供水管网水质预警、供水管网爆管侦测、预警、定位

To enhance the quality of life and ensure sustainability in crowded cities, safe management of drinking water using cutting-edge technologies is a priority. This study developed an intelligent early warning system (EWS) for alarming and controlling risks from bacteria and disinfection byproducts in a drinking water distribution system (DWDS), named BARCS (Bacterial Risk Controlling System). BARCS adopts an artificial intelligence (AI) approach to data-driven prediction and considers total chlorine (TCl) concentration as the pivot indicator for risk identification and control. First, the machine learning-based AI model in BARCS can provide a reliable prediction of TCl concentration in a DWDS, with an average R2 of 0.64 for the validation set, while offering great flexibility for BARCS to adapt to various conditions. Second, TCl concentration was proven to be a good indicator of bacterial risk in a DWDS, as well as a cost-effective surrogate variable to assess disinfection byproduct risk. Third, the robustness analysis demonstrates that with state-of-the-art water quality monitoring technologies, online implementation of BARCS in real-world settings is feasible. Overall, BARCS represents a promising solution to the safe management of drinking water in future smart cities.

Water distribution systems consistently supply high-quality water at suitable pressure and volume for human and industrial consumption. Meticulous water quality management is vital to these systems. South Korea, having established legal standards for water distribution in 1963, operates the National Auto Water Quality Monitoring System for real-time water quality monitoring and contamination warnings when levels exceed legal thresholds. The U.S. Environmental Protection Agency (EPA) points out that fixed thresholds can trigger an abundance of false-positive alarms, causing irregular hydraulic changes, and false-negative errors. This could potentially lead to a failure in detecting initial instances of pollution or micropollution that fall below the established threshold. To address this, our study developed an proactive contamination warning method for South Korea’s monitoring system, utilizing long short-term memory (LSTM) for water quality prediction. We also employed ensemble empirical mode decomposition (EEMD) in feature engineering to enhance LSTM’s prediction performance. Additionally, we devised an optimal water quality prediction model development methodology by comparing short-and long-term prediction performances. Our findings revealed that using EEMD for feature engineering improved the stability and reduced the prediction lag of LSTM, outperforming traditional methods. This refined approach offers a more reliable and efficient means of monitoring and managing water quality in distribution systems.

Urban water supply system is an important part of public facilities. The quality of water supply pipe network directly affects the living standard of residents, and the leakage of water supply pipe network will cause a lot of waste of water resources. In this article, the night minimum flow method and the normal distribution model are used to analyze the leakage of the community water supply network and determine the leakage threshold of the community. At the same time, Auto-Regressive Moving Average (ARMA) model is used to predict the state of water supply network, and leakage warning is timely. In this article, the effectiveness of the proposed model is verified by experiments, which provides a method for timely sensing the state of the community water supply pipe network and controlling leakage.

No abstract available

Water quality events within drinking water distribution systems are commonly first detected by consumers. This undermines confidence, is in contrast to the level of service consumers now expect and in extreme cases can indicate a risk to public health. Using real-world datasets, this research demonstrates how the deployment of water quality sensors within operational drinking water distribution systems enables service providers to gain understanding and control of water quality events proactively in distribution networks. Multiple analytical approaches are combined to extract actionable insights, these include the following: cross-correlation for system connectivity and transit time derivation; an innovative turbidity event scale system; and material flux analysis of discolouration material. This research demonstrates the ability to confirm and track network-wide events, determine root causes, and inform proactive management via advisory event scores. We present definitive evidence of a multiplicative increase in obtainable insights using multi-parameter multi-sensor approaches over single-parameter single-sensor equivalents. This understanding is vital to help guide future strategies for networks of water quality sensing.

Urbanization, climate change, and aging infrastructure present critical challenges for water distribution networks (WDNs). To address the dual objectives of resilience and water quality in WDNs, this work explores the potential of real-time topological adaptation in WDNs to combine the benefits of both looped and branched networks. Leveraging remote-controlled valves, the proposed methodology dynamically reconfigures looped networks into branched topologies, combining the benefits of both configurations. Applied to the EPANET Net3 benchmark model, this approach reduced water age from 33.6 h to as low as 8.6 h in critical sections, significantly improving water quality while maintaining serviceability. Critical link analysis (CLA) identified optimal pipe closures, revealing that only 7 out of 117 pipes were pivotal for quality enhancement. This study also introduces an operational planning tool for WDN operators, which can be further enhanced with emerging technologies like digital twins, to advance adaptability and performance. The findings offer a practical framework for improving WDN resilience and water quality, addressing infrastructure challenges effectively.

This paper explores the intersection of water quality management and advanced metaheuristic algorithms (MAs) by optimizing the location of water quality sensors in urban water networks. A comparative analysis of ten cutting-edge MAs, Harris Hawk Optimization (HHO), Artemisinin Optimization (AO), Educational Competition Optimizer (ECO), Fata Morgana Algorithm (FATA), Moss Growth Optimization (MGO), Parrot Optimizer (PO), Polar Lights Optimizer (PLO), Rime Optimization Algorithm (RIME), Runge Kutta Optimization (RUN), and Weighted Mean of Vectors (INFO), was conducted to determine their effectiveness in minimizing the risk of contaminated water consumption. Both benchmark and real-world water network serve as case studies to assess algorithmic performance. The optimization process focuses on reducing the volume of contaminated water by treating sensor placement as a critical design variable. EPANET 2.2 software was integrated with the optimization algorithms to simulate water quality and hydraulic behavior within the networks. The obtained results from analysis of two urban water networks revealed that the newer algorithms, such as the RIME and FATA, exhibit superior convergence rates and stability compared to traditional methods. While all tested algorithms demonstrated satisfactory performance, this study provides foundational insights for future research, paving the way for more effective algorithmic solutions in water quality management.

