilating,
and air-conditioning systems for occupant symptoms in US office buildings. Ind. Air 2008, 18, 301–316.
[CrossRef] [PubMed]
13. Qian, J.; Hospodsky, D.; Yamamoto, N.; Nazaroff, W.W.; Peccia, J. Size-resolved emission rates of airborne
bacteria and fungi in an occupied classroom. Ind Air. 2012, 22, 339–351. [CrossRef] [PubMed]
14. Wolff, C.H. Innate immunity and the pathogenicity of inhaled microbial particles. Int. J. Biol. Sci. 2011, 7, 261–268.
[CrossRef] [PubMed]
15. World Health Organization. Indoor Air Pollutants: Exposure and Health Effects; Report on a WHO Meeting,
EURO Reports and Studies no. 78; WHO Regional Office for Europe: Copenhagen, Denmark, 1983.
16. ACGIH. On-site investigation, pp. 1–8; Fungi, pp. 1–10; Bacteria, pp. 1–7. In Guidelines for the Assessment of
Bioaerosols in the Indoor Environment; Committee on Bioaerosols, Ed.; American Conference of Governmental
Industrial Hygienists: Cincinnati, OH, USA, 1989.
17. Traistaru, E.; Moldovan, R.C.; Menelaou, A.; Kakourou, P.; Georgescu, C. A comparative study on the quality
of air in offices and homes. J. Environ. Sci. Health A 2013, 48, 1806–1814. [CrossRef] [PubMed]
18. Grisoli, P.; Rodolfi, M.; Villani, S.; Grignani, E.; Cottica, D.; Berri, A.; Picco, A.; Dacarro, C. Assessment of
airborne microorganism contamination in an industrial area characterized by an open composting facility
and a wastewater treatment plant. Environ. Res. 2009, 109, 135–142. [CrossRef] [PubMed]
19. Grisoli, P.; Rodolfi, M.; Chiara, T.; Zonta, L.A.; Dacarro, C. Evaluation of microbiological air quality and of
microclimate in university classrooms. Environ. Monit. Assess. 2012, 184, 4171–4180.
20. Heildelberg, J.F.; Shahamat, M.; Levin, M.; Rahman, I.; Stelma, G.; Grim, C.; Colwell, R.R. Effect of aerosolization on
culturability and viability of Gram-Negative bacteria. Appl. Environ. Microbiol. 1997, 63, 3585–3588.
21. Macher, J.M.; Chatigny, M.A.; Burge, H.A. Sampling airborne microorganisms and aeroallergens. In Air
Sampling Instruments for Evaluation of Atmospheric Contaminants, 8th ed.; Cohen, B.S., Hering, S.V., Eds.;
American Conference of Governmental Industrial Hygienists Inc.: Cincinnati, OH, USA, 1995; pp. 589–617.
22. Dacarro, C.; Grignani, E.; Lodola, L.; Grisoli, P.; Cottica, D. Proposta di indici microbiologici per la valutazione
della qualità dell’aria degli edifici. G. Ital. Med. Lav. Ergon. 2000, XXII, 229–235.
274
Appl. Sci. 2019, 9, 1101
23. Gołofit-Szymczak, M.; Górny, R.L. Microbiological Air Quality in Office Buildings Equipped with Different
Ventilation Systems. Ind. Air 2018, 28, 792–805. [CrossRef] [PubMed]
24. Jouzaitis, A.; Willeke, K.; Grinshpun, S.A.; Donnelly, J. Impaction onto a Glass Slide or Agar
versus Impingement into a Liquid for the Collection and Recovery of Airborne Microorganisms.
Appl. Environ. Microbiol. 1994, 60, 861–870.
25. Stewart, S.L.; Grinshpun, S.A.; Willeke, K.; Terzieva, S.; Ulevicius, V.; Donnelly, J. Effect of Impact Stress on
Microbial Recovery on an Agar Surface. Appl. Environ. Microb. 1995, 61, 1232–1239.
26. Walter, M.V.; Marthi, B.; Fieland, V.P.; Ganio, L.M. Effect of Aerosolization on Subsequent Bacterial Survival.
Appl. Environ. Microb. 1990, 56, 3468–3472.
27. Flannigan, B. Indoor microbiological pollutants—Sources, species, characterization and evaluation.
In Chemical, Microbiological, Health and Comfort Aspects of Indoor Air Quality—State of the Art in SBS;
Knöppel, H., Wolkoff, P., Eds.; EDSC, EEC, EAEC: The Netherlands, 1992; pp. 73–98.
28. Miller, J.D. Fungi as contaminants in indoor air. Atmos. Environ. 1992, 6A, 2163–2172. [CrossRef]
29. Rao, C.Y.; Burge, H.A. Review of quantitative standards and guidelines for fungi in indoor air. J. Air Waste
Manag. Assoc. 1996, 46, 899–908. [CrossRef] [PubMed]
30. Commission of European Communities. Indoor Air Quality & Its Impact on Man; Report n.12;
Commission of European Communities: Luxembourg, 1993.
31. Sofuoglu, S.C.; Moschandreas, D.J. The link between symptoms of office building occupants and in-office air
pollution: The Indoor Air Pollution Index. Ind. Air 2003, 13, 332–343. [CrossRef]
32. Flannigan, B.; Miller, J.D. Microbial growth in indoor environments. In Microorganisms in Home and Indoor
Work Environments: Diversity, Health Impacts, Investigation and Control; Flannigan, B., Samson, R.A., Miller, J.D.,
Eds.; Taylor & Francis: Abingdon, UK, 2001; pp. 35–67.
33. Dacarro, C.; Picco, A.M.; Grisoli, P.; Rodolfi, M. Determination of aerial microbiological contamination in
scholastic sports environments. J. Appl. Microbiol. 2003, 95, 904–912. [CrossRef] [PubMed]
275
applied
sciences
Article
How Working Tasks Influence Biocontamination in
an Animal Facility
Anna M. Marcelloni 1 , Alessandra Chiominto 1 , Simona Di Renzi 1 , Paola Melis 1 , Annarita Wirz 2 ,
Maria C. Riviello 2,3 , Stefania Massari 4 , Renata Sisto 1 , Maria C. D’Ovidio 1 and Emilia Paba 1, *
1 Department of Occupational and Environmental Medicine, Epidemiology and Hygiene,
Italian Workers’ Compensation Authority (INAIL), Via Fontana Candida 1, Monte Porzio Catone,
00078 Rome, Italy; [email protected] (A.M.M.); [email protected] (A.C.); [email protected] (S.D.R.);
[email protected] (P.M.); [email protected] (R.S.); [email protected] (M.C.D.O.)
2 Santa Lucia Foundation IRCCS, Via Ardeatina 306, 00142 Rome, Italy; [email protected]
3 Institute of Cell Biology and Neurobiology, National Research Council (CNR), Via E. Ramarini 32,
Monterotondo, 00015 Rome, Italy; [email protected]
4 Department of Occupational and Environmental Medicine, Epidemiology and Hygiene,
Italian Workers’ Compensation Authority (INAIL), Via Stefano Gradi 55, 00143 Rome, Italy; [email protected]
Received: 30 April 2019; Accepted: 27 May 2019; Published: 29 May 2019
Abstract: The exposure to biocontaminants in animal facilities represents a risk for developing
infectious, allergic and toxic diseases. The aim of this study was to determine what factors
could be associated with a high level of exposure to biological agents through the measure and
characterization of airborne fungi, bacteria, endotoxin, (1,3)-β-d-glucan and animal allergens. Airborne
microorganisms were collected with an air sampler and identified by microscopic and biochemical
methods. Endotoxin, (1,3)-β-d-glucan, Mus m 1, Rat n 1, Can f 1, Fel d 1, Equ c 4 allergens were
detected on inhalable dust samples by Kinetic LAL, Glucatell, and ELISA assays, respectively.
Our data evidenced that changing cages is a determinant factor in increasing the concentration of
the airborne biocontaminants; the preparation of bedding and distribution of feed, performed in
the storage area, is another critical working task in terms of exposure to endotoxins (210.7 EU/m3 )
and (1,3)-β-d-glucans (4.3 ng/m3 ). The highest concentration of Mus m 1 allergen (61.5 ng/m3 ) was
observed in the dirty washing area. The detection of expositive peaks at risk of sensitization (>2 μg/g)
by Fel d 1 in animal rooms shows passive transport by operators themselves, highlighting their role
as vehicle between occupational and living environments.
Keywords: allergens; endotoxin; biological agents; laboratory animal allergy; environmental monitoring;
occupational exposure
1. Introduction
The exposure to biocontaminants is well documented and studied in several occupational settings,
and few data are available in animal facilities, although more attention has been addressed to animal
allergens. These biological agents are aeroallergens, mainly lipocalins, derived by different biological
fluids and/or tissue (saliva, serum, urine, dander, hair, fur) that remain in suspension for different times
in relation to meteorological conditions and factors influencing their dispersion [1,2]. In recent years,
the topic regarding animal allergen exposure has been treated both in pets [3–5] and experimental
animals [6–9], and attention has been also addressed to co-exposure of allergens and endotoxins in
these workplaces, although in a more restricted way [10–15]. Exposure to biocontaminants in animal
facilities in USA and in the United Kingdom has received great attention [16–19] as demonstrated
by the NIOSH publication in 1998 [20], where the so-called LAA (Laboratory Animal Allergy) was

