Energy & Buildings 283 (2023) 112849

Contents lists available at ScienceDirect

Energy & Buildings

journal homepage: www.elsevier.com/locate/enb

An experimental study of thermal performance of the radiant ceiling

cooling in office building in Thailand
Ashmin Aryal, Pipat Chaiwiwatworakul ⇑, Surapong Chirarattananon
The Joint Graduate School of Energy and Environment, CHE Center for Energy Technology and Environment, King Mongkut’s University of Technology Thonburi, 126 Pracha-Uthit
Road, Bangmod, Tungkru, Bangkok 10140, Thailand

a r t i c l e i n f o a b s t r a c t

Article history: This paper investigated the radiant ceiling cooling in office buildings in a tropical hot humid climate. Full-
Received 2 October 2022 scale experiments were conducted in an outdoor laboratory chamber to observe the indoor thermal con-
Revised 18 January 2023 ditions and the envelope heat gain from the intermittent operation of the installed radiant chilled panels
Accepted 30 January 2023
(RCP) coupled with an outdoor air unit (OAU). The radiant system was also modeled by using a simulation
Available online 2 February 2023
software namely Transient System Simulation Software (TRNSYS), and the calculation results from the
model were compared to the experimental observations. In the simulation part, the validated TRNSYS
Keywords:
model was adopted to assess the radiant system performance in a typical office environment under
Radiant ceiling cooling
Subjective thermal comfort
the four schemes of combinations of the technology options for the OAU and the building envelope.
Thermal performance The perceived comfort from the radiant cooling was assessed based on the thermal sensation of the local
Tropical hot humid climate subjects acclimatized to hot humid climate. The energy consumption of the radiant cooling was also
apprised as opposed to the conventional mixed-air system. The simulation results presented an insight
into the importance of the exhaust air heat recovery and the envelope insulation on energy saving and
human comfort improvement for the radiant cooling.
Ó 2023 Elsevier B.V. All rights reserved.

1. Introduction ing envelopes. The outdoor air units (OAUs) are installed and ded-
icated to treat all the latent load for controlling the room humidity
Provision of indoor thermal comfort by mechanical refrigera- via the dehumidified outdoor air ventilation. This load decoupling
tion is energy intensive. In Thailand where the climate is tropically allows a possible use of a low-lift chiller to produce chilled water at
hot and humid, the mixed-air system has traditionally been used to a higher temperature (e.g. 18 °C) for handling the sensible heat
condition office buildings and it contributes 50–60 % of the total gains. This offers a dramatic increase in the coefficient of perfor-
building electricity consumption [1,2]. The mixed-air system circu- mance (COP) of the mechanical refrigeration compression process.
lates the large cooled air volume from air handling units (AHUs) to In addition, the OAUs themselves can be equipped with energy-
treat both sensible and latent heat from the building spaces and efficient devices to enhance the cooling performance of the radiant
the outdoor air intake. To perform the aforementioned multiple system. Among various radiant system categories, suspended
functions, the AHUs operate by using the low chilled water supply metal ceiling presents a successful implementation in new build-
temperature (7 °C) for the heat removal, resulting in the high elec- ings. As the radiant chilled ceiling possesses design flexibility, fast
trical power demand of the chillers. In order to achieve the ulti- response to the varying thermal space loads, and ease of mainte-
mate goal of establishing an ultra-low energy building, the nance, it also creates a great opportunity for installation with exist-
means to comfort air-conditioning needs a great improvement in ing buildings in a major renovation [3].
its efficiency. Table 1 summarizes the studies of the radiant cooling and
Hydronic radiant cooling has been perceived to be a potential energy performance. Different strategies were investigated for
technology that can conserve energy without compromising ther- operating the OAUs with the radiant cooling. Stetiu [4] introduced
mal comfort. The radiant system uses large area of the cooled sur- an optimal schedule of the outdoor air ventilation to control accu-
faces such as ceilings and floors, to radiatively absorb the internal mulated moisture during non-office hours and to avoid the con-
sensible heat or load from spaces and the external heat from build- densation risk during running the radiant ceiling under the U.S.
climatic zones. Hao et al. [5] applied the displacement ventilation
⇑ Corresponding author. for offices with the radiant ceiling in China. The OAUs were run
E-mail address: [email protected] (P. Chaiwiwatworakul). intermittently by starting one hour before cooling the space by

https://doi.org/10.1016/j.enbuild.2023.112849
0378-7788/Ó 2023 Elsevier B.V. All rights reserved.
A. Aryal, P. Chaiwiwatworakul and S. Chirarattananon Energy & Buildings 283 (2023) 112849

Nomenclature s simulated value

Acronyms t global component
AHUs Air handling units d diffuse component
COP Coefficient of performance f fenestration
DOAS Dedicated outdoor air system w wall
PMV Predicted mean vote nw north wall
RCP Radiant ceiling panel ew east wall
CV-RMSD Coefficient of variance of the root mean square differ- sw south wall
ence ww west wall
SHR Sensible to total heat ratio amb ambient air
TSV Thermal sensational vote a room air
VAV Variable air volume sa supplied air
ma mixed air
Symbols ex exhaust air
dp dewpoint
I Solar radiation
m Mass flow rate o operative
Q Heat gain or heat extraction g globe
rp radiant ceiling panel
T Temperature
oau outdoor air unit
RH Relative humidity
V Velocity chs chilled water supply
U Heat transfer coefficient chr chilled water return
sen sensible component
lat latent component
Subscripts
in inner surface
out outer surface

radiant ceiling. The simulation studies of Khan et al. [6] in India the comfort classification in ISO7730:2005 [18], the capability of
and of Lim and Jeong [7] in Korea reported the significant energy the radiant systems to precisely control the indoor condition could
improvement of the radiant system when the OAUs were equipped be systematically assessed and compared [12,13].
with the energy recovery wheel (ERW). While the above studies Other than the studies of the efficient OAUs ventilation strategy,
[4–7] particularly focused on the OAUs technologies, the radiant research on building envelope has been carried out, as it has criti-
systems operated to meet the desired comfort conditions of the cal impacts both on the system performance due to the direct cou-
indoor air dry-bulb temperature and relative humidity. These con- pling of the radiative thermal exchange with the cooled surfaces
ditions were also used as criteria to evaluate the energy saving and on the thermal comfort due to the influence of the mean radi-
potential of the radiant system as compared to the mixed-air sys- ant temperature. The study results from the literature [19,20] pre-
tem. The criteria and the key findings results are given in the table. sented that the heat load from the building envelope of the radiant
However, in the thermal comfort viewpoint, the radiant system system was possibly higher than that of the mixed-air system.
employs the mean radiant temperature to take into account the According to the references [19,21,22], envelope configuration,
radiative heat transfer from human body, whilst the mixed-air sys- materials, and orientation all had strong effect on the instanta-
tem does the full negligence; the system operation based on the neous cooling load. Jordan et al. [22] reported the essence of better
same room air temperature setpoint would result in a disparate inclusive envelope designs like wall insulation and sun shading for
thermal comfort level between the two air-conditioned the proper functioning and success of the radiant air-conditioning.
environments. Higher shading coefficients of the glazing was suggested for ther-
It is perceived that the human comfort is influenced by and able mal comfort improvement purposes.
to be controlled via the four physical factors of the air temperature, The above evaluation method implied that the occupants would
relative humidity, mean radiant temperature, and air movement. express the same thermal sensation response under the controlled
According to the ASHRAE standard 55 [8], the summer neutral environment by the radiant and the convective cooling systems.
comfort zone is established and prescribed by the borderlines of However, as mentioned in [14], the thermal comfort is a psycho-
the operative temperature (To), the equivalent temperature derived logical issue and it was possibly evolved by technological develop-
based on the air dry-bulb temperature and the mean radiant tem- ment of air-conditioning system and its control. An example of
perature expressing the combined effect of the convective and evidence could be the previous work of the authors [26] conduct-
radiative heat transfer between an occupant body and the indoor ing the subjective thermal comfort survey, reporting that Thai vol-
environment. This neutral zone is corresponding to the predicted unteers who were familiarized with the mixed-air system
mean vote (PMV) index within ± 0.5 value range computed from expressed their difference in the thermal sensation response under
the Fanger’s heat balance model using the above four factors and the radiant cooling environment using suspended chilled ceiling.
the two personal factors of the clothing insulation and metabolic The neutrality of the radiant cooling was observed at 25.0 °C while
rate. By adopting the PMV or the operative temperature as the cri- that of the mixed-air system was relatively higher at 25.7 °C. The
teria with the same comfort level warranty, the energy perfor- subject sensation gradients were also different for the two systems.
mance of the different radiant technologies (Radiant ceiling/floor) To bridge the existing research gap of the radiant cooling in the
have been compared to the mixed-air system; those undertaken tropical zone, this paper investigated the promising technology
in various locations such as in India [9], in USA [10,11], in Europe integration of the radiant chilled ceiling, efficient outdoor air unit
[12–14], in Japan [15], in China [16] and in Qatar [17]. Based on and envelope insulation to achieve a low-energy office building

2
A. Aryal, P. Chaiwiwatworakul and S. Chirarattananon Energy & Buildings 283 (2023) 112849

Table 1
Radiant cooling studies on energy performance and thermal comfort.

