ArticlePDF Available

Abstract and Figures

The rise of Internet of Things (IoT), coupled with the advances in Artificial Intelligence technologies and cloud-based applications have caused fundamental changes in the way societies behave. Enhanced connectivity and interactions between physical and cyber worlds create ‘smart’ solutions and applications to serve society needs. Water is a vital resource and its management is a critical issue. ICT achievements gradually deployed within the water industry provide an alternative, smart and novel way to improve water management efficiently. Contributing to this direction, we propose a unified framework for urban water management, exploiting state-of-the-art IoT solutions for remote telemetry and control of water consumption in combination with machine learning-based processes. SMART-WATER platform aims to foster water utility companies by enhancing water management and decision-making processes, provide innovative solutions to consumers for smart water utilisation.
Content may be subject to copyright.
Making urban water smart: the SMART-WATER solution
Gerasimos Antzoulatos, Christos Mourtzios, Panagiota Stournara,
Ioannis-Omiros Kouloglou, Nikolaos Papadimitriou, Dimitrios Spyrou,
Alexandros Mentes, Efstathios Nikolaidis, Anastasios Karakostas,
Dimitrios Kourtesis, Stefanos Vrochidis and Ioannis Kompatsiaris
The rise of Internet of Things (IoT), coupled with the advances in Articial Intelligence technologies and
cloud-based applications, have caused fundamental changes in the way societies behave. Enhanced
connectivity and interactions between physical and cyber worlds create smartsolutions and
applications to serve societys needs. Water is a vital resource and its management is a critical issue.
ICT achievements gradually deployed within the water industry provide an alternative, smart and novel
way to improve water management efciently. Contributing to this direction, we propose a unied
framework for urban water management, exploiting state-of-the-art IoT solutions for remote telemetry
and control of water consumption in combination with machine learning-based processes. The
SMART-WATER platform aims to foster water utility companies by enhancing water management and
decision-making processes, providing innovative solutions to consumers for smart water utilisation.
Key words |Internet of Things, LPWAN, predictive big data analytics, smart metering, telemetry,
water management
Smart Automated Metering and Remote Control design and deploy a novel inte-
grated infrastructure that encompasses telemetry and remote-control technologies
for the efcient management of the water supply network.
Device Connectivity and Data Management develop alternative technologies for the
implementation of a xed wireless network that will support real-time telemetry and
remote-control services. The xed networks performance and reliability are tested
and evaluated in real conditions in Thessalonikis urban environment.
Data Processing and Visualisation develop data analysis and visualization tools for
complex event processing, identication of patterns behind periodic peak demands,
forecasting water consumption and demand, assisting the water utility company to
efcient water management and decision-making processes.
Water Management create new products for the consumers and the water utility
company, assisting the former to build consumption patterns and raise their water
consciousness and awareness, and the latter to sustainably manage water, including
issues concerning consumption, demand, resources as well as to detect leaks etc.
Gerasimos Antzoulatos (corresponding author)
Ioannis-Omiros Kouloglou
Efstathios Nikolaidis
Anastasios Karakostas
Stefanos Vrochidis
Ioannis Kompatsiaris
Information Technologies Institute (ITI), Centre for
Research and Technology Hellas (CERTH),
Christos Mourtzios
Nikolaos Papadimitriou
Dimitrios Kourtesis
Research and Development Department,
APIFON S.A. Telecommunications,
Panagiota Stournara
Dimitrios Spyrou
Alexandros Mentes
Thessaloniki Water Supply and Sewerage Company
S.A. (EYATH S.A.),
Tsimiski 98, 54622 Thessaloniki,
This is an Open Access article distributed under the terms of the Creative
Commons Attribution Licence (CC BY-NC-ND 4.0), which permits copying
and redistribution for non-commercial purposes with no derivatives,
provided the original work is properly cited (
2691 © 2020 The Authors Water Science & Technology |82.12 |2020
doi: 10.2166/wst.2020.391
Downloaded from
by guest
on 21 December 2020
Feedback of water usage evaluate the overall infrastructure based on operational
efciency, usability and reliability, and consumer satisfaction.
Factors such as climate change, escalating urban water
demand, customer requirements and the integration of new
technologies in customersservice have all been increasing
the onus on water providers to adopt more effective and sus-
tainable approaches for urban water management. The
combination of smart metering and Internet of Things (IoT)
allows the development of sophisticated systems to better
manage resources and develop new applications to facilitate
various parts of society; that is, citizens/consumers, public
and private organizations, businesses etc. In the water indus-
try, this combination is the key to develop smart water
management systems serving both consumers and water uti-
lity companies and fostering sustainability, through water
leak detection (Saraswathi et al. ), river water quality
monitoring in real-time (Chowdury et al. ), water ow
monitoring (Kusuma & Anil ), short-term water con-
sumption and water demand forecasting (Antunes et al.
;Benítez et al. ). Smart water systems are inherently
connected with the concept of smart cities, since they com-
prise part of the citys infrastructure. Future smart cities will
face challenges of privacy and cybersecurity that should be
taken under consideration to minimize potential risks and
strengthen the important role of smart water in urban areas
(Moy de Vitry et al. ).
The smart metering industry is rapidly evolving. The
term smartdenes any new systems employing the latest
in communication capabilities and enhanced functionalities
(Boyle et al. ). There is a plethora of innovative smart
water meters that frequently record and transmit metering
data. The absence of an open and widely accepted com-
munication standard in the smart water metering industry
led most vendors to introduce proprietary solutions at proto-
col and platform level (Diehl Stiftung & Co. KG ;
Sensus ) adding deployment complexity and resulting
in vendor and protocol lock-inissues (Alvisi et al. ).
Water management infrastructures typically make use
of traditional walk-byand drive bymethods, which
do not allow frequent data collection and cannot support
bidirectional communication (Cole et al. ). Real time
telemetry can only be feasible via a xed wireless network
consisting of data collectors at xed installation points.
Designing and implementing such a network in urban
areas is challenging. Short-range technologies like BLE
(Gomez et al. ) and ZigBee (Gislason ) can no
longer meet the growing demands of modern applications.
Alternatives based on cellular networks (2G/3G/4G) pro-
vide wider coverage with the expense of cost. Till now, the
majority of Measuring Instruments Directive (M.I.D.) certi-
ed water meters by top-notch vendors adopt short range
protocols, i.e. wM-Bus (EN 13757-4) (Carratù et al. ),
which hinder large scale deployments due to the large
number of required gateways. Low Power Wide Area net-
works (LPWAN) technologies (Sanchez & Cano )
provide increased connectivity range (>5 km), low power
consumption and low infrastructural costs. The main
LPWAN technologies used are: Sigfox (Sigfox ), LoRa
(Semtech ) and NarrowBand-IoT (NB-IoT) (GPP ).
Intelligent water metering effectively generates a big
volume of data streams that can be stored into the water uti-
lity providersinformation systems. However, the need for
real-time monitoring as well as for advanced automated
reporting tools and predictive analytical processes have
emerged. The reliable and accurate short-term water con-
sumption and demand predictions can signicantly improve
water management, thus bringing signicant nancial savings
to a water utility company. Hence, efcient demand forecast-
ing facilitates on one hand identication of water network
failure or water loss and on the other effective scheduling
of the utilisation of the energy-related operations for treat-
ment and pumping during the low-cost periods, etc. (Anele
et al. ;Antunes et al. ;Guancheng & Shuming
;Xenochristou et al. ;Benítez et al. ;Oyebode
& Ighravwe ). In the last decades, different linear and
nonlinear approaches for short-term water demand forecast-
ing have been proposed in the literature. Two well-known
representatives of the rst category are the Auto-Regressive
Moving Average (ARMA) and Auto-Regressive Integrated
Moving Average (ARIMA) approaches, which have been
used in water demand time series modeling (Donkor et al.
;Kofinas et al. ;Antunes et al. ;Anele et al.
). The nonlinear ones aim to t machine learning tech-
niques to model the intrinsic nonlinearity into the water
demand data and predict the time series behaviour, increas-
ing the forecasting accuracy. A variety of studies have
2692 G. Antzoulatos et al. |The SMART-WATER solution Water Science & Technology |82.12 |2020
Downloaded from
by guest
on 21 December 2020
explored the application of machine learning methods such
as Articial Neural Networks, Support Vector Machines for
Regression (SVR), Random Forests, Deep Neural Networks
etc. in short-term water demand forecasting or even hybrid
models that combine one or more methods (Kofinas et al.
;Walker et al. ;Anele et al. ;Antunes et al.
;Guancheng & Shuming ;Vijai & Sivakumar ;
Xenochristou et al. ;Oyebode & Ighravwe ).
In the framework of the SMART-WATER (Mourtzios
et al. ) research project, a unied system consisting of
smart water meters coupled with smart water valves and
data analysis processes were utilized in the urban area of
Thessaloniki, Greece. The proposed innovative combination
makes its application an additional challenge.
In the following sections, a brief reference to related pro-
jects is exhibited and is followed with a short description of
the SMART-WATER project. Then, the methodological fra-
mework, as well as the main methods and approaches, are
mentioned in detail. Finally, the preliminary results of the
project will be discussed in the last section of this work.
During the last decade, realizing the need to improve water
management resources along with the achievements of evol-
ving IoT technology, signicant progress has been made and
various projects have come up in the eld of water, aiming
to address the arising challenges. In the study of Tsavdaridou
et al. (), Automated Meter Reading (AMR) was utilized
with positive displacement water meters to detect leakages
experimentally in Thessaloniki, in cooperation with the
citys water utility company (EYATH S.A.).