Appl. Sci. 2019, 9, 2216
considered an occupational risk, and by Gordon in 2001 [21]. Various papers have been published in
scientific literature both as a review [22–24] and experimental studies [25–27] while in Italy the topic of
LAA is not studied carefully yet.
Health effects reported on laboratory animal workers (LAWs) are mainly represented by
asthma, rhinitis, conjunctivitis, dermatitis and anaphylaxis [8,19,28]. Discordant percentages of
LAA have been reported in the LAWs and these differences could be attributed to different evaluation
methods (questionnaires, analytical test, etc.). In this context a relevant aspect is represented by
environmental exposure to animal allergens, and although a lot of papers report the monitoring of
these biocontaminants, there is some difficulty in comparing results because of the differences in
sampling and analytical methods [29–32]. In addition to allergens, there are several other biological
agents in animal facilities with possible health consequences such as airborne bacteria, fungi and
microbial indicators (endotoxin and (1,3)-β-d-glucan) which may originate from animals (fur, epidermal
material), their waste (feces and urine) and materials used in their maintenance (food, bedding) [33].
Airborne endotoxin, an outer membrane component of Gram-negative bacteria, has been identified as
the major risk factor for nasal, chest and skin symptoms also contributing to allergy diseases [11,34].
Some authors report that fecal bacteria in soiled bedding may produce increased concentrations of
endotoxin compared to other settings and this may be an occupational health concern. The authors
concluded that the percentage of dust and endotoxin in different types of rodent bedding could be
an important factor affecting the occupational exposure of personnel working with laboratory animals
causing airway inflammation, hypersensitivity pneumonitis, organic dust toxic syndrome (ODTS),
chronic obstructive pulmonary disease (COPD) and asthma-like syndrome, as well as having harmful
effects on the animals themselves [33].
Pathogenic moulds, yeast and their products may also exist in the air of animal units, causing
further risk for personnel working there. In particular, (1,3)-β-d-glucans are non-allergenic and hardly
soluble in water glucose polymers, which consist of a part of the cellular wall of most fungi, but also
many plants and some bacteria. Due to their presence in both viable and dead cells, they may be
considered a good indicator of exposure to fungi [35]. Although little information is available on
their health effects in animal facilities, however, their association with dry cough, cough associated
with phlegm, hoarseness, and atopy, has been reported in different studies [36–38]. Very few and not
recent investigations dealing with airborne biological agents in laboratory animal settings, other than
allergens, can be found [12–15,39] to date. These issues have received little attention in our country.
The objective of this study was to determine factors (working tasks, cage changing frequency and
animal strains) were associated with the greatest level of exposure to biological agents of the personnel
working with rodents through the measure and characterization of airborne viable fungi, bacteria,
endotoxin, (1,3)-β-d-glucan and animal allergens.
2.1. Animal Facility

The study was conducted in a Biomedical Research Institute where animals are maintained in
agreement with Italian Legislative decree 26/2014 [40]. The facility includes 14 conventional animal
rooms (3 housing rats and 11 housing mice) where animals are housed in stainless steel wire cages
(on average 2–3 animals/cage) with perforated floor and trays underneath; filters on the top prevent
the distribution of airborne particles. Each cage is stacked in a rack that can hold up to 30 cages for
mice and 24 for rats. The facility houses about 7000 rodents.
The racks are washed every 4 months. The cages and their equipment (grids containing bottles
of water and feed, the bottles and soiled bedding) are changed once or twice per week depending
on experimental protocols. When cages are changed, a clean rack is brought into the room, animals
and accessories are moved to new cages and dirty cages are moved to the cage-washing area where
they are emptied and washed (dirty washing area). Subsequently, the cages are allowed to dry and
278
Appl. Sci. 2019, 9, 2216
restacked for future use (clean washing area). There is also a storage area where preparation of the
bedding and distribution of feed are carried out; these activities are performed only in the afternoon.
The staff is divided into laboratory technicians and researchers; laboratory technicians perform shifts
of 8 h and 30 min, and are engaged in changing cages, bedding and feeding animals, cleaning rooms,
washing cages; researchers are engaged in several experimental procedures such as collecting blood
and urine specimens, surgery and sacrifice. Personal protective equipment is required for all personnel
working in the facility and includes surgical cap, mask, gloves, shoe cover, and disposable coat.
Animal rooms and the storage area have a ventilation system separate from the rest of the facility.
The outside air is collected and treated by an Air Handling Unit (AHU) located on the coverage plan.
Another AHU treats the air entering the two washing areas. In each work environment the air enters
through vents located on the ceiling (two in each room), while the exhaust air is removed through
grilles located on the walls near the floor (four in each rooms) or on the ceiling in the dirty and clean
washing areas. In animal rooms 15–20 air changes per hour are guaranteed; the values of temperature
and relative humidity are maintained in a controlled range (T: 20 ◦ C ± 1 ◦ C; RH: 45–55%).
2.2. Sampling and Analysis of Airborne Bacteria and Fungi