No abstract available

No abstract available

Predicting essential water quality parameters, such as discharge, pressure, turbidity, temperature, conductivity, residual chlorine, and pH, is crucial for ensuring the safety and efficiency of water supply systems. This study employs long short-term memory (LSTM) networks to address the challenge of capturing temporal dependencies in these complex processes. Our approach, using a robust LSTM-based model, has demonstrated significant predictive accuracy, as evidenced by substantial R-squared values (e.g., 0.86 for discharge and 0.97 for conductivity). These models have proven particularly effective in handling non-linear patterns and time-series data, which are prevalent in water quality metrics. The results indicate the potential for LSTMs not only to enhance the real-time monitoring of water systems but also to aid in the strategic planning and management of water supply systems. This study’s findings can serve as a basis for further research into the integration of AI in environmental engineering, particularly for predictive tasks in complex, dynamic systems.

ABSTRACT Good water quality in a water distribution system (WDS) can be reached by using flow control valves (FCV) to reduce water residence time (WRT), and chlorine booster stations to maintain acceptable free residual chlorine concentrations (FRCC). The research developed a cost-effective, adaptative approach for managing FCV and chlorine booster stations, while maintaining acceptable pressure and FRCC ranges. The methodology is applied to a real full-scale case study considering three optimization formulations: 1) FCV operation, 2) chlorine booster station operation, and 3) combining them in an adaptive management strategy. Results revealed that chlorine booster stations individually or combined with FCV enhanced water quality significantly at an acceptable cost, while FCV optimization alone or combined with chlorine booster stations reduced WRT by, at most, 3.8%. Chlorine booster stations were thus more effective than FCV in this case study. This formulation can be applied to other WDS to improve water quality cost-effectively.

A sustainable water supply is essential for public health. However, sudden water quality accidents can occur due to changes in hydrological or meteorological conditions. The developed model provide quick preemptive response solutions with intuitive visual outputs are needed. To address this, we combined the EPANET model with MATLAB to create a graphical user interface and augmented reality model. The results demonstrated that the developed model can enhance user accessibility and visibility of model output, providing quick, preemptive response solutions for water quality accidents. We hope our research will be used as a practical response model, benefiting from the intuitiveness of our water quality simulation model.

Ensuring the safety and reliability of drinking water supply requires accurate prediction of water quality in water distribution networks (WDNs). However, existing hydraulic model-based approaches for system state prediction face challenges in model calibration with limited sensor data and intensive computing requirements, while current machine learning models are lack of capacity to predict the system states at sites that are not monitored or included in model training. To address these gaps, this study proposes a novel gated graph neural network (GGNN) model for real-time water quality prediction in WDNs. The GGNN model integrates hydraulic flow directions and water quality data to represent the topology and system dynamics, and employs a masking operation for training to enhance prediction accuracy. Evaluation results from a real-world WDN demonstrate that the GGNN model is capable to achieve accurate water quality prediction across the entire WDN. Despite being trained with water quality data from a limited number of sensor sites, the model can achieve high predictive accuracies (Mean Absolute Error = 0.07 mg L-1 and Mean Absolute Percentage Error = 10.0 %) across the entire network including those unmonitored sites. Furthermore, water quality-based sensor placement significantly improves predictive accuracy, emphasizing the importance of careful sensor location selection. This research advances water quality prediction in WDNs by offering a practical and effective machine learning solution to address challenges related to limited sensor data and network complexity. This study provides a first step towards developing machine learning models to replace hydraulic models in WDN modelling.

No abstract available

No abstract available

No abstract available

–

This paper presents the research progress on the application of fuzzy theory and artificial neural networks in water quality early-warning systems. First, it elucidates the fundamental principles of fuzzy theory and artificial neural networks. Then, it examines methods and principles for integrating both approaches to construct an effective water quality early-warning system. Finally, through specific application cases, it demonstrates the effectiveness of this approach and discusses its potential for future development. Specifically, the paper elaborates on the principles of fuzzy theory, which allows for the handling of imprecise and uncertain information, and artificial neural networks, which are inspired by the neural structure of the human brain. It then explores the detailed methods and underlying principles for combining these two advanced techniques to develop a highly efficient early-warning system. Through the presentation and analysis of detailed application cases, the study demonstrates the significant effectiveness of this combined approach. Furthermore, it highlights the promising prospects for future development in enhancing water quality monitoring and protection. This research provides a substantial contribution to the study of water quality early-warning systems and the integration of fuzzy theory with artificial intelligence methods.