Researchers Studied system Location Comparison criteria Method Key findings

Stetiu [4] RCP vs VAV USA Indoor air temperature 24 ± 1 °C Simulation 30 % energy savings and 27 % peak
demand reduction
Hao et al. [5] Radiant ceiling vs all-air system China Indoor air temperature 25 °C, Empirical energy saving by 68.5 % for chillers
relative humidity 50 % equation and by 39.0 % for AHUs
Khan et al. [6] Radiant slab vs fan coil system India Same cooling load Measurement 17.5 % energy savings which would
& simulation increase to 30.3 % when the OAUs
were equipped to the energy
recovery wheel (ERW)
Lim and Jeong [7] Radiant ceiling vs VAV system Korea Indoor air temperature 25 °C, Simulation 41.3 % energy savings
relative humidity 50 %
Khan et al. [9] Radiant slab vs VAV system India Operative temperature 27 ± 1 °C Measurement 11–27 % energy savings
& simulation
Henze et al. [10] Radiant slab vs VAV system USA Operative temperature Simulation 20 % energy saving achieved by
24.5 ± 1 °C radiant slab combined with
geothermal heat exchanger
Jeong et al. [11] Radiant ceiling vs VAV system USA Same cooling load Simulation 29 % energy saving by radiant
system as compared to mixed air
system
Fabrizio et al. [12] Radiant floor & ceiling vs VAV Europe Operative temperature Simulation 11 % and 15 % energy reduction for
system 23.0–26.0 °C radiant floor and chilled ceiling,
compared to VAV system
Salvalai et al. [13] Radiant ceiling vs all-air system Europe Comfort class Simulation 5–50 % energy saving achieved by
radiant ceiling as compared to fan
coil system
Oxizidis and Papadopoulos Various radiant systems vs all- Greece Operative temperature Simulation Radiant systems consumed 14 %
[14] air system 25.0 ± 0.5 °C less cooling energy than fan coil
system
Imanari et al. [15] Radiant ceiling vs all-air system Japan PMV ± 0.3 Numerical 10 % energy saving with RCP as
calculation compared to all-air system
Liao et al. [16] Radiant ceiling vs all-air system China Operative temperature Experimental 26.5 % energy saving achieved by
23.0 ± 1 °C measurement radiant ceiling cooling
Ismail and Ouahrani [17] Personalized cooling radiant Qatar PMV Simulation 18 % energy saving achieved by
cubicle vs all-air system radiant cubicle
Kwong et al. [23] Radiant slab vs VAV system Malaysia Radiant system setpoint (23– Measurement 34.4 % energy savings
24.8 °C) VAV setpoint (24 °C) & Simulation
Niu et al. [24] Radiant ceiling vs all-air system Hong Operative temperature 25.0 °C, Simulation 44 % energy saving achieved by
Kong relative humidity 50 % radiant ceiling
Shastry and Rumsey [25] Radiant slab vs VAV system India N/A Field Radiant slab system consumed
measurement 34 % less energy than VAV system

based on the thermal comfort sensation of the subjects in the hot ogy, King Mongkut’s University of Technology, Thonburi (Latitude
humid climate. Through physical experiments and software simu- 13.4° and longitude 100.5°). The chamber’s length-sides faced east
lations, the indoor thermal environment, the dynamic cooling load, and west. The opaque walls were a lightweight structure of fiber
and the consumed energy of the radiant system were characterized cement board as the outer layer and gypsum board as the inner
in comparison with the mixed-air system. layer. Insulation was inserted as a sandwiched layer to reduce
the external heat from solar radiation and the temperature differ-
ence between the outdoor and indoor air. Double low-e glazing
2. Experiments windows were situated on the width-side walls (north and south
walls) with a coverage area of 60 %. The solar transmittance (ssol)
Full-scale experiments were carried out to investigate the cool- and the solar heat gain coefficient (SHGC) of the glass window
ing of a radiant chilled ceiling system in an office environment were 0.544 and 0.597, respectively. External horizontal slats were
under real climate conditions. The experimental room was the equipped to intercept the beam radiation incidence on the win-
same room used to assess the thermal comfort sensation of Thai dow. The roof was aluminium sheets attached underneath with
subjects by the radiant cooling, as presented in a prior study of insulation foam. Table 2 summarizes the wall configurations,
the authors [26]. Several sensors were placed inside the room to materials, and their thermal properties. As noticed, there were
measure the indoor thermal condition and to monitor the radiant two insulation types used in this room where the fiberglass matt
system operation. The observed data were analyzed and used to layer existed since earlier, and the polystyrene foam layer was
validate the calculations of the radiant system model developed newly added during the experiment setup. Fig. 1 shows pho-
by TRNSYS software. tograph of an exterior view of the laboratory room.
The room interior was refurbished as an office-like environ-
2.1. The setup ment. Desktop computers served as the data logger and operated
continuously (0:00–24:00) during the experimentation. LED lamps
The main facilities comprised an experimental radiant cooling provided the interior lighting during 8:00–18:00. The dissipated
room, a data acquisition system, and a meteorological station. A heat from the electric devices formed part of the internal air-
detailed setup of the facilities is given below. conditioning load. As no volunteer subjects stayed in the radiant
a) Experimental room room, a wooden box, inside which fluorescent lamps were placed
The room was a 3 m by 9 m rectangular chamber located on the and turned on, was used to generate sensible heat equivalent to
roof deck of a building of the school of Bioresources and Technol- three people (252 W). The subject’s latent load was simultaneously

3
A. Aryal, P. Chaiwiwatworakul and S. Chirarattananon Energy & Buildings 283 (2023) 112849

generated by using a small humidifier (0.32 kg/hr). Both fluores- Fig. 2 shows the diagram of the system chilled water loop. The
cent lamps and humidifiers were switched on only during office chilled water from an air-cooled chiller was directly supplied to
hours using a timer. the OAU for dehumidifying the air. The chilled water was also
b) Radiant ceiling system served for the cooling of the RCP through a plate heat exchanger.
A hybrid system of radiant ceiling panels (RCP) coupled with Placement of the OAU and the RCP in the experimental room is also
outdoor air unit (OAU) was installed in the experimental room. shown in Fig. 2.

Table 2
Envelope configuration of the radiant cooling room.

Envelope Area Material layer Thickness Density Specific heat capacity (J/kg.k) Conductivity (W/m.k)
(m2) (m) (kg/m3)
Wall Uw = 0.318 7.8 (N, S) and 23.4 (E, W) Cement fiber 0.010 1900 1000 0.58
Fiberglass 0.050 29 1213 0.034
Polystyrene 0.050 100 1380 0.035
Gypsum board 0.012 600 1090 0.191
Roof Ur = 0.323 27 Aluminum 0.003 2700 860 200
Fiberglass 0.050 29 1213 0.034
Polystyrene 0.050 100 1380 0.035
Gypsum board 0.009 600 1090 0.191
Window (N and S) 4.5 Double low-e glazing, SHGC = 0.597, Uf = 1.76, ssol = 0.544

Fig. 1. The experimental room of radiant ceiling cooling.

Fig. 2. Radiant cooling system.

4
A. Aryal, P. Chaiwiwatworakul and S. Chirarattananon Energy & Buildings 283 (2023) 112849

Table 3
Specification of the installed sensors for the radiant cooling experiments.

