On Smart Water System Developments: A Systematic Review

Abstract

Water is an essential resource for life and is also a necessary resource for the sustainable economic competitiveness of any country. In recent decades, climate change, economic development, and rising population have led to water scarcity in certain regions. In response, new technologies and water management techniques have been researched and developed, which are now incorporated into the concept of smart cities. These innovations, called smart water systems, aim to enhance water management by monitoring consumption, quality, reservoir levels, leaks, and asset conditions, and optimizing water processes to maximize water system resilience. The first systems were based on smart meters and have advanced to so-called digital twins for water systems. This review aims to present a comprehensive review of smart water system developments, the geographic distribution of the works, their technological readiness level, and their implementation challenges.

1. Introduction

Water is an indispensable resource for a country to achieve development and economic competitiveness. Due to this fact, water demand is growing globally. Another factor increasing the water demand is the growing population and the issues associated with it, for example, the water required to produce food and usable household water. Worldwide, freshwater withdrawals can be roughly classified as 70% for agriculture, 18% for industry, and about 12% for domestic use according to UNESCO. From these quantities, around 25% of irrigation water and half of the water for domestic purposes is supplied by groundwater. It highlights its essentiality for the development and wellness of society [1].

Currently, approximately 50% of the world’s population live in regions experiencing periods of severe water scarcity, which can last for a few months or include all year round. Along with this issue, another important factor which affects water supplies is pollution, which reduces the freshwater availability [2]. Pollution is present in both rich and poor countries. While in lower-income countries, it is mostly due to poor management of wastewater treatment, in higher-income countries, the main problem is the runoff from agricultural activities. Additional to the aforementioned problems, climate change also contributes by means of global warming and directly affects the global water cycle, which in turn increases the frequency of droughts and floods yielding to severe weather and climate events. Clearly, all sectors of the economy can be affected by these factors, for instance, healthcare, agriculture, fisheries, industry, transportation, tourism, and recreation [3]. In this way, due to the possibility of mitigating or eliminating factors contributing to climate change (e.g., pollution, environmental degradation, poor water management), smart cities have been a hot research topic in the last few years. As a result, new paradigms on urban planning, management, and monitoring are constantly emerging.

Smart cities implement a transformation of digital tools such as sensors, IoT (Internet of Things), artificial intelligence, cloud services and computing, and cybersecurity to measure, forecast, and manage water usage in facilities in order to prevent and minimize reaction times to faults and damages in water assets (pipelines, pumps, valves, among others). A smart water system in an urban environment is a crucial component of a smart city, as it enables the management of water networks in order to monitor, supervise, and optimize resources effectively. Another important relation between smart cities and water systems is the optimization of the cleaning and distribution process of water, because providing clean drinking water is instrumental in promoting health practices that can help to reduce mortality and illness rates. This is particularly crucial for children and elderly people who are more vulnerable to contracting diseases from polluted water and it is a critical consideration in regions where water availability is limited, emphasizing the need for efficient water usage [4]. For instance, wastewater treatment plants are facilities that help to alleviate the environmental degradation caused by wastewater or sewage by removing and eliminating impurities in water that have been used in domestic, commercial, and industrial activities, returning it to the environment or reusing it in other processes [5]. For this and other reasons, water stakeholders have pursued the concept of smart cities by means of which technology and artificial intelligence can be merged to optimize processes and activities in water systems yielding to the concept of smart water systems.

In the literature, there are several definitions for a smart water system, but most of them can be condensed into the following one: a smart water system is a fusion of a network of sensors embedded with software and processing power that allows the measuring of water parameters (physical or chemical) and evaluating water network assets via measuring several data as level, flow, pressure, vibration, sounds, moisture, cameras, etc., to control, monitor or supervise, and make the data available online using IoT technology to optimize water management, distribution, and use, while increasing the benefits of the stakeholders [6].

In this work, a systematic literature review of smart water system technologies is presented using the well-known Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) protocol. The main objective of the systematic review is providing a comprehensive overview of the current state of smart water technology development and which regions or countries are having more activity in this field. The selection of the relevant literature was conducted in multiple stages, which is summarized in Figure 1. Section 3 provides further details on this process.

1.1. Research Questions

According to the objectives of the work, answering the following research questions (RQ) is fundamental.

• RQ₁: How has the state-of-the-art regarding smart water systems developed in the last 10 years?
• RQ₂: What countries have more published reports concerning smart water systems?
• RQ₃: What are the challenges in implementing a smart water system using current technology?

1.2. Contributions

The main contributions of this work are the following:

• A clear view of the evolution of smart water systems in the last 10 years. The systematic literature review presents a list of current literature on smart water systems and some statistics related to the literature found are presented and examined to gain insight into trends and technological developments and applications. Furthermore, the works are classified by their technology readiness level (TRL).
• Worldwide countries and institutions reporting more work on smart water systems are analyzed. The geographic distribution of the literature found is presented. After that, an in-depth analysis is carried out to refine the documents found, and then they are classified by the geographical localization of the authors.
• Challenges and limitations are evaluated for smart water system research and implementation. Some of the major challenges discussed in the literature are presented, e.g., challenges related to instrumentation, data processing, cybersecurity, and stakeholder engagement.

The remainder of this work is as follows. Section 2 presents a conceptual background. Section 3 details the review methodology used to develop this systematic review. Section 4 presents the results and findings to answer the research questions. Section 5 discusses the implementation challenges, limitations, and future research. Finally, Section 6 concludes the paper.

2. Background

In the previous section, a definition for a smart water system was introduced. This definition is based on those definitions existing in the literature [7,8,9,10,11,12,13,14]. In Figure 2, the ecosystem for smart water systems is shown. Roughly, it is three-layered, the physical world where the physical process takes place, the middle layer or the cyber-physical world, which is the interface between the physical world, and the last layer, the digital world. On the one hand, the cyber-physical world consists of all the hardware and software required to carry out the data measurement and knowledge collection from the physical system to the digital world. On the other hand, the digital world is the ecosystem consisting of algorithms, data lakes or big data, analytics tools, and others that allow processing of data to obtain insights about the physical system. These data can be delivered to the user in order to make data-based decisions or be used in control algorithms in order to take an automatic action. Also included in this world is a recent entity called digital twin, which is able to process the information collected in real-time from the physical system using mathematical models and algorithms to make predictions about different scenarios in the real plant such as anomalies, faults, perturbations, in order to be prepared for unexpected behaviors; this class of systems can help to save resources, money, and ensure people’s safety.

2.1. Technology Readiness Level (TRL)

Technology readiness level (TRL) is an index that is useful for measuring or assessing the maturity of a particular technology; it was introduced by the European Commission [15]. Any technological project can be evaluated using the parameters for each technology level and assigning its corresponding TRL according to its progress. The measurement system has nine technological readiness levels and goes from TRL 1, which is the lowest level, to TRL 9, which is the highest level in this scale. Each TRL is shown and described in Figure 3.

