A Low-Cost System for Real Time Monitoring and

Assessment of Potable Water Quality at Consumer

Sites
Theofanis P. Lambrou and Christos G. Panayiotou Christos C. Anastasiou
KIOS Research Center, Dept. of Electrical and Computer Engineering Dept. of Civil Engineering
University of Cyprus Frederick University
Nicosia, Cyprus. Nicosia, Cyprus.
Email: {faniseng,christosp}@ucy.ac.cy Email: [email protected]

Abstract— This paper presents the design and development of can take frequent samples of the water quality at a very
a low cost system for real time monitoring of drinking water limited number of locations. However, substantial proportion
quality at consumer sites. The system consists of several in-pipe of contamination problems is attributable to problems within
electrochemical and optical sensors and emphasis is given on low
cost, lightweight implementation and reliable long time operation. distribution systems and due to the limited spatio-temporal
Such implementation is suitable for large deployments enabling sampling, it is impossible for the water companies and con-
a sensor network approach for providing spatiotemporally rich sumers to know the quality of potable water delivered to
data to water consumers, water companies and authorities. consumer households.
Extensive literature and market research is performed to identify Given the size of the water distribution networks (pipe
low cost, on-line sensors that can reliably monitor several
parameters which can be used to infer the water quality. Based length) and the number of households served, we consider
on selected parameters a sensor array is developed along with important that the spatio-temporal sampling is significantly
several microsystems for analog signal conditioning, processing, increased, thus it is necessary to collect water quality samples
logging, and remote presentation of data. Finally, an algorithm at significantly more locations (if possible at all consumer
for fusing on-line multi sensor measurements is developed to sites). Therefore, the main challenge is to develop reliable and
assess the water contamination risk.
low-cost sensing systems for accurate and continuous in-pipe
water quality monitoring. The system developed is intended
I. I NTRODUCTION
to be used as a part (node) of a low cost water sensor network
Clean drinking water is a critical resource, important for to provide water quality information to consumers, water
the health and well-being of all humans. Several experimental companies and authorities. The spatio-temporal data provided
studies [1], [2] indicate the need for continuous on-line water by such network can support complex decisions concerning
quality monitoring with efficient spatio-temporal resolution the quality of drinking water, including the detection of the
and demonstrate that the conventional reagent-based water location/source of hazardous agents and pathogens, raising
quality methods fail to satisfy this requirement due to higher awareness, and encouraging better water-handling and man-
labor and operational cost. US Environmental Protection agement.
Agency has carried out an extensive experimental evaluation Therefore, the contribution of this paper is twofold. First,
[3] of water quality sensors to assess their performance on sev- it proposes the need for a shift in the current monitoring
eral contaminations. The main conclusion was that many of the paradigm and propose the idea of monitoring the quality of
chemical and biological contaminants used have an effect on water delivered to consumers, using low cost, low performance
many water parameters monitored including Turbidity (TU), and tiny sensors. We argue that this approach can achieve
Oxidation Reduction Potential (ORP), Electrical Conductivity more reliable quality monitoring due to the large spatially
(EC) and pH. Thus, it is feasible to monitor and infer the water distributed deployment and the possibility of correlating the
quality by detecting changes in such parameters. quality measurements from various consumers. Second, it
A limited number of on-line, reagent-free water monitor- presents the first step towards this goal which is the design
ing systems exist (e.g. Hach HST GuardianBlue [4], JMAR and development of a low cost system that can be used
BioSentry [5]), but these systems are bulky (sensors are at the premises of every consumer. The embedded systems
installed in flow cells located in cabinets) and remain cost developed can also be used in a consumer-oriented manner
prohibitive for large scale deployments (cost tens of thousands to continuously monitor qualitative water parameters and fuse
of dollars per unit). It is worth mentioning that cost is multi-parametric sensor response in order to assess the water
mostly attributed not to sensing probes but to instrumentation- consumption risk at consumer level, locally and independently
automation controllers (analyzers) and panels. Such systems from other consumer measurements.
 The remaining of this paper is organized as follows. Sec- which can consume as high as 50% of the operational budgets.
tion II presents the methodology and justification for the selec- Conventionally, ultrasonic, brush, water-jet, or chemical type
tion of water quality parameters to be monitored. Section III of automatic cleaners [15] are used to remove coatings from
presents the experimental implementation of the hardware the sensor probes. Recently, several alternative cost effective
and software modules and validates the performance of the methods have been proposed that can either actively remove
developed system. Finally the paper ends with the conclusion. fouling (e.g. electrolysis) or passively prevent fouling. Flat
measuring surface probe method [16] is the most cost effective,
II. M ETHODS passive self-cleaning method and is based on the mechanical
Drinking water quality standards are determined according package and design of the probe. When the electrode’s flat
to World Health Organization [6] guidelines for drinking-water measuring surface is exposed to turbulent flow, the resulting
quality as well as other pertinent organizations (i.e. EU [7], scrubbing action provides a self-cleaning effect in most appli-
USEPA [8]). These organizations set the standards for drinking cations under medium range flows. The flat sensing surface
water quality parameters and indicate which microbiological, virtually eliminates deposits that can foul the electrode and
chemical and indicator parameters must be monitored and significantly reduces necessary maintenance. This simple, but
tested regularly in order to protect the health of the consumers effective method has no moving parts, requires no power and
and to make sure the water is wholesome and clean. also prolongs electrode life and eliminates breakage.
The selection of the physicochemical parameters to be
monitored was based on extensive scientific literature review III. R ESULTS
[3], [9], [10], [11], [12], [13] on the relation between cer- A modular but holistic approach is adopted for the design
tain physicochemical parameters and chemical or biological and development of the system. Modularity enables the si-
contaminations that present in water. Table I enumerates the multaneous sampling of all parameters and the decomposition
suggested parameters to be monitored from high to low cor- of several operations like calibration, upgrades and repair
relation significance when interpreting water contaminations or replacement of faulty parts. Holisticity focuses on the
(assess hazard). Table I also presents the measurement cost (for interdependence of all parts (interfaces and communication
purchase and maintenance) associated with these parameters protocols) to pursue the objectives of low cost and assessment
based on recent review [14] of measurement and instrumen- of drinking water quality.
tation methods, compensation and calibration procedures and
probe lifetime concerning these parameters. Thus, the param- ZigBee
USB
eters selected to monitored are the following: 1) Turbidity, 2) Module
Mux/
Demux
Oxidation Reduction Potential (instead of Free Chlorine), 3)
Temperature 4) pH, and 5) Electrical Conductivity. It is noted
USB SPI Rx Tx Rx Tx Rx Tx Rx Tx Rx Tx
that Free Chlorine concentration can be estimated as a function pH ORP TU
Central Measuring Node EC
of the ORP, pH and temperature measurements. Nitrates, PIC32 MCU MCU MCU MCU MCU