Triplicate air samples were collected using a portable microbiological sampler (SAS Super ISO,
PBI International, Milan, Italy) at a flow rate of 100 L min−1 , from 10 workplaces: 7 animal rooms
randomly selected (3 housing rats and 4 housing mice) (see Table 1), 2 washing areas (dirty and
clean), and the storage area. Two offices were investigated: the technician’s and manager’s offices.
All measurements were taken in the middle. In order to evaluate the influence of changing cages on
environmental contamination, in animal rooms air samples were taken over three consecutive days:
the day of changing cages, the one before, and the one after. Regarding the storage area, air sampling
was carried out in the morning (without activity) and in the afternoon.
Table 1. Animals housed in each investigated room during the study and frequency of changing cages.
Rooms Strains Total Number Male Female Pups Changing Cages

A Rats 149 63 86 0 Bi-weekly
B Mice 494 328 166 0 Weekly
C Mice 682 311 267 104 Weekly
D Mice 613 225 283 105 Bi-weekly
E Rats 155 85 60 10 Weekly
F Mice 723 316 380 27 Bi-weekly
G Rats 377 37 85 255 Bi-weekly
Air volumes of 300 L were sampled. Total cultivable and Gram-negative bacteria were impacted
on Tryptone Soy Agar (TSA) and Mac Conkey 3 (McC3 ) plates, respectively, fungi on Malt Extract Agar
(MEA) and Dichloran Glycerol Agar (DG18) (all from Oxoid S.p.A, Milan, Italy). TSA and McC3 plates
were subsequently incubated for 1–2 days at 37 ◦ C (mesophilic bacteria), DG18 and MEA at 25 ◦ C for
7–10 days.
The bacterial isolates were identified with microscopic and biochemical methods by API 20E and
NE tests (bioMerieux, Marcy l’Etoile, France) and Microstation ID instrument (Biolog Inc., Hayward,
CA, USA). Identification of moulds and yeasts were accomplished via macroscopic and microscopic
examination, referring to the manual Medically Important Fungi: A guide to Identification [41],
while biochemical identifications were performed using the Microstation ID instrument. Data are
expressed as CFU/m3 . The limit of detection (LOD) was 2 CFU/m3 . The uncertainty assessment results
are presented in Table 2.
279
Appl. Sci. 2019, 9, 2216
2.3. Inhalable Dust Sampling and Analysis of Endotoxin, (1,3)-β-d-glucan and Allergens
Stationary inhalable dust samples were collected using airChek2000 pumps (SKC Inc., Eighty
Four, PA, USA), at a sampling flow rate of 2 L·min−1 , equipped with stainless IOM sampler and
glass (GF, pore size 1.6 μm) and polycarbonate (PC, pore size 0.8 μm) for respectively endotoxin and
(1,3)-β-d-glucan analysis.
Closed-face cassette with support pad and mixed cellulose ester (MCE, diameter 33 mm, 0.8 μm
porosity), at a flow rate of 2.5 L·min−1 , were used for allergen sampling. Samplers were placed in
the middle of the room at 1.5 m above the floor to simulate human breathing zone. A field blank
was deployed on each sampling. The sampling strategy was the same as environmental monitoring
(three consecutive days in animal rooms, in the morning and in the afternoon inside the storage area).
The sampling time was 3 h.
GF extracts (5 mL of 0,05% Tween 20 in Pyrogen Free Water) were centrifuged at 1000× g rpm for
20 min. and analysed in duplicate with Kinetic LAL method (QCL-LAL assay Lonza Walkersville,
MD, USA). Concentrations are reported in EU/m3 . The LOD was 0.005 EU/mL.
For (1,3)-β-d-glucan, PC extracts (5 mL of 0.85% NaCl and 0.05% tween 80) were vigorously
shaken for 30 at room temperature (250 rpm) and analyzed with the (1,3)-β-d-glucan specific
Kinetic Chromogenic LAL assay (Glucatell, Associates of Cape Cod), as previously described [42,43].
(1,3)-β-d-glucan concentrations are presented in ng/m3 . The LOD in suspension was 2.53 pg/mL.
MCE filters were extracted by submerging in 1 mL phosphate saline buffer (PBS plus 0.01% Tween
20) in a 15 mL centrifuge tube, vortexed at room temperature and shaken for 2 h. The eluate was
collected by compressing the filter in a plastic syringe, 1% of human serum albumin was added for
protein stabilization and stored at −20 ◦ C until assay. ELISA kits for Mus m 1, Rat n 1, Can f 1, Fel d 1,
Equ c 4 were performed following the protocols of analysis (Indoor Biotechnologies Ltd., Manchester,
UK.) with read of the plates at 405 nm optical density. Results are expressed as ng/m3 . The uncertainty
assessment results are reported in Table 2.
(*) Type A standard uncertainty with a confidence level 1-α > 0.95 using the Student’s T to find the
appropriate coverage factor based on degrees of freedom. (**) Measures are detected in the morning
(without activity) and in the afternoon (during activity).
Table 2. Uncertainty measurement (*) for sampled biological agents by working areas.
Mesophilic Bacteria Fungi Endotoxin (1,3)-β-d-glucan Rat n 1 Mus m 1

Workplace
(UFC/m3 ) (UFC/m3 ) (EU/m3 ) (ng/m3 ) (ng/m3 ) (ng/m3 )
Animal rooms ±52.26 ±0.58 ±1.35 ±0.59 ±0.16 ±4.56
Washing area clean ±21.18 ±16.91 ±9.25
Washing area dirty ±434.69 ±66.18 ±6.54
Storage area (**) ±44.77 ±94.89 ±163.54 ±4.57 ±0.00 ±8.32
(*) Type A standard uncertainty with a confidence level 1-α > 0.95 using the Student’s T to find the appropriate
coverage factor based on degrees of freedom. (**) Measures are detected in the morning (without activity) and in the
afternoon (during activity).
2.4. Statistical Analysis

One way ANOVA and Students’t-test were considered to evaluate significant differences in
environmental biocontaminant concentrations in various work tasks and areas. Univariate and
multivariate, both fixed and mixed effect linear regression models were used, in which the concentration
of the different biological agents acts as outcome variable.
Data were analyzed using the statistical software R (R Foundation for Statistical Computing,
Vienna, Austria—ISBN 3-900051-07-0, URL http://www.R-project.org).
3. Results
Table 1 shows strains and number of animals (divided into males, females and pups) housed
in investigated rooms (A–G) and the frequency of changing cages. Concentrations (mean value and
280
Appl. Sci. 2019, 9, 2216
standard deviation) of mesophilic bacteria, fungi, endotoxin, (1,3)-β-d-glucan, Rat n 1 and Mus m
1 detected in each working areas are reported in Table 3. Regarding animal rooms, data are also
presented by sampling room and the number of rodents is indicated.
3.1. Airborne Bacteria and Fungi