2.2. Wireless Communication Technologies

Currently, wireless communication technologies can be classified depending on their transmission ranges. They can fall into two classes, short-range [16,17] and long-range [17,18,19,20,21]. These technologies and their classification are shown in Figure 4. The selection of an appropriate technology depends on the pursued application, e.g., if the smart water system involves monitoring several ponds in a vast extension or long irrigation lands, long-range technology such as LoRa [22] or cellular network [23] may be required. On the other hand, if the smart water system is devoted to monitoring a small irrigation system, then short-range technology, such as WiFi, may be more appropriate [24]. For more details, the interested reader is referred to [25] for Zigbee, [16,26,27] for WiFi, [17,28,29,30,31,32] for cellular communication, [17,18,19,33] for LoRa and Sigfox, [20] for LTE-M, and [21,34] for NB-IoT.

3. Methodology

This section is devoted to detail the procedure for performing the systematic review. The review has been carried out following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [35], which ensures an updated state-of-the-art revision of smart water system developments in the last ten years, as well as transparency and reproducibility. Next, the considered steps are briefly described.

Data Collection from Bibliographic Databases

In order to collect the bibliographic data from the literature, the database SCOPUS was used. The query employed to perform the search is based on the keywords of interest and it was defined as TITLE-ABS-KEY (“smart water system” OR “smart water systems” OR “hydroinformatics” OR “IoT for water” OR “Internet of things for water” OR “digital twins for water” OR “digital twin for water” OR “smart campus” OR “optimization of water systems” OR “prediction of water composition”) AND PUBYEAR > 2013 AND PUBYEAR < 2025; notice that a little refinement to include plural and singular cases in the keywords was considered, e.g., “smart water system” and “smart water systems”. Moreover, the search was performed by analyzing the title, abstract, and keywords of journal articles, conferences, reviews, and chapters, among others. The search found 1739 documents from 2014 to 2024. Figure 5 shows the historical generation of documents during this period and

Table 1. Documents by year related to the established keywords from 2014 to 2024.

Year	2014	2015	2016	2017	2018	2019	2020	2021	2022	2023	2024
Documents	44	32	51	97	150	145	202	229	279	257	253

gives the specific number of documents by year. Additionally, Figure 6 shows the document distributions for each type of literature. Conference papers comprise the most of the works, followed by journal articles.

The main goal of this work is to analyze the development of smart water technologies in the last 10 years; therefore, the above results do not suffice. Thus, a refined filter process was conducted. Namely, excluding irrelevant and duplicate works and performing a more in-depth qualitative analysis. So, an in-depth analysis of the documents was performed by searching for published works containing the specific keywords in their keyword list. The detailed PRISMA flowchart is shown in Figure 7. The results are shown in Figure 8, which reveals that the most related area to the use of smart water technologies is smart campuses, where researchers and developers are more active. Moreover, analyzing the production of documents by country, Figure 9 shows that China produces more than the remaining countries. It does not suffice to gain insight into “smart water systems”. Thus, another filter was applied to analyze the documents found by searching for documents containing the keyword “smart water” in the title or the abstract; 122 papers were obtained. Finally, it was necessary to evaluate the full texts. From this one-by-one revision, 31 papers were obtained. They include articles in journals, conferences, and technical reports; reviews and books have been dismissed. These documents are gathered in Table 2.

4. Results

This review aims to survey smart water technology in the last 10 years, to identify how this research is geographically distributed, and to identify current challenges related to smart water systems and their implementation. To do that, it is required to answer the research questions RQ_i, $i\in \{1,2,3\}$ based on the chosen literature; this section is devoted to answering these.

4.1. How Has the State of the Art Regarding Smart Water Systems Developed in the Last 10 Years?

Exploring the literature, there are several predominant topics which can be arranged as shown in Figure 10. In turn, these topics can be arranged in the following clusters: IoT-based sensors for monitoring and control, leak detection and loss management, energy optimization and smart control, IoT-based smart water systems to improve security and protection of a network, digital twins and simulation, smart water systems to optimize irrigation processes, some miscellaneous applications, and theoretical models. These clusters will be discussed in what follows to answer RQ₁: How has the state of the art regarding smart water systems developed in the last 10 years?

4.1.1. IoT-Based Sensors for Monitoring and Control

One of the most common devices employed in smart water systems is the IoT-based smart water meter, due to its role in understanding consumption in water systems to improve their operations. The project iWIDGET, presented in [36], is an effort to create a web-based application programming interface (API) for help in the development and implementation of test beds or prototypes. This includes applications and user interfaces that work on platforms such as Mac OS, Linux, iOS, Windows, Android, and others. The software architecture used by the API is the Representational State Transfer (REST) combined with a web address or a Uniform Resource Identifier (URI). Similarly, in [37], the IoT is used to develop a system for resource monitoring and management. It is composed of several layers. The first layer is the equipment perception layer; it consists of the sensors required for water monitoring such as water level and flow, and water quality sensors. The second layer is the information transmission layer; it gathers the communication technology, protocols, and the required software, e.g., technology based on GPRS, operating systems, databases, programming languages, etc. The third layer is the application layer; it consists of data acquisition and storage, data observation, and data application. Using the proposed architecture, the authors propose a safety analysis methodology to monitor the operation and emergency response of the system. The aim of using IoT-based monitoring is to gain insight into the system using accurate and comprehensive real-time information. Following this path, ref. [38] implements a class of smart water meters to monitor water consumption to control its usage and prevent waste. The smart water system has an architecture of a system of several layers such as the device layer, communication layer, logic layer, functional layer, and user application layer.

Some studies highlight the importance and underline the limitations of data acquisition systems when different types of signals are used from different sensors. In [39], the authors proposed a smart sensor interface for smart water quality monitoring. The proposal is based on the use of an Intel Galileo Gen 2 Board as an interface device to collect data from measuring instruments, e.g., pH, dissolved oxygen concentration, turbidity, etc. Galileo is also used as a data transmitter through WiFi communication. The system was thought to be used in cities located on river banks to monitor pollution. It can be useful for Municipal Corporations to take preventive measures. Another study dealing with this issue is that presented in [40], which focuses on the deployment and use of smart city technologies and services to apply them on a university campus in order to achieve real-time water consumption and temperature monitoring. Other miscellaneous services, such as online resources, canteen management, notification, and others, are also considered. These resources are networked and communicated via IoT technology, and the information is concentrated on an Android-based platform to reach final users.