though considered as an important parameter for human health RTCC ADC

Analog Signal
ADC
Analog
ADC
Analog
Signal
ADC
Analog
Signal
ADC
Analog
Signal
Unit Signal
is not selected because measurement methods are subjected to Unit Unit Unit Unit

failures (Ion-Selective Electrodes) or are cost prohibitive (UV

Spectrophotometric Method). Finally, dissolved oxygen is not
Conductivity pH ORP Turbidity Flow
selected due to several compensations and frequent membrane Temp

replacement needed for accurate measurements.

.
Parameter Units Quality Range Meas. Cost (a) Central measurement node architecture.
RJ45
1 Turbidity NTU 0–5 Medium Ethernet
USB
ZigBee
LCD
Module Module
2 Free Residual Chlorine mg/L 0.2 – 2 High
3 ORP mV 650 – 800 Low
4 Nitrates mg/L <10 High Router
5 Temperature oC – Low
6 pH pH 6.5 – 8.5 Low Control Node SPI Digital Ports
7 Electrical Conductivity µS/cm 500 – 1000 Low ARM MCU
Notification Node
8 Dissolved Oxygen mg/L – Medium WEB Linux OS
sms
PIC MCU
www SQL
email
TABLE I
S UGGESTED PARAMETERS TO BE M ONITORED . Q UALITY RANGE IS (b) Control node architecture. (c) Notification node
architecture.
SUGGESTED BY WHO [6] GUIDELINES AND EU STANDARDS [7].
Fig. 1. System architecture.