In the IceWater project, which is presented in Fantozzi
et al. (), focusing on improving energy efciency and
leak detection, ICT solutions were utilized. The possibilities
offered by using AMR systems to collect data at the consumers
level were taken into account in Candelieri & Archetti ().
In this work, an approach for water demand forecasting on an
hourly basis was developed, aiming at optimizing operations of
the water network in order to reduce energy costs.
Another project, the iWIDGET project (Makropoulos
et al. ) proposed two web-based platforms developed
mainly for consumers: (a) a web-based platform providing
consumers with information for domestic water consumption
accessible through the respective platform (Kossieris et al.
a). This web tool made use of time series charts helping
consumers gain deeper insight into their consumption pro-
le and past water consumption data, (b) an educational
e-Learning platform, which was presented in Kossieris et al.
(b), providing various applications, aiming to motivate
and help consumers improve their water consumption prole.
The iWIDGET project aimed at developing a collaboration
between consumers and water utilities, exploring huge data
sets from smart water meters (Ribeiro et al. ).
In the WATERNOMICS project, a system infrastructure
and platform were developed using ICT technologies for
water resource management aiming at increasing water ef-
ciency. It proposed three (3) water consumption levels,
namely domestic, corporate and municipal. The platform
was the main tool to achieve improvements in water con-
sumption behaviour (Perfido et al. ;Kouroupetroglou
et al. ). Another smart water metering approach is pre-
sented in Mudumbe & Abu-Mahfouz , including a web-
based visualization tool for real-time and historical water
consumption data. In Singapore, a Smart Water Grid is
intended by its Public Utility Board for real-time monitoring
of pressure and water quality throughout the network as
well as for real-time water consumption with the use of Auto-
mated Meter Reading (AMR) (PUB ). Furthermore, the
SmartH2O platform was developed (Novak et al. ;Riz-
zoli et al. ) aiming at water consumption reduction
through behavioural change in consumerswater consump-
tion. It was based on the combination of smart metering
data and visualization tools contributing to creating feedback
between the consumers and the water utility sector.
SMART-WATER is an ongoing pilot project that puts for-
ward a novel design of an infrastructure that utilizes
modern telemetry and remote-control technologies to pro-
vide innovative services to consumers and water utility
companies. Custom-made gateways constitute the backbone
of a xed wireless and multiprotocol network that makes
real time telemetry and remote control viable. The gateway
functions overcome lock-inrestrictions since they can
interface with any wM-Bus and LoRa sensors. SMART-
WATERs platform can decrypt and manage end device
data from water valves and a wide range of different
vendor water meters. State-of-the-art data analytics tools
and techniques from the elds of statistics and Machine
Learning are applied to predict consumption and identify
patterns behind periodic peak demands. Finally, two web-
based platform applications are developed to provide end
users, consumers and water utility companies with real
time telemetry and remote-control services. Specically,
2693 G. Antzoulatos et al. |The SMART-WATER solution Water Science & Technology |82.12 |2020
Downloaded from
by guest
on 21 December 2020
the goals of the SMART-WATER project can be summarised
according to the ve directions, as illustrated in the Figure 1.
These goals have been contextualized in a ve-layer uni-
ed generic architecture (Figure 2). The proposed platform
supports the pipeline process from automatic metering of
water consumption to the water management and decision
support process by the water utility company as well as by
the consumers.
End Device Layer: The primary goals of SMART-WATER
project are the frequent recording and wireless transmission
of water consumption (smart automated telemetry) data com-
bined with remote control of water supply in almost real time.
Device Connectivity Layer: This layer is in charge of
the bidirectional connection between end devices and the
central server infrastructure using multiple wireless telecom-
munication protocols.
Device and Data Management Layer:At the 3rd level of
the network architecture, the reception, decoding, decryp-
tion and storage of the data transmitted by the water
meters, as well as the management of the actuation com-
mands to the valves, take place.
Data Processing and Analytics Layer: In this level, the
storage and processing of the metering data is realised. This
layer employs state-of-the-art machine learning techniques
to perform predictive analysis on water consumption and
demand, identify patterns and trends in time series data, and
detect abnormal water consumption behaviour of consumers.
User Interaction Layer: The systems end-users are water
utility company authorized personnel and consumers. Two
different User Interfaces (U.I.) were designed and developed
covering each groups needs for real time water consump-
tion monitoring and water supply control.
The main contribution of this work is to focus on the
backbone aspects of the SMART-WATER project, by illustrat-
ing the proposed approaches, in each one of the above layers,
which are able to tackle the aforementioned limitations and
challenges. Specically, in this work, a holistic framework
is proposed that encapsulates innovative smart metering
and remote-control services in real time, designs and deploys
anovelxed wireless network to connect the smart devices
with the central infrastructure and develops state-of-the-art
methods to manage and analyse the obtained data, aiming
to assist the water utility company to achieve efcient water
management and provide high quality services to consumers.
In this section, the methods, tools and techniques that were
designed, implemented and applied in the eld through the
whole life span of the project are exhibited thoroughly, by
following the ordering of the architecture layers.
Figure 1 |SMART-WATER ve main goals.
2694 G. Antzoulatos et al. |The SMART-WATER solution Water Science & Technology |82.12 |2020
Downloaded from
by guest
on 21 December 2020
End device layer
The Sensors layer is the source of data to the SMART-WATER
system infrastructure. Smart water meters consist a novel
metering technology. Built-in communication systems
enable efcient remote reading while sophisticated software
alerts effectively for leak detection, bursts or other irregulari-
ties such as tampering attempts and reverse ows. The
Figure 2 |SMART-WATER ve layered architecture.
2695 G. Antzoulatos et al. |The SMART-WATER solution Water Science & Technology |82.12 |2020
Downloaded from
by guest
on 21 December 2020
prerequisite features they should demonstrate to constitute
the building blocks of a water management infrastructure
are M.I.D. certication, high metering accuracy (ultrasound
or electromagnetic), wirelessly transmitting metering data
at dense regular intervals and ultra-low consumption (10þ
years autonomy). Based on the aforementioned features,
water meters from ve different vendors were chosen for
the pilot testing in order to assess their metering and telecom-
munication performance. Water meters use wM-Bus or LoRa
telecommunication protocols.
The second category of end devices are the smart wireless
water valves, which make it possible to control water supply
and comprise one of the main innovative features of the pro-
posed system. The wireless valves should exhibit certain
features: remotely controlled via a wireless and bidirectional
telecommunication protocol, ultra-low consumption (10
years autonomy) and tamper alert. Telecommunications-
wise, water valves use LoRaWAN protocol.
Device connectivity layer
Multi-Protocol gateways (MPG)
An important challenge tackled during the project was the
lack of any commercially available gateway that can simul-
taneously communicate bidirectionally with wM-Bus and
LoRa end devices and at the same time establish
bidirectional communication with the central infrastructure.
In the literature, there have been proposed IoT gateways
that utilize multiple protocols. Amiruddin et al. ()pro-
posed a gateway that could interface with Zigbee και BLE
sensors while Kaur & Singh  and Gunasagarana et al.
 proposed similar telecommunication nodes that could
communicate with RF, Zigbee and Bluetooth sensors.
Water meters and water valves operate on different tele-
communication protocols. An additional problem is that the
two protocols used by end devices, namely wM-Bus and
LoRaWAN, are not internet protocols, which does not
allow end devices to communicate directly to the server.
All the above issues led us to design and manufacture an
electronic device that: (a) communicates bidirectionally
and simultaneously with wM-Bus and LoRaWAN technol-
ogy devices and (b) communicates bidirectionally via at
least one internet protocol with the central server.
The Printed Circuit Board (PCB) (Figure 3), together with-
the electronic components (modules/boards), comprise
the multi-protocol gateway, which is placed in an IP55
plastic enclosure. The breakout boards are placed on the
female pinheads, allowing the user to easily add-drop them
depending on the connectivity scenario. The gateway uses
Raspberry Pi Zero W (1 GHz, single-core CPU, 512MB
RAM) as its main processing unit.
The Multi-Protocol Gateway performs numerous func-
tions simultaneously. Firstly, it collects wM-Bus telegrams
Figure 3 |PCB with all modules attached on female pinheads. Dimensions: 22 cm(L) ×14 cm (W) ×5 cm(H).
2696 G. Antzoulatos et al. |The SMART-WATER solution Water Science & Technology |82.12 |2020
Downloaded from
by guest
on 21 December 2020
(T, S, C modes) transmitted by wM-Bus water meters within
range. Secondly, it communicates bidirectionally with LoR-
aWAN end devices. To overcome the issue that the
telecommunications protocols used by end devices are not
IP protocols, a GSM module was integrated in each gateway
to allow two-way communication with the server. The gate-
ways use NB-IoT network as a secondary two-way protocol
in premises where there is limited GSM coverage or as a fail-
over solution in cases where GSM outage take place.
Hybrid xed network
One of the most challenging tasks of the SMART-WATER
project was the design and implementation of the bidirec-
tional and wireless xed network, which is necessary for
real time telemetry and remote-control services. The dense
urban environment and reinforced structural materials in
Greek cities, combined with the limitations of RF technol-
ogies, made network design and deployment more
complicated. Figure 4 depicts the architecture of the xed
wireless network, realized through different connectivity
scenarios. SMART-WATERs hybrid network makes use
of multiple telecommunication protocols, both legacy
(wM-Bus, GSM) and LPWAN (LoRa, NB-IoT), employing
different gateway types; for example, off-the-shelf outdoor
LoRaWAN Gateways, wM-Bus-to-LoRa Bridges, LoRa repea-
ters (the last two are realized through MPG), to ensure the
optimal connectivity conditions and minimum packet loss.