Inspired by the proper use of water for residential communities, the work in [41] proposes a system to monitor water flow and level in real-time using a mobile application and an IoT-based network. The mobile application enables users with privileges to manage some control tasks, whilst the remaining users can access only to see valuable information concerning water contamination, water leakage, low water pressure and flow, and other water information. In the same way, ref. [42] proposes a software to “easily” implement a smart water system, minimizing the cost and centralizing the information to monitor and control of watersheds in real-time; this is inspired by the scarce proven case studies and the lack of guidelines for building smart water systems. The software includes a set of robust sensors and actuators, cloud service platforms, and a comprehensive web-based “how-to” guide. In addition, two case studies are presented in this work: The first is to illustrate the detection and communication of flood hazards at the level of individual roads and the second is to show a real-time stormwater control network.

There are other classes of IoT-based sensors that focus on monitoring water quality. For example, ref. [43] develops a sensor network for a pond to monitor water parameters such as pH, water temperature, etc. The network combines IoT and analysis tools to assess the condition of water quality. The stakeholders, authorities and any user can access data to monitor water health, but, if an unhealthy water condition is detected, a notification is sent only to the authorities. This enables the responsible authorities to take early actions to reduce any harmful effects that can occur from a delay in the response. Similarly, ref. [44] proposes a sensor prototype to monitor water parameters such as pH, turbidity, flow, etc. The aim of this work is to prevent stakeholders when an anomaly is found in the system, thus ensuring a safe water supply.

A broader vision exploring smart sensors combined with water balance models, IoT technologies, and artificial intelligence is explored in [22,45]. On the one hand, the authors in [45] deployed a cyber-physical system in a university campus; this is based on IoT technologies and artificial intelligence (AI) and aims to construct a smart campus. The system is tested in a smart building for optimizing water consumption and controlling the processes. The authors argue that “smart campuses are a realistic representation of more complex systems” and allow testing of different technologies before being deployed on a large-scale scenario. This proposal merges several services such as data visualization, consumption modeling, alert generation, and others. Water consumption predictions are achieved using Gaussian mixture models and can be used in any facility to reduce water wastage. On the other hand, in [22], different classes of smart sensors are used along with water balance models, optimization algorithms, and weather forecasts to optimize dynamic irrigation in public green spaces. This system is developed using the open-source FIWARE ecosystem. The proposal is tested in a French Mediterranean coastal city saving about 35% and 60% of irrigation water during the dry season and the dry–wet transition season, respectively.

4.1.2. Leaks Detection and Losses Management

A lot of efforts have been made to minimize water wastage. In this sense, some works are focused in leak detection and loss management to minimize water loss, e.g., using smart water technology [51]. In this work, the minimum night flow approach and traditional water balance are used. The methodology is applied to a University campus with the size of a small town and the real-time monitoring is achieved via a set of sensors that measure the hydraulic parameters of the water network. The minimum night flow-based approach is employed to determine flow thresholds to activate an alarm in case of detecting a level loss in the network, which means that there exists a leak in the network. It allows the users to take quick actions to minimize water loss due to leakage and decrease the non-revenue water. Similarly, ref. [52] introduces a system to monitor leakages in a water network. It demonstrates its functionality by collecting information from a sensor system that measures the hydraulic parameters of the water; the data can be processed through several algorithms and water loss is detected. This task can be achieved in two ways: on the one hand, by monitoring the rate of flow inside a building, or, on the other hand, by computing the minimum night flow for an area. The proposal is tested using data collected from a university campus. Following the line of leak detection, ref. [53] focused on optimizing the use of pressure-reducing valves and pumps jointly; the proposal is proven using the well-known urban D-Town benchmark with 388 nodes, 429 pipes, 13 pumps, 4 pressure reduction valves, 1 reservoir, and 7 tanks at the simulation level. The authors claim that using this approach to manage the pressure of the water network can save 50% of energy and achieve a significant leak reduction. The proposal is based on the use and optimization of a model for water demand forecast close to real-time.

Other works focus on detecting anomalies in smart water systems using IoT technology; this helps to minimize non-revenue water loss. The result presented in [54] uses the information collected from smart sensors, which is assumed to be a time-series signal. These data enable the comparison between 2 different methods for anomaly detection, namely, an ARIMA-based framework and the HOTSAX technique; some applications for which this approach can be useful include fraud detection, surveillance, diagnosis, etc. Another important source of problems that cause leaks or water loss is corrosion in the water network. In this sense, the work [55] investigates changes in the minimum-night flow rate and its relationship with this phenomenon in water systems. To do that, some mathematical models are constructed to investigate the unreported loss of water due to corrosion by using regression analysis along with the Geographic Information System (GIS). This helps reduce non-revenue water, preserve drinking water, and to gain insight into the water distribution network to identify its useful life span.

4.1.3. Energy Optimization and Smart Control in Smart Water Systems

Energy optimization is paramount in smart water systems because they should be able to operate in complex situations and hard environments, such as changing weather and varying demands for both water and power systems. The work [56] deals with this problem in a smart water system for a public service company in Italy. The objectives of the work are optimizing energy consumption, reducing water losses, centralizing the monitoring and control room, and improving data quality and reliability; it helps to save approximately 20% of energy consumption. Another work, saving 50% energy in a smart water system, is that reported in [53], where the results were obtained by managing the water network distribution pressure.

In the area of smart campus, there have been some efforts to optimize energy in smart water systems. In [57], an architecture based on learning automata for utility management (LAUM) was proposed. The objective of this tool is to help solve problems concerning water and energy resources, deployment simplification, integration of smartness, minimization of resource waste, and improving user satisfaction. The LAUM is tested in two scenarios, the first case controlling streetlights and the second one controlling water tanks. In the same context of a smart campus, in [46], the potential of different machine-learning techniques and models is explored for bath water demand forecast in shower rooms in a university of China. This helps to reduce water consumption and save energy used for pumps and boilers. Similarly, in [45], a cyber-physical system is deployed on a university campus, based on IoT technologies and artificial intelligence (AI); the system is tested in a smart building to optimize water consumption and control processes. This proposal merges several services such as data visualization, consumption modeling, alert generation, and others. Some other works such as [24] are devoted to integrating IoT and AI to optimize the control task in smart water systems.