One of the most effective ways to minimize polluted water consumption is to arrange quality sensors properly in the water distribution networks (WDNs). In this study, the NSGA-III algorithm is developed to improve the optimal locations of sensors by balancing four conflicting objectives: (1) detection likelihood, (2) expected detection time, (3) detection redundancy, and (4) the affected nodes before detection. The research procedure proposed the dynamic variations of chlorine between defined upper and lower bounds, which were determined utilizing the Monte Carlo simulation model. For selecting a contamination matrix with the same characteristics and effects of all possible events, a heuristic method was applied. The coefficients of importance are introduced in this study for the assessment of contamination events and network nodes. The Pareto fronts for each of the two sets of conflicting objectives were computed for benchmark and real water distribution networks using the proposed simulation–optimization approach. Results indicated that sensors should be installed downstream of the network to maximize sensor detection likelihood; however, this increases detection time. For the benchmark network, maximum and minimum detection likelihoods were calculated as 92.8% and 61.1%, respectively, which corresponded to the worst detection time of 11.58 min and the best detection time of 5.06 min. So, the position of sensors regarding the two objective functions conflicts with each other. Also, the sensitivity analysis related to the number of sensors illustrated that the Pareto fronts became a more efficient tool when the number of sensors increased. The best pollution detection likelihood in the real water network increased by 18.93% and 24.66% by incrementing the number of sensors from 5 to 10 and 5 to 15, respectively. Moreover, adding more than 10 sensors to the benchmark network and more than 15 to the real system will provide little additional detection likelihood.

No abstract available

Water chlorination is the most used disinfection method in water distribution networks (WDNs). Nonetheless, water quality parameters, including chlorine concentration, are not available at every point of the WDN, although such information is of direct relevance to drive the operation at water treatment plants to keep the correct chlorine residual through the system. This work proposes the use of data-driven models, i.e., Artificial Neural Networks and Evolutionary Polynomial Regression, to predict the water quality parameters in most areas of a WDN, using water quality data measures at few sampling points. The study is demonstrated on the case studies of the trunk network of Bogota’s water distribution system.

No abstract available

No abstract available

Contamination events in water distribution networks (WDNs) can have a huge impact on water supply and public health; increasingly, online water quality sensors are deployed for real-time detection of contamination events. Machine learning has been used to integrate multivariate time series water quality data at multiple stations for contamination detection; however, accurate extraction of spatial features in water quality signals remains challenging. This study proposed a contamination detection method based on generative adversarial networks (GANs). The GAN model was constructed to simultaneously consider the spatial correlation between sensor locations and temporal information of water quality indicators. The model consists of two networks—a generator and a discriminator—the outputs of which are used to measure the degree of abnormality of water quality data at each time step, referred to as the anomaly score. Bayesian sequential analysis is used to update the likelihood of event occurrence based on the anomaly scores. Alarms are then generated from the fusion of single-site and multi-site models. The proposed method was tested on a WDN for various contamination events with different characteristics. Results showed high detection performance by the proposed GAN method compared with the minimum volume ellipsoid benchmark method for various contamination amplitudes. Additionally, the GAN method achieved high accuracy for various contamination events with different amplitudes and numbers of anomalous water quality parameters, and water quality data from different sensor stations, highlighting its robustness and potential for practical application to real-time contamination events.

Ensuring access to good quality water is crucial for sustainable development, particularly in developing nations. However, the lack of affordable and reliable solutions for monitoring water quality remains a significant challenge that the LOTUS sensor meets: it is a compact and versatile multiparametric sensor designed for real-time monitoring of chlorine, pH, temperature, and conductivity in potable water. The proposed solution features a cylinder-like structure, measuring 21 cm long and 3.5 cm in diameter. It integrates temperature sensors, conductivity sensors, and a $2 \mathrm{x}10$ sensor array of multi-walled carbon nanotube (CNT) chemistors. The CNTs, arranged in random networks between interdigitated electrodes, can remain non-functionalized or be functionalized with a dedicated polymer to modulate their sensitivity. Six LOTUS sensors were tested in a 44m-long water loop under a flow rate of 0.3m/s and pressure of 1 bar to evaluate the system performance. To manage the high noise levels caused by strong electromagnetic interferences in the facility, particularly under flowing water conditions, a strong effort was put into data denoising techniques. After extensive denoising and calibration model optimization, temperature, conductivity, chlorine, and pH estimation were achieved with mean absolute error (MAE) as low as $0.34^{\mathrm{o}}\mathrm{C}, 73.2\mu \mathrm{S}/\text{cm}, 0.13\text{mg}/\mathrm{L}$ and 0.12 pH unit in flowing water (achieved on different sensors due to dataset limitations). Notably, the dataset also demonstrates the role of CNT functionalization in chemical sensing, with the selected polymer modulating the sensitivity to $\text{pH}$ by 50%.

Pipe leakage and bursts are the primary contributors to water losses in water distribution networks (WDNs). However, the use of object detection techniques for identifying such failures is underexplored. This study proposes a novel deep-learning-based framework for pipe burst detection and localization (PBD&L) within WDNs. The framework employs spatial encoding of pressure fields obtained from hydraulic simulations of normal and burst scenarios. These encoded images serve as inputs to a faster region-based convolutional neural network (Faster R-CNN) object detection model, specifically designed for infrastructure monitoring. The framework was tested on three WDNs—Fossolo, PB23, and CM53—under varying sensor coverages (100%, 75%, and 50%). The results indicate that the model consistently achieves high detection accuracy across different network configurations, even with limited sensor availability. For Fossolo and PB23, the model demonstrated stable performance; however, for the CM53 network, accuracy decreased at full sensor coverage, possibly owing to overfitting or signal redundancy. Overall, the proposed method presents a robust solution for PBD&L in WDNs, showcasing significant practical applicability. Its ability to maintain high performance under partial observability and diverse network conditions demonstrates its potential for integration into real-time smart water management systems, enabling automated monitoring, rapid response, and improved operational efficiency.