Instrument Measurement range Accuracy Measured variable

Paddle wheel sensor (UN-FM100) 20–4200 L/minute Flow: ±2% mw,oau, mw,rp
Anemometer (KIMO CTV200) 0.0–1.0 m/s Velocity: ±3% Va
Platinum RTD sensor (PT100) 50–300 °C Temperature: ±0.1 °C Tew,i, Tew,o, Tww,i, Tww,o, Tswi, Tsw,o, Tnw,i, Tnw,o,
Tchs-oau, Tchr-oau, Tchs,rp, Tchwr,rp, Trp
Temperature & humidity sensor (HMP45A) Temperature: (–32.9)-60 °C Temperature: ±0.2 °C Tamb, RHamb, Tsa, RHsa, Tma, RHma, Tex, RHex
Humidity: 0.8–100 % Humidity: 0.05 %RH/°C
Temperature & humidity of indoor condition Globe temperature: 0–70 °C Air temperature: ±0.3 °C Ta, Tg, RHa
(WBGT-2010SD) Humidity: 0–100 % Globe temperature: ±0.2 °C
Air humidity: 3 %
Heat flow sensor (EKO MF-180 M) N/A Sensitivity: 25 lV/W/m2 Qew,, Qww, Qnw, Qsw,

According to the ASHRAE standard 62.1 prescriptions for office A three-dimensional (3D) model of the experimental room was
ventilation [27] and the assumed occupancy rate of 0.1 person/m2, created using Google SketchUp software, and it was then imported
the outdoor air intake in the experimental room was set at 50 CFM. to TRNSYS by using TRNSYS 3D Plugin. This geometrical model
An additional 50 CFM of the returned air from the room was circu- enabled TRNSYS to determine the view factor among the interior
lated through the OAU to treat the moisture from air infiltration room surfaces for accurate calculation of the radiative heat
and to increase the indoor air movement. The OAU cooling coil exchange among the surfaces.
was equipped with a heat-pipe wrapped around coil to enhance With the standard module of multi-zone building ‘‘Type 56”,
the air dehumidification. A heat-pipe heat recovery unit was also the envelope components of walls, windows, and roof of the room
installed to precool the outdoor air by using the exhaust air. The as described in Table 2 were configured into the module. In mod-
dehumidification rate of the OAU was regulated by an automatic eling the radiant chilled ceiling, the Layer Type Chilled Ceiling
valve to maintain the room air dewpoint temperature. was added to the model. The Layer Type Chilled Ceiling required
The radiant ceiling comprised five metal panels each of which various chilled ceiling parameters that were the pipe spacing
was connected in parallel by the cooling water circuit. In the dia- (0.15 m), pipe inside diameter (0.012 m), specific normal power
gram, a circulation pump circulated the cooling water at a constant (351.36 kJ/hm2), specific normal flow (33.15 kg/hm2), norm area
rate of 0.175 kg/s. A manifold was used to distribute the cooling (3.8 m2), and the difference between mean fluid temperature and
water equally to each radiant panel. A motorized value (located mean surface temperature of norm condition (3.5 K). The numeric
in the chilled water circuit) was used to regulate the RCP surface values in the brackets were the given input to the chilled ceiling
temperature to maintain the room operative temperature accord- parameters. The ceiling coverage ratio of the radiant panels was
ing to the setting point. 0.75 for this experimental room. Other room characteristics of air
c) Data acquisition system infiltration and internal heat sources estimated from measure-
Several sensors were installed to monitor the operation of the ments were also input into ‘‘Type 56” for the heat balance
radiant system and the indoor thermal conditions. The sensor sig- calculation.
nals were transmitted and stored in a computer for every minute. The data reader ‘‘Type 9e” was used to read the weather file pre-
Table 3 lists the installed sensors and their specifications in brief. pared from the records of the meteorological station at the study
The models of the sensors appear in the brackets. site. The records consisted of global radiation, diffused radiation,
d) Meteorological station ambient air temperature and relative humidity, sky temperature,
A nearby meteorological station measured the outdoor solar and ground reflectance. Radiation processor ‘‘Type16i” was used
radiation including the global, the diffuse horizontal, and the beam for trans-positioning the horizontal radiation data to the radiation
normal components. The vertical radiation on the north, south, on the tilted surfaces (walls and windows) and determining the
east, and west facades was also measured. All radiation sensors corresponding incidence angle of the beam radiation.
were categorized as the secondary standard of the ISO 9060 and The OAU in the experimental room was modeled using a cooling
the world meteorological organization (WMO). Other measure- coil model ‘‘Type 508a”, a wrapped-around coil model ‘‘Type 689”,
ments were the ambient air temperature (Tamb) and relative and an air-to-air heat recovery model ‘‘Type 667”. The OAU circu-
humidity (RHamb). lated the constant air flow through its coil with a rate of 50 CFM of
All measured data were subject to data quality control and were the outdoor air and 50 CFM of the returned air from the room. The
one-minute records. The recording time of the experimental room mixing of the room air and the outdoor air was done by the air
was synchronized to that of the station. All the data served as the mixing valve ‘‘Type 648”. The cooling coil ‘‘Type508a” was
input of the TRNSYS model for the radiant cooling simulations. equipped with a heat-pipe round around coil ‘‘Type 689” to
enhance the dehumidification performance. A heat-recovery unit
‘‘Type 667” was used for precooling the outdoor air by using the
2.2. Radiant system modelling exhaust air from the experimental room.
Forcing function module ‘‘Type14h” was used for controlling
TRNSYS can perform transient simulations of buildings and the components; psychometrics module ‘‘Type 33c” for determin-
HVAC (heating, ventilation, and air conditioning) systems. It is a ing the processed air properties; online graphical plotter ‘‘Type
popular tool for modeling radiant cooling systems [7,13,22,28]. 65c” for printing the different observed variables in the system.
The library of HVAC models and building models available from In the simulation studio setting, the TRNSYS model was run to
TRNSYS is extensive and uses a component-based approach to determine the room indoor conditions from the operation of the
building simulation, which allows for more flexibility and cus- RCP and the OAU using the records of the varying outdoor climate
tomization in modeling different types of systems. TRNSYS is uti- conditions during the experimentations. The coefficient of varia-
lized in this study as the modeling tool for both the building and tion of root mean square difference (CV-RMSD) between the exper-
the integrated radiant cooling system. imental values and the simulated values were evaluated to
5
A. Aryal, P. Chaiwiwatworakul and S. Chirarattananon Energy & Buildings 283 (2023) 112849

Fig. 3. Experiment of the OAU dehumidification.

6
A. Aryal, P. Chaiwiwatworakul and S. Chirarattananon Energy & Buildings 283 (2023) 112849

Table 4
CV-RMSD calculated for the OAU dehumidification experiment.

Parameter Tsa RHsa Tma RHma Tex RHex Qoau

CV-RMSD 11.32 10.24 1.55 11.28 1.39 2.89 15.15

ascertain the accuracy of the developed TRNSYS model and the sys- a) Outdoor conditions
tem parameter settings. During the experiment (10–15 January 2022), the sky was clear,
and the global solar radiation (It) was intense. Over the days, the
2.3. Experimental results ambient air temperature (Tamb) ranged between 22 °C and 34 °C,
and the relative humidity (RHamb) was 40–80 %, portraying the
Two experiments were carried out to investigate the air dehu- hot-humid climate in the tropics. The variations of the outdoor
midification of the OAU and the radiative cooling of the RCP. The conditions were shown in Fig. A.1 in Appendix.
measured data were used to evaluate the radiant system perfor- b) Indoor conditions
mance and to validate the TRNSYS radiant system modeling. Indoor conditions of the experimental radiant room were
described by the room air temperature (Ta), operative temperature
2.3.1. Air dehumidification of the OAU (To), mean radiant temperature (Tmrt), and dewpoint temperature
The OAU operated to control the room humidity at the dew- (Tdp), and relative humidity (RHa). As observed in Fig. 3(a), by run-
point temperature setpoint of 16.0 °C and to prevent condensation ning the OAU from 6:00, the room dewpoint temperature was
on the radiant panels during the office hours. reduced and could be maintained around the setpoint of 16.0 °C

Fig. 4. Experiment of the radiant ceiling panels coupled with OAU.