4.1.4. IoT-Based Smart Water Systems to Improve Security and Protection of the Network

Security is not trivial in smart water systems based on IoT technologies and implemented in smart cities; this is a hot topic of research in smart water systems and other areas. For instance, LoRaWAN-based smart water systems are exposed to some attacks, such as, Denial-of-Service (DoS) attacks, eavesdropping, replay attacks, physical tampering, among others. An interesting work presenting this kind of attack along with a proof-of-concept experiment can be found in [61]. In [58], an effort is made to improve the security of this class of networks, where an architecture is presented to integrate the smart water system into IoT and discuss the challenges in the security and protection of the network. These challenges increase considerably as a consequence of the expansion of the attack surface; other factors such as heterogeneity and the number of interconnected devices also increase the complexity of guaranteeing the system’s security. Motivated by this, the authors propose a framework to enable developers to deal with the issues related to each layer of the IoT layers and integrate functions and services (layer-to-layer) to minimize the related security risks. The proposed architecture has 4 layers which are IoT end devices, communications, services, end users, and applications. Roughly, the proposed smart water security development framework consists of five systematic steps for each layer: (1) specify the model (characterization), (2) attack surface identification, (3) impact evaluation of possible attacks, (4) analysis and proposal of mitigation actions, and (5) priority assignment of services according to possible attacks. In the same way, ref. [47] introduces the concept of the Blockchain of Things (BCoT); which is a combination of technology based on blockchain and the Internet of Things (IoT) to improve the performance of the smart water system and enhance its security towards a water 4.0 ecosystem. The BCoT is presented in a 4-layer architecture: physical layer, network layer, blockchain layer, and application layer; the proposal is tested using the Ethereum blockchain platform to simulate a real-world situation. Recent works explore critical scenarios where the smart water system is under attack through false data injection [62], and propose detection schemes to identify this kind of attack. Other works, such as [63], explore schemes focused on multi-modal detection frameworks for different classes of failures in smart water systems.

4.1.5. Digital Twins and Simulation

In the last few years, the digital twin concept has gained popularity. A digital twin is a piece of software used to represent and simulate cyber-physical systems, allowing for the study of their characteristics, monitoring of their status, and prediction of their future behavior. On the one hand, in [48], a digital twin called Digital Hydraulic Simulator (DHALSIM) is proposed to monitor some specific water distribution systems. Their control algorithms and their physical and network processes can be simulated in order to gain insight into this system. To do that, a co-simulation between a hydraulic simulator for Water Network for Resilience (WNTR) and a simulator for industrial network (MiniCPS) is implemented. The proposal is tested in the benchmark case study of C-Town under cyber-attack scenarios. On the other hand, ref. [49] implements a digital twin to monitor water quality. This system uses sensory data from the water system to run simulations to gain insight into the water quality conditions to forecast possible scenarios concerning changes in water quality under certain conditions. Machine learning is used to perform data analysis and to detect anomalies or abnormal contamination in water bodies. The platform is tested by employing a use case to create a dataset. Digital twin technology for smart water systems is still emerging and needs further research and development to establish a favorable balance between cost and benefits, allowing resistance to implementation in real-world smart water systems to be overcome.

IoT-based monitoring is an excellent tool for reactive control, providing real-time visibility for diagnostics in smart water systems. In contrast, digital twins use this foundational data to create forecasting/predictive models enabling proactive control, optimization, and diagnostics. While IoT-based monitoring is a cost-effective option for digitalization, it lacks predictive power, whilst digital twins are more powerful and efficient, allowing for advanced simulations at a greater complexity and cost. In summary, IoT-based monitoring systems are the foundational layer of a digital twin.

4.1.6. Smart Water Systems to Optimize Irrigation Processes

An irrigation system that considers measures such as soil moisture, humidity, temperature, image processing, and rainfall to control the process can be called a smart irrigation system. The work in [24] belongs to this class of system and provides a systematic procedure to implement the smart water system into the smart irrigation one. The control task is performed via an IA-based algorithm; it integrates IoT and AI to optimize traditional watering operations to contribute to the Sustainable Development Goals (SDGs) and it is implemented in the irrigation of two green zones on a university campus. In the same context, ref. [59] developed a wireless sensor node based on IoT for sensing conditions in the soil and in the air in an irrigation space, this information is used to control a solenoid valve to regulate the irrigation process optimizing plant grow conditions and reducing the water and energy use. The case study presented is a greenhouse used to cultivate tomatoes in a real environment and the results are compared against other irrigation techniques showing the advantage of the proposal.

Motivated by the water and energy used in agricultural processes, a smart water system is proposed for smart farms in [50]. The proposed algorithm effectively manages these smart water systems to minimize water usage. This approach combines modeling and deep learning (AI-NN) with a wireless sensor network that monitors soil moisture levels. In [22], meanwhile, different classes of smart sensors are used along with water balance models, optimization algorithms, and weather forecasts to optimize dynamic irrigation in public green spaces. The system is developed using the open-source FIWARE ecosystem. Proposal [50] provides results in simulation and proposal [22] is tested in a French Mediterranean coastal city saving about 35% and 60% of irrigation water during the dry season and the dry–wet transition season, respectively. Recent works, such as [64,65], use machine learning and artificial intelligence combined with IoT technology to improve the water management system through data-driven decision making.

4.1.7. Miscellaneous Applications

Following the line of smart water systems, ref. [23] implements a stormwater control system for an urban watershed. This system utilizes a smart water management approach for controlling water flow downstream, minimizing erosion, and reducing pollutants. The control task consists of a wave-based control of water flow to maintain it on a desired set-point according to the hydrography; it is achieved using internet-controlled valves. The case study focuses on an urbanized creekshed located in a city in Michigan. Moving the spotlight to India, ref. [66] presents results at the simulation level to monitor water flow using the consumption of some regions of the country. The proposal consider some protocols to prioritize water flow according to its priority value, e.g., emergency belt, residential belt, industry belt, and others. The proposed model is tested in simulation using the SCI-WSN simulator to illustrate the automatic water channelization. Other cases, such as [60], are focused on proposing alternatives to diminish the use of one single use plastics. Therein, the design and implementation of a smartbottle ecosystem is proposed; the smartbottle is an interactive and reusable water bottle, which can communicate with a smart water refill station. It is based on IoT and ICT technologies. As mentioned before, the aid of the proposal is to eliminate the use of single-use plastic water bottles in the Instituto Politécnico de Viana do Castelo.

Recent works have explored the Cyanobacterial Bloom Concentration prediction to minimize the associated risks to public health and aquatic ecosystems. On the one hand, ref. [67] proposes a solution to the persistent challenge of data scarcity by using machine-learning techniques to train a prediction model. The proposal consists of taking a water body with rich historical data and training a model, then this model is transferred to another water body where data are sparse, and at the final step, the trained model is fine-tuned with representative data of the target water body. This technique is called the transfer learning model. On the other hand, ref. [68] implements a harmful algal bloom prediction system by deploying a set of buoys equipped with low-cost water quality sensors, which feed an embedded AI model capable of monitoring and alerting in case a specific parameter (chlorophyll-a) rises above some predefined thresholds.

