Article

Use of Machine Learning for Leak Detection and Localization

in Water Distribution Systems
Neda Mashhadi, Isam Shahrour *, Nivine Attoue, Jamal El Khattabi and Ammar Aljer

Civil and Geo-Environmental Engineering Laboratory (LGCgE), Lille University, 5900 Lille, France;
[email protected] (N.M.); [email protected] (N.A.); [email protected] (J.E.K.);
[email protected] (A.A.)
* Correspondence: [email protected]

Abstract: This paper presents an investigation of the capacity of machine learning methods (ML) to
localize leakage in water distribution systems (WDS). This issue is critical because water leakage
causes economic losses, damages to the surrounding infrastructures, and soil contamination. Pro-
gress in real-time monitoring of WDS and ML has created new opportunities to develop data-based
methods for water leak localization. However, the managers of WDS need recommendations for the
selection of the appropriate ML methods as well their practical use for leakage localization. This
paper contributes to this issue through an investigation of the capacity of ML methods to localize
Citation: Mashhadi, N.; Shahrour, I.; leakage in WDS. The campus of Lille University was used as support for this research. The paper is
Attoue, N.; El Khattabi, J.; Aljer, A.
presented as follows: First, flow and pressure data were determined using EPANET software; then,
Use of Machine Learning for Leak
the generated data were used to investigate the capacity of six ML methods to localize water leak-
Detection and Localization in Water
age. Finally, the results of the investigations were used for leakage localization from offline water
Distribution Systems. Smart Cities
flow data. The results showed excellent performance for leakage localization by the artificial neural
2021, 4, 1293–1314.
https://doi.org/10.3390/smartcities40
network, logistic regression, and random forest, but there were low performances for the unsuper-
40069 vised methods because of overlapping clusters.

Academic Editor: Pierluigi Siano Keywords: EPANET; flow; localization; machine learning; pressure; leak

Received: 8 August 2021

Accepted: 28 September 2021
Published: 1 October 2021 1. Introduction
Water leakage constitutes an important issue in managing water distribution systems
Publisher’s Note: MDPI stays neu-
because it causes economic losses, damages to the surrounding soil and infrastructures,
tral with regard to jurisdictional
claims in published maps and institu-
and soil contamination. According to the World Bank [1], the non-revenue water (NRW)
tional affiliations.
level in developing countries ranges from 40% to 50% of the water pumped into the dis-
tribution systems. The American Water Works Association Research Foundation (AW-
WARF) estimated that water utilities in the United States suffered from 250,000 to 300,000
main breaks per year, causing approximately USD 3 billion in annual damages [2]. In [3],
Copyright: © 2021 by the authors. Li- it was reported that the water losses from leakage in some countries in the Middle East
censee MDPI, Basel, Switzerland. represented 50% of the water supply. Reference [4] identified leakage as one the most
This article is an open access article common operational problems in the water distribution system of Athens.
distributed under the terms and con- Relevant research has been conducted for the development of methods for water
ditions of the Creative Commons At- leakage detection. These methods can be classified into hardware- and software-based
tribution (CC BY) license (https://cre- methods. The first category uses various technologies such as acoustic monitoring [5], gas
ativecommons.org/licenses/by/4.0/). injection [6], thermography [7], ground-penetrating radar [8], and free-swimming sys-
tems [9]. Acoustic monitoring includes technologies for listening sticks, leak noise corre-
lation, and leak noise loggers. These methods have high performances but suffer from the
high cost. The gas injection method injects a non-toxic, water-insoluble, and lighter-than-
air gas into water pipes. The leak can be detected by scanning the ground surface using
gas detectors. This method is characterized by speed tracing, but its high cost reduces its
practical use. The ground-penetrating radar method is based on tracking the reflection of

Smart Cities 2021, 4, 1293–1314. https://doi.org/10.3390/smartcities4040069 www.mdpi.com/journal/smartcities

Smart Cities 2021, 4, 69 2 of 23

electromagnetic waves generated at the ground surface. It provides information about the
presence of anomalies in the subsoil. Water leaks can be detected by identifying soil voids
created by water leaks or by detecting sections of pipes that appear deeper than they ac-
tually are due to the increase in the dielectric properties of the surrounding saturated soils.
This method can be used for metallic or plastic pipes, but it is expensive and time-con-
suming. The free-swimming systems methods are based on introducing the water pipes
of capsules with an embedded power source, electronic components, and instrumentation
(acoustic sensor, accelerometer, magnetometer, GPS synchronized ultrasonic transmitter,
and temperature sensor). These capsules record the internal environment of the pipes and
send the recorded data to a server. The analysis of registered data permits detection and
localize anomalies related to water leakage. This method is well adapted for pipes with
large diameters.
The second category of water leakage detection methods is based on analyzing data
related to the water operation system. It includes statistical methods [10,11], the water
balance method [12], the minimum night flow method [6], the real-time transient model-
ing [13], and the negative pressure wave [14]. Leak detection using statistical methods is
based on the determination of the statistical characteristics of the water flow and pressure
in the water network and the determination of the outliers, which could be related to wa-
ter leakage. The efficiency of these methods is related to the quality of the recorded data
and the regularity of the consumption patterns. The water balance method relies on the
principle of mass conservation. A leak is identified if the difference between the amount
of water put into the water network and the sum of water consumption and usage exceeds
an established tolerance. The efficiency of this method depends on the quality of the mon-
itoring system and the knowledge of the water usage in the water network. The MNF
method is based on water flow analysis when the water demand is low and the water
pressure is high. A leak alarm is generated when the MNF exceeds a threshold, depending
on the water network’s characteristics and usage. This method is widely used; its effi-
ciency depends on the quality of the water network monitoring and the regularity of the
water usage. The real-time transient modeling method is based on comparing the hydrau-
lic recorded data with the results of hydraulic models. The efficiency of this method de-
pends on the quality of recorded data and the quality of the hydraulic models and their
calibration. The harmful pressure wave method is based on tracking acoustic waves cre-
ated by the water pressure drop resulting from the water leak. Pressure sensors are in-
stalled at the beginning and the end stations of the pipeline. The record of the generated
waves allows for the detection and localization of water leakage. This method is efficient
but suffers from high operating costs.
The large variety of developed methods highlights the great difficulty of detecting
and localizing water leakage in urban water distribution systems because of the complex-
ity.
The recent progress in the real-time monitoring of the water distribution systems has
been offering new opportunities to develop data-based methods for water leakage detec-
tion and localization. Machine learning-based methods have been widely used to detect
and localize water leakage in water distribution systems.
Caputo and Pelagagge [15] used artificial neural networks (ANNs) to detect and lo-
calize the water leak in water distribution systems. Data were generated using a hydraulic
model of the water network for various operating conditions and cases with different lo-
cations and amounts of the water leak. The method detected leaks correctly in small water
distribution systems. Salam et al. [16] used the radial basis function neural network
method for leak detection. The hydraulic software, EPANET, was used for data genera-
tion. The pressure variations in the water network were used as input data for the ANN
model, while the leak intensity and locations constituted the output parameters. The au-
thors showed that the method could detect the magnitude and the location of leakage with
a 98% accuracy. Mounce et al. [17] used the ANN method to identify anomalies in the
water distribution time series data in a pattern matching-based approach. This method
Smart Cities 2021, 4, 69 3 of 23