The gateways installation points need to be carefully
selected in terms of connectivity requirements and cost
constraints. In this direction, a custom testerdevice was
designed and assembled in order to perform on site connec-
tivity tests, which dened the optimal gateway installation
point inside consumers premises and veried the connec-
tivity scenarios described in Figure 4.
Device and data management layer
At the 3rd level of the proposed system architecture, the
reception, decoding, decryption and storage of metering
data, as well as the management of the downlink com-
mands, take place. The server unit can simultaneously
process and decode/decrypt payload data from water
meters and water valves. This level hosts three main
services: (a) end device management, (b) gateway monitor-
ing and (c) LoRa and wM-Bus servers.
In the proposed architecture (Figure 2), the distributed
layers from end devices to UIs are depicted. An MQTT
Figure 4 |Fixed networks connectivity scenarios and gateway types.
2697 G. Antzoulatos et al. |The SMART-WATER solution Water Science & Technology |82.12 |2020
Downloaded from
by guest
on 21 December 2020
Broker (Mosquito Broker) has been used in each distributed
unit of the system and an AMQP (Advanced Message Queu-
ing Protocol) Broker (RabbitMQ Broker) in the central
server. In addition, the IoT Core tool has been added to
the system, a managed cloud service provided by Amazon
Web Services (AWS), used to synchronize all MQTT Bro-
kers in the system. Two cloud-based Linux computers are
used through Amazons online services (AWS) and speci-
cally through the Elastic Compute Cloud (EC2) service.
The server uses a noSQL database (MongoDB) both for
end device management and for data storage. Servers API
(Application Programming Interface) is a RESTful API,
with token-based authentication which was developed to
support consumers and administrators applications and
provide them with the necessary data and metadata originat-
ing from the networks devices (end devices and gateways).
Device management
Given that the end devices are remotely installed, it is cru-
cial to develop a system that allows their monitoring and
registration. The Device Management Systemallows the
network administrator to register new end devices with all
necessary information (e.g. Device id,Customer id,
decryption keys,installation address,coordinates).
Gateway monitoring
MPGs are remotely controlled and monitored. Virtual Pri-
vate Network (VPN) technology is used for remote
management and monitoring of all gateways. The user can
perform routine debugging actions, software updates,
system monitoring etc. The gateways additionally transmit
frequent heartbeatuplink messages to the server through
the packet forwarder services.
Lora server/ WM-Bus server
When an uplink message from a LoRa end device reaches
the central server, it is directed to the LoRaWAN server
as shown in Figure 2. The LoRa server is an open source
application (ChirpStack ) developed according to
the LoRaWAN protocol in the Go programming language.
wMbus Serveris a custom application developed for
SMART-WATER. wM-Bus messages are directly processed
by the AMQP API service and then transferred to the
wM-Bus server in order to be decrypted and decoded.
Each wM-Bus end device has a unique decryption key
stored in a database upon subscription through the Device
Management platform. wM-Bus server was developed in
the Python programming language.
Data processing and analytics layer
The main objective of this layer is to analyse the acquired
time-series data in order to make short-term predictions of
the water consumption in a specic urban region of interest.
To achieve this goal, various methodologies from statistics
and machine learning elds are implemented and evaluated.
The proposed SMART-WATER analytical approach consists
of the following seven (7) steps:
1. Data acquisition: the water consumption data for
specic time period and region of interest will be
acquired from the SMART-WATER data storage.
2. Visualise the Time Series: visualising the insight charac-
teristics in the time series, such as the trend, seasonality,
existence of outliers etc., is an essential step prior to
deeper analysis and building a more rened model. The
SMART-WATER dashboard aims towards this direction,
providing ltering, searching and zooming functionalities
to easily explore time-series data.
3. Preprocessing: Depending on the dataset, the steps of
preprocessing will be dened, including the creation of
timestamps, the detection and elimination of outliers,
the stationary test, the normalisation process etc.
iOutlier detection and removal: Outliers are obser-
vations that are very different from the majority of
the observations in the time series (Hyndman &
Athanasopoulos ). They may be errors, due to
hardware issues, sometimes extremely high values of
water consumption appear, or they may simply be
unusual. These values should be removed because
the majority of analytical techniques are not able to
deal with them efciently. A common and simple
way to detect outliers is using the z-score method.
ii Test for Stationary Time Series: A well-known test for
stationarity is the Augmented Dickey Fuller Test
(ADF) which is a form of Hypothesis Testing. The
null hypothesis of the ADF test is that the time series
is non-stationary. If the ADF Test Statistic is less
than the Critical Value, the null hypothesis should
be rejected, concluding that the series is stationary
and it is expected with a very high probability that
its previous behaviour will be continued in the future
iii Normalization: Data normalization involves convert-
ing time series data into a given range. It is needed
for the application of machine learning models,
2698 G. Antzoulatos et al. |The SMART-WATER solution Water Science & Technology |82.12 |2020
Downloaded from
by guest
on 21 December 2020
Neural Networks. A common technique is the Min-
Max Normalization, in which the original data are
transformed linearly.
4. Choose the input data features: a crucial step for the qual-
ity of the forecasting process is the right choice of input
features, especially during the creation of a model. Each
feature represents an input variable that highly affects the
outcome of the forecast (Benitez et al. ). In the
SMART-WATER framework, it is considered that
previous water consumption observations affect the short-
term water demand forecast. Although all the range of
data has to be utilised during the training phase, only a
small portion has a direct inuence on each stepof the fore-
casting process. Thus, the predicted water consumption
for the timestamp T þ1, WCF(T þ1), (e.g. day, hour etc.)
can be predicted by employing the past observations
OTn,...,OTin the timestamps (T-n, ,T), namely
WCF(Tþ1) ¼M(OT,OT1,...,OTn), where M is the
5. Training phase: The main goal of this step is the creation
of a forecasting model that will enable correct prediction
of the future water consumption. Thus, the appropriate
method should be chosen along with splitting the time-
series dataset for training and testing purposes.
iDevelop base-line approach: create/train ARIMA and
SARIMA models; namely, the appropriate models
parameters (p, d, q, P, D, Q, m) should be determined.
ii Develop machine learning approaches: create/train
machine learning models such as Support Vector
Machines (SVMs), Articial Neural Networks
(ANNs), Long Short-Term Memory (LSTM) Neural
6. Testing phase: The trained model is evaluated in terms of
its ability to correctly predict future water consumption.
The testing set is employed and its performance is
estimated using suitable metrics such as Mean Absolute
Percentage Error (MAPE), Root Mean Square Error
(RMSE) and Coefcient of determination (R
7. Compare the results and choose the most accurate
model: the best model is chosen in terms of the perform-
ance indicators over the testing set. It should be noted
that better forecasting models are extracted as the
MAPE and RMSE values tend to zero (0) and the R
values approach unity (1).
The proposed schema is easily customised and self-
learning as it has been designed and developed to be appli-
cable both at the individual consumerslevel as well as for
clusters of consumers with similar water consumption
patterns. Furthermore, it is completely data-driven and inde-
pendent of the machine learning model employed for
forecasting. In this work, the ARIMA and LSTM methods
are employed and their forecasting performance evaluated.
A brief mathematical formalization of these methods is pre-
sented as follows.
Auto-Regressive Integrated with Moving Average Model
(ARIMA)isanefcient approach for short-term forecasting
(Hyndman & Athanasopoulos ). Usually, it denotes as
ARIMA(p, d, q) model, where pis the order of Autoregres-
sion (AR) terms (number of time-step lags to be used),
which represents the relation between past and current
observations, dis the order of differencing for attaining sta-
tionarity (integrated) and qis the order of Moving Average
(MA) terms, which means the size of the MA window. In
the ARIMA model, the forecast value is a linear combi-
nation of past observations (linear combination lags of y
up to p) and past errors (linear combination of lagged fore-
cast errors, up to q), expressed as follows:
yt¼cþφ1yt1þ...þφpytpþθ1εt1þ... þθqεtqþεt
For the estimation of the p,dand qparameters, an itera-
tive strategy is applied consisting of the identication step,
where a sub-class of models that might better represent the
data is chosen, the estimation step, where the parameters
of the model are trained using the historical data and nally
the diagnostic checking, where the trained model is vali-
dated using a disposal dataset.
LSTM stands for Long Short-Term Memory; Neural
Network is an enhancement of the Recurrent Neural
Network (RNN) in relation to their ability to capture the
long-term dependencies (Hochreiter & Schmidhuber ).
Specically, RNNs use previous information in the present
task allowing historical information to be stored in the net-
works internal state and mapping them to the nal
output. RNNs suffer from short term memory due to the
vanishing gradient problem (Masum et al. ;Hua et al.
). Alternatively, the LSTMs overcome this limitation as
they are equipped with unique Gatesenabled to avoid
the long terms for vanishing (Figure 5).
Particularly, an LSTM memory block is a core com-
ponent that encompasses a memory cell and gates, namely
an Input Gate, an Output Gate and a Forget Gate. It com-
putes the mapping from an input sequence x¼(x1,...,xt)
to an output sequence y¼(y1,...,yt). The following steps
are executed inside the memory cell:
2699 G. Antzoulatos et al. |The SMART-WATER solution Water Science & Technology |82.12 |2020
Downloaded from
by guest
on 21 December 2020
1. In the Forget Gate, the input sequence from the current
input xt, is combined with the previous hidden state
ht1, through the sigmoid function σ(x)¼1
which returns values in the range between 0 and 1.