4.1.8. Theoretical Models and Approaches in the Field of Smart Water Systems

Modeling a system is paramount to perform some tasks, such as model-based simulations, analysis, and design of controllers, and other interesting tools, such as predictor models [67,68] and digital twins [48,49]. Smart water systems are not the exception. Exploring the chosen literature, there are several kinds of modeling techniques. Roughly, in the selected literature, they can be arranged as: optimization models and algorithms, machine learning and AI models, and cyber-physical and digital twins models.

• Optimization models and algorithms. In [14], a formal Lyapunov-based analysis was employed to design a sustainable water sensing system to balance harvested energy and monitoring tasks. The authors propose an asymptotically optimal scheme for Data Transmission Scheduling (DTS). Then, that proposal is improved to a faster one (FAST-DTS) using a lightweight online algorithm, which allows it to adapt its behavior to complex dynamics. The proposed model is an adaptive one based on lightweight auto-regressive models; thus, it avoids using complex hydraulic models and making stochastic assumptions. Works such as [53] implement models based on metaheuristics algorithms such as Particle Swarm Optimization (PSO), and the combination of non-linear auto-regressive with exogenous input artificial neural networks (NARX) and unscented Kalman filter (UKF). This kind of model allows for maximizing the efficiency point of pumps, considering the water demand and hydraulic characteristics of the smart water system; it helps to reduce energy consumption. In the field of data-driven control, we can find works such as [69], where hydraulic modeling and Gaussian processes are combined to propose a data-driven predictive control approach, which is put to the test in a smart water system managing wastewater and stormwater networks. A more recent work [22] proposes the integration of water balance models with optimization algorithms and smart sensors to create a model for optimizing the irrigation process of public green spaces and diminishing water consumption. The water balance model allows forecasting the humidity of the soil, while the optimization algorithms allow scheduling the irrigation of the spaces with an adequate water volume, according to the forecasting model. Other works employing water balance models and minimum night flow (MNF) [51] are devoted to leak detection in water distribution networks. The hydraulic parameters of the water network are obtained in real-time by means of IoT-based sensors. The model allows for quick detection when a pipe has burst, which enables a quick action from the user to minimize the water wastage. Work [51] proposes a model and [52] presents the detailed implementation in a large-scale smart water system demonstrator called SunRise. In this last work, some improvements, such as adding the k-mean algorithm, have been made. In the same context, models combining the MNF and IoT technology have been proposed for dealing with corrosion phenomena in water networks [55]. In the arena of stormwater systems, some attempts have been made to design control models to shape streamflow within an urban watershed and reduce risks [23,42].
  These works show the evolution of some models based on optimization techniques or algorithms combined with mathematical models such as water balance, from the energy and sustainability focus to an integral area where we have multi-objective models considering energy and water saving as well as safety of the smart water system.
• Machine-learning and AI models. In the arena of machine-learning and AI models, some efforts have been made to construct models based on AI to detect leaks in pipes using classification algorithms such as The Radial Basis Function Neural Network (RBF-NN) [31]. According to the authors, it was able to detect the magnitude and location of leaks with a 98% accuracy. Other works, such as [54], tackle the problem of leak detection using anomaly detection models and water consumption in the form of time series. These series are analyzed via an ARIMA-based framework to forecast the leaks and via a technique called Heuristically Order Time series-Symbolic Aggregate Approximation (HOT-SAX) to detect irregular water consumption. Dealing with the same problem, leak detection, ref. [27] uses Support Vector Machines (SVM) to forecast water consumption and detect any anomaly in the water consumption measurement provided from IoT-based sensors to detect leakages in the water network in real-time.
  Other kinds of models based on deep-learning algorithms and mathematical models are those that allow the monitoring of the soil moisture in smart farms where the smart water systems are used to monitor and control the irrigation task in order to minimize the energy and cost related with this activity depending on the type of plants; these are very important things in agriculture [50]. Another effort made in this area is presented in [59], where, as before, a monitoring system in a smart farm was deployed using IoT sensors to monitor soil moisture and temperature. These data are used by a fuzzy logic control system to perform the control task using an optimal decision-making scheme based on fuzzy logic.
  In the area of smart campuses, some works have been focused on creating models for short-term bath water demand forecasting using the well-known ARIMA, ARIMAX, Random Forest (RF), long short-term memory (LSTM), and Neural Basis Expansion for Interpretable Time Series Forecasting (N-BEATS) models [46], or for monitoring and scheduling irrigation processes using AI modeling techniques and IoT-based sensors and actuators (e.g., water valves) [24].
  Other works, such as [67,68], are focused on developing prediction models for avoiding risky situations in aquatic ecosystems and human health. These models are based on deep-learning techniques, convolutional neural networks, and long short-term memory architectures. An important contribution to modeling techniques is made in [67], where AI bloom prediction models are constructed using water bodies rich in data. Then, the model is tuned using data from water bodies where the data are sparse. This enables users to implement prediction models of cyanobacterial bloom concentration in water bodies, where getting information from them is a hard or very complex task. The transfer learning approach for deep learning is also considered in [70], but is employed for water consumption forecasting. It is trained and evaluated using a real-world dataset. Another work dealing with forecasting water consumption is that in [71], where models based on machine learning are constructed using key factors, such as, temperature, precipitation, and time (hours, days, months).
  From the above works, it can be seen that efforts are made for developing models based on machine learning and AI to forecast water consumption and prevent energy and water wastage, with special focus on real-time monitoring and optimization of smart water systems. Other important efforts are those for water quality monitoring of aquatic ecosystems.
• Cyber-Physical and Digital Twin Models. Raising technology allows us to construct and simulate more complex models that are capable of emulating real-world systems. Moreover, if we tune them properly and feed those systems with real-time data collected from IoT sensors, we have a digital twin. This kind of system employs data coming from cyber-physical systems and allows us to monitor dynamical systems in real-time or to use historical datasets for forecasting future system behaviors. Some works, such as [45,48,49], have made an effort to create this kind of model for smart water systems. Their evolution starts with systems designed for emulating cyber-physical systems in real-time, but through the years, those models have increased in sophistication to include tasks such as smart building and water management for optimizing consumption and sustainability, and to predict and mitigate risks.

It is clear that control and hydraulic models are a fundamental piece of smart water systems. This models are the basis for new developments and improvements using edge-technology based on artificial intelligence, IoT technology, optimization techniques, and digital twins.

4.2. What Countries Have More Published Reports Concerning Smart Water Systems?

The geographical research distribution by country is shown in Figure 11. When a work has authors from several countries, all of them are considered in the statistics, e.g., if a document has authors from Spain and Brazil, thus the paper is included in the statistics of both countries. Therefore, analyzing Figure 11 and Figure 12, where more details are provided, the answer of RQ₂ is India, i.e., the country where the majority of publications associated with smart water systems are concentrated is India.