7
A. Aryal, P. Chaiwiwatworakul and S. Chirarattananon Energy & Buildings 283 (2023) 112849

throughout the office period. However, the variation of the room a) Outdoor conditions
air temperature and its arising to 31.0 °C in the afternoon indicated The outdoor conditions during this experiment are given in
the OAU was not capable to handle the apparent sensible load. The Fig. A2(a) in the appendix, showing that the solar radiation and
measurements also showed the room operative temperature and the ambient air conditions were comparable to that of the OAU
the mean radiant temperature varying correspondingly with the experiment. The incident solar radiation on walls facing north
room air temperature. (Inw), south (Isw), east (Iew), and west (Iww) measured from the sta-
Fig. 3(b) exhibits the measured OAU processed air conditions. tion are presented in Fig. A2(b).
The lowest temperature of the supply air (Tsa) from the OAU during As the experiments were conducted in January when the sun
the start-up period implied the full operation of the machine to traveled due south, the south-facing radiation was higher than that
remove the accumulated moisture inside the room overnight. of the north. The east radiation was relatively high in the morning
The supply air relative humidity (RHsa) was observed to be rather while the opposite was true for the west radiation. Superimposed,
low due to the reheating process of the wrapped around coil. In the calculated radiations on the plots indicated that the TRNSYS
the figure, the calculation results from the TRNSYS model were radiation processor could well determine the incident radiations
superimposed for the model validation purpose. on the outer surfaces of the experimental room walls.
According to ASHRAE Guideline 14-2014 [31], the coefficients of b) Indoor conditions
variance of the root mean square difference (CV-RMSD) was To condition the room, the radiant system started running the
adopted to validate the TRNSYS model accuracy. As suggested in OAU to dehumidify the indoor air from 6:00. The radiant ceiling
[32], the CV-RMSD values derived from the hourly data should panels simultaneously cooled the room with a constraint of its sur-
not exceed 2 % for temperature, and not exceed 30 % for energy; face temperature of at least 1 °C above the room dewpoint temper-
the CV-RMSD from our experiments shown in Table 4 indicated ature. Fig. 4(a) shows that the room dewpoint temperature was
that the accuracy of the OAU model by TRNSYS was acceptable. around 16 °C. The room operative temperature could be main-
It should be remarked that the experimental data were one- tained at 24 °C with a small variation of ±0.2 °C by regulating the
minute interval, not the hourly data. surface temperature of the radiant ceiling panels between 20 °C
c) Cooling load and 22 °C. No condensation formed on the ceiling panel. Observing
The OAU cooling load (Qoau) was calculated from the measured that the room air temperature was 0.5 °C higher than the operative
chilled water data (the chilled water flow rate, and its supply and temperature while the mean radiant temperature was lower by
return temperatures). The measurements of the processed air con- 0.5 °C. Fig. 4(b) shows the operating conditions at the radiant ceil-
ditions were used to segregate the OAU load into the sensible (Qoau, ing panels. The temperature of the supplied and returned chilled
sen) and latent (Qoau,lat) components. water were 2–3 °C lower than the ceiling temperature. The simu-
Fig. 3(c) shows that the calculations from the chilled water side lation results from TRNSYS model superimposed on the plots exhi-
did well agree with that of the air side. The analysis also reported bits good agreement with the experimental measurements.
that the sensible to total heat ratio (SHR) was as low as 0.35 when c) Wall heat gains
the OAU started, and it would increase in the afternoon when the Heat transfer through the north (Qnw), south (Qsw), east (Qew),
outdoor air temperature and the room sensible load were higher. and west (Qww) walls of the experimental room were measured;
however, the result of the south wall was only given as an example.
The corresponding outer and inner surface temperatures of the
2.3.2. Radiative cooling of the chilled ceiling
wall are also presented.
In this experiment (21–26 January 2022), the chilled ceiling
As the room walls were insulated, the wall heat gains were
panels coupled with OAU was demonstrated to control the opera-
lower than 20 W/m2. The inner surface temperatures were rather
tive temperature of the experimental room at 24 °C and the room
constant around 25 °C while that of the outer surfaces were as high
dewpoint temperature at 16 °C.

Fig. 5. Heat gain through south wall.

8
A. Aryal, P. Chaiwiwatworakul and S. Chirarattananon Energy & Buildings 283 (2023) 112849

Fig. 6. System cooling load.

Table 5
CV-RMSD calculated for the radiant cooling experiment.

Parameter Inw Iew Isw Iww Qew Qnw Qsw Qww

CV-RMSD 20.1 12.0 4.6 7.5 19.4 36.3 14.5 16.0
Parameter Tnw,out Tew,out Tsw,out Tww,out Tnw,in Tew,in Tsw,in Tww, in
CV-RMSD 3.4 15.8 8.4 9.6 10.9 3.4 5.2 3.3
Parameter Ta RHa Tmrt To Tdp Trp Qoau Qrp
CV-RMSD 1.2 1.3 6.5 5.0 0.7 6.0 18.8 19.5

as 45 °C. Fig. 5 presents the comparison of the measured results tem was appraised for the office hour period. Influences of different
with the calculations from the TRNSYS model. It is obvious that OAU configuration options and building envelope features were
both were quite consistent. also assessed in the simulations. Finally, the energy-saving poten-
d) Cooling load tial of the radiant ceiling system under the intermittent operation
Fig. 6 shows the radiant system loads at the ceiling panels and scheme was evaluated with a reference case of the conventional
at the OAU. The panels had its load peak when the operation mixed-air system. The detailed descriptions of the simulation
started by cooling the panels itself. The load tended to have the schemes are given as follows.
second peak around noon. The ceiling panels had largely variable
load as compared to the OAU. It also shared 45.6 % of the total sys-
tem load. The relatively low load sharing might result from the low 3.1. Typical office
heat gain from the insulated envelope and good sun-shaded win-
dow. Total system cooling load intensity could be calculated at Office buildings in Thailand could be characterized by air-
111 Wth/m2. conditioned space enclosed by non-insulated envelope. Opaque
Based on the sensors accuracy in Table 3 and the uncertainty walls were presented by thin layers, and medium weight building
analysis in the reference [29,30], the absolute and relative uncer- structures. Past energy auditing indicated that although the opa-
tainty calculated for the radiant ceiling panel heat extraction were que walls were constructed by different walling materials, their
0.115 kW and 5 % respectively. overall heat transfer coefficient (Uw) was approximately 2.65–
Table 5 presents the CV-RMSD of the experimental parameters 2.90 W/m2/K [33–35]. Single pane, heat reflective glazes with dark
including the incident solar radiation on the outer surface of the colors were quite common for windows [36]. The ratio of the win-
room walls, the wall heat gain, indoor conditions, and cooling load dow area to the total wall area was averaged at 0.40. Fig. 7(a) exhi-
of the RCP and the OAU. The accuracy of the TRNSYS model was bits a typical office building form in the simulation study. The
acceptable. building was a rectangular shape of 20 m in width by 40 m in
length, comprising 15 stories with a total floor area of 12,000 m2
(Table 6).
3. Simulation study Typical interior illumination was designed based on the uni-
form and direct lighting concept. LED lamps were currently domi-
The experimentally validated TRNSYS model was now used to nant due to their excellent luminous efficacy. Daylighting for the
simulate the air-conditioning of the radiant ceiling in a typical building’s interior illumination was still quite limited by the favor-
office building in the tropical climate. The radiant system operating able use of heat reflective glasses and indoor curtains. According to
parameters in the simulations were set identical to that of the the lighting practice, the illuminance on the work plane level was
experiments. The perceived thermal comfort from the radiant sys- designed at 400 lx.
9
A. Aryal, P. Chaiwiwatworakul and S. Chirarattananon Energy & Buildings 283 (2023) 112849

Fig. 7. Typical office building model.

Table 6
The features of typical office building in the simulation study.

System Features
Envelope Wall:
 Opaque section of 0.15 m cement block with mortar plastering
 Glazing section of single heat reflective glass
 Heat transfer coefficient (Uw) of the opaque wall was 2.97 W/(m2.K)
 Heat transfer coefficient (Uf) of glazing was 5.73 W/(m2.K)
 Solar heat gain coefficient (SHGC) of glazing was 0.527 W/(m2.K)
 Glazing thermal transmittance (s) was 0.463
Roof:
 0.15 m concrete slab with 0.05 m-thick fiberglass insulation
 Heat transfer coefficient (Ur) of the roof was 0.69 W/(m2.K)
Lighting  Uniform lighting of 400 lx on the work plane level
LED lamps
 Lighting power density 7 W/m2
Air-conditioning  Supply chilled water was at 7 °C
The room air temperature setpoint was at 24 °C
 Outdoor air for ventilation: 932.8 CFM/floor (1.17 CFM/m2)
 The indoor air movement at 0.15 m/s