was based on the similarity research between new events and profiles established from
past events. This research allowed the classification of the new events and consequently
to identify abnormal events, which could be related to leak. Recently, Rojek and
Studzinski [18] used the ANN method to detect and localize water leakage in the water
distribution systems. Tests on real off-line data showed that the ANN method correctly
identified the localization of simulated leaks.
Zhang et al. [19] used the multiclass support vector machine method (SVM) for leak-
age detection in a large-scale water distribution network. First, the method K-means clus-
tering was used to subdivide the water network into leakage zones. Then, data with leak-
age events were generated using the Monte Carlo method together with the hydraulic
model. The authors showed that the multiclass SVM could identify the leakage zone using
flow and pressure data. However, Chan et al. [20] reported that this method faced a sig-
nificant challenge concerning determining the number of clusters and the high impact of
the random determination of the first cluster on the clustering process.
Soldevila et al. [21] used the K-nearest neighbors to classify data generated by the
hydraulic model EPANET from the simulation of leakage events at the totality of the
nodes of the water distribution network. Data were then used to train the K-Nearest
Neighbors model to localize the leakage area. The good performance of this method in the
localization of one water leak was assessed on three examples.
Ciupke [22] used the regression tree method to detect water leakage. Alerts were es-
tablished when the water flow exceeded the normal water flow range. The method was
tested on real examples and gave very good results, even for detecting small leaks.
Van der Walt et al. [23] analyzed the capacity of Bayesian probabilistic analysis, the
support vector machine, and an artificial neural network to detect and localize water leak-
age from pressure and flow data. These methods were compared to data generated from
numerical modeling and laboratory tests. Since analysis showed that the performances of
these methods depend on the complexity of the water network and the amount of availa-
ble data, the authors did not propose general recommendations for the use of the machine
learning methods for leak detection.
This literature review shows that intensive research has been conducted to use ma-
chine learning methods for leakage localization. However, the literature is still missing a
comparison of the different categories of machine learning methods to localize water leak-
age in the same water distribution system. This paper proposes to fill this gap by compar-
ing the capacities of various categories of machine learning methods to localize leakage in
a complex water distribution system based on the water network of the scientific campus
of Lille University in France.

2. Materials and Methods

2.1. Research Methodology
This research aimed at investigating the capacity of machine learning methods to lo-
calize the position of leakages in water distribution systems using flow and water pressure
data. Following the methodology proposed by different scholars [15,16,19,23], the hydrau-
lic software EPANET was used to create data related to different scenarios of leakage in
the water distribution system of the scientific campus. For each leakage scenario, EPANT
provided the water flow from the supply sections and the pressure in five hydraulic areas
of the campus. The generated data were then used for training and testing six machine
learning methods. The tests were first conducted with water flow and pressure data.
The performances of the machine learning methods were investigated using the pa-
rameters accuracy, precision, recall, and F1-score, which are determined from the confu-
sion matrix (Table 1):
𝑇𝑟𝑢𝑒𝑝𝑜𝑠𝑖𝑡𝑖𝑣𝑒
𝑃𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛 = (1)
𝑇𝑟𝑢𝑒𝑝𝑜𝑠𝑖𝑡𝑖𝑣𝑒 + 𝐹𝑎𝑙𝑠𝑒𝑝𝑜𝑠𝑖𝑡𝑖𝑣𝑒
Smart Cities 2021, 4, 69 4 of 23

𝑇𝑟𝑢𝑒𝑝𝑜𝑠𝑖𝑡𝑖𝑣𝑒
𝑅𝑒𝑐𝑎𝑙𝑙 = (2)
𝑇𝑟𝑢𝑒𝑝𝑜𝑠𝑖𝑡𝑖𝑣𝑒 + 𝐹𝑎𝑙𝑠𝑒𝑛𝑒𝑔𝑎𝑡𝑖𝑣𝑒

𝑃𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛 𝑅𝑒𝑐𝑎𝑙𝑙
𝐹1 − 𝑠𝑐𝑜𝑟𝑒 = 2 (3)
𝑃𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛 + 𝑅𝑒𝑐𝑎𝑙𝑙

𝑇𝑟𝑢𝑒𝑝𝑜𝑠𝑖𝑡𝑖𝑣𝑒 + 𝑇𝑟𝑢𝑒𝑛𝑒𝑔𝑎𝑡𝑖𝑣𝑒
𝐴𝑐𝑐𝑢𝑟𝑎𝑐𝑦 = (4)
𝑇𝑟𝑢𝑒𝑝𝑜𝑠𝑖𝑡𝑖𝑣𝑒 + 𝐹𝑎𝑙𝑠𝑒𝑝𝑜𝑠𝑖𝑡𝑖𝑣𝑒 + 𝑇𝑟𝑢𝑒𝑛𝑒𝑔𝑎𝑡𝑖𝑣𝑒 + 𝐹𝑎𝑙𝑠𝑒𝑛𝑒𝑔𝑎𝑡𝑖𝑣𝑒
The following sections present the generated data and the machine learning methods
used in this research.

Table 1. Confusion matrix.

Prediction
Positive Negative
Actual
Positive True Positive False Negative
Negative False Positive True Negative

2.2. Data generation

Data were generated using the software EPANET, developed by the Water Supply
and Water Resources Division (Formerly the Drinking Water Research Division) of the US
Environmental Protection Agency.
The water distribution network of the scientific campus of Lille University was used
as support for this research. This campus represents a small town with approximately 150
buildings and 25,000 users including students, faculty members, and technical, and ad-
ministrative staff [24]. Figure 1 illustrates the water distribution network of the campus
[25,26]. The water network is composed of 15 km of strongly meshed pipes. The water
company supplies the campus with water in three sections, located in the North, West,
and South of the campus (Figure 1).

Figure 1. The water distribution system of the scientific campus [25] (the red circles indicate the
water supply of the campus).