4.3. What Are the Challenges in Implementing a Smart Water Systems Using Current Technology?

Smart water systems are still facing several challenges. One of the major challenges is the lack of systematic and standardized benchmarking associated with the lack of consensus on what a smart water system is. These do not a unique definition in the literature and it depends on the point of view from where the problem or task is tackled [34] and also depends on the application field [72]. Therefore, there is a lack of generality in current applications because they are designed ad-hoc. In addition, the problem of scalability and real-time processing of data appears and makes the need for systematic and standardized benchmarking more critical, making its integration into actual operations more difficult [73]. As a consequence, there is no sufficient evidence concerning the benefits of the implementation of smart water systems.

A second challenge concerns the collection and analysis of water consumption data to gain insights and there is room for improvement concerning privacy preservation while ensuring security. Additionally, since the smart water system is a network connecting a large number of devices, the vulnerability of the network is a recurrent problem because it is prone to malicious attacks; attackers aim to manipulate sensible data or inject anomalies to possibly cause dangerous situations [74].

A third challenge for the implementation of smart water systems is the required instrumentation, e.g., the sensors. The implementation of a smart water system with reliable sensors, infrastructure upgrades, and the computational capacity required generally can face financial and logistical barriers due to the associated costs because of the high costs of the sensors and sometimes the unavailability of them for some specific measuring tasks. Moreover, this becomes a more complex process when the smart water system should be integrated with existing infrastructure and legacy systems, which typically is the case. The complexity increased due to cybersecurity, data management, system modeling, and others.

A fourth challenge is about the resilience of the smart water system. The system should be able to operate in complex situations and hard environments, such as changing weather and varying demands for both water and power systems. This becomes a challenge mainly for resilient systems based on renewable energies, which are designed to be autonomous. They should be able to work and harvest energy from their environments, e.g., harvesting energy from water, wind, sun, and others. In this field, there is much room for improvement in optimizing the harvesting methods and optimization of the resources and operations [75].

A fifth challenge is about the stakeholder’s engagement, i.e., there is a need for high engagement from the private, public, and research sectors. Smart water systems are based on a multidisciplinary approach, and they require working hand-by-hand in order to utilize cutting-edge technology which can be developed by investing in research after the implementation of the cutting-edge system. This requires the water operators’ and users’ care and adoption to keep it working in optimal conditions. To achieve it, a cultural business shift from reactive to proactive, educational programs, and new government policies about water are required [34,73,76].

Table 2 presents key data concerning the found literature, such as if the work was simulated or implemented, the type of sensors and actuators being used, the communication technology, as well as the communication protocol if it uses cloud services, and finally, its TRL classification.

Table 2. Key data concerning the found literature involving technology, communication protocols, cloud services, TRL, etc.

Work	Simulation	Implementation	Microcontroller	Sensors	Actuator	Communication Technology	Communication Protocol	Cloud Service	TRL
[36]	✓	-	-	Smart water meters	-	No specified	HTTP	iWIDGET	3
[37]	✓	✓	-	Level sensor Flow sensor Water quality sensor	-	GPRS	TCP/IP	-	3
[39]	-	✓	Intel Galileo Gen 2	Level sensor pH electrode TSD10 turbidity sensor Temperature sensor	-	WiFi	IEEE 802.11	-	3
[57]	-	✓	-	Light sensor Soil-humidity sensor Level sensor	Streets lights Pumps Irrigation valves	Zigbee	IEEE 802.15.4	-	3
[58]	-	✓	-	Smart water meter	Valves	WiFi	IEEE 802.11	-	4
[51]	-	✓	-	Automated meter readers, Bulk meters, customer sub-meters, flow and pressure sensors	Valves	GPRS	TCP/IP	-	5
[77]	✓	✓	-	Pressure and flow sensors	Valves	Zigbee	IEEE 802.15.4	-	4
[40]	-	✓	ESP12ENode Mcu V3	Temperature sensor (LM35), flow sensor (YF-S201)	-	WiFi	IEEE 802.11	-	3
[52]	-	✓	Embedded-no specified	Automatic meters, piezoresistive pressure sensors	Isolation valves	WiFi	IEEE 802.11	-	5
[56]	✓	-	-	-	-	GPRS	TCP/IP	Amazon Web Services	8
[41]	-	✓	Arduino with an PIC16f877a	Water level	Motors	WiFi	IEEE 802.11	-	3
[53]	✓	-	-	-	Pressure-reducing valves and pumps	-	-	-	4
[42]	-	✓	PSOC5-LP (Cypress)	Ultrasonic and water level sensors; soil moisture sensors; optical rain gauges; dissolved oxygen, pH, ORP, temperature.	Butterfly valves	2G, 3G, 4G	TCP/IP	Amazon Web Services-Microsoft Azure	5
[23]	-	✓	ARM Cortex-M3 (Cypress)	Ultrasonic sensor (Maxbotix MB7384); Measurement station USGC 04174518 (water level, rainfall forecast)	Butterfly valve (Dynaquip MA44); gate valve Valterra 6912 with a linear actuator AEI 6112CH	2G, 3G	TCP/IP	Amazon Web Services	4
[50]	✓	-	ARM Cortex	Moisture	Pumps	2G	TCP/IP	-	3
[54]	✓	-	-	Flow meters; pressure sensors; water quality sensors	-	-	-	-	3
[66]	✓	-	-	Sensor-based smart valves	Smart valves	-	-	-	3
[48]	✓	-	-	-	-	-	-	-	3
[59]	-	✓	Nano Arduino with nRF24L01 module; Raspberry Pi	Soil moisture sensor YL-69; DHT22 sensor (temperature); capacitive moisture sensor; thermistor	Solenoid valve	WiFi	IEEE 802.11	-	8
[69]	✓	✓	-	Level sensor	Pumps	-	-	-	4
[60]	-	✓	-	RFID tags	-	RFID	-	FIWARE	8
[46]	✓	-	-	Flow water meters; Meteorological station	A flow control device	IoT Network (no detailed)	-	-	3
[43]	-	✓	AtMega328p with ESP wireless module	pH sensor; Turbidity sensor; Rain condition sensor	-	WiFi	IEEE 802.11	GoogleCloud	5
[78]	-	✓	Arduino mega 2560 R3 with SIM 800L module	Temperature sensor; Turbidity sensor; TDS sensor; pH sensor	-	2G	TCP/IP	-	2
[55]	-	✓	-	Pressure and flow sensors	Pressure regulation valves	IoT wireless (no detailed)	-	-	3
[44]	-	✓	Arduino ATmega328 with ESP8266 module	pH sensor; turbidity sensor; temperature sensor; ultrasonic sensor	-	WiFi	IEEE 802.11	-	3
[24]	-	✓	ESP8266	Moisture sensor; Temperature sensor; Humidity sensor; Rainfall sensor; IP camera	Valves controlled by relays	WiFi	IEEE 802.11	IoTtalk	8
[38]	-	✓	ESP32 MCU with WiFi + BT + E32 LoRa module	Water flow sensor	-	LoRa; WiFi; Bluetooth	LoRaWAN; IEEE 802.11; IEEE 802.15.1	-	3
[45]	✓	✓	-	IWM-PL3; electronic pulse emitter module for multi jet water meters	-	LoRa; WiFi	LoRaWAN; IEEE 802.11	-	5
[49]	✓	-	-	-	-	-	-	-	3
[47]	✓	✓	Raspberry Pi 4; ESP32; Raspberry Pi Pico W	Flow meters; Ultrasonic sensor; Total dissolved solids sensor; BME280 sensor	Solenoid valves	WiFi	IEEE 802.11	Node-RED; Ethereum	3
[22]	✓	✓	-	Temperature and volumetric sensor; flowmeter; external weather API	Control valve	LoRa	LoRaWAN	FIWARE-Grafana	5