A total of 306 air samples were analysed. Mean mesophilic bacteria levels ranged from 33.8 to
480 CFU/m3 with a peak value of 885 CFU/m3 found in an animal room during the changing cages.
Gram-negative concentrations were below the LOD (2 CFU/m3 ) except for animal rooms, during the
changing cages, and for the storage area where a mean value of 3 CFU/m3 was recorded. Among
a total of 18 genera and species, the most frequently isolated were Staphylococcus warneri (23.9%),
Staphylococcus xylosus (21.4%) Micrococcus spp. (14.5%), Acinetobacter schindleri (10.4%).
Colonies of Escherichia coli, Enterobacter cloacae and Enterococcus faecalis (risk group 2, according to
Annex XLVI of the Italian Legislative Decrees [44,45], were isolated in four animal rooms during the
changing cages.
Mean concentrations of moulds and yeasts were low (flora was dominated by moulds such
as Alternaria, Ulocladium, Cladosporium, Penicillium, Eurotium, Aspergillus, Scopulariopsis, Phoma,
with occasional isolates of yeasts. No pathogen was identified.
3.2. Endotoxin, (1,3)-β-d-glucan and Allergens

The mean concentrations of endotoxin, (1,3)-β-d-glucan and allergens (See Table 3) were obtained on
89 samples taken in different occupational environments. Statistically significant differences were found
between offices and animal rooms, these last having a higher concentration of endotoxin (p = 0.00616) and
between the animal rooms and the storage area in the concentration of (1,3)-β-d-glucan (p = 0.038).
The evaluation of allergens was performed on 181 air samples for ELISA tests. Mus m 1 allergens
were detected in all working areas except for offices, with the highest value (61.5 ng/m3 ) in the dirty
washing area while Rat n 1 allergens were absent except for in animal rooms. Can f 1 and Equ c 4
allergens were absent while Fel d 1 allergen was detected in rat rooms with the highest levels during
the changing cages activity (range 0.4–3 ng/m3 ).
Table 3. Concentrations (mean value and standard deviation) of mesophilic bacteria, fungi, endotoxin,
(1,3)-β-d-glucan, Rat n 1 and Mus m 1 by working areas (number of animals is indicated for each room).
Mesophilic Bacteria Fungi Endotoxin (1,3)-β-d-glucan Rat n 1 Mus m 1

Workplace
(UFC/m3 ) (UFC/m3 ) (EU/m3 ) (ng/m3 ) (ng/m3 ) (ng/m3 )
Animal rooms 130.7 ± 207.5 1.6 ± 3.04 5.9 ± 3.6 1.5 ± 1.3 0.1 ± 0.4 4.3 ± 11.0
Room A (149) 63.1 ± 92.7 2.1 ± 4.4 7.8 ± 0.7 1.8 ± 1.2 0.0 ± 0.0 0.0 ± 0.1
Room B (494) 122.9 ± 110.2 2.0 ± 1.7 7.2 ± 4.2 3.6 ± 2.3 0.0 ± 0.0 6.1 ± 9.9
Room C (682) 112.4 ± 115.7 1.1 ± 2.4 3.6 ± 3.5 0.9 ± 0.6 0.1 ± 0.3 2.7 ± 4.2
Room D (613) 64.8 ± 83.5 1.3 ± 2.7 0.1 ± 0.0 1.9 ± 0.8 0.5 ± 0.9 0.3 ± 0.2
Room E (155) 317.0 ± 427.0 3.3 ± 5.3 7.3 ± 3.2 0.1 ± 0.0 0.3 ± 0.5 0.1 ± 0.2
Room F (723) 66.7 ± 75.8 0.4 ± 1.1 11.1 ± 4.8 1.3 ± 0.8 0.0 ± 0.0 20.6 ± 24.1
Room G (377) 220.5 ± 232.5 1.7 ± 1.7 5.2 ± 2.6 0.9 ± 0.3 0.0 ± 0.0 0.1 ± 0.2
Washing area clean 78.3 ± 2.4 19.3 ± 13.6 7.4 ± 1.8 1.3 ± 0.7 0.4 ± 0.0
Washing area dirty 211.1 ± 175.0 107.8 ± 63.1 16.7 ± 4.7 1.6 ± 0.9 61.5 ± 0.0
Storage area * 266.7 ±36.1 112.5 ± 76.4 124.2 ± 102.8 3.9 ± 1.7 1.5 ± 0.8
Offices 308.0 ± 91.2 60.0 ± 25.9 2.4 ± 0.8 1.4 ± 1.1
* Mean concentrations measured in the storage area refer to samplings performed the morning and in the afternoon.
3.3. Effect of the Working Activities

One of our main questions was whether or not specific working tasks may cause an increased risk
of biological exposure for personnel working in animal facilities. The hypothesis that the airborne
concentration of biocontaminants could increase during the changing cages was tested. To point out
the increase in level of (1,3)-β-d-glucan the time was considered a two-level factor: sampling during
281
Appl. Sci. 2019, 9, 2216
the changing cages was compared to the sampling before and after. The β-glucans concentration
“during” increases on average by a factor of 1.3 with respect to the concentration “before” and “after”
and the difference is statistically significant (p = 0.04). For endotoxins, a mixed effect linear regression
model was used in which animal room was introduced as a random effect variable. The concentrations
increased on average by a factor of 1.8 with respect to the levels before or after (p = 0.04). The highest
mean concentrations of endotoxin (210.7 EU/m3 ) and (1,3)-β-d-glucan (4.3 ng/m3 ) were measured in
the afternoon during the preparation of bedding and distribution of feed (storage area) (See Table 4).
The allergenic concentration was treated as a continuous variable. As in the case of endotoxin,
to increase the statistics, the time was considered a two-level factor and a mixed effect linear regression
model was used with the animal room treated as a random effect variable. The concentration level
of the mouse allergenic contaminant increases on average by a factor of 9.3 with respect to the levels
before or after the changing cages (p = 0.037).
Table 4. Concentrations (mean value and standard deviation) of inhalable endotoxin, (1,3)-β-d-glucan,
Rat n 1 and Mus m 1 by working activities.
Endotoxin (1,3)-β-d-glucan Rat n 1 Mus m 1

Workplace
(EU/m3 ) (ng/m3 ) (ng/m3 ) (ng/m3 )
Animal rooms
Before 5.2 ± 3.7 0.8 ± 0.5 0.1 ± 0.1
During 7.1 ± 3.8 2.4 ± 1.9 0.4 ± 0.6 10.5 ± 17.4
After 5.4 ± 3.5 1.4 ± 0.6 2.2 ± 5.3
Storage area
Morning 37.8 ± 37.3 3.5 ± 2.4 0.9 ± 0.0
Afternoon 210.7 ± 20.0 4.3 ± 1.6 2.2 ± 0.0
3.4. Effect of Animal Strains and Changing Cages Frequency