No abstract available

Change point detection (CPD) has proved to be an effective tool for detecting drifts in data and its use over the years has become more pronounced due to the vast amount of data and IoT devices readily available. This study analyzes the effectiveness of Cumulative Sum (CUSUM) and Shewhart Control Charts for identifying the occurrence of abrupt pressure changes for pipe burst localization in Water Distribution Network (WDN). Change point detection algorithms could be useful for identifying the nodes that register the earliest and most drastic pressure changes with the aim of detecting pipe bursts in real-time. TSN et, a Python package, is employed in order to simulate pipe bursts in a WDN. The pressure readings are served to the pipe burst localization algorithm the moment they are available for real-time pie burst localization. The performance of the pipe burst localization algorithm is evaluated using a key metric such as localization accuracy under different settings to compare its performance when paired with either CUSUM or Shewhart. Results show that the pipe burst localization algorithm has an overall better performance when paired with CUSUM. Although, it does show great accuracy for both CPD algorithms when pressure readings are being continuously made available without a big gap between time steps. The proposed approach however still needs further experiments on different WDNs to assess the performance and accuracy of the algorithm on real-world WDN models.

This paper presents a comparative study of hyperparameter optimization techniques - Particle Swarm Optimization (PSO) and Population-Based Training (PBT) - for fully-linear deep learning architectures, specifically FL-ResNet and FLDenseNet, applied to pipe burst localization in Water Distribution Networks (WDNs). Accurate and efficient burst detection in WDNs is critical due to the potential for substantial water loss and service disruptions, yet traditional detection methods are often costly, labour-intensive, or limited in scalability. Recent advancements in deep learning offer enhanced diagnostic and predictive capabilities for WDN management. This study leverages PSO and PBT to optimize the performance of FL-ResNet and FL-DenseNet models. Results indicate that PSO consistently achieves higher accuracy with lower variance and faster convergence compared to PBT, with PSO-FL-ResNet reaching a mean accuracy of 98.92 % in significantly fewer training epochs. Statistical analysis validates PSO's superior stability and efficiency across both models. These findings suggest that PSOoptimized models could provide a more reliable and resource-efficient solution for pipe burst localization in WDNs.

Pipe bursts in water distribution networks (WDNs) pose significant threats to the safety of distribution networks, driving attention to deep learning-based burst detection and localization. However, the applicability of different pressure features still needs to be compared and verified. A large number of nodes challenges deep learning with the excessive number of classification categories and low recognition accuracy. To address these problems, this paper extracts different burst pressure features, including pressure value, pressure difference, and pressure fluctuation ratio, and inputs one of these features into a Burst Diagnosis Multi-Stage Model (BDMM) based on three CS-LSTMs (a combination of the Cuckoo Search algorithm and a long short-term memory network). The first model addresses a binary classification problem, outputting labels indicating whether a pipe burst has occurred. The second one solves a multi-classification problem, outputting the label of the burst partition, and the third model also solves a multi-classification problem, outputting the ID of the bursting junction. The model is tested on a real network and outperforms ELM. For basic burst identification tasks using CS-LSTM, differences among the three features are minimal, while pressure difference and pressure fluctuation ratio exhibit superior performance to pressure value when resolving more complex problems like burst junction localization.

No abstract available

No abstract available

No abstract available

Pipe bursts cause a considerable loss of treated water, increase the risks of environmental contamination and are a health hazard for the end-user as they can create a passage for contaminants to enter water distribution networks (WDN). Identifying pipe burst locations will help water service providers promptly repair them. Many methods have been proposed to locate pipe bursts. But none have proved to produce results accurate enough for water service providers to heavily rely on. Therefore, this paper aims to develop a fully-linear ResNet to accurately locate pipe bursts using pressure sensors. The performance of the model was compared to the fully-linear DenseNet and ResNet methods. The proposed FL-ResNet had a mean accuracy of 98.23% which is comparable to the 99% achieved by FL-DenseNet and it required significantly less training time compared to the FL-DenseNet. As part of future work, transfer learning could be explored, by training the model on simulated data and then on real-world data.

An acoustic method for simultaneous condition detection, localization, and classification in air-filled pipes is proposed. The contribution of this work is threefold: (1) a microphone array is used to extend the usable acoustic frequency range to estimate the reflection coefficient from blockages and lateral connections; (2) a robust regularization method of sparse representation based on a wavelet basis function is adapted to reduce the background noise in acoustical data; and (3) the wavelet components are used to localize and classify the condition of the pipe. The microphone array and sparse representation method enhance the acoustical signal reflected from blockages and lateral connections and suppress unwanted higher-order modes. Based on the sparse representation results, higher-level wavelet functions representing the impulse response are used to localize the position of the sensor corresponding to a blockage or lateral connection with higher spatial resolution. It is shown that the wavelet components can be used to train and to test a support vector machine (SVM) classifier for the condition identification more accurately than with a time domain SVM classifier. This work paves the way for the development of simultaneous condition classification and localization methods to be deployed on autonomous robots working in buried pipes.