As Thailand is situated in the tropical region surrounding the requirements of building envelope, lighting, and air-conditioning
earth’s equator with a hot humid climate, building cooling is systems. Existing buildings in the major renovation are recom-
required for thermal comfort provision all year round. Air- mended to be complied with the code, as well. Based on the system
conditioning in a typical office uses a mixed-air system. Air han- features in Table 4 and the code calculation procedure, the overall
dling units (AHUs) supply cooled air to control the room tempera- thermal transfer value (OTTV) of the building envelope was
ture according to the temperature setpoint of thermostats. The obtained at 59.8 Wth/m2. The coefficient of performance (COP) of
room humidity is indirectly affected by the AHU operation and the water chiller and the air-conditioning system were 3 and
variable outdoor air conditions. Outdoor air units (OAU) intro- 2.34, respectively [1].
duced and pre-conditioned the outside air for ventilation to The typical office was modeled and simulated in TRNSYS with
achieve good indoor air quality as per prescribed in ASHRAE 62.1 the solar radiation and weather data from our meteorological sta-
(2.5 L/s/person and 0.3 L/s/m2 for office space). As shown in tion. The annual simulation reported that the energy intensity of
Fig. 7(b), an air-cooled chiller supplied chilled water with a tem- the typical office model was at 108 kWh/m2/year where air-
perature of 7 °C to both the AHUs and OAUs. The indoor air was conditioning contributed 59 %, lighting 16 %, and equipment 25 %
directly exhausted to the outdoor without energy recovery. which were consistent with that reported from the country energy
Building use patterns are described by the hourly profiles of audit [37].
occupancy density, lighting power density, and equipment power
density as shown in Fig. 8. The building model set out the occu- 3.2. Simulation schemes
pancy density based on a design norm of 10 m2/person. The occu-
pants performed sedentary activity with a metabolic rate of 1.2 In comparison of the two air-conditioning systems, four simula-
met according to the ASHRAE thermal comfort standard. The rated tion schemes were established from combinations of the technol-
lighting and equipment power densities were at 7 W/m2 and 10 W/ ogy options of the building envelope i.e. insulation and sun
m2, respectively. The building was assumed to operate for five shading, and of the OAU i.e. wrapped around coil and heat recovery
working days from Monday to Friday, and including Saturday, unit; all of them could enhance the radiant cooling application,
8:00–18:00. Table 4 summarizes the physical features and system potentially leading the considered building to an ultra-low energy
performances of the typical office building. In the annual simula- building.
tion of the typical building, the infiltration rate was defined to a) Envelope features
0.3 ACH. The effect of thermal bridging is not considered in this For office buildings in Thailand, opaque walls were non-
study. insulated, and glazed windows were not effectively shaded. The
In Thailand, new commercial buildings with the floor area of building envelope contributed approximately 60 % of the total
more than 10,000 m2 are subject to the compliance with building cooling load of the conventional mixed-air system [37]. The high
energy code (BEC) prescribing a set of the minimum performance interior surface temperature of the walls and windows caused
10
A. Aryal, P. Chaiwiwatworakul and S. Chirarattananon Energy & Buildings 283 (2023) 112849

Fig. 8. Usage profiles of the typical office building.

Table 7 with heat recovery wheel for pre-cooling the outdoor air by the
The simulation schemes of the radiant ceiling cooling. exhaust air. The OAU cooling coil equipped with a wrap-around
Scheme Description coil can improve the dehumidification of the main coil and effec-
I: R1- Radiant ceiling coupled with the OAU which has only a main tively decouple the sensible and latent loads. However, their appli-
B1 cooling coil (R1). The system is operated in the typical office (B1). cations were considered climate-dependent [39–41].
II: R1- The radiant system in Scheme I (R1) operated in the typical office Two OAU configurations were considered in this study where
B2 with enhanced building envelope performance (B2). the one represented a conventional OAU having only a cooling coil,
III: R2- Radiant ceiling coupled with the OAU that is equipped with wrap-
denoted as R1. The other configuration was the OAU which was
B1 around coil and heat recovery (R2). The system is operated in the
typical office (B1). integrated with a wrap-around coil and heat recovery unit,
IV: R2- The radiant system in Scheme III (R2) operated in the typical office denoted as R2. Fig. 12 presents the two radiant system configura-
B2 with enhanced building envelope performance (B2). tions. The RCP was supplied with higher-temperature chilled water
by a low-lift chiller and the OAU was supplied with 7 °C chilled
the possible thermal discomfort to the building occupants, even water with a separate chiller. Table 7 summarizes the four schemes
though the room temperature was well-controlled according to of radiant cooling in the simulations. In the table, B1 represents the
the setpoint. Furthermore, it was reported from prior studies typical office building envelope and B2 represents the energy-
[19,20,38] that the direct radiative heat transfer from the building efficient envelope of the insulated walls and the well-shaded
envelope to the cooled surfaces could cause the instantaneous windows.
cooling load of the radiant system larger than the mixed-air c) Subjective thermal comfort
system. Indoor thermal comfort is a criterion for assessing the operation
In this simulation, the insulated walls and shaded windows of an air-conditioning system [8]. It is also used as a reference
were established as the scheme of an energy-efficient envelope when the energy performance and consumption of different air-
for the radiant ceiling cooling. The exterior walls attached inside conditioning cases are being compared. Predicted mean vote
with 0.05 m thick polystyrene foam could reduce the heat transfer (PMV) is the index most frequently used to determine the indoor
coefficient to 1.09 W/m2/K. The windows equipped with external comfort of air-conditioning. Based on the four physical indoor con-
shading devices were assumed, hence reducing the window shad- dition parameters (i.e. room temperature, room humidity, mean
ing coefficient (SC) from 0.8 to 0.5. In this study, the shading coef- radiant temperature, and air velocity) and the two personal factors
ficient is defined as the ratio of transmitted to the corresponding (i.e. subject metabolic rate, and clothing), the PMV value rates the
incident solar radiation. thermal sensation of a large group of people into seven scales from
b) OAU configurations ‘‘-3” of ‘‘too cool” to ‘‘+3” of ‘‘too hot”, where ‘‘0” indicates the ther-
The OAU is essential to control the room humidity to prevent mal neutrality.
moisture condensation on the cooled ceiling surface. The OAU The ISO7730: 2005 [18] standard recommends the four comfort
can enhance the radiant system performance when it is equipped categories for air-conditioning design and energy calculation. The

11
A. Aryal, P. Chaiwiwatworakul and S. Chirarattananon Energy & Buildings 283 (2023) 112849

Fig. 9. The radiant system configurations in the simulation study.

Fig. 10. Thermal comfort of Thai subjects for the radiant and the convective cooling.