In order to determine the effect of the animal strains on environmental biocontamination,
the concentration measured in animal rooms was normalized according to the number of animals.
This normalized quantity was related to the strain treated as a two level factor. In the case of the
mesophilic bacteria and endotoxin, the average level was lowered in the case of mice with respect to
the case of rats by a factor of −0.5070 (p = 0.020) and by a factor of −0.032 (p = 1.5·10-8) respectively.
The result was not statistically significant for Gram-negative bacteria, fungi and (1,3)-β-d-glucan.
To test if animal species have an effect on the allergenic pollutants concentration, a Pearson’s
Chi-squared with Yates’ continuity correction test was applied. The test was statistically significant in
the case of the Mus m 1 and Fel d 1 (p = 0.004 and p = 0.045 respectively). The positive result of the
Mus m 1 is statistically associated with the mouse species whilst the Fel d 1 is statistically associated
with the rat.
In regards to the cage changing frequency, the concentrations normalized to the number of animals
have been compared with once per week and twice per week cage changing frequency, treated as
a two-level factor. A statistically significant difference has been found only in the case of mesophilic
bacteria concentration. The twice per week frequency has on average a bacteria concentration reduced
by a factor of 0.13 with respect to the level of rooms with a once per week frequency (p = 0.019).
4. Discussion
To the best of our knowledge, this is the first Italian study which assesses the exposure to
biological agents in a conventional animal facility, taking into account the biocontamination of different
occupational areas in conjunction with the effect of specific working tasks, cage changing frequency
and animal strains on it.
The most frequent genera identified in this study are common skin inhabitants and are thus easily
shed into the environment on desquamated epithelial skin cells. Some bacterial species have been
282
Appl. Sci. 2019, 9, 2216
reported to cause a variety of diseases, ranging from pneumonia (Acinetobacter) to intestinal infections
(E. coli) (https://www.cdc.gov). Other microorganisms, such as Enterobacter and Pseudomonas, exhibit
a high degree of biological activity, are producers of endotoxin, and often show strong allergenic
properties that increase the risk of disease in workers [46,47]. However, colonies of bacterial species
belonging to risk group 2, according to the Italian Legislative Decrees 81/2008 and 106/2009 [44,45],
have been isolated in some animal rooms during the changing cages. With regard to endotoxin,
no regulatory value is currently available, but various recommendations are available. Among them,
Dutch Expert Committee on Occupational Standards recommends an occupational exposure limit
of 90 EU/m3 [48]. Endotoxin concentrations measured in this study were below this standard in
all settings except for the storage area during activities where a mean value of 210.6 EU/m3 was
observed. Our data confirm what was reported by Kaliste et al., 2002 [33], that bedding material
could be a significant source of accumulation and release of bacterial endotoxins. This evidence was
demonstrated also by Whiteside et al., 2010 [49] who measured a range of endotoxins between 3121
and 5401 EU/g in hardwood rodent bedding.
Endotoxin concentrations measured in this study are higher than those reported by Hwang et al.,
2016 [50], who found a mean value of 0.14 EU/m3 (range = 0.03–0.60 EU/m3 ) in an animal laboratory
housing mice, where the main tasks were feeding/weighing with air exchange rates comparable to ours,
and then those reported by Pacheco et al., 2006 [13] who measured a mean endotoxin concentration
of 315 pg/m3 (about 3.15 EU/m3 ) with a peak value of 678 pg/m3 (corresponding about 6.78 EU/m3 )
during the changing cages in a mouse facility.
Our data are also slightly higher than those reported by Ooms et al., 2008 [39] (mean value
= 4.7 EU/m3 ; range = 3.4–6.3 EU/m3 ) in rabbit rooms; this difference may be attributed to higher
mechanical air-change rates (30.3 air changes/h) in the animal housing space. Finally, results of the
present study are quite lower than those shown by Lieutier-Colas et al., 2001 [15], who found the
highest concentration (15.4 ng/m3 = 154.7 EU/m3 ) associated with cage cleaning and feeding in animal
rooms housing rats.
The detection of endotoxin in the technician’s office (mean value = 2.3 EU/m3 ) suggests that these
components may be carried on clothes and shoes or on airborne particles like allergens.
The highest levels of (1,3)-β-d-glucan were found in the storage area, indicating that activities
performed in this environment were also associated with an increased exposure to fungi, some of
which belong to strongly allergenic and mycotoxin producers genera (Alternaria spp., Penicillium spp.).
Regarding animal allergens, no occupational environmental limits are available, partly due to
the lack of standardized measurement methods and partly because of complex exposure-response
relationships and the influence of genetic susceptibility; it is important to keep the allergen exposure as
low as possible.
We found the highest concentration for Mus m 1 in the dirty washing area (mean value = 61.5 ng/m3)
and its concentrations increased significantly during the changing cages on average by a factor of 9.3 with
respect to the levels before or after (p = 0.037). Straumfors et al., [16] in a recent paper report that cage
emptying and cage washing in the cage washroom represented the highest exposure with a median value of
3.0 ng/m3 for Mus m 1. The lower concentration found by these authors could be explained by the use of
individually ventilated cages in respect to cage-rack systems and by the different sampling methodologies
(stationary versus personal monitoring).
Our data regarding the highest values of Mus m 1 in respect to Rat n 1 in animal facilities are in
agreement with the ones reported in literature (range: 17–564 ng/m3 for Mus m 1; 0.43–27.36 ng/m3 for
Rat n 1) [1,15,51].
For cat allergens (Fel d 1), limit values between 2 and 8 μg/g Fel d 1 for sensitization and above
8 μg/g for the development of acute asthma attacks, respectively, have been proposed [52]. In this work
we measured concentrations at risk of sensitization (>2 μg/g) in two rat rooms.
The detection of cat allergens (Fel d 1) in animal rooms is unexpected, along with the presence of
rat and mouse allergens (Figure 1), showing a cross-contamination likely due to the passive transport
283
Appl. Sci. 2019, 9, 2216
by operators themselves; this highlights their role as vehicle between occupational areas and living
environments [53–56].
Figure 1. Aeroallergen levels over three consecutive days in mouse and rat rooms: before, during and
after changing cages (Each circle represents the observed value and dimension/diameter indicates the
concentration).
5. Conclusions