Figure 2 shows the EPANET hydraulic model of the campus. It includes 45 pipes and
33 junctions. The water network was divided into five hydraulic zones as indicated in
Smart Cities 2021, 4, 69 5 of 23

Figure 2. Data were generated by modeling the water leakage according to 215 leak sce-
narios (leak configuration), summarized in Table 2 and Figure 2. Zone 1 was the largest
and most complex zone. It included 62 leakage scenarios. Zones 4, 5, 2, and 3 had 47, 41,
35, and 41 leakage scenarios.
For each leak scenario, EPANET was used to determine the water flow from the three
supply sections (FL1, FL2, and FL3) and the pressure values at the five observation nodes
(Table 3, Figure 2).
Each leak scenario was modeled under two conditions. The first condition concerned
a constant pressure at the water supply sections, which were considered tanks with a con-
stant water height (H = 40 m). The second condition concerned the water leak, which was
considered by the following condition between the pressure (P) and water:
Q = CxPa (5)

The parameters C and a characterize the water leakage, which designate the emitter
coefficient and emitter exponent, respectively. Simulations were conducted with a = 0.5
and C = 1.
Table 4 provides a statistical analysis of the generated leak data. It shows that tank 1
provided the highest water supply rate (supply flow rate = 0.41), followed by tank 2 (flow
rate = 0.35). This means that the water supply of the campus was mainly provided from
the north and west of the campus, where the construction density was higher than that in
the south of the campus. The highest average pressure is observed in zone 3, located in
the South of the campus (average pressure approximately 35 m), followed by zones 5 and
4 (average pressure approximately 30 m). The average pressure in zones 1 and 2 was ap-
proximately 28 m.

Figure 2. EPANET model of the water network ono the scientific campus (45 pipes, 33 junctions).
Ln designates the position of leak n, Pzm designates the position of the pressure sensor number m.
The black squares indicate the supply section.
Smart Cities 2021, 4, 69 6 of 23

Table 2. Leakage scenarios were used for the generation of data (leak nodes are given in Figure 2).

Zone Position of Water Leak

L1, L2, L4, L7, L8, L10, L13, L18, L20
L10 + L11; L10 + L12; L10 + L13; L10 + L18; L10 + L2; L10 + L20; L10 + L3; L10 + L4; L10 + L7; L10 + L8; L11 +
1
L12; L11 + L13; L11 + L18; L11 + L2; L11 + L20; L11 + L3: L11 + L4; L11 + L7; L11 + L8; L12 + L13; L12 + L18;
(62 leak
L12 + L2; L12 + L20; L12 + L4; L12 + L7; L12 + L8; L13 + L18; L13 + L2; L13 + L20; L13 + L3; L13 + L4; L13 +
scenarios)
L7; L13 + L8; L18 + L2; L18 + L20; L18 + L3; L18 + L4; L18 + L7; L18 + L8; L2 + L20; L2 + L4; L2 + L7; L2+ L8;
L1 + L10; L1 + L11; L1 + L12; L1 + L13; L1 + L18; L1 + L2; L1 + L20; L1 + L3; L1 + L4; L1 + L7; L1 + L8
L5, L6, L9, L22, L23, L24, L25, L30
2
L23 + L22; L24 + L22; L24 + L23; L25 + L22; L25 + L23; L25 + L24; L30 + L22; L30 + L23; L30 + L24; L30 + L25;
(35 leak
L5 + L22; L5 + L23; L5 + L24; L5 + L25; L5 + L30; L6 + L22; L6 + L24; L6 + L25; L6 + L30; L6 + L5; L9 + L22; L9
scenarios)
+ L23; L9 + L24; L9 + L25; L9 + L30; L9 + L5; L9 + L6:
L41, L42, L43, L44, L46, L47, L48, L49, L50
3
L41 + L44; L41 + L46; L41 + L47; L41 + L48; L41 + L50; L42 + L44; L42 + L46; L42 + L47; L42 + L48; L42 +
(30 leak
L50; L43 + L44; L43 + L46; L43 + L47; L43 + L48; L43 + L50; L44 + L46; L44 + L47; L44 + L48; L44 + L50; L46 +
scenarios)
L47; L47 + L48; L47 + L50; L48 + L49; L50 + L49
L31, L32, L33, L34, L35, L36, L37, L38, L39, L40, L45
4 L31 + L36; L31 + L37; L31 + L38; L31 + L39; L31 + L45; L32 + L36; L32 + L37; L32 + L38; L32 + L39; L32 + L45;
(47 leak L33 + L36; L33 + L37; L33 + L38; L33 + L39; L33 + L45; L34 + L36; L34 + L37; L34 + L38; L34 + L39; L34 + L45;
scenarios) L35 + L36; L35 + L37; L35 + L38; L35 + L39; L35 + L45; L36 + L37; L36 + L38; L36 + L39L36 + L45;L37 + L38;
L37 + L39; L37 + L45; L38 + L39; L38 + L45; L39 + L45; L40 + L45
L14, L16, L17, L19, L21, L26, L27, L28, L29
5 L14 + L21; L14 + L26; L14 + L27; L14 + L28; L14 + L29; L15 + L21; L15 + L26; L15 + L27; L15 + L28; L15 + L29;
(41 leak L16 + L21; L16 + L26; L16 + L27; L16 + L28; L16 + L29; L17 + L21; L17 + L26; L17 + L27; L17 + L28; L17 + L29;
scenarios) L19 + L21; L19 + L26; L19 + L27; L19 + L28; L19 + L29; L21 + L26; L21 + L27; L21 + L28; L21 + L29;L26 + L27;
L26 + L28; L26 + L29; L27 + L28; L27 + L29; L28 + L29

Table 3. Pressure observation nodes (positions are given in Figure 2).

Zone 1 2 3 4 5
Pressure node PZ1 PZ2 PZ3 PZ4 PZ5

Table 4. Statistical descriptive parameters of the pressure and flow rate data.

Minimum Maximum Average Standard Deviation

FL1 (%) 0.13 0.71 0.4 0.15
FL2 (%) 0.23 0.62 0.35 0.82
FL3 (%) 0.60 0.59 0.23 0.11
PZ1 (m) 2.0 39.7 28.4 9.4
PZ2 (m) 1.4 39.4 27.4 9.3
PZ3 (m) 11.0 39.8 35.6 4.5
PZ4 (m) 1.8 39.2 29.2 10.2
PZ5 (m) 4.9 39.4 30.0 7.4