12
A. Aryal, P. Chaiwiwatworakul and S. Chirarattananon Energy & Buildings 283 (2023) 112849

Table 8 With this regard, the total amount of the building cooling load and
The indoor comfort classification of Thai subjects for the assessment of the radiant the air-conditioning energy consumption could be obtained by
ceiling and the mixed-air systems.
multiplying the simulated results of the middle floor by thirteen,
Class TSV range Operative temperature range (°C) and then summed with the results of the ground floor and the
Radiant cooling Convective cooling top floor.
Class A 0.5  TSV  0.5 23.7  To  26.6 25.0  To  26.3
Class B 0.7  TSV  0.7 23.1  To  27.2 24.8  To  26.3 4.1. Indoor conditions and comfort
Class C TSV  -0.7 or To  23.1 or To  24.8 or
TSV  0.7 To  27.2 To  26.3 Only the indoor conditions and comfort of the top floor were
presented, as the space was exerted by the most drastic outdoor
conditions due to the solar radiation incidence on the roof deck.
categories specify the sufficient level of the indoor thermal comfort (a) Reference case: The mixed-air system
as per the defined PMV value ranges: Category I as 0.2 < PMV < 0. As exhibited in the reference case in Fig. 11(a), the mixed-air
2, Category II as 0.5 < PMV < 0.5, Category III as 0.7 < PMV < 0.7, system regulated the supply air temperature around 16–18 °C,
and Category IV as PMV < 0.7 or PMV > 0.7 [42]. The standard cat- resulting the room air temperature varying with a broad range
egories also relate to the expected percentage of thermally dissat- from 21 °C to 26 °C to meet the subject neutral comfort. The mean
isfied occupants and the allowable range of variable operative temperature value of the room air was calculated at 23.5 °C, well
temperature. agreeing with the design guideline of comfort air-conditioning
Even though air-conditioning systems allow to schedule the for office buildings in Southeast Asian countries [40,42,43]. As
indoor condition set-points (e.g. room temperature and humidity, expected for the convective air-conditioning, the room mean radi-
operative temperature, etc.); the system operation behaviors could ant temperature was higher than that of the room air. It was
lead to the controlled indoor conditions deviating from the set- remarked that for the tropical office buildings of which the envel-
points [12]. Further, the deviation was also caused by the configu- ope is non-insulated, the mean radiant temperature could be as
rations of the air-conditioned buildings. In some studies [12,13], high as 27–30 °C for more than 80 % of the office hours.
the EN15251 comfort categories were adopted as the criterion to The distribution plot shows a broad variation of the room oper-
assess the perceived indoor comfort from the radiant system with ative temperature: a range of 25–27 °C shared about 70 % of the
respect to the convective mixed-air system. The assessments were office hours, and a range of 23–25 °C shared another 20 %. Based
made based on the percentage of occurrence of the individual com- on the thermal comfort classification in Table 8, The subjects, at
fort category (or the PMV range) during the air-conditioned period. the center of the office space (Position I), would feel neutral (Class
By this assessment means, it was postulated that the thermal A) around 40 % of the time. The large variation in the.
response of subjects should be identical for the two distinct subject thermal sensation is obvious. The large difference
systems. between the mean radiant temperature and the room air temper-
In our prior study [26], the thermal comfort of Thai people in ature by up to 4.0 °C could be remarked as another potential factor
the tropical climate was experimentally evaluated under the two in the thermal discomfort.
environments of the radiant and the convective cooling. The anal- (b) Radiant ceiling scheme I (R1-B1)
ysis results reported the difference in their thermal responses as Scheme I present the cooling of the radiant system as illustrated
shown in Fig. 10(a). The subjects met the neutral sensation by in Fig. 9 (a). The distribution plot in Fig. 11(b): Scheme I shows that
the radiant ceiling cooling at a lower operative temperature value the chilled ceiling cooled the space by varying its surface temper-
than the convective cooling. In addition, their thermal response in ature from 20 °C to 24 °C, resulting in the room air could be main-
the radiant environment was less sensitive than in the convective tained at 24–26 °C for most of the time (more than 80 %). The
cooling case. Concerning the sensation difference, this study corresponding mean radiant temperature and the operative tem-
adopted the thermal sensation vote (TSV) relationship in Fig. 10 perature ranged from 23 °C to 27 °C, and from 24 °C to 26 °C,
(a) and the comfort classification in Table 8 as the criterion for respectively. Both the mean radiant temperature and the operative
comparing the radiant ceiling system with the mixed-air system. temperature varied within a rather narrow range close to the room
Analogous to the EN15251 standard, three comfort classes were air temperature. The radiant cooling provided more precise control
arranged where Class A corresponded with 0.5 < TSV < 0.5, Class of the indoor conditions than the mixed-air system. It is obvious
B with 0.7 < TSV < 0.7, and Class C with TSV < 0.7 or TSV > 0.7. In that the radiantly cooled space had higher indoor air temperature
Table 8, the corresponding operative temperature range of the than the space with the mixed-air system. The opposite was true
individual class are given for the two systems. for the mean radiant temperature.
In order to assess the indoor comfort, this study made the cal- The distribution plot of Scheme I shows that the air humidity in
culation at the center of the building floor (Point I in Fig. 10(b)). the space could be well controlled by the OAU at the dewpoint
Moreover, as the influence of the building envelope on the thermal temperature of 15–17 °C, enough to prevent the condensation on
comfort was considered, another four points (Points II-IV) in the the radiant ceiling. In the comfort viewpoint, the thermal sensation
perimeter area measured two meters apart from the walls were vote distribution indicated the local Thai subjects could meet their
included in the assessment. thermal neutrality (Class A) as high as 95 % of the office hours. The
comfort achievement by radiant cooling was more substantial than
the conventional mixed-air case.
4. Results and discussion (c) Radiant ceiling scheme II (R1-B2)
Scheme II assessed the radiant system to condition the office
The four radiant cooling schemes were simulated using the building with the wall insulation and better window shading.
annual weather records of Bangkok. Although the building model The distribution plot in Fig. 11(c) shows that the radiant ceiling
comprised fifteen stories, the simulations were made only for three could operate at a higher surface temperature than Scheme I. The
thermal zones: the ground floor, the middle floor, and the top floor. air temperature in the radiant room could be more precisely con-
Here, the middle floor was the representative zone of the second trolled. This would result from the reduced amount of heat gain
floor up to the fourteenth floor each of which was assumed to be through the insulated walls and solar transmission from the
identical in the thermal environment and the boundary conditions. shaded window.
13
A. Aryal, P. Chaiwiwatworakul and S. Chirarattananon Energy & Buildings 283 (2023) 112849

Fig. 11. Indoor conditions of mixed air system as compared to schemes of radiant system.

Scheme II had the room mean radiant temperature quite com- the interior room surfaces. The resultant operative temperature
parable to Scheme I. As the radiant ceiling could operate at a higher could be well maintained due to the smaller variation of the room
temperature in Scheme II; this implied the reduced temperature of air temperature. Finally, the subject comfort level of Scheme II was
14
A. Aryal, P. Chaiwiwatworakul and S. Chirarattananon Energy & Buildings 283 (2023) 112849

superior to Scheme I. The simulation results of Scheme II demon- the sensible heat was 47.7 kWhth/m2/year (51.7 %), and the latent
strated that the improved envelope could enhance the operation heat was 44.7 kWhth/m2/year (48.3 %). The large proportion of the
of the radiant ceiling system and better control the thermal indoor latent to the total OAUs load was due to the dehumidification of
conditions. the outdoor air to treat the whole system latent load.
(d) Radiant ceiling scheme III (R2-B1) The simulated AHUs load was 160.7 kWhth/m2/year, constitut-
This scheme simulated the indoor conditions of the radiant sys- ing 90 % of the system sensible heat and 10 % of the system latent
tem where the OAU was equipped with wrapped around coil and heat. The results pointed out that the typical OAUs operation that
heat recovery unit to enhance the dehumidification and the system cooled the outdoor air to 13 °C was not sufficient to treat the whole
thermal performance. In this scheme, the improved radiant system system latent load.
operated to condition the typical office with non-insulated envel- For the radiant cooling Scheme I, the total system load was
ope, so the simulated results were logical to be compared with that 246.1 kWhth/m2/year. The OAUs were responsible for 122.6 kWhth/
of Scheme I. m2/year or 49.8 % of the total load. The OAUs’ sensible and latent
In Fig. 11(d), the distribution plot of Scheme III shows that the components were 66.6 kWhth/m2/year and 56.0 kWhth/m2/year,
radiant ceiling had a surface temperature lower than that of respectively. Thus, the OAUs of the radiant system took 20.3 % lar-
Scheme I. This consequence was from the increase of the room ger latent load than the OAUs of the mixed-air system. The simu-
air temperature when the outdoor air was reheated by wrapped lation shows that the radiant ceiling had a sensible load of 123.5
around coil before it was supplied into the space. The lower ceiling kWhth/m2/year. For this radiant system configuration, the load
surface temperature led the mean radiant temperature to margin- sharing of the chilled ceiling and the OAUs were comparable.
ally decrease to compensate the higher room air temperature. For Scheme II where the building walls were insulated, the
However, the compensation was not enough to maintain the oper- reduced sensible heat decreased the radiant ceiling load by
ative temperature as low as that of Scheme I. Thus, the time per- 23.7 %, the OAUs load by 0.68 %, and the system load by 14.6 %,
centage to meet the thermal neutrality of Scheme III was slightly as compared to Scheme I. The major load the radiant ceiling was
lower. The humidity level in the room was not different from other from the envelope heat gain. The proportion of the radiant ceiling
schemes. to the OAUs load was now modified to 43.8: 56.2.
(e) Radiant ceiling scheme IV (R2-B2) For Scheme III, the OAUs of the radiant system were equipped
As compared to Scheme III, the chilled ceiling of Scheme IV with a wrapped around coil and a heat recovery unit. Fig. 12 shows
could raise its surface temperature as the interior wall temperature that the integration of the two devices reduced the system load by
decreased from the improved building envelope performance by 8.1 % and the OAUs load by 40 % with respect to that of Scheme I.
insulation and sun shading. Both the room air temperature and As the wrapped around coil reheated the OAU supply air, the
mean radiant temperature of Scheme IV were lower than that of higher air temperature magnified the sensible load of the radiant
Scheme III. As a result, the thermal comfort condition of the radiant chilled ceiling. The proportion of the radiant ceiling load to OAUs
room was improved as can be seen in Fig. 11(e). load was 67.6: 32.4 for this radiant system configuration.
For the last scheme (Scheme IV), the influence of the reduced
4.1.1. Cooling load envelope heat gain on the radiant system was similar to that in
Based on the perceived thermal comfort reported in the previ- Scheme II. The system load of Scheme IV was 22.3 % lower than
ous section, the cooling loads of the four radiant cooling schemes that of Scheme I.
were compared with that of the reference case: the mixed-air sys-
tem. The loads segregated by the sensible and the latent compo- 4.1.2. Energy consumption
nents of each system device are given in Fig. 12. As shown, the Energy consumption of the two air-conditioning systems was
loads are presented in unit of kWhth/m2/year. assessed from the major machines i.e. chillers, chilled water
The mixed-air system possessed the largest building cooling pumps, and fan motors. The chiller power was calculated using
load of 253.0 kWhth/m2/year of which the AHUs and the OAUs the generic air-cooled chiller model expressed as a function of
were responsible for 63.5 % and 36.5 %, respectively. At the OAUs, the chiller’s load, chilled water supply temperature, and ambient

Fig. 12. Cooling loads of different air-conditioning schemes.