Methods for determining pipe burst (PB) locations based on supervisory control and data acquisition systems (SCADAs) have been widely studied. Existing location methods either only detect PB areas or are only applied to relatively simple pipe networks. Thus, it is necessary to develop a PB localization method that can both accurately locate the PB and be used for complex pipe networks. Accordingly, in this study, a complex experimental pipe network with a high-frequency SCADA was built, and PB events were constructed under different conditions. After they were analyzed, the water pressure and flow anomalous features of the PBs were summarized. According to the propagation laws and damping theories of transient flow (TF), a PB localization method, that is, the TF simulation (TFS) method, was first proposed and developed after a TF model of the experimental pipe network was established. Using this method, two PB events were recognized, the burst pipes were identified, and the PBs were accurately located. Compared with the negative pressure wave method, the TFS method exhibits a higher location accuracy and a wider applicability to pipe networks. Through error analyses, it is shown that the TFS method has small errors. To reduce cost, the sampling frequency was decreased to 50 Hz, that is, the minimum required sampling frequency for this study, and under this condition, PB localization was successfully realized. An additional three PB localizations illustrate that they have acceptable errors and that the TFS method has good repeatability.

No abstract available

No abstract available

No abstract available

No abstract available

We propose a new method for leak detection and localization in water pipes based on a mathematical model that describes the flow dynamics by two coupled linear first order hyperbolic partial differential equations. Using the modulating function approach, a system of auxiliary PDEs is derived and solved in order to obtain appropriate modulating functions. This allows estimating the leak size and the leak position, resorting to algebraic I/O equations only. For this purpose, no spatial discretization of the PDE model is needed. The theoretical results are validated with experimental data from a water pipe prototype and the performance of the proposed approach is evaluated in comparison to an existing late lumping model-based leak detection system.

These days, pipelines are vital to industry since they are an important component of both the distribution of water resources and the transmission of energy vectors. The biggest challenges in monitoring the pipes are the size of the infrastructures, their age, and the unpredictability of external events. The possibility of leaks in this situation might result in significant financial losses, leak detection and localization are both extremely difficult yet crucial tasks. Many of the shortcomings of the current approaches are that they need system shutdowns, involve invasive procedures or function with short pipelines only. A non-intrusive technique based on acoustic signal processing is presented in this paper. The innovative part of this method is the use of phase diagram representation and the diagram-based entropy computation to determine the position of leaks. We performed a continuous monitoring with our system on a pipeline and the results were compared with the results from the video inspection-based method. Regardless of how long the pipe is, the results show that our proposed technique has a high accuracy.

Water distribution networks are often susceptible to pipeline leaks caused by mechanical damages, natural hazards, corrosion, and other factors. This paper focuses on the detection of leaks in water distribution networks (WDN) using a data-driven approach based on machine learning. A hybrid autoencoder neural network (AE) is developed, which utilizes unsupervised learning to address the issue of unbalanced data (as anomalies are rare events). The AE consists of a 3DCNN encoder, a ConvLSTM decoder, and a ConvLSTM future predictor, making the anomaly detection robust. Additionally, spatial and temporal attention mechanisms are employed to enhance leak localization. The AE first learns the expected behavior and subsequently detects leaks by identifying deviations from this expected behavior. To evaluate the performance of the proposed method, the Water Network Tool for Resilience (WNTR) simulator is utilized to generate water pressure and flow rate data in a water supply network. Various conditions, such as fluctuating water demands, data noise, and the presence of leaks, are considered using the pressure-driven demand (PDD) method. Datasets with and without pipe leaks are obtained, where the AE is trained using the dataset without leaks and tested using the dataset with simulated pipe leaks. The results, based on a benchmark WDN and a confusion matrix analysis, demonstrate that the proposed method successfully identifies leaks in 96% of cases and a false positive rate of 4% compared to two baselines: a multichannel CNN encoder with LSTM decoder (MC-CNN-LSTM) and a random forest and model based on supervised learning with a false positive rate of 8% and 15%, respectively. Furthermore, a real case study demonstrates the applicability of the developed model for leak detection in the operational conditions of water supply networks using inline sensor data.

The problem of water scarcity affects many areas of the world due to water mismanagement and overconsumption and, more recently, to climate change. Monitoring the integrity of distribution systems is, therefore, increasingly important to avoid the waste of clean water. This paper presents a new signal processing technique for enhancing the performance of the methodology of leak detection in water distribution pipes based on time domain reflectometry (TDR). The new technique is based on a particular kind of TDR inversion (spatial TDR) based on a “gray-box” lumped parameter model of the system. The model does not include, e.g., radiative phenomena, non-TEM (transverse electromagnetic) modes etc. but is capable of reproducing accurately the complicated reflectograms obtained by a TDR leak detection system assuming a proper profile of capacitance per unit length along the sensing element. Even more importantly, the model is identified using only the reflectograms taken by the system with very little prior information about the system components. The developed technique is able to estimate with good accuracy, from reflectograms with unclear or ambiguous interpretation, the position and the extension of a region where water is located. The measurement is obtained without prior electromagnetic characterization of the TDR system components and without the need of modeling or quantifying a number of electromagnetic effects typical of on-site measurements.

No abstract available

The increasing demands on urban water infrastructure have made efficient and accurate detection of defects in water pipe networks critical for ensuring operational reliability and resource management. This study presents a novel approach integrating Convolutional Neural Network (CNN), YOLOv5, and Faster R-CNN models to localize and classify water pipe defects. By leveraging a curated dataset of water pipe images, the CNN model demonstrated strong feature extraction capabilities, achieving an accuracy of 89.4%. YOLOv5 balanced detection accuracy (90.1%) and real-time processing (30 images per second), while Faster R-CNN excelled in defect localization with 87.1% accuracy. Compared to traditional edge detection and morphological methods, the proposed system significantly improves robustness, accuracy, and real-time applicability. The findings underscore the potential of deep learning in automating water pipe maintenance, contributing to smarter infrastructure management.