Figure 3 illustrates the impact of the leakage position on the flow rate ratios FL1, FL2,
and FL3. It shows that leakage in zones 1 and 2 caused a high flow rate from tank 1 (FL1),
a medium flow rate from tank 2 (FL2), and a low flow rate from tank 3 (FL3). Leakage in
zone 3 caused a high flow rate from tank 3 (FL3), medium flow from tank 2 (FL2), and low
flow from tank 1 (FL1). Leakage in zone 4 caused a high flow rate from tank 2 (FL2) and
low to medium flow from tank 3 (FL3). Finally, leakage in zone 5 caused a high flow rate
from tank 1 and tank 2 (FL1 and FL2) and flow from tank 3 (FL3). Table 5 summarizes the
impact of the leakage position on the water flow rate from the three supply sections. It can
Smart Cities 2021, 4, 69 7 of 23

be observed that a high flow rate from tank 2 (FL2) could be attributed to leakage in zone
4, and a high flow rate from tank 3 (FL3) could be attributed to leakage in zone 3. A high
flow rate from tank 1 (FL1) could be attributed to leakage in zone 1. The medium flow rate
from tanks 1 and 2 could be related to leakage in zones 2 and 5.
Figure 4 illustrates the impact of the leakage position on the pressures from PZ1 to
PZ5. It shows that leakage in each zone caused a significant drop in the pressure in the
leakage zone. It also shows a significant impact of some leakages in a zone on the pressure
in other zones, such as the impact of (i) leakage in zone 1 on the pressure in zone 2 (ii)
leakage in zone 2 on the pressure in zone 1, (iii) leakage in zone 3 on the pressure in zone
4, and (iv) leakage in zone 5 on the pressure in zone 2.

0,9

0,8
FL1 FL2 FL3

0,7

0,6

0,5
Flow rate

0,4

0,3

0,2

0,1

0
0 1 2 3 4 5 6
Zone

Figure 3. Impact of the leak localization on the water supply flow rate.

Table 5. Impact of the leak position on the water supply rate.

Leak Zone FL1 FL2 FL3

1 Strong Medium Low
2 Medium Medium Low to medium
3 Low Medium Strong
4 Low Strong Low to medium
5 Medium Medium Low to medium
Smart Cities 2021, 4, 69 8 of 23

45
PZ1 PZ2 PZ3 PZ4 PZ5
40

35

30
Pressure (m)

25

20

15

10

0
0 1 2 3 4 5 6
Zone

Figure 4. Impact of the leak localization on the pressure values at the observation points.

2.3. Use of Machine Learning Methods

Analyses were conducted with three supervised machine learning methods (logistic
regression, decision tree, and random forest), two unsupervised methods (hierarchical
classification and a combination of the principal component analysis (PCA) and the K-
means methods, and an artificial neural network (ANN). In addition, simulations were
conducted using the Kaggle platform (https://www.kaggle.com, accessed on 20 Septem-
ber 2021). The following sections briefly present the methods used in this research.
The logistic regression is used for binary classification [27]. The method used in this
work waws based on the functions:
h_(θ) (x) = g(θ^(T)x) (6)

g(z) = 1/(1 − e^(−z) ) (7)

where x is the input data, and θ is the parameter determined by the minimization of the
cost function.
The decision tree method is based on applying a series of questions to determine the
model response [28,29]. This method classifies a population into branch-like segments that
construct a tree with a root node, internal nodes, and leaf nodes. The model generates a
flowchart (tree), where each internal node (represented by a question) tests some features
and guides down through the branches (the result of the splitting) with a “gini” coeffi-
cient, which is defined as follows:

𝐺= 𝑝 𝑖 𝑥 1−𝑝 𝑖 (8)

The parameter c designates the number of total classes; p(i) is the probability of pick-
ing a data point with class i.
Random forest methods are used for both classification and regression by combining
randomized decision trees [30,31]. Each decision tree gives a vote for a target variable. The
random forest algorithm chooses the combination that obtains the highest vote. This
method has high predictive accuracy; it is efficient on large data sets and works well with
missing data. However, it suffers from interpretation difficulties and overfitting in the
case of noisy data.
The hierarchical classification method is used to build a hierarchy of clusters. The
results of clustering are usually presented in a dendrogram. Hierarchical classification
Smart Cities 2021, 4, 69 9 of 23

could be conducted using (i) a “bottom-up” approach, where each observation starts in
its cluster, and pairs of clusters are merged as one move up the hierarchy or (ii) a “top-
down” approach, where all observations start in one cluster; splits are performed when
moving down the hierarchy. The Ward method is used in this analysis [32]. The PCA
method is used to reduce the input data dimension by focusing on the principal compo-
nents [33].
K-means clustering is a type of unsupervised learning. It aims at partitioning n ob-
servations into k clusters [34]. Initially, K initial means are randomly generated. Then, K
clusters are created by associating each observation with the nearest centroid. Next, the
objective function, the sum of the distance, is optimized until the best cluster centers can-
didates are found. Finally, data points are clustered based on feature similarity.
The ANN is inspired from the human brain functioning [35]. It transforms the input
data (input layer) through a series of neural layers (hidden layers) to output data (output
layer). The transformation is based on weights, which are adjusted by optimizing the pre-
diction of a training data set. The Sigmoid function is used in data transformation.

2.4. Input and Output Parameters

Machine learning methods were used with the following input and output parame-
ters:
Input parameters: water flow at the three supply sections (FL1, FL2, FL3) and water
pressure values at five observation points (Pz1, Pz2, Pz3, Pz4, Pz5)
Output parameter: number of the campus zone (Z1, Z2, Z3, Z4, Z5).

3. Results
3.1. Supervised Methods
The training phase of the supervised methods was conducted with 80% of the data,
while 20% was used for the testing phase.
Table 6 summarizes the results obtained with the water supply flow data. It shows
that both the logistic regression and random forest methods gave excellent results with an
accuracy = 1.0, precision = 1.0, recall = 1.0, and F1-score = 1. The decision tree method gave
very good results with an accuracy = 0.95, precision = 0.96; recall = 0.95, and F1-score =
0.95.
Figure 5 shows the confusion matrix of the decision tree method. It indicates excellent
performances for zones 1, 2, and 5. For zone 3, the precision was equal to 0.78, and for
zone 4, the recall was equal to 0.75.

Table 6. Classification report for the supervised methods—flow data.

Method Accuracy Precision Recall F1-Score

Logistic Regression 1.0 1.0 1.0 1.0
Decision Tree 0.95 0.96 0.95 0.95
Random Forest 1.0 1.0 1.0 1.0
Smart Cities 2021, 4, 69 10 of 23

1 12

6
2

3 7

4 2 6

5 5

1 2 3 4 5

Figure 5. Confusion matrix for the decision tree method—flow data.

Table 7 summarizes the results obtained with the pressure data. It can be observed
that both the logistic regression and the random forest methods gave excellent results with
an accuracy = 1.0, precision = 1.0, and a recall = 1.0. The decision tree method gave good
results with an accuracy = 0.88, precision = 0.91, recall = 0.94, and F1-score = 0.91. Figure 6
shows the confusion matrix for the decision tree method. It indicates excellent perfor-
mances for all the zones, except for zone 1 (recall = 0.70) and the zone 2 (precision = 0.54).
The bad results for zones 1 and 2 could be related to the spatial proximity of these zones
and their hydraulic interaction (Figures 2 and 4).