The closer to 0 means to forget, and the closer to 1
means to keep. In this way, the Forget Gate assists in remov-
ing information from the cell state.
2. The Input Gate aims to store new information into the
memory cell state. Initially, a sigmoid layer determines
what will be updated via the formula:
Secondly, the input will be updated by employing the
tanh layer, by creating a vector of new candidate values zt
that can be added to the memory cell state, as stated in
the following equation:
where φ() denotes the tanh function, φ(x)¼2σ(2x)1.
The values uctuate between -1 and 1.
3. The new state of the memory cell, ct, should be updated.
The estimated values, ft,it,zt, are fused along with the
old memory cell state ct1. Specically:
4. The output of the memory cell is realised by element-wise
multiplication between the value obtained from a tanh
function of ctand the output of a sigmoid layer, i.e., the
Output Gate activation vector ot, as follows:
ht¼ot×tanh (ct)
The weight matrices Wf,Wi,Wz,Wo,Uf,Ui,Uz,Uo
and Voalong with the bias vectors bf,bi,bz,bocan be
learnt in the training stage estimated. Through the
cooperation between the memory cell and the gates,
LSTM is endowed with a powerful ability to predict
time series with long-term dependences (Hochreiter &
Schmidhuber ;Masum et al. ;Hua et al. ).
User interaction layer
Water utility company (administrator) Web application
The Main Dashboard (Figure 6) is the cornerstone com-
ponent of the Water Utility Company Web Application, as
it encompasses crucial functionalities enabling the Water
Utility organisation to improve its daily operations. There-
fore, end devices are mapped based on their geographical
location. Furthermore, it enables the operators to gain a gen-
eral glance and overview of the water consumption by
region of interest as well as monitor hourly water consump-
tion in near real-time.
Figure 5 |An illustration of LSTM memory block.
2700 G. Antzoulatos et al. |The SMART-WATER solution Water Science & Technology |82.12 |2020
Downloaded from
by guest
on 21 December 2020
The Single Consumer Dashboard (Figure 7)enables
the practitioners to monitor the water consumption of a
specic consumer in a particular time horizon, to com-
pare his/her consumption with personal past or average
consumptions of the consumers region. The interactive
visualisations fuel the operators with functionalities to
navigate in specic time periods, aggregate and illustrate
time series data hourly, every six hours, daily and
The Alert Dashboard (Figure 8) noties in a concise way
the alerts that are received from the end devices. The alerts
are categorized into 14 classes, each one indicating a
specic malfunction of the end device.
Consumers Web application
The consumers application is a web based and responsive
application. After logging in (Figure 9), the consumer is
Figure 6 |Water Utility Company Web Application dashboards: Main Dashboard.
Figure 7 |Water Utility Company Web Application dashboards: Single Consumer Dashboard.
2701 G. Antzoulatos et al. |The SMART-WATER solution Water Science & Technology |82.12 |2020
Downloaded from
by guest
on 21 December 2020
directed to the consumer Main Homepage (Figure 10)in
case he/she owns more than one pair of end devices. Other-
wise, he/she is directed to the Main Dashboard (Figure 11)
where he/she can track water consumption by hour/day/
week/month and actuate on water valves. Information
about former bills is also available and a prediction based
on current consumption, for the next billing is also depicted
(Figure 10).
This section summarizes ndings based on pilot tests in real
conditions. Having installed the majority of end devices (75
water meters and valves out of 100) and MPGs, useful con-
clusions were drawn regarding SMART-WATER systems
infrastructure. The system underwent rigorous stress tests
to validate its stability across all architecture layers and
the ability to serve multiple users simultaneously. Further-
more, the data analysis processes for short-term
forecasting of water consumption in the region of interest
were assessed in terms of their accuracy.
Time performance
The server infrastructure was initially assessed in terms of
its API performance and secondly for the latency of data
acquisition. In the SMART-WATER case, users are consu-
mers and water utility companys administrators who access
the respective web applications. It is of prime importance to
design and develop a reliable and robust server system
which can respond to multiple user calls simultaneously. In
that direction, the servers API was tested using performance
tool Apache Jmeter ( Several
critical scenarios were simulated in order to extract Response
Time (ms) (amount of time needed to complete a call, Wescott
) and Throughput (kΒ/s) (number of calls per second,
Wescott ) as key performance metrics for the API. The
APIs performance was assessed in different cases realizing
12 scenarios. The scenarios entail both GET and POST
requests spanning from simple Loginrequests to more
demanding ones. For example, Scenario 10 describes the
case of 200 (virtual) users doing consecutive calls (6 GET
calls) for 200 seconds in the Consumersapplication home
page. The description of the scenarios, including their specic
features as well as the response times and throughput simu-
lated results per scenario, are listed in Table 1.
It is evident that as the number of users increase, a
decrease in throughput is observed as expected due to the
additional load created. Scenarios 1, 2 and 3 exhibit the
lowest response time (<100 ms) while Scenario 11 the high-
est one exceeding 1.8 sec. Particularly, it represents the most
challenging test case, which is also reected in the lowest
throughput (53 kB/s). Scenario 1 exhibits the highest
Figure 8 |Water Utility Company Web Application dashboards: Alert Dashboard functions and utilities.
2702 G. Antzoulatos et al. |The SMART-WATER solution Water Science & Technology |82.12 |2020
Downloaded from
by guest
on 21 December 2020
simulated throughput reaching 103 kB/s (Figure 12). The
results are more than acceptable, since scenarios describe
challenging cases hardly possible to occur with real users
which means that the API server is designed to serve numer-
ous users at the same time without any signicant delay.
Secondly, data acquisition times were recorded to
evaluate the systems processing speed and trafc load
management. The data acquisition pipeline was divided
into three fundamental stages. Each stage is assigned to a
respective time slot, namely Gateway to Serverstage
(time for a message to reach the server upon reception at
the gateway), Processingstage (time for decoding, decryp-
tion, authentication processes after the message reaches the
server) and Persiststage (time for data aggregation). The
total statistical sample was 3,000 received uplink messages
from water meters installed in consumers premises. The
total simulation time was 2.8 days. The time data points
of the three stages are presented in respective box plots
(Figure 13). For the Gateway to serverplot, the median
time (50% percentile) value is 0.87 s (orange line in the
box) while the inter-quartile range (IQR) of the distribution
extends from 0.66 s (percentile 25%) to 0.99 s (percentile
75%). For the second stage of the data acquisition pipeline,
Processing, the respective values are 0.00312 s (percentile
25%), 0.00319 s (median), and 0.00329 s (percentile 75%).
Finally, the statistics for the Persisttime samples are
0.0334 s (percentile 25%), 0.034 s (percentile 50%), and
0.036 s (percentile 75%).
Figure 9 |SMART-WATER Consumers web application: Log-in page.
2703 G. Antzoulatos et al. |The SMART-WATER solution Water Science & Technology |82.12 |2020
Downloaded from
by guest
on 21 December 2020
The longest timeslot among the three stages is the
Gateway to servercase, which also exhibits the greatest
deviation. Οver-the-internet transmission introduces the
longest time delay, due to the limited connectivity con-
ditions at the gatewaysinstallation points. The
remarkable uctuation in time values is related to RF
Figure 10 |SMART-WATER Consumers web application: Main Homepage.
Figure 11 |SMART-WATER Consumers web application: Main Dashboard.
2704 G. Antzoulatos et al. |The SMART-WATER solution Water Science & Technology |82.12 |2020
Downloaded from
by guest
on 21 December 2020
propagation issues introduced by the dense urban
environment. The time periods of the other two stages
are considerably lower and the respective deviations
denote the merest uctuation. This stability implies that
the system can handle all trafc generated by the end
devices without delay (near real time).
Moreover, Runtime Performance, one of Google
Chromes Developer Tools, was employed to effectively
evaluate the response times for the two web applications
(the water utility companys and the consumers). More
specically, the above tool describes and analyzes the way
that a website appears and loads. In order to measure the
response time of the consumer application, the browsing
conditions on a mobile device should be simulated. There-
fore, 4xCPU slow down for throttling and the Fast 3G for
the network were selected.
Table 1 |Description of the critical scenarios and APIs performance in terms of Response time (ms) and Throughput (kB/s) using the Apache Jmeter performance tool
Scenario No. Scenario Description GET/POST No. Users Duration (sec.) Response time (ms) Throughput (kB/s)
Scenario 1 Consumer App home page 6 GET calls 10 60 96 103
Scenario 2 Login request 1 POST call 10 60 98 102
Scenario 3 Valve request 1 POST call 10 60 93 99
Scenario 4 Consumer App home page 6 GET calls 60 120 581 77
Scenario 5 Login request 1 POST call 60 120 578 78
Scenario 6 Valve request 1 POST call 60 120 571 79
Scenario 7 Consumer App home page 6 GET calls 100 120 848.8 69
Scenario 8 Login request 1 POST call 100 120 798 69
Scenario 9 Valve request 1 POST call 100 120 917 62
Scenario 10 Consumer App home page 6 GET calls 200 200 1,752 57
Scenario 11 Login request 1 POST call 200 200 1,869 53
Scenario 12 Valve request 1 POST call 200 200 1,749 58
Figure 12 |Response time (ms) and throughput (kB/s) estimations using the Apache Jmeter performance tool, for each of the simulated test scenarios and the number of users.