In line water sensors illustrate the need for efficient and The overall system architecture under discussion in pre-
periodic probe cleaning to maintain reliable measurements. sented in Fig. 1 and is comprised of following three subsys-
Cleaning mechanisms constitute an important cost parameter tems: a central measurement node (PIC32 MCU based board)
 Parameter Measurement principle Units Range Resolution Accuracy Quality Range
Turbidity Optical/infrared backscattering NTU 0 – 100 0.1 ±0.5 0–5
ORP Galvanic cell, platinum electrode mV -2000 – 2000 2 ±10 600 – 800
pH Galvanic cell, glass electrode pH 0 – 14 0.05 ±0.1 6.5 – 8.5
Conductivity Conductive cell µS/cm 100 – 20000 10 5% 500 – 1000
Temperature NTC resistance oC -5 – 70 0.1 ±0.5 –
TABLE II
S PECIFICATIONS AND ACCOMPLISHED PERFORMANCE FOR EACH MONITORED PARAMETER .

that collects water quality measurements from sensors, imple- sensors have flat measuring surfaces for cost effective self-
ments the algorithm to assess water quality and transmits data cleaning. Considerable attention is given to the analog signal
to other nodes, an optional control node (ARM/Linux web- conditioning circuits, calibration and compensation procedures
server based platform) that stores measurement data received to reduce noise and attain high resolution and accuracy.
from the central measurement node in a local database and
provides gateway to the internet, visualize data (charts), and
sends email/sms alerts and finally a tiny notification node(s)
(PIC MCU based board) that receives information from the
central measurement node through an interconnected ZigBee
RF transceiver and provides local near-tap notifications to
the user (water consumer) via several interfaced peripherals
(LED, LCD, Buzzer). A photo of the central measurement,

Fig. 4. Complete System Photo.

Cf

Rf Rf
EC Probe

Fig. 2. Photos of system’s nodes. Rl

2 R 2
TU Probe
1 1
LED D Vout Vout
control and notification node is presented in Fig. 2 while 3
Vexc
3

Fig.3 presents the multi-parameter sensor array comprised of

(a) Turbidity preamplifier (b) Conductivity preamplifier
the Turbidity(TU), ORP, pH, Electrical Conductivity(EC) and
Precision current
Temperature probes mounted in a plastic pipe. The complete source <1mA