Table 7. Classification report for the supervised methods—pressure data.

Method Accuracy Precision Recall F1-Score

Logistic Regression 1.0 1.0 1.0 1.0
Decision Tree 0.88 0.91 0.94 0.91
Random Forest 1.0 1.0 1.0 1.0

1 12 5

2 6

3 7

4
8

5
5

1 2 3 4 5

Figure 6. Confusion matrix for the decision tree method—pressure data.

Smart Cities 2021, 4, 69 11 of 23

Pressure and flow data were used with only the decision tree method. The logistic
regression and the random forest methods gave excellent results with either the flow or
pressure data. Table 8 summarizes the classifications report for the decision tree. It shows
that this method gives excellent results with an accuracy of 0.98, precision of 0.97, recall
of 0.97, and F1-score of 0.96. It can be observed that the performance obtained with the
flow and pressure data was better than that obtained with the flow data (Table 6) and
pressure data (Table 7). Figure 7 shows the confusion matrix for the decision tree method.
It indicates excellent performances for all the zones, except for zone 2 (recall = 0.83) and
the zone 5 (precision = 0.83).

Table 8. Classification report for the decision tree method—flow and pressure data.

Method Accuracy Precision Recall F1-Score

Decision Tree 0.98 0.97 0.97 0.96

17
1

5 1
2

7
3

4 8

5 5

1 2 3 4 5

Figure 7. Confusion matrix for the decision tree method—flow and pressure data.

3.2. Unsupervised Methods

Figure 8 shows the results obtained with the hierarchical classification method with
the pressure data. It shows the existence of three groups: the first group, G1, was com-
posed of the pressures in zones 3 and 4; the second group, G2, concerned the pressure in
zone 1; the third group, G3, included the pressures in zones 2 and 5.
Figure 9 illustrates the classification with water flow rate and pressure data. It shows
the presence of fiive groups: G1 with the pressure in zone 4; G2 with the flow rate FL1 and
the pressure in zone 3, G3 with the pressures in zones 2 and 5; G4 with the flow rate FL2,
G5 with the flow rate FL3 and the pressure in zone 1.
Smart Cities 2021, 4, 69 12 of 23

1.6

Distance
0.8

0
PZ3 PZ4 PZ1 PZ2 PZ5

Figure 8. Hierarchical Classification method-results with the pressure data.

2.0
Distance

1.0

0
PZ4 FL1 PZ3 PZ2 PZ5 FL2 FL3 PZ1

Figure 9. Hierarchical classification method—results with the flow rate and pressure data.

Figure 10 illustrates the results obtained by applying the PCA and K-means methods
on the flow rate data for k = 5 clusters. The component PC1 shows three clusters and two
partially overlapping clusters. The component PC2 indicates two clusters and three par-
tially overlapping clusters. Thus, in the (PC1, PC2) plan, the five clusters could be well
distinguished.
Smart Cities 2021, 4, 69 13 of 23

Figure 10. PCA and K-means clustering—flow rate data.

Figure 11 illustrates the results obtained with the pressure data for k = 5 clusters. Both
PC1 and PC2 showed significant overlapping clusters. In the (PC1, PC2) plan, the five
clusters were not well distinguished.

Figure 11. PCA and K-means clustering-Pressure data.

Figure 12 illustrates the results obtained with flow rate and pressure data for k = 5
clusters. Both PC1 and PC2 showed clusters overlapping. In the (PC1, PC2) plan; only
three clusters could be well distinguished.
Smart Cities 2021, 4, 69 14 of 23

Figure 12. PCA and K-means clustering—flow rate and pressure data.

3.3. Artificial Neural Network

Analyses were conducted with a multilayer backpropagation neural network model.
Figure 13 shows the results of the application of the ANN method with the water supply
data. It indicates a rapid convergence of the ANN model. Indeed, a good convergence was
observed with approximately 20 epochs. The ANN model gave excellent results with an
accuracy = 1.0, precision = 1.0, recall = 1.0, and F1-Score = 1.0 (Table 9).

Figure 13. Application of the ANN method with the water supply data.

Figure 14 illustrates the results obtained with the pressure data. The convergence of
the training phase was obtained with approximately 150 epochs, while the convergence
of the validation phase was achieved with around 50 epochs. The model gave excellent
results with an accuracy = 1.0, precision = 1.0, recall = 1.0, and F1-Score = 1.0 (Table 9).
Figure 15 illustrates the results obtained with flow and pressure data. It indicates a
convergence of the training stage with approximately 100 epochs and convergence of the
validation stage with approximately 50 epochs. The model gave excellent results with an
accuracy = 1.0, precision = 1.0, recall = 1.0, and F1-Score = 1.0 (Table 9).
Smart Cities 2021, 4, 69 15 of 23

Figure 14. Application of the ANN method with pressure data.

Figure 15. Application of the ANN method with flow and pressure data.

Table 9. Results obtained with the ANN.

Data Accuracy Precision Recall F1-Score

Flow data 1.0 1.0 1.0 1.0
Pressure data 1.0 1.0 1.0 1.0
Flow and pressure data 1.0 1.0 1.0 1.0

3.4. Analysis of the Water Leak in the Scientific Campus of Lille University
This section presents an analysis of the leak in the scientific campus of Lille Univer-
sity. The analysis was based on daily flow data collected in 2015 at the three supply sec-
tions: FL1 in the North, FL2 in the west, and FL3 in the South. The year 2015 was selected
because of the availability of data for this year and the observation of several abnormal
events in the water consumption, related to water leakage. The water usage in the campus
concerns mainly domestic activities in the students’ residences, academic activity, and
buildings’ cleaning. Water is not used for irrigation. Since the water usage is related to
regular activities, the water consumption at the daily scale is expected to be regular.
The following sections present a successively analysis of the daily water consump-
tion and leakage detection and localization.

3.5. Analysis of the Daily Water Consumption (Qd)

Figure 16 illustrates the variation of the daily water consumption of the campus (Qd).
It indicates missing data in the period from May 3 to May 27. This period is not considered
Smart Cities 2021, 4, 69 16 of 23

in the analysis. This figure shows a significant variation in Qd. The minimum daily con-
sumption was equal to 414 m3, while the maximum was equal to 1680 m3 and the average
consumption was equal to 890 m3. Low daily consumption values could be attributed to
the vacation periods, while the high daily consumption values could be associated with
water leakage.
Figures 17 and 18 illustrate the repartition of the daily water supply among the three
supply sections. They show that the daily water supply from the North (F1D) was higher
than those from the west and south campus. It also had the most significant variation
(Table 10): the minimum daily supply was equal to 100 m3, while the maximum was equal
to 772 m3 and the average was equal to 442 m3; compared with F2D (Resp F3D): minimum
= 197 m3 (Resp. 51 m3), maximum = 772 m3 (Resp. 454 m3) and average = 251 m3 (Resp. 197
m3). Thus, the water supply F1D accounted for about 50% of the total water supply, while
F2 accounted for 28% and F3 for 22% of the campus water supply.