2705 G. Antzoulatos et al. |The SMART-WATER solution Water Science & Technology |82.12 |2020
Downloaded from
by guest
on 21 December 2020
Runtime performance analysis is categorized into four
main groups of functionalities: Scripting,Rendering,Painting
and Loading. Scripting is the time for JavaScript analysis and
evaluation. Rendering and Painting are related to HTML
compilation and CSS editing on objects that can be displayed
on the users screen. Finally, Loading refers to the interaction
of the website with the network as well as the loading of
HTML, CSS and Javascript (
In Figure 14(a),Response Time for the water utility com-
panys web application main components, measured over
the group of functionalities. As expected, scripting is a sig-
nicantly time-consuming operation in all the main
components and it is followed by Rendering. Similarly, the
Response Time of the main components of the consumer
application is presented in Figure 14(b). The consumers
application includes the Consumer Homepage and Consu-
mer Dashboard, consisting of the billing information page,
water consumption and water valve actuation option.
Experiments and results of water consumption
Multi-step water consumption forecasting is performed
using ARIMA and LSTM models. The goal is to compare
the performance of the models over their ability to make
accurate short-term predictions of total water consumption.
The available set of data is related to the water consumption,
corresponding to the period between 2020-02-01 T00:00:00
and 2020-05-18 T23:00:00 local time. The collected data
were provided in ow rates measured in an hourly frequency
from 75 AMRs (total: 2,587 hourly observations). Using the
z-score approach, 30 outliers were detected and their values
were substituted by propagating the last known observation
each time. The average consumption is around 263.75 l/h
and the maximum consumption ranged to 1,209 l/h. The dis-
tribution of the time series water consumption data grouped
by days and hours exhibits high variability over the specic
period (Figure 15).
The stationary ADF test exhibits that the time series is
stationary as its value (3.87) is less than the critical values
25, 50, and 75% (96, 209, and 371.5 respectively). The p-
value is 0.002 less than 0.05, which indicates to reject the
null hypothesis (H0), the data does not have a unit root
and the time series is stationary. This is rational, as the lim-
ited size of the dataset did not imply any trend or seasonality.
In order to t forecasting models for water consumption,
the selected dataset needs to be transformed. The data should
be re-scaled to values between 0 and 1 as described above.
Applying the LSTM model, it is required that the data be
Figure 13 |Box plots for (a) Gateway to serverstage time; (b) Processingstage time; (c) Persiststage time.
Figure 14 |Response times per main group of functionalities over the (a) Water Utility Dashboard; (b) consumer application components.
2706 G. Antzoulatos et al. |The SMART-WATER solution Water Science & Technology |82.12 |2020
Downloaded from
by guest
on 21 December 2020
within the scale of the activation function of the model. Then,
the time series dataset was divided into input and output sub-
sets enabling supervised learning to be undertaken. The past
(T-n) time steps were used as input in order to predict the
water consumption in the time step T þ1. Furthermore, the
time series data was divided into 83.30% (2,155 hourly obser-
vations) for training and the remaining 16.70% (432 hourly
observations) for validation/testing the modelsperformance
in terms of the RMSE, MAPE and R
evaluation measures.
A series of experiments were realised to ne tune the
training parameters of the LSTM network, namely the
epochs, the number of neurons and the activation function.
The RMSE is employed to evaluate the performance of the
LSTM network for each set of parameters over the testing
set. The best set of parameters, minimizing the RMSE,
is {Epochs ¼50, #neurons ¼4, activation function ¼sig-
moid}. The performance of the LSTM network in terms of
the RMSE over the training and testing set after 10 iterations
uctuated from the training set from 129.13 to 135.59 with
mean ±standard deviation value equal to 130.96 ±1.761
and for the testing set from 161.83 to 177.08 with mean ±
standard deviation value equal to 166.55 ±4.668. Similarly,
the R
uctuates from 53% to 58% in the training set (mean
±standard deviation ¼56.40% ±1.35) and in the testing set
from 57% to 64% (mean ±standard deviation ¼62.20%
±2.15). Finally, the minimum MAPE value is 45.19% for
training and 38.52% for testing and the maximum values
are 65.17% and 50.21% respectively. The average MAPE is
54.59% ±7.186 for training and 43.23% ±4.401 for testing.
In the following gure (Figure 16), the performance of
the LSTM network in the training and testing sets is exhib-
ited in terms of its ability to learn the water consumption
pattern hidden in the time series data (training) as well as
to forecast the hourly water consumption in a forecasting
horizon of the 18 days.
In order to compare the performance of the LSTM
model with an ARIMA approach, we need to select an opti-
mal time series model determining the appropriate set of
Figure 15 |Boxplots, which indicate the distribution of water consumption per day and hour in the examined time period.
Figure 16 |Water consumption observed in the period from 2020-02-01 T00:00:00 to 2020-05-18 T23:00:00. Forecasts given by LSTM ({Epochs ¼50, #neurons ¼4, activation function
¼sigmoid}) in the period from 2020-05-01 T00:00:00 to 2020-05-18 T23:00:00.
2707 G. Antzoulatos et al. |The SMART-WATER solution Water Science & Technology |82.12 |2020
Downloaded from
by guest
on 21 December 2020
parameters (p, d, q). Using the auto_arima function from the
python pmdarima package, the ARIMA(6,0,0) model attains
to minimise the Akaike information criterion (AIC). How-
ever, the performance of this model over the testing set is
poor, as the RMSE was around 321.17 and MAPE climbed
to 82.28%. Hence, the LSTM model seems to outperform
the baseline models such as ARIMA in this set of time
series data (Figure 17).
In this work, the main aspects of the on-going research pro-
ject, called SMART-WATER, is presented. Specically,
SMART-WATER is a holistic water management solution
that aims to offer innovative services to consumers and
water utility companies. Pilot tests of the SMART-WATER
system took place in the wider area of Thessaloniki,
Greece. Real time telemetry and remote control of water
supply are feasible utilizing state-of -the-art end devices on
axed bidirectional network. High precision, wireless and
M.I.D. certied water meters are used in conjunction with
wireless water valves. Real time telemetry and remote-
control services are viable through a xed wireless and
bidirectional network, which can manage messages from
multiple protocols and standards (wM-Bus, LoRaWAN,
NB-IoT, GSM). One of the key elements of the proposed
system are the custom-made multi-protocol gateways
(MPG), which unlock protocol and vendor binding issues.
They can simultaneously interface with a wide set of end
devices (water meters and water valves) of different vendors
and protocols, providing bidirectional connectivity. Given
the limited indoor penetration of the wM-Bus protocol, the
gateways need to be located close to the end devices but
at the same time remain connected to the central infrastruc-
ture via GSM or NB-IoT. Field tests also denoted that
LoRaWAN technology is signicantly superior to legacy
wM-Bus technology regarding connectivity range.
At the other side of the architecture, a universal server
can decode and decrypt wireless messages from subscribed
wM-Bus and LoRa end devices regardless of the internet
protocol used. The performance of the server and its API
was tested in terms of processing time, response time and
throughput. Simulation campaigns from real metering data
demonstrated excellent data processing time upon gateway
reception to aggregation (less than a second on average).
Response time and throughput for different scenarios
render the infrastructure able to serve numerous users simul-
taneously. As expected, a larger number of users introduces
increased response time and moderate throughput values.
The short-term water demand forecasting approach
adopts a seven-step schema, from the acquisition of the
time-series data to the evaluation of the forecasting perform-
ance. The approach has been already tested on real data
retrieved from the end devices installed in the eld.
Although the preliminary results for the operation of the
Figure 17 |Performance comparison between LSTM and ARIMA (6,0,0) models.
2708 G. Antzoulatos et al. |The SMART-WATER solution Water Science & Technology |82.12 |2020
Downloaded from
by guest
on 21 December 2020
system are encouraging, however, further work should be
done in order to evaluate its performance in the entire pipe-
line. Especially, in the advanced analytical perspective, as
all the end devices will be deployed and start to collect
data seamlessly, more robust, statistically sound and safe
conclusions will emerge in terms of the short-term water
demand/consumption estimations. Furthermore, clustering
techniques and forecasting water consumption models will
be applied, aiming at the determination of consumption pro-
les of the consumers.
This research has been co-nanced by the European Union
and Greek national funds through the Operational Program
Competitiveness, Entrepreneurship and Innovation, under
code: T1EDK- 04337).
Data cannot be made publicly available; readers should con-
tact the corresponding author for details.
3GPP Org.  Standardization of NB-IOT completed. pp. 1.
nb_iot_complete (accessed 19 February 2016).
Alvisi, S., Casellato, F., Franchini, M., Govoni, M., Luciani, C.,
Poltronieri, F., Riberto, G., Stefanelli, C. & Tortonesi, M. 
Wireless middleware solutions for smart water metering.
Sensors 19 (8), 1853. doi:10.3390/s19081853.
Amiruddin, A., Agung, A., Ratna, P., Harwahyu, R. & Sari, R. F.
 Secure multi-protocol gateway for Internet of Things.In
2018 Wireless Telecommunications Symposium (WTS), pp.
18. doi:10.1109/WTS.2018.8363934.
Anele, A., Hamam, Y., Abu-Mahfouz, A. & Todini, E. 
Overview, comparative assessment and recommendations of
forecasting models for short-term water demand prediction.
Water 9(11), 887. doi: 10.3390/w9110887.
Antunes, A., Andrade-Campos, A., Sardinha-Lourenço, A. &
Oliveira, M.  Short-term water demand forecasting using
machine learning techniques.Journal of Hydroinformatics
20 (6), 13431366. doi: 10.2166/hydro.2018.163.
Benítez, R., Ortiz-Caraballo, C., Preciado, J. C., Conejero, J. M.,
Sánchez Figueroa, F. & Rubio-Largo, A.  A short-term
data based water consumption prediction approach.Energies
12 (12), 2359. doi:10.3390/en12122359.