Fire water systems serve as a core component of the city's lifeline program. Its safety and reliability are directly related to the safety of people's lives and property and the operational order of the city. However, firefighting pipelines are often at risk of leakage due to aging, mechanical damage, corrosion, and circumferential weld cracking. Therefore, effective methods for detecting and locating pipeline leaks are essential. Existing methods are often complex, time-consuming, and unreliable. This paper proposes an integrated approach for leak localization in fire-distance fire protection pipelines by constructing a light gradient boosting machine-based model. The model simultaneously addresses the challenges of multi-categorical leak localization with unbalanced data and enhances the accuracy and reliability of leakage detection in long-distance pipelines. The model identifies leak locations through a two-step process and demonstrates high performance, achieving an accuracy rate of over 90%. Furthermore, it exhibits strong generalization capability and robustness. This method significantly enhances the accuracy of leak detection in long-distance fire protection pipelines, improves the operational reliability of fire water networks, and enables staff to make prompt maintenance decisions.

No abstract available

Researchers and engineers employ machine learning (ML) tools to detect pipe bursts and prevent significant non-revenue water losses in water distribution systems (WDS). Nonetheless, many approaches developed so far consider a fixed number of sensors, which requires the ML model redevelopment and collection of sufficient data with the new sensor configuration for training. To overcome these issues, this study presents a novel approach based on Long Short-Term Memory neural networks (NNs) that leverages transfer learning to manage a varying number of sensors and retain good detection performance with limited training data. The proposed detection model first learns to reproduce the normal behavior of the system on a dataset obtained in burst-free conditions. The training process involves predicting flow and pressure one-time step ahead using historical data and time-related features as inputs. During testing, a post-prediction step flags potential bursts based on the comparison between the observations and model predictions using a time-varied error threshold. When adding new sensors, we implement transfer learning by replicating the weights of existing channels and then fine-tune the augmented NN. We evaluate the robustness of the methodology on simulated fire hydrant bursts and real-bursts in 10 district metered areas (DMAs) of the UK. For real bursts, we perform a sensitivity analysis to understand the impact of data resolution and error threshold on burst detection performance. The results obtained demonstrate that this ML-based methodology can achieve Precision of up to 98.1% in real-life settings and can identify bursts, even in data scarce conditions.

Data-driven methods based on samples from a supervisory control and data acquisition system have been widely applied in water-supply-network burst detection to save unexpected economic and labor costs. However, the class imbalance problem in actual on-site monitoring needs to be revised to improve the performance of data-driven methods. In this study, we proposed a domain adaptation method to generate minor-category samples (pipeline-burst samples in general) of arbitrary pipe networks utilizing theoretical hydraulic models. The proposed method transferred pipeline-burst data generated from a random water supply network with theoretical hydraulic models to an actual imbalanced dataset. Accordingly, we established a machine learning model exploring a mapping matrix between two domains for minority-category data transfer. The experimental validation first verified the effectiveness and reliability of the proposed method between two customized water supply networks in terms of their bust recognition accuracy, model parameter sensitivity and time efficiency. Then, an actual monitoring dataset from a working water supply network was used to prove the suitability and compatibility of the proposed method. A bust-point location method was also provided based on the detection results of pipeline-bursting events. The validations show the superiority of our proposed approach for the imbalance data problem in pipe burst detection.

In order to improve the accuracy of pipe burst detection in water distribution networks (WDNs), a novel small sample diagnosis method (SSDM) based on the head loss ratio (HLR) method and deep transfer learning (DTL) method has been proposed. In this paper, the burst state was quickly detected through the limited data analysis of pressure monitoring points. The HLR method was introduced to enhance data features. DTL was introduced to improve the accuracy of small sample burst detection. The simulated data and real data were enhanced by HLR. Then, the model was trained and obtained through the DTL. The performance of the model was evaluated in both simulated and real scenarios. The results indicate that the leaked features can be improved by 350% by the HLR. The accuracy of SSDM reaches 99.56%. The SSDM has been successfully applied to the detection of real WDNs. The proposed method provides potential application value for detecting pipe bursts.

The companies responsible for water supply networks are making significant efforts to improve leakage detection efficiency. This study proposes a novel leak localization method. First, the pressure-driven background leakage model (PDBLM) is established using a nonlinear genetic algorithm that considers the relationship between background leakage and pressure. Second, a fuzzy c-means clustering algorithm optimized by a genetic simulated annealing algorithm (FCMGSAA) and sparrow search algorithm optimized probabilistic neural networks (SSAPNNs) are chosen for the leak location model. To reduce the output classes, a large water supply network is divided into several areas using an FCMGSAA. The areas are used as learning labels, reducing the output classes of the SSA. Then, the leak location model is used to detect the leak area based on a real case in Qingdao, China. Compared with generalized regression neural networks (GRNNs), support vector machine (SVM), convolutional neural networks (CNNs) and back-propagation neural networks (BPNNs), the accuracy of the proposed method is improved by 70, 43.33, 10 and 6.67%, respectively. The proposed method can improve leakage detection efficiency, reduce water loss and contribute to the effective management of water resources.