1800

1600

1400

1200

1000
D ia ly C o n su m p tio n (m 3 )

800

600

400

200

0
0 50 100 150 200 250 300 350 400
Day

Figure 16. Variation of the daily water consumption (Qd) of the Scientific Campus in 2015.

900

800 F1D
F3D
700 F2D

600

500
D ia ly C o n su m p tio n (m 3 )

400

300

200

100

0
0 50 100 150 200 250 300 350 400
Day

Figure 17. Variations in the repartition of the daily water supply on the scientific campus in 2015.
Smart Cities 2021, 4, 69 17 of 23

Scattergrams

F2D

F3D

F1D

0 100 200 300 400 500 600 700 800

Strip plots

F2D

F3D

F1D

0 100 200 300 400 500 600 700 800

Figure 18. Scatter grams and strip plots of the repartition of the daily water supply on the scientific
campus in 2015.

Table 10. Statistical descriptive analysis of the daily water supply of the campus.

F1D F2D F3D Total (Qd)

m3/Day m3/Day m3/Day m3/Day
Minimum 100 197 51 414
Maximum 772 462 454 1680
Average 442 251 197 890
Standard deviation 143 33 59 219

3.6. Leakage Analysis

The identification of leakage events was based on the observation of abnormal water
consumption. Figures 19 and 20 show the events with water consumption exceeding 1200
m3/h (average water consumption + 1.5 standard deviation). We observed the presence of
five groups of events, which are summarized in Table 11. The first group (G1) corresponds
to day 76 with consumption exceeding by approximately 464 m3 of the water consumption
average (Qav), followed by day 86 (G2) which exceeded Qav by 326 m3. The third group
corresponds to days 260 and 261, with water consumption exceeding Qav by 390 and 467
m3. The fourth group is related to days 264–275, with water consumption exceeding Qav
by values included in the interval 311–790 m3. The last leak (G5) occurred on day 327, with
consumption exceeding Qav by 318 m3.
Figure 21 and Table 12 show the repartition of the water supply ratios related to leak-
age events. It shows that the ratio associated with FL1 was higher than those associated
with FL2 and FL3. FL1 accounted for 56% of the total water supply for groups G1, G2, and
G5, while FL2 and FL3 accounted for approximately 22% each. For groups G3 and G4, FL1
Smart Cities 2021, 4, 69 18 of 23

accounted for approximately 46% of the water supply, while FL2 and FL3 accounted for
approximately 27% each.

1800

1600

1400

1200

1000
Dialy Consumption (m3)

800

600

400

200

0
0 50 100 150 200 250 300 350 400
Day

Figure 19. Variation of the daily water consumption of the campus–Events with consumption exceeding 1200 m3 (average
water consumption + 1.5 standard deviation) could be related to leakage, (line in the red color).

Scattergram (Ftot)

400 600 800 1000 1200 1400 1600 1800

Ftot

Strip plot (Ftot)

400 600 800 1000 1200 1400 1600 1800

Ftot

Figure 20. Scatter grams and strip plots of the distribution of the daily water consumption of the
campus; events with consumption exceeding 1200 m3 (average water consumption + 1.5 standard
deviation) could be related to leakage (line in the red color)

Table 11. Leak events in the water distribution of the scientific campus.

Day Group Qd (m3/Day) Qd-Average (m3/Day)

76 G1 (76) 1354 464
86 G2 (86) 1216 326
260 G3 (260. 261) 1280 390
261 G3 (260. 261) 1357 467
264 G4 (264–2675) 1383 493
265 G4 (264–2675) 1463 573
266 G4 (264–2675) 1477 587
267 G4 (264–2675) 1404 514
268 G4 (264–2675) 1396 506
269 G4 (264–2675) 1217 327
270 G4 (264–2675) 1201 311
Smart Cities 2021, 4, 69 19 of 23

271 G4 (264–2675) 1435 545

272 G4 (264–2675) 1459 569
273 G4 (264–2675) 1455 565
274 G4 (264–2675) 1624 734
275 G4 (264–2675) 1680 790
327 G5 (327) 1208 318

1800

1600

1400

1200

1000

800

600

400

200

0
G1 (76)
G3 (260, 261)
G4 (264-2675)
G4 (264-2675)
G4 (264-2675)
G4 (264-2675)
G4 (264-2675) Ftot
F1D
G4 (264-2675) F2D
G5 (327) F3D

Figure 21. Repartition of the water supply ratios related to leak events.

Table 12. Repartition of the water supply ratios related to leak events.

Day Groupe FL3 (%) FL2 (%) FL1 (%)

76 G1 (76) 20 23 57
86 G2 (86) 21 22 57
260 G3 (260, 261) 26 26 48
261 G3 (260, 261) 26 26 48
264 G4 (264–2675) 28 26 46
265 G4 (264–2675) 27 26 47
266 G4 (264–2675) 28 27 45
267 G4 (264–2675) 29 27 44
268 G4 (264–2675) 28 27 44
269 G4 (264–2675) 28 29 42
270 G4 (264–2675) 29 30 41
271 G4 (264–2675) 28 27 46
272 G4 (264–2675) 28 26 46
273 G4 (264–2675) 28 26 46
274 G4 (264–2675) 27 26 47
275 G4 (264–2675) 27 27 45
327 G5 (327) 22 23 55

3.7. Leakage Localization

For the localization of leakage events G1 to G5, the water supply ratios corresponding
to the leakage events are reported in Figure 22. The ratios of water flow for the events G1,
G2 and G5 are indicated by the water flow ratios (FL1, FL2, and FL3) for zone 1, while
Smart Cities 2021, 4, 69 20 of 23

those related to the events G3 and G4 are reported by the water flow ratios for zone 2.
Therefore, it could be observed that leakages G1, G2, and G5 well matched with the water
flow repartition for leakages in zone 1, while leakages G3 and G4 well matched with the
water flow repartition for leakages in zone 2. This observation indicates that leakages G1,
G2, and G5 could be attributed to zone 1, while leakages G3 and G4 could be attributed to
zone 2.

Figure 22. Localization of the leakage events (G1 to G5) on the campus.