The environmental monitoring of airborne biocontaminants has been confirmed as a valid tool to
identify the working tasks more critical in terms of exposure to biological agents in animal facilities.
Statistical analysis proves that the changing of cages is a determinant factor in increasing the
concentration of all the airborne biocontaminants and in releasing biological agents that could risk the
health of workers. It can also be pointed out that the concentrations of the biocontaminants increase
significantly during this task with respect to the other days, but return to the background levels
indicating the effectiveness of the ventilation system. Our data also identify the preparation of bedding
and distribution of feed performed in the storage area as critical working tasks in terms of exposure to
endotoxins, fungi and (1,3)-β-d-glucans.
The frequency of the changing of cages does not seem to affect the microbiological contamination;
a statistical difference has been found only in the case of mesophilic bacteria levels. In regard to the
animal strains, rats seem to influence only mesophilic bacteria and endotoxin levels.
Our data can be particularly useful for increasing knowledge about the biological risk in this
occupational sector and to support all actors of prevention, employers, health and safety representatives
and workers in all phases of risk management.
Simple preventive measures must be adopted to control the biological risk in this occupational setting.
Author Contributions: Environmental monitoring and laboratory procedures, A.M.M. and A.C.; Laboratory
procedures, S.D.R.; Data set built, P.M.; Logistics support A.W. and M.C.R.; Statistical analysis, writing-review and
editing, S.M. and R.S.; Conceived and designed the research topic, writing original draft, revising, editing and
supervision, M.C.D. and E.P.
Acknowledgments: The authors are grateful to the workers who collaborated in our experimental phases of
the study.
References
1. Zahradnik, E.; Raulf, M. Respiratory allergens from furred mammals: Environmental and occupational
exposure. Vet. Sci. 2017, 4, 38. [CrossRef] [PubMed]
2. Zahradnik, E.; Raulf, M. Animal allergens and their presence in the environment. Front. Immunol. 2014, 3, 76.
[CrossRef] [PubMed]
3. Liccardi, G.; Calzetta, L.; Baldi, G.; Berra, A.; Billeri, L.; Caminati, M.; Capano, P.; Carpentieri, E.; Ciccarelli, A.;
Crivellaro, M.A.; et al. Italian Allergic Respiratory Diseases Task ForceAllergic sensitization to common pets
(cats/dogs) according to different possible modalities of exposure: An Italian Multicenter Study. Clin. Mol. Allergy
2018, 16, 3. [CrossRef]
284
Appl. Sci. 2019, 9, 2216
4. Liccardi, G.; Salzillo, A.; Piccolo, A.; Calzetta, L.; Rogliani, P. Dysfunction of small airways and prevalence,
airway responsiveness and inflammation in asthma: Much more than small particle size of pet animal
allergens. Ups. J. Med. Sci. 2016, 121, 196–197. [CrossRef] [PubMed]
5. Patelis, A.; Dosanjh, A.; Gunnbjörnsdottir, M.; Borres, M.P.; Högman, M.; Alving, K.; Janson, C.;
Malinovschi, A. New data analysis in a population study raises the hypothesis that particle size contributes
to the pro-asthmatic potential of small pet animal allergens. Ups. J. Med. Sci. 2016, 121, 25–32. [CrossRef]
6. Stave, G.M. Occupational animal allergy. Curr. Allergy Asthma Rep. 2018, 18, 11. [CrossRef] [PubMed]
7. Mason, H.J.; Willerton, L. Airborne exposure to laboratory animal allergens. AIMS Allergy Immunol. 2017, 1,
78–88. [CrossRef]
8. Simoneti, C.S.; Ferraz, E.; Menezes, M.B.; Bagatin, E.; Arruda, L.K.; Vianna, E.O. Allergic sensitization
to laboratory animals is more associated with asthma, rhinitis, and skin symptoms than sensitization to
common allergens. Clin. Exp. Allergy 2017, 47, 1436–1444. [CrossRef]
9. Bush, R.K.; Stave, G.M. Laboratory animal allergy: An update. ILAR J. 2003, 44, 28–51. [CrossRef]
10. Lai, P.S.; Allen, J.G.; Hutchinson, D.S.; Ajami, N.J.; Petrosino, J.F.; Winters, T.; Hug, C.; Wartenberg, G.R.;
Vallarino, J.; Christiani, D.C. Impact of environmental microbiota on human microbiota of workers in
academic mouse research facilities: An observational study. PLoS ONE 2017, 12, e0180969. [CrossRef]
[PubMed]
11. Oppliger, A.; Barresi, F.; Maggi, M.; Schmid-Grendelmeier, P.; Huaux, F.; Hotz, P.; Dressel, H. Association of
endotoxin and allergens with respiratory and skin symptoms: A descriptive study in laboratory animal
workers. Ann. Work. Expo. Health 2017, 61, 822–835. [CrossRef] [PubMed]
12. Samadi, S.; Heederik, D.J.; Krop, E.J.; Jamshidifard, A.R.; Willemse, T.; Wouters, I.M. Allergen and endotoxin
exposure in a companion animal hospital. Occup. Environ. Med. 2010, 67, 486–492. [CrossRef] [PubMed]
13. Pacheco, K.A.; McCammon, C.; Thorne, P.S.; O’Neill, M.E.; Liu, A.H.; Martyny, J.W.; Vandyke, M.;
Newman, L.S.; Rose, C.S. Characterization of endotoxin and mouse allergen exposure in mouse facilities and
research laboratories. Am. Occup. Hyg. 2006, 50, 563–572.
14. Platts-Mills, J.; Custis, N.; Kenney, A.; Tsay, A.; Chapman, M.; Feldman, S.; Platts-Mills, T. The effects
of cage design on airborne allergens and endotoxin in animal rooms: High-volume measurements with
an ion-charging device. Contemp. Top. Lab. Anim. Sci. 2005, 44, 12–16. [PubMed]
15. Lieutier-Colas, F.; Meyer, P.; Larsson, P.; Malmberg, P.; Frossard, N.; Pauli, G.; de Blay, F. Difference in exposure
to airborne major rat allergen (Rat n 1) and to endotoxin in rat quartes according to tasks. Clin. Exp. Allergy
2001, 31, 1449–1456. [CrossRef] [PubMed]
16. Straumfors, A.; Eduard, W.; Andresen, K.; Sjaastad, A.K. Predictors for increased and reduced rat and mouse
allergen exposure in laboratory animal facilities. Ann. Work Expo. Health 2018, 62, 953–965. [CrossRef]
17. Feary, J.; Cullinan, P. Laboratory animal allergy: A new world. Curr. Opin. Allergy Clin. Immunol. 2016, 16,
18. Simoneti, C.S.; Freitas, A.S.; Barbosa, M.C.; Ferraz, E.; de Menezes, M.B.; Bagatin, E.; Arruda, L.K.; Vianna, E.O.
Study of risk factors for atopic sensitization, asthma, and bronchial hyperresponsiveness in animal laboratory
workers. J. Occup. Health 2016, 58, 7–15. [CrossRef]
19. Kampitak, T.; Betschel, S.D. Anaphylaxis in laboratory workers because of rodent handling: Two case reports.
J. Occup. Health 2016, 58, 381–383. [CrossRef]
20. National Institute for Occupational Safety and Health (NIOSH). Preventing Asthma in Animal Handlers;
NIOSH: Washington, DC, USA, 1998; pp. 97–116.
21. Gordon, S. Laboratory animal allergy: A British perspective on a global problem. ILAR J. 2001, 42, 37–46.
[CrossRef]
22. Corradi, M.; Ferdenzi, E.; Mutti, A. The characteristics, treatment and prevention of laboratory animal allergy.
Lab. Anim. 2012, 42, 26–33. [CrossRef] [PubMed]
23. D’Ovidio, M.C.; Martini, A.; Melis, P.; Signorini, S. Value of the microarray for the study of laboratory animal
allergy (LAA). G. Ital. Med. Lav. Ergon. 2011, 33, 109–116. [PubMed]
24. Folletti, I.; Forcina, A.; Marabini, A.; Bussetti, A.; Siracusa, A. Have the prevalence and incidence of