system photo is shown in Fig 4. Turbidity sensor is constructed R1 R2 R1 R2

Vout

2 2 Temp Probe
ORP Probe pH Probe NTC or RTD
1 1
Vout Vout
R 3 R 3

(c) ORP preamplifier (d) pH preamplifier (e) Temperature

Fig. 5. The first stage of analog signal conditioning circuitry for each
parameter
(a) In-pipe multi-parameter sensor (b) Probes with flat sensing surfaces.
array. A dedicated PIC based microsystem is developed for each
Fig. 3. Sensor photos. TU, ORP, pH have flat sensing surfaces for cost parameter to accomplish this task. The first stage of analog
effective self-cleaning. signal conditioning circuitry for each parameter is presented
in Fig. 5 while Table II shows the results regarding labora-
from scratch based on our previous work [17] whether the tory evaluation (using standard buffer solutions and reference
other sensor probes obtained from SensoreX Corpr . Electrical instruments) of each parameter along with the quality range
Conductivity sensor embeds an NTC thermistor which is suggested by WHO guidelines and EU standards. The com-
used for temperature sensing and temperature compensation ponents for the complete system prototype cost approximately
of conductivity and pH measurements. Turbidity, ORP, pH e500 which is at least an order of magnitude less expensive
than commercially available multi-parameter instruments. IV. C ONCLUSION
The software platform developed for the control node is In this paper, the design and development of a low cost
illustrated in Fig. 6. This platform enables real time measure- system for real time monitoring of drinking water quality
ment charts of monitored parameters, real time assessment at consumer sites is presented. The proposed system consist
of water quality and sensor calibration instructions through of several in-pipe water quality sensors with flat measuring
a Graphical User Interface (GUI). It also logs sensor data in probes and unlike commercially available on-line analyzers, it
a local database and posts data to web using Pachube open is low cost, lightweight and capable of processing, logging,
source web platform. and remote presentation of data. Such implementation is suit-
able for large deployments enabling a sensor network approach
for providing spatiotemporally rich data to water consumers,
water companies and authorities.
In the future, we plan to investigate the performance of the
fusion algorithm on intentional contamination events (biologi-
cal, chemical, etc) and install the system in several locations of
the water distribution network to collect spatiotemporally rich
water quality data and characterize system/sensors response in
real field deployments.
ACKNOWLEDGMENT
This work is partly supported by the Cyprus
Research Promotion Foundation under grants
Fig. 6. Software Platform. NEA ΥΠO∆/ΣTPAT/0308/26 and EΠIX/ΠPOION/0609/87
and co-funded by the Republic of Cyprus and the European
Finally, an algorithm for fusing on-line multi-sensor mea- Regional Development Fund.
surements to assess the water contamination risk is developed. R EFERENCES
This algorithm enables the system to act as an ”early warn-
[1] A. Aisopou, I. Stoianov, N. Graham, In-pipe water quality monitoring in
ing system” for possible potable water quality deterioration water supply systems under steady and unsteady state flow conditions: A
at homes. A flowchart of the risk assessment algorithm is quantitative assessment, Water Research, 46(1), pp. 235-246, 2012
illustrated in Fig. 7. The algorithm is implemented as follows: [2] S. Panguluri, G. Meiners, J. Hall, J.G. Szabo, Distribution System Water
Quality Monitoring: Sensor Technology Evaluation Methodology and
For every measured parameter i baseline data is estimated Results, U.S. Environmental Protection Agency, EPA/600/R-09/076, 2009
(mean µi and standard deviation σi ) over a moving time [3] J. Hall, A.D. Zaffiro, R.B. Marx, P.C. Kefauver, E.R. Krishnan, R.C.
window. Then, for each new parameter measurement Si , the Haught, and J.G. Herrmann, On-line Water Quality Parameters as Indi-
normalized sensor output Ni = Siσ−µ i cators of Distribution System Contamination, Journal AWWA, 2007
i
is estimated to filter [4] Hach HST, GuardianBlue Early Warning System Brochure.
baseline fluctuations. If normalized sensor outputs exhibit [5] JMAR, BioSentry Contamination Warning System Technical Overview.
sudden and significant changes from baseline data (given [6] World Health Organization Guidelines for drinking-water quality, fourth
edition, 2011
some predefined thresholds τi , currently under investigation) a [7] European Communities Drinking Water Regulations (No.2), 2007
trigger is activated and the water consumption risk is estimated [8] U.S. Environmental Protection Agency 2012 Edition of the Drinking
as a function of the euclidian distance ∥N − τ ∥. The signature Water Standards and Health Advisories, EPA 822-S-12-001
[9] C.N.Haas, M.A. Meyer, and M.S. Paller, Microbial alterations in water
of all normalized sensor outputs can be further processed to distribution systems and their relationship to physical-chemical charac-
minimize false alarms and identify possible contaminations, teristics, in Journal of American Water Works Association , 1983
given that a contamination library is available/developed. [10] K.N. Power and L.A. Nagy, Relationship between bacterial regrowth
and some physical and chemical parameters within Sydney’s drinking
water distribution system, in Water Research, 1999
[11] J. B. Cook, J. F. Byrne, R. C. Daamen and E. A. Roehl, Distribution
System Monitoring Research at Charleston Water System, in 8th Annual
Water Distribution Systems Analysis Symposium, 2006
[12] Y.J. Yang, R.C Haught and J.A Goodrich, Real-time contaminant de-
tection and classification in a drinking water pipe using conventional
water quality sensors: Techniques and experimental results, in Journal of
Environmental Management, 2009
[13] C.C. Anastasiou, P. Grafias, T.P. Lambrou, A. Kalli and C. Onisiphorou,
Use of holding tanks and their effect on the quality of potable water in
households; the case study of Cyprus, in PRE XI, 2012
[14] T.P. Lambrou, A Technical Report for Water Qualitative and Quantitative
Sensors, University of Cyprus, 2011
[15] A. Whelan and F.Regan, Antifouling strategies for marine and riverine
sensors, Journal of Environmental Monitoring, 2006
[16] SensoreX Corp Flat Surface Operating Principles.
[17] T.P. Lambrou, C.C. Anastasiou , C.G. Panayiotou, A Nephelometric
Turbidity System for Monitoring Residential Drinking Water Quality, in
Fig. 7. Risk Assessment Algorithm. Sensor Networks Applications, Experimentation and Logistics, 2009