4. Discussion
This research concerned the detection and localization of leaks in urban water distri-
bution networks. This issue is of significant concern in the management of the water dis-
tribution systems, because leaks in the water distribution system cause substantial eco-
nomic, social, and environmental impacts and severe damages to the surrounding soils
and infrastructures.
Despite the important research on the development and use of hardware- and soft-
ware-based methods for the detection, localization, and localization of water leaks, pro-
fessionals still need efficient and cost-effective methods to detect water leaks in complex
water distribution systems.
The recent progress in smart monitoring and artificial intelligence provides signifi-
cant opportunities to develop data-based methods for leak detection and localization. The
literature review showed an important concern in the use of these methods. However, on
the one hand, the majority of the applications using artificial intelligent methods remain
at the research stage. On the other hand, the literature review revealed a lack of compre-
hensive use of these methods. This research aimed to fill the gap in this area by thoroughly
investigating the machine learning methods to detect and localize leaks in the water dis-
tribution system.
The water network of the scientific campus of Lille University was used as support
for this research. This use was motivated by the campus’ representativity of a small town,
the complexity of the water network and the availability of data about the water network
asset and water consumption. The water network is monitored by approximately 93 au-
tomated meter readings (AMRs) that record the water supply and consumption in the
main buildings at an hourly time interval.
The physical water network was completed by constructing a Lab pilot of this net-
work to investigate, under well-controlled conditions, the impact of the position of a leak
on the water flow rates. Results of experiments showed an evident influence of the leak
position leak on the water supply flow rates when the leak was in the proximity of the
water supply. However, the impact is unclear for other locations, which means that the
Smart Cities 2021, 4, 69 21 of 23

leak position could not be systematically determined from only the supply flow rates. In
the future, it could be interested to monitor the pilot with pressure cells to investigate the
possibility of improving the leak localization using the water supply flow rates and the
pressure variation in the water network.
A large data set was built regarding the impact of leaks on the water network of the
scientific campus on the variations in the water supply flow rates and the pressure in five
campus zones. This data set was constructed using the hydraulic software EPANET. The
data set included the responses of the water network to 215 individual and double leaks.
The data set was used for training and testing the following six machine learning
methods:
• Three supervised methods: logistic regression, decision tree, and random forest;
• Two unsupervised methods: The hierarchical classification method and a combina-
tion of the PCA and K-means classification method;
• The ANN
The results of the tests conducted on these methods showed:
• Excellent performance of the supervised methods in the localization of leaks in the
water network. Both the logistic regression and the random forest predicted the po-
sition of the leak with an accuracy = 1.0. In contrast, the decision tree predicted leaks
with an accuracy = 0.98 with pressure and flow data;
• Excellent performances by the ANN for the localization of water leaks in the water
network (accuracy = 1.0);
• Some difficulties in exploiting the clustering capacity of the unsupervised methods
in the leak localization because of overlapping clusters.
The results of this research were used to investigate the position of water leaks in the
campus using water flow data rates recorded in 2015. Unfortunately, difficulties were en-
countered in the determination of the position of leaks because of a lack of pressure data.
Therefore, in the future, we recommend extending the monitoring of the campus water
network by adding cell pressure on the campus and flow rates in critical sections of the
water network.

5. Conclusions
This paper presented an investigation of the use of machine learning methods to local
leakage in the water distribution network. Leakage localization was based on the creation
of hydraulic zones in the water distribution network. For each zone, sensors are used to
measure the water supply variations and the water pressure. Collected data were then
used for the construction of the machine learning models.
This methodology was used to investigate the capacity of six machine learning meth-
ods to localize leaks in the water distribution network of the scientific campus of Lille
University. Data were generated using EPANET software. The investigation showed (i)
excellent performance from the supervised methods, in particular, the logistic regression
and random forest; (ii) excellent performances by the artificial neural network; (iii) diffi-
culties in the exploitation of the clustering capacity of the unsupervised methods in leak
localization because of clusters’ overlapping. Offline water supply flow data were then
used for the localization of water leakage in the scientific campus. The results gave some
indications about the localization of the water leakage.
This paper shows that the ANN and the supervised logistic regression and random
forest methods performed well in the localization of the water leakage in the water distri-
bution systems, mainly when using both water flow and pressure data. These results are
based on data generated using the software EPANET. Therefore, they should be con-
firmed on data collected from complex water networks, including water supply flow and
pressure data in the subzones of the water network, and the localization of leakage events.
Smart Cities 2021, 4, 69 22 of 23

Author Contributions: N.M., I.S., and J.E.K. conceived the research idea; A.A. and I.S. established
the research methodology; N.M., N.A., and I.S. conducted the data analysis and discussed the re-
sults. All authors have read and agreed to the published version of the manuscript.
Funding: This research did not beneficiate from specific funding support.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Data sharing not applicable.
Conflicts of Interest: All other authors have no conflict of interest.