Boyle, T., Giurco, D., Mukheibir, P., Liu, A., Moy, C., White, S. &
Stewart, R.  Intelligent metering for urban water: a
review.Water 5(3), 10521081. doi:10.3390/w5031052.
Candelieri, A. & Archetti, F.  Identifying typical
urban water demand patterns for a reliable short-term
forecasting the icewater project approach.Procedia
Engineering 89, 10041012.
Carratù, M., Ferro, M., Paciello, V., Pietrosanto, A. & Sommella,
P.  Performance analysis of wM-bus networks for smart
metering.IEEE Sensors Journal 17 (23), 78497856. doi: 10.
ChirpStack.  (accessed 5/1/2020).
Chowdury, M. S. U., Emran, T. B., Ghosh, S., Pathak, A., Alam, M.
M., Absar, N., Andersson, K. & Hossain, M. S.  IoT-based
real-time river water quality monitoring system.Procedia
Computer Science 155, 161168. doi:10.1016/j.procs.2019.
Cole, G., OHalloran, K. & Stewart, R. A.  Time of use tariffs:
implications for water efficiency.Water Science and
Technology 12 (1), 90100. doi: 10.2166/ws.2011.123.
Diehl Stiftung & Co. KG. 
en/portfolio/solutions/ (accessed 5/5/2020).
Donkor, E. A., Mazzuchi, T. A., Soyer, R. & Roberson, J. A. 
Urban water demand forecasting: review of methods and
models.Journal of Water Resources Planning and
Management 140 (2), 146159.
Fantozzi, M., Popescu, I., Farnham, T., Archetti, F., Mogre, P.,
Tsouchnika, E., Chiesa, C., Tsertou, A., Castro Gama, M. &
Bimpas, M.  ICT for efficient water resources
management: the ICeWater energy management and control
approach.Procedia Engineering 70, 633640. doi: 10.1016/
Gislason, D.  Zigbee Wireless Networking. Elsevier Inc.,
Burlington, MA, USA.
Gomez, C., Oller, J. & Paradells, J.  Overview and evaluation
of bluetooth low energy: an emerging low-power wireless
technology.Sensors 12 (9), 1173411753. doi: 10.3390/
Guancheng, G. & Shuming, L.  Short-term water demand forecast
based on deep neural network. In 1st International WDSA/
CCWI 2018 Joint Conference,July2325, 2018, Canada.
Gunasagarana, R., Kamarudina, L. M., Zakariaa, A., Kanagaraja,
E., Alimona, M. S. A. M., Shakaffa, A. Y. M., Ehkanb, P.,
Visvanathana, R. & Razalia, M. H. M.  Internet of
Things: Sensor to Sensor Communication.In2015 IEEE
SENSORS. doi:10.1109/ICSENS.2015.7370448.
Hochreiter, S. & Schmidhuber, J.  Long short-term memory.
Neural Computation 9(8), 17351780.
Hua, Y., Zhao, Z., Li, R., Chen, X., Liu, Z. & Zhang, H.  Deep
learning with long short-term memory for time series
prediction.InIEEE Communications Magazine 57 (6),
114119. doi: 10.1109/MCOM.2019.1800155.
Hyndman, R. J. & Athanasopoulos, G.  Forecasting: principles
and practice. In: OTexts, 2nd edn (R. J. Hyndman ed.).
Melbourne, Australia. (accessed on 02/02/
2709 G. Antzoulatos et al. |The SMART-WATER solution Water Science & Technology |82.12 |2020
Downloaded from
by guest
on 21 December 2020
Kaur, J. & Singh, M.  Multiprotocol gateway for wireless
communication in embedded systems. International Journal
of Computer Applications (09758887) 72 (18), 2731.
Kofinas, D., Mellios, N., Papageorgiou, E. & Laspidou, C. 
Urban water demand forecasting for the Island of Skiathos.
Procedia Engineering 89, 1023. doi: 10.1016/j.proeng.2014.
Kossieris, P., Kozanis, S., Hashmi, A., Katsiri, E., Vamvakeridou-
Lyroudia, L. S., Farmani, R., Makropoulos, C. & Savic, D. a
A web-based platform for water efficient households.Procedia
Engineering 89, 11281135. doi: 10.1016/j.proeng.2014.11.234.
Kossieris, P., Panayiotakis, A., Tzouka, K., Gerakopoulou, P., Rozos,
E. & Makropoulos, C. bAn eLearning approach for
improving household water efficiency.Procedia Engineering 89,
11131119. doi: 10.1016/j.proeng.2014.11.232.
Kouroupetroglou, C., van Slooten, J., Clifford, E., Coakley, D.,
Curry, E., Smit, S. & Perfido, D.  WATERNOMICS:
Serving diverge user needs under a single water information
platform. In e-proceedings of the 36th IAHR World Congress,
The Hague, The Netherlands.
Kusuma,S.S.&Anil,G.N. An IoT based water supply
monitoring and controlling system. International Research
Journal of Engineering and Technology (IRJET) 5(2), 857860.
Makropoulos, C., Kossieris, P., Kozanis, S., Katsiri, E. &
Vamvakeridou-Lyroudia, L.  From Smart Meters To
Smart Decisions: Web-Based Support For The Water
Efficient Household. In 11th International Conference on
Hydroinformatics,New York City, USA.
Masum, S., Liu, Y. & Chiverton, J.  Multi-step Time Series
Forecasting of Electric Load Using Machine Learning
Models. In: Artificial Intelligence and Soft Computing (L.
Rutkowski, P. Scherer, M. Korytkowski, W. Pedrycz, R.
Tadeusiewicz & J. Zurada, eds), ICAISC 2018. Lecture Notes
in Computer Science, vol 10841. Springer, Cham,
Mourtzios, C., Kourtesis, D., Papadimitriou, N., Antzoulatos, G.,
Kouloglou, I. O., Vrochidis, S. & Kompatsiaris, I.  Work-
In-Progress: SMART-WATER, a Νovel Τelemetry and
Remote Control System Infrastructure for the Management
of Water Consumption in Thessaloniki. In: Proceedings of the
13th International Conference on Interactive Mobile
Communication Technologies and Learning, p. 1192, eBook.
Moy de Vitry, M., Schneider, M. Y., Wani, O., Manny, L., Leitão,
J. P. & Eggimann, S.  Smart urban water systems: what
could possibly go wrong? Environmental Research Letters 14,
Mudumbe, M. J. & Abu-Mahfouz, A. M.  Smart water meter
system for user-centric consumption measurement. In: IEEE
13th International Conference on Industrial Informatics
(INDIN),Cambridge, pp. 993998.
Novak, J., Melenhorst, M., Micheel, I., Pasini, C., Fraternali, P. &
Rizzoli, A. E.  Integrating behavioural change and
gamified incentive modelling for stimulating water saving.
Environmental Modelling & Software 102, 120137. doi: 10.
Oyebode, O. & Ighravwe, D. E.  Urban water demand
forecasting: a comparative evaluation of conventional and
soft computing techniques.Resources 8(3), 156. doi: 10.
Perfido, D., Messervey, T., Kouroupetroglou, C. & Costa, A. 
WATERNOMICS (ICT for water resource management)
methodology and water information platform. Procedia
Environmental Science, Engineering and Management 2(4),
Public Utilities Board Singapore (PUB).  Managing the water
distribution network with a smart water grid.Smart Water 1,
113. doi: 10.1186/s40713-016-0004-4.
Ribeiro, R., Loureiro, D., Barateiro, J., Smith, J., Rebelo, M.,
Kossieris, P., Gerakopoulou, P., Makropoulos, C., Vieira, P.
& Mansfield, L.  Framework for technical evaluation of
decision support systems based on water smart metering: the
iWIDGET case.Procedia Engineering 119, 13481355. doi:
Rizzoli, A. E., Castelletti, A., Fraternali, P. & Novak, J. 
Demo abstract: smartH2O, demonstrating the impact of
gamification technologies for saving water.Comput Sci Res
Dev 33, 275276. doi: 10.1007/s00450-017-0380-5.
Sanchez, R. I. & Cano, M. D.  State of the Art in LP-WAN
solutions for industrial IoT services.Sensors 16 (5), 708.
Saraswathi, V., Rohit, A., Sakthivel, S. & Sandheep, T. J. 
Water leakage system using IoT. IJIREM 5(2), 6769.
Semtech Corporation.  AN1200.22 LoRa
Modulation Basics Application Note Revision 2 (accessed
Sensus, Xylem brand. 
sensus-technologies/sensus-rf/ (accessed 15/04/2020).
SigFox.  (accessed 15/11/2019).
Tsavdaridou, A. D., Kolokytha, E. G. & Mentes, A. K.  Urban
water consumption and new metering technologies:
Application in the city of Thessaloniki. In: Proceedings of the
XI Protection and Restoration of the Environment (PRE)
Conference, 203212.
Vijai, P. & Sivakumar, P.  Performance comparison of
techniques for water demand forecasting.Procedia Computer
Science 143, 258266. doi: 10.1016/j.procs.2018.10.394.
Walker, D., Creaco, E., Vamvakeridou-Lyroudia, L., Farmani, R.,
Kapelan, Z. & Savic, D.  Forecasting domestic water
consumption from smart meter readings using statistical
methods and artificial neural networks.Procedia Engineering
119, 14191428. doi: 10.1016/j.proeng.2015.08.1002.
Wescott, B.  Every Computer Performance Book: How to
Avoid and Solve Performance Problems on The Computers
You Work With, 1st edition. Create Space Independent
Publishing Platform, SC, USA.