15
A. Aryal, P. Chaiwiwatworakul and S. Chirarattananon Energy & Buildings 283 (2023) 112849

Fig. 13. Electricity consumption of different air-conditioning schemes.

temperature [7,11]. The pump and fan powers were derived from tial of integrating the heat recovery unit and wrapped around coil
the motor performance models [7] based on the fluid flow rate with the OAU of the radiant system was evaluated by Scheme III.
and the pressure drop of the designed piping and air duct The integration of the devices could reduce the system energy by
networks. 8.6 % compared to that of Scheme I. Since the system load was also
For the mixed-air system, the chilled water was supplied at 7 °C modified, the energy of the OAUs chiller decreased, but that of the
to the AHUs and OAUs. With the cooling load from the TRNSYS radiant ceiling chiller increased with a smaller amount.
simulations, the total system energy was calculated at 108.0 Scheme IV presents the energy consumption of the radiant sys-
kWhelec/m2/year (Fig. 13), which well agreed with that of the sur- tem with the improved OAUs and building envelope. As the system
veyed energy auditing report [37]. The chiller presented the largest cooling load of Scheme IV was reduced by around 22.3 % of that in
energy consumer of 83.4 kWhelec/m2/year. The energy consump- Schemes I and II, the radiant system energy was 42 % less com-
tions of the OAU and AHU fan motors, and of the chilled water pared to the reference case and 20.9 % less with respect to
pump were 18.9 kWhelec/m2/year and 5.64 kWhelec/m2/year, Scheme I.
respectively. The percentage share of the energy consumption
was 77.2 % for the chiller, 17.5 % for fans, and 5.3 % for pumps. 5. Conclusion
Distinct from the mixed-air system, the radiant system utilized
the chilled water at two temperature levels of 7 °C for the OAUs to This experimental and simulation study verified that the radi-
treat latent load and 17 °C for the radiant ceiling to treat sensible ant chilled ceiling, as compared to the mixed air system, provided
load. The production of the higher temperature chilled water could the better indoor comfort and conserved the electrical energy for
improve the chiller and the whole system’s performance. Fig. 13 office cooling in tropical hot humid climate. Key findings from
presents the energy consumption of the machines in the radiant the study could be summarized below:
system. The chiller energy was separated into the one of 7 °C
chilled water and the other one of 17 °C chilled water. Although  Through the experiments, the radiant ceiling of suspended
the radiant cooling Scheme I had a comparable cooling load with metal panels demonstrated the sufficient heat extraction rate
the mixed-air system, it consumed the total energy of 79.4 with fast thermal response to treat the varying sensible load
kWhelec/m2/year, 27.0 % less than the mixed-air system. The per- of the office space with good insulation and sun shading. The
formance improvement of the chilled water production resulted intermittent operation of the radiant ceiling coupled with
in the chiller energy being reduced by 15.5 %. However, significant OAU by controlling the operative temperature could achieve
energy saving was achieved by reducing the fan power which was the subject neutral comfort during the office hours.
about 89.3 % in our simulation case. It is observed that the pump  Based on the localized TSV comfort indicator, the simulations
energy of the radiant system was increased by 21.7 % in reference showed that even though the mixed-air system precisely con-
to the mixed-air system, as the temperature difference of chilled trolled the room air temperature, the neutral comfort could be
water supply and return for the radiant ceiling was just 2–3 °C, met below 40 % of the office hours due to the high, and large-
smaller than that of the mixed-air system. varying mean radiant temperature from the poor performance
Considering Scheme II, the energy consumption of the radiant of uninsulated envelope characterized by the typical office
system was further reduced by the performance improvement of buildings.
the building envelope by 9.1 % compared to Scheme I and by  As the radiant chilled ceiling effectively controlled and reduced
33.2 % compared to the reference case. Noted that the majority the room’s mean radiant temperature, the building occupants
of energy-saving was contributed by the reduced energy from could achieve their comfort neutrality as high as 95 % of the
the chiller of the radiant ceiling cooling. The energy-saving poten- office hours.

16
A. Aryal, P. Chaiwiwatworakul and S. Chirarattananon Energy & Buildings 283 (2023) 112849

 Although the radiant system essentially met the thermal neu- Declaration of Competing Interest
trality at about 1 °C lower operative temperature than the
mixed-air system; wall insulation and window sun-shading The authors declare that they have no known competing finan-
were the key measures to lower the envelope heat gain and cial interests or personal relationships that could have appeared
the energy consumption of the radiant system. to influence the work reported in this paper.
 At the OAU, the wrapped around coil could be equipped with
the chilled water coil to enhance the load decoupling, and the
heat recovery unit could be used to reduce the load of the out-
door air. Accompanied by the high performance building envel- Acknowledgement
ope, the radiant ceiling cooling could save energy as large as
40 % as opposed to the mixed-air system. The financial support from the National Science and Technology
Development Agency (NSTDA), grant no. P-16-51993 through this
research project is gratefully acknowledged.

Data availability
Appendix
The authors are unable or have chosen not to specify which data has
been used.

Fig. A1. Solar radiation and ambient air during the OAU experiment.

17
A. Aryal, P. Chaiwiwatworakul and S. Chirarattananon Energy & Buildings 283 (2023) 112849

(a) Solar radiation and ambient condition

(b) Incident solar radiation on the experimental room walls

Fig. A2. Outdoor conditions of the radiative cooling experiment.

References [7] H. Lim, J.W. Jeong, Energy saving potential of thermoelectric radiant cooling
panels with a dedicated outdoor air system, Energy Build. 169 (2018) 353–365,
https://doi.org/10.1016/j.enbuild.2018.03.062.
[1] S. Chirarattananon, P. Chaiwiwatworakul, V.D. Hien, P. Rakkwamsuk, K.
[8] ASHRAE. ANSI/ASHRAE Standard 55-2017, Thermal Environmental Conditions
Kubaha, Assessment of energy savings from the revised building energy code
for Human Occupancy, Atlanta: American Society of Heating Refrigerating and
of Thailand, Energy 35 (2010) 1741–1753, https://doi.org/10.1016/j.
Air-Conditioning Engineers Inc., 2017.
energy.2009.12.027.
[9] Y. Khan, M. Bhandari, J. Mathur, Energy-saving potential of a radiant cooling
[2] P. Vangtook, S. Chirarattananon, Application of radiant cooling as a passive
system in different climate zones of India, Sci. Technol. Built Environ. 24 (2018)
cooling option in hot humid climate, Build. Environ. 42 (2007) 543–556,
356–370, https://doi.org/10.1080/23744731.2017.1348140.
https://doi.org/10.1016/j.buildenv.2005.09.014.
[10] G.P. Henze, C. Felsmann, D.E. Kalz, S. Herkel, Primary energy and comfort
[3] B. Junasová, M. Krajčík, O. Šikula, M. Arıcı, M. Šimko, Adapting the construction
performance of ventilation assisted thermo-active building systems in
of radiant heating and cooling systems for building retrofit, Energy Build. 268
continental climates, Energy Build. 40 (2008) 99–111, https://doi.org/
(2022), https://doi.org/10.1016/j.enbuild.2022.112228.
10.1016/j.enbuild.2007.01.014.
[4] C. Stetiu, Energy and peak power savings potential of radiant cooling systems
[11] J.W. Jeong, S.A. Mumma, W.P. Bahnfleth, Energy conservation benefits of a
in US commercial buildings, Energy Build. 30 (1999) 127–138, https://doi.org/
dedicated outdoor air system with parallel sensible cooling by ceiling radiant
10.1016/S0378-7788(98)00080-2.
panels, ASHRAE Trans. 109 PART 2 (2003) 627–636.
[5] X. Hao, G. Zhang, Y. Chen, S. Zou, D.J. Moschandreas, A combined system of
[12] F.C.& M.F. Enrico Fabrizio , Stefano P. Corgnati, To, Numerical comparison
chilled ceiling, displacement ventilation and desiccant dehumidification, Build.
between energy and comfort performances of radiant heating and cooling
Environ. 42 (2007) 3298–3308, https://doi.org/10.1016/j.
systems versus air systems, HVAC&R Res. ISSN. 18 (2008) 692–708. 10.1080/
buildenv.2006.08.020.
10789669.2011.578700.
[6] Y. Khan, V.R. Khare, J. Mathur, M. Bhandari, Performance evaluation of radiant
[13] G. Salvalai, J. Pfafferott, M.M. Sesana, Assessing energy and thermal comfort of
cooling system integrated with air system under different operational
different low-energy cooling concepts for non-residential buildings, Energy
strategies, Energy Build. 97 (2015) 118–128, https://doi.org/10.1016/j.
Convers. Manag. 76 (2013) 332–341, https://doi.org/10.1016/j.
enbuild.2015.03.030.
enconman.2013.07.064.