References
1. Kingdom, B.; Liemberger, R.; Marin, P. The Challenge of Reducing Non-Revenue (NRW) Water in Developing Countries. How the
Private Sector Can Help: A Look at Performance-Based Service Contracting; Water Supply and Sanitation (WSS) Sector Board Discus-
sion Paper N. 8; The World Bank: Washington, DC, USA, 2006.
2. Thornton, J.; Sturm, R.; Kunkel, G. Water Loss Control, 2nd ed.; McGraw Hill: New York, NY, USA, 2008; ISBN 9780071499187.
3. Renzetti, S.; Dupont, D. Buried Treasure: The Economics of Leak Detection and Water Loss Prevention in Ontario. Environmen-
tal Sustainability Research Centre (ESRC) Working Paper Series. 2013. Available online: http://hdl.handle.net/10464/4279 (ac-
cessed on 30 June 2021).
4. Kanakoudis, V.K. A troubleshooting manual for handling operational problems in water pipe networks. Water Supply Res. Tech-
nol.-AQUA 2004, 53, 109–124.
5. Hunaidi, O.; Wang, A.; Bracken, M.; Gambino, T.; Fricke, C. Acoustic methods for locating leaks in municipal water pipe net-
works. In Proceedings of the International Conference on Water Demand Management, Dead Sea, Jordan, 30 May–3 June 2004;
pp. 1–14.
6. Adegboye, M.A.; Fung, W.-K.; Karnik, A. Recent Advances in Pipeline Monitoring and Oil Leakage Detection Technologies:
Principles and Approaches. Sensors 2019, 19, 2548. https://doi.org/10.3390/s19112548
7. Fahmy, M.; Moselhi, O. Detecting and locating leaks in underground water mains using thermography. In Proceedings of the
26th International Symposium on Automation and Robotic in Construction (ISARC), Austin, TX, USA, 24–27 June 2009; pp. 61–
67.
8. Ayala–Cabrera, D.; Herrera, M.; Izquierdo, J.; Ocaña–Levario, S.; Pérez–García, R. GPR-Based Water Leak Models in Water
Distribution Systems. Sensors 2013, 13, 15912–15936.
9. Noran, P.; Obenauf, P. Asset Management of a Failing 36" Ductile Iron Sewage Force Main. In: Pipelines 2010; American Society
of Civil Engineers: Reston, WV, USA, 2010; pp 566-576, doi:doi:10.1061/41138(386)55.
10. Zhang, X. Statistical leak detection in gas and liquid pipelines. Pipes Pipelines Int. 1993, 38, 20–26.
11. Buchberger, S.G.; Nadimpalli, G. Leak Estimation in Water Distribution Systems by Statistical Analysis of Flow Readings. J.
Water Resour. Plan. Manag. 2004, 130, 321–329, doi:10.1061/(ASCE)0733-9496(2004)130:4(321).
12. Lambert, A. International report: Water losses management and techniques. Water Sci. Technol. Water Supply 2002, 2, 1–20.
13. Billmann, L.; Isermann, R. Leak detection methods for pipelines. Automatica 1987, 23, 381–385.
14. Silva, R.; Buiatti, C.; Cruz, S.; Pereira, J. Pressure wave behaviour and leak detection in pipelines. Comput. Chem. Eng. 1996, 20,
S491–S496.
15. Caputo, A.C.; Pelagagge, P.M. Using Neural Networks to Monitor Piping Systems. Process Saf. Prog. 2003, 22, 119–127.
https://doi.org/10.1002/prs.680220208.
16. Salam, A.E.U.; Tola, M.; Selintung, M.; Maricar, F. On-line monitoring system of water leakage detection in pipe networks with
artificial intelligence. ARPN J. Eng. Appl. Sci. 2014, 9, 1817–1822.
17. Mounce Stephen, R.; Mounce, R.B.; Boxall, J.B. Novelty detection for time series data analysis in water distribution systems
using support vector machines. J. Hydroinformatics 2011, 13, 672–686. https://doi.org/10.2166/hydro.2010.144.
18. Rojek, I.; Studzinski, J. Detection and localization of water leaks in water nets supported by an ICT system with artificial intel-
ligence methods as a way forward for smart cities. Sustainability 2019, 11, 518, https://doi.org/10.3390/su11020518.
19. Zhang, Q.; Wu, Z.Y.; Zhao, M.; Qi, J.; Huang, Y.; Zhao, H. Leakage Zone Identification in Large-Scale Water Distribution Sys-
tems Using Multiclass Support Vector Machines. J. Water Resour. Plan. Manag. 2016, 142, 4016042.
https://doi.org/10.1061/(ASCE)WR.1943-5452.0000661.
20. Chan, T.K.; Chin, C.S.; Zhong, X. Review of Current Technologies and Proposed Intelligent Methodologies for Water Distrib-
uted Network Leakage Detection. IEEE Access 2018, 6, 78846–78867. https://doi.org/10.1109/ACCESS.2018.2885444.
21. Soldevila, A.; Blesa, J.; Tornil-Sin, S.; Duviella, E.; Fernandez-Canti, R.M.; Puig, V. Leak localization in water distribution net-
works using a mixed model-based/data-driven approach. Control Eng. Pract. 2016, 55, 162–173.
https://doi.org/10.1016/j.conengprac.2016.07.006.
Smart Cities 2021, 4, 69 23 of 23

22. Ciupke K. Leak Detection Using Regression Trees. In: Timofiejczuk A., Łazarz B., Chaari F., Burdzik R. (eds) Advances in Tech-
nical Diagnostics. ICDT 2016. Applied Condition Monitoring, vol 10. Springer, Cham. 2018, https://doi.org/10.1007/978-3-319-
62042-8_28.
23. van der Walt, J.C.; Heyns, P.S.; Wilke, D.N. Pipe network leak detection: Comparison between statistical and machine learning
techniques. Urban Water J. 2018, 15, 953–960. https://doi.org/10.1080/1573062X.2019.1597375.
24. Shahrour, I.; Abbas, O.; Abdallah, A.; AbouRjeily, Y.; Afaneh, A.; Aljer, A.; Ayari, B.; Farrah, E.; Sakr, D.; Al Masri, F. Lessons
from a Large-Scale Demonstrator of the Smart and Sustainable City BT—Happy City—How to Plan and Create the Best Livable Area for
the People; Brdulak, A.; Brdulak, H., Eds.; Springer International Publishing: Berlin/Heidelberg, Germany, 2017; pp. 193–206.
https://doi.org/10.1007/978-3-319-49899-7_11.
25. Farah, E.; Shahrour, I. Leakage Detection Using Smart Water System: Combination of Water Balance and Automated Minimum
Night Flow. Water Resour. Manag. 2017, 31, 4821–4833. https://doi.org/10.1007/s11269-017-1780-9.
26. Farah, E.; Abdallah, A.; Shahrour, I. SunRise: Large scale demonstrator of the smart water system. Int. J. Sustain. Dev. Plan. 2017,
12, 112–121. https://doi.org/10.2495/SDP-V12-N1-112-121.
27. Harrell, F.E. Regression Modeling Strategies; Springer: Berlin/Heidelberg, Germany, 2001; ISBN 0-387-95232-2.
28. Breiman, L.; Friedman, J.H.; Olshen, R.A.; Stone, C.J. Classification and Regression Trees; Wadsworth, Inc.: Belmont, CA, USA,
1984.
29. Lin, N.; Noe, D.; He, X. Tree-based methods and their applications In Springer Handbook of Engineering Statistics; Pham, H., Eds.;
Springer: London, UK, 2006; pp. 551–570.
30. Prinzie, A.; Van den Poel, D. Random Forests for multiclass classification: Random MultiNomial Logit. Expert Syst. Appl. 2008,
34, 1721–1732, doi:10.1016/j.eswa.2007.01.029.
31. Breiman, L. Random Forests. Mach. Learn. 2001, 45, 5–32, doi:10.1023/A:1010933404324.
32. Ward, J.H., Jr. Hierarchical Grouping to Optimize an Objective Function. J. Am. Stat. Assoc. 1963, 58, 236–244.
33. Mohamad Asri, M.N.; Mat Desa, W.N.S.; Ismail, D. Combined Principal Component Analysis (PCA) and Hierarchical Cluster
Analysis (HCA): An efficient chemometric approach in aged gel inks discrimination. Aust. J. Forensic Sci. 2020, 52, 38–59,
doi:10.1080/00450618.2018.1466913.
34. Hartigan, J.A.; Wong, M.A. Algorithm AS 136: A K-Means Clustering Algorithm. J. R. Stat. Soc. Ser. C 1979, 28, 100–108.
35. Dave, V.S.; Dutta, K. Neural network-based models for software effort estimation: A review. Artif. Intell. Rev. 2014, 42, 295–307.