occupational asthma and rhinitis because of laboratory animals declined in the last 25 years? Allergy 2008,
63, 834–841. [CrossRef]
285
Appl. Sci. 2019, 9, 2216
25. D’Ovidio, M.C.; Wirz, A.; Zennaro, D.; Masssari, S.; Melis, P.; Peri, V.M.; Rafaiani, C.; Riviello, M.C.;
Mari, A. Biological occupational allergy: Protein microarray for the study of laboratory animal allergy (LAA).
AIMS Public Health 2018, 5, 352–365. [CrossRef] [PubMed]
26. Larese Filon, F.; Drusian, A.; Mauro, M.; Negro, C. Laboratory animal allergy reduction from 2001 to 2016:
An intervention study. Respir. Med. 2018, 136, 71–76. [CrossRef] [PubMed]
27. Tafuro, F.; Selis, L.; Goldoni, M.; Stendardo, M.; Mozzoni, P.; Ridolo, E.; Boschetto, P.; Corradi, M. Biomarkers of
respiratory allergy in laboratory animal care workers: An observational study. Int. Arch. Occup. Environ. Health
2018, 91, 735–744. [CrossRef]
28. Bhabha, F.K.; Nixon, R. Occupational exposure to laboratory animals causing a severe exacerbation of atopic
eczema. Austr. J. Dermatol. 2012, 53, 155–156. [CrossRef] [PubMed]
29. Jones, M. Laboratory animal allergy in the modern era. Curr. Allergy Asthma Rep. 2015, 15, 73. [CrossRef]
30. Muzembo, B.A.; Eitoku, M.; Inaoka, Y.; Oogiku, M.; Kawakubo, M.; Tai, R.; Takechi, M.; Hirabayashi, K.;
Yoshida, N.; Ngatu, N.R.; et al. Prevalence of occupational allergy in medical researchers exposed to
laboratory animals. Ind. Health 2014, 52, 256–261. [CrossRef]
31. Acton, D.; McCauley, L. Laboratory animal allergy: An occupational hazard. AAOHN J. 2007, 55, 241–244.
[CrossRef]
32. Elliott, L.; Heederik, D.; Marshall, S.; Peden, D.; Loomis, D. Incidence of allergy and allergy symptoms among
workers exposed to laboratory animals. Occup. Environ. Med. 2005, 62, 766–771. [CrossRef] [PubMed]
33. Kaliste, E.; Linnainmaa, M.; Meklin, T.; Nevalainen, A. Airborne contaminants in conventional laboratory
rabbit rooms. Lab. Anim. 2002, 36, 43–50. [CrossRef]
34. Freitas, A.S.; Simoneti, C.S.; Ferraz, E.; Bagatin, E.; Brandão, I.T.; Silva, C.L.; Borges, M.C.; Vianna, EO.
Exposure to high endotoxin concentration increases wheezing prevalence among laboratory animal workers:
A cross-sectional study. BMC Pulm. Med. 2016, 16, 69. [CrossRef] [PubMed]
35. Douwes, J. (1–>3)-Beta-D-glucans and respiratory health: A review of the scientific evidence. Indoor Air
2005, 15, 160–169. [CrossRef] [PubMed]
36. Paba, E.; Chiominto, A.; Marcelloni, A.M.; Proietto, A.; Sisto, R. Exposure to airborne culturable microorganisms
and endotoxin in two Italian poultry slaughterhouses. J. Occup. Environ. Hyg. 2014, 11, 469–478. [CrossRef]
37. Rylander, R.; Lin, R.H. (1–>3)-beta-D-glucan-relationship to indoor air-related symptoms, allergy and asthma.
Toxicology 2000, 152, 47–52. [CrossRef]
38. Rylander, R.; Norhall, M.; Engdahl, U.; Tunsater, A.; Holt, P.G. Airways inflammation, atopy, and (1–>3)-beta-D-
glucan exposure in two schools. Am. J. Respir. Crit. Care Med. 1998, 158, 1685–1687. [CrossRef]
39. Ooms, T.G.; Artwohl, J.E.; Conroy, L.M.; Schoonover, T.D.; Fortman, J.D. Concentration and emission of
airborne contaminants in a laboratory animal facility housing rabbits. J. Am. Assoc. Lab. Anim. Sci. 2008, 47,
39–48. [PubMed]
40. Implementation of the Directive 2010/63/EU on the protection of animals used for scientific purposes, Italian
Legislative Decree 4 march 2014, n. 26. Off. J. 2014, 60.
41. Larone, D.H. Medically Important Fungi: A guide to Identification, 3rd ed.; ASM: Washington, DC, USA, 1995.
42. Lee, T.; Grinshpun, S.A.; Kim, K.Y.; Iossifova, Y.; Adhikari, A.; Reponen, T. Relationship between indoor
and outdoor airborne fungal spores, pollen, and (1→3)-β-D-glucan in homes without visible mold growth.
Aerobiologia 2006, 22, 227–236. [CrossRef]
43. Iossifova, Y.Y.; Reponen, T.; Bernstein, D.I.; Levin, L.; Kalra, H.; Campo, P.; Villareal, M.; Lockey, J.;
Hershey, G.K.; LeMasters, G. House dust (1-3)-beta-D-glucan and wheezing in infants. Allergy 2007, 62,
44. Implementation of the article 1 of the law 3 august 2007, no. 123 concerning the protection of health and
safety in the workplaces, Italian Legislative Decree no. 81/2008. Ordinary Suppl. No.108 Off. J. 2008, 101.
45. Supplementary and corrective provisions of the legislative decree 9 April 2008, no. 81, concerning the
protection of health and safety in the workplaces, Italian Legislative Decree no. 106/2009. Ordinary Suppl. No.
142 Off. J. 2009, 180.
46. Duchaine, C.; Grimard, Y.; Cormier, Y. Influence of building maintenance, environmental factors, and seasons
on airborne contaminants of swine. Am. Ind. Hyg. Assoc. J. 2000, 61, 56–63. [CrossRef]
47. Mandryk, J.; Alwis, K.U.; Hocking, A.D. Effects of personal exposures on pulmonary function and
work-related symptoms among sawmill workers. Ann. Occup. Hyg. 2000, 44, 281–289. [CrossRef]
286
Appl. Sci. 2019, 9, 2216
48. Health Council of the Netherlands. Endotoxins-Health-Based Recommended Occupational Exposure Limit;
Publication no. 2010/04OSH. 010; Health Council of the Netherlands: The Hague, The Netherlands, 2010.
49. Whiteside, T.E.; Thigpen, J.E.; Kissling, M.G.; Grant, M.G.; Forsythe, D.B. Endotoxin, coliform, and dust
levels in various types of rodent bedding. J. Am. Assoc. Lab. Anim. Sci. 2010, 49, 184–189. [PubMed]
50. Hwang, S.H.; Park, D.J.; Park, W.M.; Park, D.U.; Ahn, J.K.; Yoon, C.S. Seasonal variation in airborne endotoxin
levels in indoor environments with different micro-environmental factors in Seoul, South Korea. Environ. Res.
2016, 145, 101–108. [CrossRef]
51. Ohman, J.L.; Hagberg, K.; MacDonald, M.R.; Jones, R.R.; Paigen, B.J.; Kacergis, J.B. Distribution of airborne
mouse allergen in a major mouse breeding facility. J. Allergy. Clin. Immunol. 1994, 94, 810–817.
52. Liccardi, G.; Triggiani, M.; Piccolo, A.; Salzillo, A.; Parente, R.; Manzi, F.; Vatrella, A. Sensitization to common
and uncommon pets or other furry animals: Which may be common mechanisms? Transl. Med. UniSa 2016,
14, 9–14.
53. Liccardi, G.; Salzillo, A.; Piccolo, A.; D’Amato, M.; D’Amato, G. Can the levels of Can f 1 in indoor environments be
evaluated without considering passive transport of allergen indoors? J. Allergy Clin. Immunol. 2013, 131, 1258–1259.
[CrossRef]
54. Krop, E.J.; Doekes, G.; Stone, M.J.; Alberse, R.C.; van der Zee, J.S. Spreading of occupational allergens:
Laboratory animal allergens on hair-covering caps and in mattress dust of laboratory animal workers.
Occup. Environ. Med. 2007, 64, 267–272. [CrossRef] [PubMed]
55. Pereira, F.L.; Silva, D.A.; Sopelete, M.C.; Sung, S.S.; Taketomi, E.A. Mite and cat allergen exposure in Brazilian
public transport vehicles. Ann. Allergy Asthma Immunol. 2004, 93, 179–184. [CrossRef]
56. Ferrari, M.; Perfetti, L.; Moscato, G. A case of indirect exposure to cat at school. Monaldi Arch. Chest. Dis.
2003, 59, 169–170. [PubMed]
287
MDPI
St. Alban-Anlage 66
4052 Basel
Switzerland
Tel. +41 61 683 77 34
Fax +41 61 302 89 18
www.mdpi.com
Applied Sciences Editorial Office