Table 2. Key data concerning the found literature involving technology, communication protocols, cloud services, TRL, etc.

Work	Simulation	Implementation	Microcontroller	Sensors	Actuator	Communication Technology	Communication Protocol	Cloud Service	TRL
[36]	✓	-	-	Smart water meters	-	No specified	HTTP	iWIDGET	3
[37]	✓	✓	-	Level sensor Flow sensor Water quality sensor	-	GPRS	TCP/IP	-	3
[39]	-	✓	Intel Galileo Gen 2	Level sensor pH electrode TSD10 turbidity sensor Temperature sensor	-	WiFi	IEEE 802.11	-	3
[57]	-	✓	-	Light sensor Soil-humidity sensor Level sensor	Streets lights Pumps Irrigation valves	Zigbee	IEEE 802.15.4	-	3
[58]	-	✓	-	Smart water meter	Valves	WiFi	IEEE 802.11	-	4
[51]	-	✓	-	Automated meter readers, Bulk meters, customer sub-meters, flow and pressure sensors	Valves	GPRS	TCP/IP	-	5
[77]	✓	✓	-	Pressure and flow sensors	Valves	Zigbee	IEEE 802.15.4	-	4
[40]	-	✓	ESP12ENode Mcu V3	Temperature sensor (LM35), flow sensor (YF-S201)	-	WiFi	IEEE 802.11	-	3
[52]	-	✓	Embedded-no specified	Automatic meters, piezoresistive pressure sensors	Isolation valves	WiFi	IEEE 802.11	-	5
[56]	✓	-	-	-	-	GPRS	TCP/IP	Amazon Web Services	8
[41]	-	✓	Arduino with an PIC16f877a	Water level	Motors	WiFi	IEEE 802.11	-	3
[53]	✓	-	-	-	Pressure-reducing valves and pumps	-	-	-	4
[42]	-	✓	PSOC5-LP (Cypress)	Ultrasonic and water level sensors; soil moisture sensors; optical rain gauges; dissolved oxygen, pH, ORP, temperature.	Butterfly valves	2G, 3G, 4G	TCP/IP	Amazon Web Services-Microsoft Azure	5
[23]	-	✓	ARM Cortex-M3 (Cypress)	Ultrasonic sensor (Maxbotix MB7384); Measurement station USGC 04174518 (water level, rainfall forecast)	Butterfly valve (Dynaquip MA44); gate valve Valterra 6912 with a linear actuator AEI 6112CH	2G, 3G	TCP/IP	Amazon Web Services	4
[50]	✓	-	ARM Cortex	Moisture	Pumps	2G	TCP/IP	-	3
[54]	✓	-	-	Flow meters; pressure sensors; water quality sensors	-	-	-	-	3
[66]	✓	-	-	Sensor-based smart valves	Smart valves	-	-	-	3
[48]	✓	-	-	-	-	-	-	-	3
[59]	-	✓	Nano Arduino with nRF24L01 module; Raspberry Pi	Soil moisture sensor YL-69; DHT22 sensor (temperature); capacitive moisture sensor; thermistor	Solenoid valve	WiFi	IEEE 802.11	-	8
[69]	✓	✓	-	Level sensor	Pumps	-	-	-	4
[60]	-	✓	-	RFID tags	-	RFID	-	FIWARE	8
[46]	✓	-	-	Flow water meters; Meteorological station	A flow control device	IoT Network (no detailed)	-	-	3
[43]	-	✓	AtMega328p with ESP wireless module	pH sensor; Turbidity sensor; Rain condition sensor	-	WiFi	IEEE 802.11	GoogleCloud	5
[78]	-	✓	Arduino mega 2560 R3 with SIM 800L module	Temperature sensor; Turbidity sensor; TDS sensor; pH sensor	-	2G	TCP/IP	-	2
[55]	-	✓	-	Pressure and flow sensors	Pressure regulation valves	IoT wireless (no detailed)	-	-	3
[44]	-	✓	Arduino ATmega328 with ESP8266 module	pH sensor; turbidity sensor; temperature sensor; ultrasonic sensor	-	WiFi	IEEE 802.11	-	3
[24]	-	✓	ESP8266	Moisture sensor; Temperature sensor; Humidity sensor; Rainfall sensor; IP camera	Valves controlled by relays	WiFi	IEEE 802.11	IoTtalk	8
[38]	-	✓	ESP32 MCU with WiFi + BT + E32 LoRa module	Water flow sensor	-	LoRa; WiFi; Bluetooth	LoRaWAN; IEEE 802.11; IEEE 802.15.1	-	3
[45]	✓	✓	-	IWM-PL3; electronic pulse emitter module for multi jet water meters	-	LoRa; WiFi	LoRaWAN; IEEE 802.11	-	5
[49]	✓	-	-	-	-	-	-	-	3
[47]	✓	✓	Raspberry Pi 4; ESP32; Raspberry Pi Pico W	Flow meters; Ultrasonic sensor; Total dissolved solids sensor; BME280 sensor	Solenoid valves	WiFi	IEEE 802.11	Node-RED; Ethereum	3
[22]	✓	✓	-	Temperature and volumetric sensor; flowmeter; external weather API	Control valve	LoRa	LoRaWAN	FIWARE-Grafana	5

5. Discussion

Smart water systems are rapidly evolving and represent a multidisciplinary and multi-domain research area; they encompass real-time monitoring systems combined with IoT technologies, artificial intelligence, secure communication, and digital twins. It is important to follow the technological progress of this class of systems, but also it is important to identify the regions where research is particularly active for possible collaborations between different communities. This is also key for identifying challenges that may impact the implementation and acceptance of this class of systems by the stakeholders. The following section provides a list of countries where the research in smart water systems is concentrated as well as the main challenges identified from the reported studies up until the year 2024.