Xenochristou, M., Kapelan, Ζ, Hutton, C. & Hofman, J.  Smart
water demand forecasting: Learning from the data.In13th
International Conference on Hydroinformatics (HIC 2018).
EasyChair Publications, Vol. 3, pp. 23512358. doi: 10.
First received 31 May 2020; accepted in revised form 29 July 2020. Available online 18 August 2020
2710 G. Antzoulatos et al. |The SMART-WATER solution Water Science & Technology |82.12 |2020
Downloaded from
by guest
on 21 December 2020
... The Figures 5 and 6 are based on an extensive literature [8]. As validation, real-world implementations for the two examples are shown: (W)M-Bus is used for read-out of water meters [12,13], LoRa is applied for large-scale monitoring of urban drainage and water distribution networks [13][14][15] as well as for smart rainwater harvesting [16]; and GPRS is commonly utilised for single measurement points [17,18], which is in concordance with the recommendations of the presented framework. ...
... The Figures 5 and 6 are based on an extensive literature [8]. As validation, real-world implementations for the two examples are shown: (W)M-Bus is used for read-out of water meters [12,13], LoRa is applied for large-scale monitoring of urban drainage and water distribution networks [13][14][15] as well as for smart rainwater harvesting [16]; and GPRS is commonly utilised for single measurement points [17,18], which is in concordance with the recommendations of the presented framework. ...
Conference Paper
In urban water infrastructure, information and communication technology are presently concentrated on central facilities (e.g., treatment plant) or are installed at main points in the urban drainage or water distribution network, e.g., inlet points of district meter areas or combined sewer overflow structures. In this regard, the Internet of Things concept as part of smart city development enables a large-scale implementation of measuring equipment and allows the integration of decentralised elements into an overall controlled system. Consequently, reliable and suitable information and communication technology is a key element for the exchange of measurement and control data as well as for the success of these systems. From a water engineering perspective, it is often difficult to choose the right information and communication technology for the intended urban water infrastructure application. Likewise, from an Internet of Things perspective, it is often unclear what kind of urban water infrastructure applications are feasible, making them difficult to efficiently implement. Aim of this work is to develop a first-decision making tool, which can be used by network operators, researcher, and stakeholders to support supervisory control and data acquisition development and to realise an efficient information and communication technology system in the field of network-based urban water infrastructure. In contrast to existing recommendations, our approach is based on a comprehensive review of required spatial and temporal resolution of measurement and control data for a wide range of different network-based urban water infrastructure applications. Subsequently, this enables a targeted coordination with the properties of communication technologies (e.g., data rate, range, and quality of service) and leads to a significantly improved and integrative decision-making tool. Subsequently, we tested the functionality of the framework on two exemplary applications in urban water infrastructure, namely (1) determining suitable communication technologies for an early warning system for leakage detection and localisation in water distribution networks (e.g., (Wireless)Meter-Bus for water meters, long range wide area network for water pressure sensors, and global system for mobile communications at inlet points of district meter areas, and (2) identifying feasible applications for an existing long range wide area network (e.g., monitoring micro-climate and automatic irrigation at nature-based solutions). Results from the framework application have been evaluated through a literature review on used communication technologies, and are found to be consistent with real-word applications. As conclusion, different communication technologies are necessary to satisfy different requirements associated with an integrative management of urban water infrastructure.
... Brears [13] added that actions must be designed in an integrated and coordinated way to conduct an effective management of urban water and water in nature with institutional policies at individual scales. Antzoulatos et al. [14] argued that traditional water management was inadequate as new technologies were available to revolutionize the management of urban water systems. A sustainable urban water system must privilege contemporary socio-economic development without compromising the future supply [15]. ...
... Sustainability 2022, 14, 5695 ...
Full-text available
The objective of this study was the definition and determination of sustainable water consumption indicators for activity categories, as well as the evolution of water consumption in commercial buildings. These indicators were determined through statistical analyses using Shewhart charts. Within a broader scope, the research proposed a methodology to automate sustainable management of water consumption in building operation using BIM–IoT–FM integration. The scientific rigor of the methodology was based on the precepts of design science research. The methods proposed for the construction of functionalities and the application of the reference indicators provided an optimized analysis of water consumption and the detection of excess consumption and leaks. The methodology, implemented in an online prototype, AquaBIM, could deliver a significant advance for building management. A conceptual test of AquaBIM evaluated the consumption indicators and validated our methodology through its application in a commercial building. The building consumption analyses showed a potential for approximately 15% savings. In addition, five requirements of the international sustainability certification AQUAHQE were met. The results of our research provide an innovative approach for the automation of sustainable building management and could be expanded to monitor and report on the consumption of other critical resources such as electricity and gas.
... On the other hand, the web-based data management application incorporates intelligent functions that enable the efficient management of water resources and consumers by supporting a two-way communication between the water company and consumers. An overview of the methodological framework and main methods and approaches of the Smart-Water project is given in Antzoulatos et al. (2020) and Mourtzios et al. (2021). ...
... However, at the same time, the integration of Internet of Things (IoT) devices in such Critical Infrastructures (CIs) has led to an escalation of cyberattacks and incidents in recent years [14] [15]. As intelligent devices and SCADA systems are used in highly critical domains such as power, water, and nuclear plants, malicious actions can have catastrophic consequences if they succeed [16] [17]. The impact of cyberattacks can range from loss of revenue to power outages and possibly nuclear disasters, as the Ukrainian power plant incident [18] [19] and STUXNET [20][21] [22] have indicated in the past. ...
Conference Paper
Full-text available
The integration of intelligent devices in the industry allows the automation and control of industrial processes, in an efficient and effective manner. Such systems have contributed to the rapid evolution of production infrastructures, increasing the reliability, reducing production costs, and automating the entire manufacturing operations. However, the utilization of intelligent devices has led to an increased attack surface in critical infrastructures, threatening to compromise regular operations. Attacks against such environments can have disastrous consequences in case their goal is achieved, due to the critical nature of such infrastructures. Thus, the timely identification of vulnerable spots through high-quality risk assessment, is considered highly important for avoiding or mitigating potential risks. In this paper, we focus on Distributed Network Protocol 3 (DNP3), a protocol with high utility in smart grids. Specifically, we investigate, identify and describe the vulnerabilities-by-design of DNP3 through 8 DNP3-centered cyberattacks. In addition, we present a novel method for conducting risk assessment, stemming from the combination of two techniques, namely, Attack Defence Trees (ADTs) and Common Vulnerability Scoring System v3.1 (CVSS). Through our proposed technique, the risk of a cyberattack occurring is calculated, thus contributing to securing the critical infrastructure.
... For example, most intelligent water management systems use IoT sensors in the process of streamlining water consumption. In [5], the authors describe an IoT network that monitors water consumption in residential areas, including solutions for analyzing the data collected. Thus, this platform has multiple benefits for consumers and water suppliers, being a consistent basis for supporting the decision-making process regarding the control of water use in urban households. ...
Full-text available
Water resource management represents a fundamental aspect of a modern society. Urban areas present multiple challenges requiring complex solutions, which include multidomain approaches related to the integration of advanced technologies. Water consumption monitoring applications play a significant role in increasing awareness, while machine learning has been proven for the design of intelligent solutions in this field. This paper presents an approach for monitoring and predicting water consumption from the most important water outlets in a household based on a proposed IoT solution. Data processing pipelines were defined, including K-means clustering and evaluation metrics, extracting consumption events, and training classification methods for predicting consumption sources. Continuous water consumption monitoring offers multiple benefits toward improving decision support by combining modern processing techniques, algorithms, and methods.
... As one of the most popular DL algorithms for processing time-series data, the LSTM is equipped with a memory cell and several gates to overcome the gradient explosion or gradient disappearance problem of RNN. Figure 4 shows the cell structure of the LSTM, and Equations (10)- (15) give the basic mathematical computational steps of the algorithm [90]. First, the forgetting gate f t , input gate i t , and candidate value z t are calculated by using the current input x t and the previous hidden state h t−1 . ...
Full-text available
As one of the most critical elements in the hydrological cycle, real-time and accurate rainfall measurement is of great significance to flood and drought disaster risk assessment and early warning. Using commercial microwave links (CMLs) to conduct rainfall measure is a promising solution due to the advantages of high spatial resolution, low implementation cost, near-surface measurement, and so on. However, because of the temporal and spatial dynamics of rainfall and the atmospheric influence, it is necessary to go through complicated signal processing steps from signal attenuation analysis of a CML to rainfall map. This article first introduces the basic principle and the revolution of CML-based rainfall measurement. Then, the article illustrates different steps of signal process in CML-based rainfall measurement, reviewing the state of the art solutions in each step. In addition, uncertainties and errors involved in each step of signal process as well as their impacts on the accuracy of rainfall measurement are analyzed. Moreover, the article also discusses how machine learning technologies facilitate CML-based rainfall measurement. Additionally, the applications of CML in monitoring phenomena other than rain and the hydrological simulation are summarized. Finally, the challenges and future directions are discussed.
... Therefore, vendors tend to adopt different solutions for data representation and communication technologies, thus increasing the heterogeneity and the complexity of metering platforms. For example, Fig. 1 shows that metering devices and IoT sensors can leverage several low-power wireless protocols: Wireless M-Bus, LoRa, and NarrowBand IoT (NB-IoT) [9], which might present different performance depending on radio frequency propagation issues. ...