E-mail: [email protected]
www.mdpi.com/journal/applsci
MDPI
St. Alban-Anlage 66
4052 Basel
Switzerland
Tel: +41 61 683 77 34
Fax: +41 61 302 89 18
www.mdpi.com ISBN 978-3-03943-704-7

Indoor Air Quality Book 2021

Uploaded by

Copyright:

Available Formats

Indoor Air Quality Book 2021

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Indoor Air Quality Book 2021

Uploaded by

Copyright:

Available Formats

What is the purpose of this publication?

What is the purpose of this publication?

What topics are covered in the articles?

What topics are covered in the articles?

Indoor Air Quality

ISBN 978-3-03943-703-0 (Hbk)

Cover image courtesy of Dikaia E. Saraga.

About the Editor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii

Myung Eun Cho and Mi Jeong Kim

Jongseong Gwak, Motoki Shino, Kazutaka Ueda and Minoru Kamata

Michał Piasecki and Krystyna Barbara Kostyrko

Zhengguo Yang, Yuto Lim and Yasuo Tan

Anna Mainka and Elwira Zajusz-Zubek

Mehmet Taştan and Hayrettin Gökozan

Gonçalo Marques and Rui Pitarma

Thomas Maggos, Vassilios Binas, Vasileios Siaperas, Antypas Terzopoulos, Panagiotis

Pietro Grisoli, Marco Albertoni and Marinella Rodolﬁ

Received: 8 February 2020; Accepted: 11 February 2020; Published: 22 February 2020

Appl. Sci. 2020, 10, 1501; doi:10.3390/app10041501 1 www.mdpi.com/journal/applsci

Received: 31 May 2019; Accepted: 27 June 2019; Published: 4 September 2019

Keywords: particulate matter; perception; response behavior; psychological attribute

Appl. Sci. 2019, 9, 3660; doi:10.3390/app9183660 5 www.mdpi.com/journal/applsci

2. Status and Risks of Particulate Matter in Korea

2.1. Characteristics of Particulate Matter Generation in Seoul

Figure 1. Annual mean particulate matter (PM) concentration in Seoul (2009–2018).

Figure 3. Seasonal dust concentrations in Seoul.

2.2. Domestic Particulate Matter Standards and Health Protection

Table 1. Comparison of air quality standards for PM (Source: Ministry of Environment).

PM Standards Standard Achievement

yearly 15 10 12, 15 25 65.0%

Table 2. Particulate matter forecast grade (2018).

PM Concentration (μg/m3 , Average for 24 h) Good Normal Bad Very Bad

Table 3. Particulate matter warning and alarm issuing grade (2018).

Category Warning Issuing Alarm Issuing

3. Materials and Methods

3.1. Participants and Questionnaire Design

3.2.1. Measuring Response Behaviors to Particulate Matter

Table 4. Responsive behaviors to particulate matter.

Question: What do you do when Particulate Matter Occurs?

3.2.2. Measuring Psychological Attributes that Hinder Reactions to Particulate Matter

Table 5. Psychological attributes.

Question: What Are your Thoughts in Response to Particulate Matter?

2. If I continue to be exposed to particulate matter, I will be damaged in a few years

Question: What Are your Thoughts in Response to Particulate Matter?

Temporal 8. The risk of particulate matter is present very often or constantly

Externalizing 15. The problem of particulate matter is beyond my ability to solve

4.1. Perceived Indoor Air Quality and Overall Satisfaction

Table 6. Perceptions of and satisfaction with air quality.

4.2. Particulate Matter Environmental Exposures and Perceived Symptoms

Table 7. Average daily external activity time.

Group 1 Group 2 Group 3

Table 8. Health-related symptoms experienced by residents as a consequence of particulate matter.

I have recently been taking medication

Figure 4. Physical symptoms of particulate matter experienced by residents.

4.3. Behavioral Changes in Response to Particulate Matter

Table 9. Residents’ response behaviors to particulate matter.

Figure 5. Residents’ response behaviors to particulate matter.

4.4. Barriers to Mild Behavioral Changes as a Response to Particulate Matter

4.4.1. Lack of Knowledge

Table 10. Residents’ knowledge of particulate matter.