5.1. Development of Smart Water Systems

Developments in smart water systems have led to current research into smart water meters, web-based application development for water monitoring, integration of legacy systems with IoT-based smart technologies, cyber-security, privacy, artificial intelligence applied to smart water systems, digital twins, and others. This has helped developers in implementing testbeds or prototyping smart water systems (SWS) in real-time applications. One enabling technology for achieving a lot of advances in implementing SWS is IoT-based technology as argued [37]. This helps to develop systems for monitoring and managing systems for gaining insight into it in real-time to data-based decision making. Some works, such as [39], target their efforts for highlighting the importance and underlying limitations of data acquisition systems in environments where a lot of types of sensor signals co-exist and should be concentrated, processed, and communicated in a platform to reach the final users, e.g., the integration of several solutions into a SWS using a SCADA solution, Modbus protocols, and IoT-based sensors. Concerning the last ones, there have been efforts to design and develop IoT-based wireless sensors [44,59] for measuring conditions in the soil and in the air in irrigation spaces, for monitoring water parameters such as pH, turbidity, dissolved oxygen concentration, flow, pressure, conductivity, and others. Currently, IoT technology is present in every SWS enabling its “smartness”. Nevertheless, despite all the advantages of “smartness”, there are remaining problems concerning the standardization of smart water system architectures and several frameworks have been proposed [37,38,58], which slows its spreading in real-world applications. In [37], a three-layered architecture is presented, it consists of an equipment perception layer, an information transmission layer, and an application layer. After that, ref. [58] presented a four-layered architecture with the layers: IoT end devices, communications, services, end users, and applications. In [38], a five-layered architecture is presented, which introduces the device layer, the communication layer, the logic layer, the functional layer, and the user application layer. These architecture refinements are due to the edge-technology used in SWS and some issues concerning it, for instance, the integration between legacy and SWS, privacy concerns, cyber-security, cloud services, and others. As technology advances, it is possible that more refinement would be required for SWS architecture to ensure security and privacy concerns. The most recent architecture was presented in [47], it consists of 4 layers: physical layer, network layer, blockchain layer, and application layer. It is focused on improving the system’s performance and enhancing its security by combining IoT and blockchain technology towards a water 4.0 ecosystem. This ecosystem involves all the water assets and stakeholders of SWS, including cities located on banks of river [37], smart water technology in universities smart campuses [51,52,55], residential communities [41], public services [56], stormwater management mechanisms, wastewater networks management, and other watersheds [24,42,50]. Findings show that research is focused on water applications such as monitoring consumption, pollution concentration, parameters of quality, pressure in pipes and other assets, levels, and others. Some other typical tasks are controlling water-related processes and management, such as fault detection, fraud detection, leak detection, energy and water consumption using optimization, and others. In recent years, a concept that has become popular is the digital twin, which is used to represent and simulate cyber-physical systems in order to study their nature, monitor, and predict their future behavior [48,49]. When this class of systems is combined with artificial intelligence, they become a powerful tool for analyzing, predicting, and detecting the normal behavior of the SWS or the case of anomalous behavior [48,49]. Research in smart water systems is in its infancy and a lot of research should be done still. Research in this area requires engagement from the stakeholders in researching, designing, and implementing SWS using edge-technology to contribute to the Sustainable Development Goals (SDGs) while ensuring economic feasibility. An important aspect requiring the focus of the community research is reaching a consensus about what a smart water system means, defining an architecture for the smart water system, ensuring the security and privacy for the SWS users, and designing instrumentation to be economically feasible. Currently, research and development communities are tackling issues such as security and privacy by technology such as LoRa, or by employing communication protocols such as Modbus combined with Sigfox, cellular networks, or WiFi. Nevertheless, more research is required to show the benefits of implementing this class of solutions or to propose other solutions where a good tradeoff between security, privacy, performance, and cost should be sought. On the other hand, an in-depth investigation, implementation, and analysis are desired concerning digital twins for smart water systems to show and highlight their vast applications and benefits towards the water 4.0 ecosystem.

5.2. Distribution of Documents by Countries

As shown in Figure 11, the review found that the majority of research in smart water systems is concentrated in India, followed by the USA and Spain. This could be for several reasons, including the country’s priority agenda on water management or the availability of research resources, such as funds, infrastructure, an academic-industry-government sector alliance, and others. From these findings, clearly, it is paramount to motivate the community to focus its efforts on smart water system research in order to reach the SDGs and mitigate the effects and consequences of climate change and population growth concerning this vital and non-replaceable resource, water.

5.3. Challenges in the Implementation of Smart Water Systems

Currently, there are many challenges in smart water systems. However, from the reviewed studies, it can be seen that, without a doubt, an important challenge is to obtain a systematic and standardized benchmarking for SWS. Other persistent challenges that can be identified from the review process are those concerning privacy, security, and economic feasibility. All of them could be tackled if an appropriate alliance was made between the academic, industrial, and government sectors. As long as a consensus is not reached between these sectors, smart water systems will continue advancing, but slowly.

5.4. Mechanisms to Overcome Resistance to Implementation

Considering the above sections, several mechanisms can be proposed to overcome resistance in implementing smart water systems. Pilot demonstrators in controlled areas, such as university campuses or districts, can be used to validate the benefits of implementing smart water systems and build stakeholders’ confidence. Another mechanism that can be used to overcome resistance is integrating new solutions to take advantage of the current assets and infrastructure, by upgrading pumps, valves, and SCADA systems instead of replacing the current system. Another issue to be addressed is creating robust security frameworks such that cyber risks can be tackled, improving trust. Equally important is to demonstrate economic savings and environmental efficiency, such as non-revenue water reduction and energy savings, among others. Finally, with the rise of digital twins and simulation tools, some specific mechanisms are required to test innovations before under controlled conditions before deploying them on a big scale and in a real-world scenario. This reduces the associated risks and makes transitions to real-world operation considerably smoother.

6. Conclusions

This work presented a systematic review of developments in smart water systems over the last decade. It examines the geographic location of the generated documents and the current implementation challenges faced by these systems. Our findings show that initial efforts primarily focused on smart watering technology, but currently encompass the development of digital twins for water. The review also finds that more collaboration between stakeholders is required to overcome current challenges in its implementation. Regarding smart water system resilience, it is paramount to increase the autonomy of the nodes while maintaining a good trade-off between efficiency, efficacy, and cost. Finally, a key aspect is to foster a sustainable consciousness in the community to ensure the adoption of smart water systems.