Conference Paper
Smart metering represents an interesting field, where IoT can bring huge benefits to collect and analyze water and environmental data. However, it presents well-known issues such as the presence of high heterogeneity at the hardware, software, and network layers. This is even amplified by the vendors' tendency to adopt proprietary solutions and different low-power wireless communication protocols for IoT sensors and metering devices. Realizing an interoperable platform for the collection and analysis of smart Water and environmental data metering thus becomes a complex task, that needs to address multiple requirements at several levels, starting from the collection of data on the field (IoT sensors, smart meters) and its processing on Cloud Computing platforms. The Water 4.0 project, which involves a collaboration between universities and private companies, including Dipietro Group, aims at addressing these challenges. This paper presents the comprehensive smart environmental data metering solution that we realized within Water 4.0, that enables data collection, data-processing, and Over-The-Air (OTA) for IoT devices, with the ultimate goal of reducing water losses and monitoring water quality.
Conference Paper
Many homes in the country of Peru, especially those located in unofficial settlements, are not connected to public service networks, and in the case of residential water, require tanker truck delivery. However, this water has often been contaminated from the upstream storage, conveyance and delivery systems that provide it, and thus will not comply with government water quality standards, ultimately compromising the health of the people who rely on it. While the topic of quality monitoring in traditional water networks has been studied, research has not focused on water quality control in under-developed and under-served unofficial settlements. This study introduces an IoT architecture and web-based system for real-time monitoring of the key water quality parameters for the residents of unofficial settlements in Peru. The system proposed was implemented across five layers: capture, communication, processing, storage and presentation. Two experiments were conducted in a residential home with real time measurement of temperature, turbidity, TDS y pH. When comparing the results of both experiments, the pH parameter had a better precision with a 2% error rate. In addition, the survey results showed that the experts agree with the proposal.
Full-text available
Previous studies have shown that soft computing models are excellent predictive models for demand management problems. However, their applications in solving water demand forecasting problems have been scantily reported. In this study, feedforward artificial neural networks (ANNs) and a support vector machine (SVM) were used to forecast water consumption. Two ANN models were trained using different algorithms: differential evolution (DE) and conjugate gradient (CG). The performance of these soft computing models was investigated with real-world data sets from the City of Ekurhuleni, South Africa, and compared with conventionally used exponential smoothing (ES) and multiple linear regression (MLR). The results obtained showed that the ANN model that was trained with DE performed better than the CG-trained ANN and other predictive models (SVM, ES and MLR). This observation further demonstrates the robustness of evolutionary computation techniques amongst soft computing techniques.
Full-text available
Urban water systems are being redefined for a digital age, promising substantial advantages for service users and providers, and for society as a whole. However, beside the much-discussed benefits of smart urban water systems in future smart cities, the transition will also bring unique challenges, particularly with respect to privacy and cybersecurity. [...]
Full-text available
A smart water network consists of a large number of devices that measure a wide range of parameters present in distribution networks in an automatic and continuous way. Among these data, you can find the flow, pressure, or totalizer measurements that, when processed with appropriate algorithms, allow for leakage detection at an early stage. These algorithms are mainly based on water demand forecasting. Different approaches for the prediction of water demand are available in the literature. Although they present successful results at different levels, they have two main drawbacks: the inclusion of several seasonalities is quite cumbersome, and the fitting horizons are not very large. With the aim of solving these problems, we present the application of pattern similarity-based techniques to the water demand forecasting problem. The use of these techniques removes the need to determine the annual seasonality and, at the same time, extends the horizon of prediction to 24 h. The algorithm has been tested in the context of a real project for the detection and location of leaks at an early stage by means of demand forecasting, and good results were obtained, which are also presented in this paper.
Full-text available
Current water quality monitoring system is a manual system with a monotonous process and is very time-consuming. This paper proposes a sensor-based water quality monitoring system. The system consists of several sensors which is used to measure physical and chemical parameters of the water. The main components of Wireless Sensor Network (WSN) include a microcontroller for processing the system, communication system for inter and intra node communication and several sensors. Real-time data access can be done by using remote monitoring and Internet of Things (IoT) technology. Data collected at the apart site can be displayed in a visual format on a server PC with the help of Spark streaming analysis through Spark MLlib, Deep learning neural network models, Belief Rule Based (BRB) system and is also compared with standard values. If the acquired value is above the threshold value automated warning SMS alert will be sent to the agent. The uniqueness of our proposed paper is to obtain the water monitoring system with high frequency, high mobility, and low powered. Therefore, our proposed system will immensely help Bangladeshi populations to become conscious against contaminated water as well as to stop polluting the water.
Full-text available
While smart metering applications have initially focused on energy and gas utility markets, water consumption has recently become the subject of increasing attention. Unfortunately, despite the large number of solutions available on the market, the lack of an open and widely accepted communication standard means that vendors typically propose proprietary data collection solutions whose adoption causes non-trivial problems to water utility companies in term of costs, vendor lock-in, and lack of control on the data collection infrastructure. There is the need for open and interoperable smart water metering solutions, capable of collecting data from the wide range of water meters on the market. This paper reports our experience in the development and field testing of a highly interoperable smart water metering solution, which we designed in collaboration with several water utility companies and which we deployed in Gorino Ferrarese, Italy, in collaboration with CADF (Consorzio Acque Delta Ferrarese), the water utility serving the city. At the core of our solution is SWaMM (Smart Water Metering Middleware), an interoperable wireless IoT middleware based on the Edge computing paradigm, which proved extremely effective in interfacing with several types of smart water meters operating with different protocols.
Full-text available
There is an ever growing demand of water due to the factors like global warming, urbanization and population growth. The situation demands to use more efficient planning which can be attained by technological advancement like Internet of things and smart systems. The cost related to water management system can be optimized by using prediction. The future demand for water could be better modeled with forecasting techniques. A collection of techniques (Artificial Neural Network (ANN), Deep Neural Network (DNN), Extreme Learning Machines (ELM), Least Square Support Vector Machine (LSSVM), Gaussian process regression (GPR), Random Forest (RF), multiple regression have been applied to analyze the performance in water demand forecasting using the common evaluation criteria. The work is aimed at short term prediction using hourly and daily intervals. A good performance was obtained through the ANN model for all short term predictions.
Full-text available
Nowadays, a large number of water utilities still manage their operation on the instant water demand of the network, meaning that the use of the equipment is conditioned by the immediate water necessity. The water reservoirs of the networks are filled using pumps that start working when the water level reaches a specified minimum, stopping when it reaches a maximum level. Shifting the focus to water management based on future demand allows use of the equipment when energy is cheaper, taking advantage of the electricity tariff in action, thus bringing significant financial savings over time. Short-term water demand forecasting is a crucial step to support decision making regarding the equipment operation management. For this purpose, forecasting methodologies are analyzed and implemented. Several machine learning methods, such as neural networks, random forests, support vector machines and k-nearest neighbors, are evaluated using real data from two Portuguese water utilities. Moreover, the influence of factors such as weather, seasonality, amount of data used in training and forecast window is also analysed. A weighted parallel strategy that gathers the advantages of the different machine learning techniques is suggested. The results are validated and compared with those achieved by ARIMA also using benchmarks.
Conference Paper
Water scarcity and water stress issues pose a serious threat to the global population. Climate change, drought, population growth and consolidation in urban centres have all been increasing the pressure on water service providers to deploy more sustainable approaches in urban water management. Real-time monitoring and control of water consumption are key ingredients for a smart water management system which will raise consumers’ environmental awareness and reduce costs. This paper introduces a smart infrastructure system which enables remote telemetry and control of water consumption via a web application. The bidirectional and reliable communication between terminal devices (smart meters and valves) and the end-user (consumers and water utility operators) is realized through a fixed hybrid network which uses multiple telecommunication protocols. The metering data is collected and further processed through modern data analytics tools to give a deeper insight on water consumption and behavioral patterns.
Time series prediction can be generalized as a process that extracts useful information from historical records and then determines future values. Learning long-range dependencies that are embedded in time series is often an obstacle for most algorithms, whereas LSTM solutions, as a specific kind of scheme in deep learning, promise to effectively overcome the problem. In this article, we first give a brief introduction to the structure and forward propagation mechanism of LSTM. Then, aiming at reducing the considerable computing cost of LSTM, we put forward a RCLSTM model by introducing stochastic connectivity to conventional LSTM neurons. Therefore, RCLSTM exhibits a certain level of sparsity and leads to a decrease in computational complexity. In the field of telecommunication networks, the prediction of traffic and user mobility could directly benefit from this improvement as we leverage a realistic dataset to show that for RCLSTM, the prediction performance comparable to LSTM is available, whereas considerably less computing time is required. We strongly argue that RCLSTM is more competent than LSTM in latency-stringent or power-constrained application scenarios.
Short-time water demand forecasting is essential for optimal control in a water distribution system (WDS). Current methods (e.g., time-series models and conventional artificial neural networks) have limited power in practice due to the nonlinear nature of changes in water demand. In particular, 15-min time-step forecasting may not be accurate when using conventional models. To tackle this problem, this paper investigates the potential of deep learning in short-term water demand forecasting, developing a gated recurrent unit network (GRUN) model to forecast water demand 15 min and 24 h into the future with a 15-min time step. The performance of GRUN was compared with a conventional artificial neural network (ANN) model and seasonal autoregressive integrated moving average (SARIMA) model. A correction module was used to reduce the cumulative error. The results show that the deep learning method improves the performance of water demand prediction. The correction module enhances the performance of ANN and GRUN models. In general, deep neural network models like GRUN outperform the ANN and SARIMA models for both 15-min and 24-h forecasts. These findings can provide more flexible and effective solutions for water demand forecasting.