This paper reports on the use of a circular microphone array to analyze the reflections from a pipe defect with enhanced resolution. A Bayesian maximum a posteriori algorithm is combined with the mode decomposition approach to localize pipe defects with six or fewer microphones. Unlike all previous acoustic reflectometry techniques, which only estimate the location of a pipe defect along the pipe, the proposed method uses the phase information about the wave propagated in the form of the first non-axisymmetric mode to estimate its circumferential position as well as axial location. The method is validated against data obtained from a laboratory measurement in a 150 mm diameter polyvinyl chloride pipe with a 20% in-pipe blockage and 100 mm lateral connection. The accuracy of localization of the lateral connection and blockage attained in this measurement was better than 2% of the axial sensing distance and 9° error in terms of the circumferential position. The practical significance of this approach is that it can be implemented remotely on an autonomous inspection robot so that accurate axial location and circumferential position of lateral connections and small blockages can be estimated with a computationally efficient algorithm.

AI-based sensor placement for anomaly detection in water distribution systems (WDS) enhances system reliability and efficiency. Advanced WDS are attractive targets for malicious actors who may gain control over these systems to disrupt or manipulate water services for personal gain. Monitoring various points in the network is crucial for effective anomaly detection. However, increasing the number of sensors or placing them inaccurately can raise computational costs, increase management overhead, and degrade detection performance. In this study, we address sensor placement optimization (SPO) and anomaly detection (AD) in WDS to improve performance while reducing computational and management overhead by focusing on critical sensor locations. We use the BATADAL dataset, which includes key WDS components such as pump units, pressure junctions, and tanks. Our implementation identifies the most critical pump unit locations for monitoring flow rates, allowing us to focus only on essential points to lower computational costs and management overhead in anomaly detection. For SPO, we implement Graph Neural Networks (GNN) to extract relationships between pump units and other components, identifying the most significant pump units for anomaly prediction. GNN-based SPO (GSPO) selects critical locations, and we then use flow rate data from these selected sensors to train the AD model. We employ a Long Short-Term Memory (LSTM)-based AD approach (LAD) to capture sequential patterns in the sensor data. The LAD approach achieved significant results with an accuracy of 0.925 and an F1-score of 0.75. To further enhance model performance over time, we adopt a continuous learning approach. We fine-tune the LSTM model monthly with new incoming data while retaining previous knowledge using a weight interpolation approach, ensuring the model remains up-to-date and continues to improve in detecting anomalies.

Effective management of water distribution networks is necessary, particularly in regions that are experiencing climate stress. Water losses in Chrysoupoli, Northern Greece, can amount to 62% because of deteriorating infrastructure and seasonal variations in demand. 22 Internet of Things (IoT)-enabled Local Monitoring Stations (LMS) that measure flow, pressure, and water quality are used in this study's integrated remote sensing strategy for leakage monitoring. The STAGON platform receives real-time data transmissions via SCADA. Dynamic simulation, anomaly detection, and improved leak response are all supported by a digital twin created with MIKE+. Accuracy, speed, and cost-efficiency are greatly increased when automated leak detection is combined with remote sensing, providing a scalable solution for Smart Water Governance.

This research work proposes a robust and efficient real-time water leak detection system by using an integrated approach. By harnessing the capabilities of water flow sensors, microcontrollers, and LoRa transmitters, it provides comprehensive monitoring and instant alerts. The sensors continuously track water flow, transmitting data to the microcontroller for analysis. Any anomaly triggers the microcontroller to relay critical information to the LoRa receiver, initiating an urgent response. This prompts the I2C LCD display to visually notify users of a leak detection in real-time, while a buzzer makes an audible alarm. The synchronized visual and auditory alerts enable users to promptly intervene, mitigating potential water damage. The proposed system architecture ensures seamless coordination between its components for rapid leak identification. The two water flow sensors swiftly detect irregularities, communicating this data to the microcontroller. It then relays the information via the LoRa transmitter to the receiver. On obtaining the transmission, the receiver conveys the signal back to the microcontroller to activate the output interfaces. The I2C LCD display and buzzer then provide instant visual and audio cues to alert users. This end-to-end integration facilitates real-time communication, analysis, and response, empowering timely action against water leaks. Hence, this study proposes an effective solution through its robust architecture. By leveraging key technologies like sensors, microcontrollers, and transmitters, it enables comprehensive real-time monitoring and instant alerts against water leaks.

The implementation of a smart water network (SWN) is viewed as a strategic approach to address many challenges faced by water utilities, such as pipe leak detection and main break prevention. This paper develops a convolutional neural network (CNN)–based model to classify acoustic wave files collected by the South Australian Water Corporation’s (SA Water’s) SWN over the city of Adelaide. The VGGish model (VGG refers to the team who developed the model—Visual Geometry Group) is selected as a suitable transfer learning model to extract features from wave files. The CNN model classifies an acoustic wave file as an anomaly or other background or environmental noise. Identification of a wave file as an anomaly triggers a Siamese CNN model to determine whether it is related to a regular/irregular scheduled event (for example, irrigation system near public parks or water consumption by large buildings). A field investigation is initiated if a wave file is classified as an anomaly and it is not related to a scheduled event. The developed models have been validated using data that is recorded by SWN in Adelaide. This validation data set comprises 1098 wave files, which are recorded by 34 accelerometers and are associated with 32 known leaks. The validation results shown that accuracy of alarms generated by the developed models is 92.44%. The validations confirm the developed models as an effective tool for water pipeline leak and crack detection, which, in turn, enables proactive management of the pipeline assets.