18
A. Aryal, P. Chaiwiwatworakul and S. Chirarattananon Energy & Buildings 283 (2023) 112849

[14] S. Oxizidis, A.M. Papadopoulos, Performance of radiant cooling surfaces with [29] A. Koca, Y. Karakoyun, O. Acikgoz, M. Dogu, A.S. Dalkilic, An experimental
respect to energy consumption and thermal comfort, Energy Build. 57 (2013) investigation on the radiant cooled wall’s heat transfer characteristics in a
199–209, https://doi.org/10.1016/j.enbuild.2012.10.047. fully conditioned real-sized living environment, Energy Build. 277 (2022),
[15] T. Imanari, T. Omori, K. Bogaki, Thermal comfort and energy consumption of https://doi.org/10.1016/j.enbuild.2022.112578.
the radiant ceiling panel system, Energy Build. (1999), https://doi.org/10.1016/ [30] A. Koca, Experimental examination of heat transfer coefficients in hydronic
S0378-7788(98)00084-X. radiant wall cooling systems, J. Build. Eng. 60 (2022), https://doi.org/10.1016/j.
[16] W. Liao, J. Peng, Y. Luo, Y. He, N. Li, Y. a,, Comparative study on energy jobe.2022.105209.
consumption and indoor thermal environment between convective air- [31] ASHRAE. ASHRAE Guideline 14-2014, Measurement of Energy and Demand
conditioning terminals and radiant ceiling terminals, Build. Environ. 209 Savings; American Society of Heating, Ventilating, and Air Conditioning
(2022), https://doi.org/10.1016/j.buildenv.2021.108661. Engineers, Inc.: Atlanta, GA, USA, 2014.
[17] N. Ismail, D. Ouahrani, Modelling of cooling radiant cubicle for an office room [32] V.G. González, G.R. Ruiz, C.F. Bandera, Empirical and comparative validation
to test cooling performance, thermal comfort and energy savings in hot for a building energy model calibration methodology, Sensors (Switzerland).
climates, Energy 244 (2022), https://doi.org/10.1016/j.energy.2022.123185. 20 (2020) 1–29, https://doi.org/10.3390/s20175003.
[18] ISO, International Standard 7730, Moderate Thermal environment: [33] S. Chirarattananon, V.D. Hien, P. Tummu, Thermal performance and cost
Determina- tion of PMV and PPD Indices and Specification of the Conditions effectiveness of wall insulation under Thai climate, Energy Build. 45 (2012)
for Thermal Environment, International Organization for Standardization, 82–90, https://doi.org/10.1016/j.enbuild.2011.10.025.
Geneva, 2005. [34] K. Lohwanitchai, D. Jareemit, Modeling energy efficiency performance and
[19] B. Ning, Y. Chen, H. Jia, Cooling load dynamics and simplified calculation cost-benefit analysis achieving net-zero energy building design: case studies
method for radiant ceiling panel and dedicated outdoor air system, Energy of three representative offices in thailand, Sustainability 13 (2021), https://doi.
Build. 207 (2020), https://doi.org/10.1016/j.enbuild.2019.109631. org/10.3390/su13095201.
[20] J. Feng, S. Schiavon, F. Bauman, Cooling load differences between radiant and [35] N. Inklab, Technology assessment of dedicated outdoor air systems for air-
air systems, Energy Build. 65 (2013) 310–321, https://doi.org/10.1016/j. conditioned buildings in tropical climate, The Joint Graduate School of Energy
enbuild.2013.06.009. and Environment, King Mongkut’s University of Technology Thonburi., 2014.
[21] R. Talami, J.A. Jakubiec, Early-design sensitivity of radiant cooled office [36] A.M. Fathoni, P. Chaiwiwatworakul, V. Mettanant, Energy analysis of the
buildings in the tropics for building performance, Energy Build. 223 (2020), daylighting from a double-pane glazed window with enclosed horizontal slats
https://doi.org/10.1016/j.enbuild.2020.110177. in the tropics, Energy Build. 128 (2016) 413–430, https://doi.org/10.1016/j.
[22] S. Jordan, J. Hafner, T.E. Kuhn, A. Legat, M. Zbašnik-Senegačnik, Evaluation of enbuild.2016.06.034.
various retrofitting concepts of building envelope for offices equipped with [37] S. Chirarattananon, Building for Energy Efficiency, Energy Conservation
large radiant ceiling panels by dynamic simulations, Sustainability 7 (2015) Promotion Fund (ENCON Fund), 2005.
13169–13191, https://doi.org/10.3390/su71013169. [38] J. Woolley, S. Schiavon, F. Bauman, P. Raftery, J. Pantelic, Side-by-side
[23] Q.J. Kwong, S.J. Kho, J. Abdullah, V.R. Raghavan, Evaluation of energy laboratory comparison of space heat extraction rates and thermal energy
conservation potential and complete cost-benefit analysis of the slab- use for radiant and all-air systems, Energy Build. 176 (2018) 139–150, https://
integrated radiant cooling system: a Malaysian case study, Energy Build. 138 doi.org/10.1016/j.enbuild.2018.06.018.
(2017) 165–174, https://doi.org/10.1016/j.enbuild.2016.12.014. [39] T.P. Mcdowell, S.J. Emmerich, FOR DIFFERENT CLIMATES Thermal Energy
[24] J.L. Niu, L.Z. Zhang, H.G. Zuo, Energy savings potential of chilled-ceiling System Specialists LLC, Madison, Wisconsin, USA, 2014.
combined with desiccant cooling in hot and humid climates, Energy Build. 34 [40] N. Inklab, P. Chaiwiwatworakul, S. Chuangchote, P. Rakkwamsuk, S.
(2002) 487–495, https://doi.org/10.1016/S0378-7788(01)00132-3. Chirarattananon, Performance assessment of dedicated outdoorair systems
[25] G. Sastry, P. Rumsey, VAV vs. Radiant, ASHRAE J. 56 (2014) 17–24. for office building in Thailand, Elsevier B.V. (2015), https://doi.org/10.1016/
[26] A. Aryal, P. Chaiwiwatworakul, S. Chirarattananon, W. Wongsuwan, Subjective j.egypro.2015.11.555.
assessment of thermal comfort by radiant cooling in a tropical hot humid [41] P.J. Martínez, P. Martínez, V.M. Soto, J. Rodriguez, L.A. Bujedo, Analysis of a
climate, Energy Build. 254 (2021), https://doi.org/10.1016/j. DOAS operation in different Spanish climates using an experimentally
enbuild.2021.111601. validated TRNSYS model, Int. J. Vent. 19 (2020) 97–111, https://doi.org/
[27] ASHRAE, ANSI/ASHRAE Standard 62.1-2015, Ventilation for acceptable indoor 10.1080/14733315.2019.1579485.
air quality, 2015, American Society of Heating, Refrigerating, and Air- [42] R.J. de Dear, K.G. Leow, S.C. Foo, Thermal comfort in the humid tropics: Field
Conditioning Engineers, Inc., Atlanta, GA. experiments in air conditioned and naturally ventilated buildings in
[28] Y. Hu, X. Xia, J. Wang, Research on operation strategy of radiant cooling system Singapore, Int. J. Bio Meteorol. 34 (1991) 259–265.
based on intermittent operation characteristics, J. Build. Eng. 45 (2022), [43] S.A. Damiati, S.A. Zaki, S. Wonorahardjo, N.H. Wong, H.B. Rijal, Comfort
https://doi.org/10.1016/j.jobe.2021.103483. temperature in air conditioned office buildings: case study of Indonesia and
Singapore, Malaysia-Japan Jt. Int. Conf. 2015 (2015) 1–6.

19