Water supply network monitoring data usually contains a large number of anomalous data, which is difficult to reflect the real working condition of the water supply pipe, affecting the accuracy of the wind power prediction, and thus causing certain economic losses. In order to solve this problem, we analyze the characteristics of the abnormal data and propose a data cleaning method for water supply pipe network based on support vector machine, and use the seasonal differential auto regressive moving average model as a comparison. The seasonal differential auto regressive moving average model is used as a comparison to determine the advantages of its high accuracy rate of anomaly identification, which can better repair the errors and missing data in the pipe network monitoring data and provide a more reliable basis for analyzing and making decisions. In this paper, the monitoring data of water supply pipe network in a southern city were collected and preprocessed, including the filling of missing values and the detection of data outliers. Next, we modeled and predicted the data using support vector machines to identify potential outliers. Then, for potential outliers, we compare them with the original data for secondary determination. Finally, we use support vector machine for data filtering and use the cleaned data for the operation and management of the pipe network system. The effectiveness of the method in improving the quality and accuracy of the monitoring data was demonstrated by testing and analyzing the actual data. Future research could further explore and optimize the method to improve the operational efficiency and reliability of the pipe network system.

Underground pipeline leaks and infiltrations pose significant threats to water security and environmental safety. Traditional manual inspection methods provide limited coverage and delayed response, often missing critical anomalies. This paper proposes AquaSentinel, a novel physics-informed AI system for real-time anomaly detection in urban underground water pipeline networks. We introduce four key innovations: (1) strategic sparse sensor deployment at high-centrality nodes combined with physics-based state augmentation to achieve network-wide observability from minimal infrastructure; (2) the RTCA (Real-Time Cumulative Anomaly) detection algorithm, which employs dual-threshold monitoring with adaptive statistics to distinguish transient fluctuations from genuine anomalies; (3) a Mixture of Experts (MoE) ensemble of spatiotemporal graph neural networks that provides robust predictions by dynamically weighting model contributions; (4) causal flow-based leak localization that traces anomalies upstream to identify source nodes and affected pipe segments. Our system strategically deploys sensors at critical network junctions and leverages physics-based modeling to propagate measurements to unmonitored nodes, creating virtual sensors that enhance data availability across the entire network. Experimental evaluation using 110 leak scenarios demonstrates that AquaSentinel achieves 100% detection accuracy. This work advances pipeline monitoring by demonstrating that physics-informed sparse sensing can match the performance of dense deployments at a fraction of the cost, providing a practical solution for aging urban infrastructure.

Maintaining high-quality water resources is essential for sustainable urban water resource management and public health. This paper introduces the GRU-PINN model, developed based on the Gated Recurrent Unit (GRU) network and integrated with a Physics-Informed Neural Network (PINN), to analyze real-world monitoring data from a water treatment company. By embedding domain-specific physical constraints into the loss function, the model enhances interpretability and reduces false alarms. The proposed approach begins with feature engineering on the raw dataset, including missing value imputation, normalization, trend feature extraction, and rolling feature computation. Feature selection is then performed based on feature importance ranking and mutual information analysis. The GRU-PINN model is subsequently employed for anomaly detection in the dataset. The performance of the model is evaluated using the F1-score, precision, and recall. The F1-score, which represents the harmonic mean of precision and recall, is particularly suitable for imbalanced datasets, as the dataset used in this paper contains very few anomaly instances. Experimental results demonstrate that the proposed GRU-PINN model outperforms traditional models by achieving a higher F1-score, thereby improving anomaly detection performance.

As urban populations grow and aging infrastructure undergoes increasing strain, the need to keep Water Distribution Systems (WDS) as efficient, resilient, and sustainable as possible has become a priority for modern cities. In order to do so, this paper presents a holistic framework that blends Digital Twin (DT) with Machine Learning (ML) for intelligent monitoring and controlling of Water Distribution Systems (WDS). This system is based on the construction of an online virtual model of a water network using EPANET simulations, and additionally, it carries out anomaly detection, demand forecasting, leak identification, and pressure control through the implementation of machine learning algorithms. The datasets include real sensor data (pressure, flow rate, turbidity, pH), SCADA logs, and synthetic anomalies generated from simulations. An Isolation Forest model achieved 97.8% accuracy for identifying system anomalies, while a Long Short-Term Memory (LSTM) model returned an $R^{2}$ of $\mathbf{9 4 \%}$ in forecasting water demand. As a result, a Deep Q-Learning agent reduced zone pressure variance by 66.3 %, thus enabling performance improvements and a reduction in energy consumption. This end-to-end use of cloud and edge technologies builds an integrated DT-ML pipeline that utilizes real-time data streaming, analytics, and control actions deployed through a feedback loop. The results are visualized via Grafana dashboards to assist utility operators with intuitive monitoring. The study demonstrates the transformative potential of combining digital technologies and machine learning to shift from reactive monitoring to proactive and autonomous water management. The proposed approach ensures both service availability and optimal resource utilization, and lays the foundation for future scalability of autonomous infrastructure systems in alignment with Industry 4.0 goals.

Effective water management is essential for ensuring sustainability, reducing resource waste, and addressing challenges such as water scarcity and pollution. This paper presents an Artificial Intelligence of Things (AIoT) driven smart water monitoring system, integrating Internet of Things (IoT) technologies and machine learning (ML) to optimize water consumption and detect anomalies in real-time. Implemented at the IFSP Catanduva Campus in Brazil, the system monitors key water parameters, identifies irregularities like leaks, and enables proactive management through data-driven decision-making. Case studies demonstrate its effectiveness, showcasing improvements such as swift leak detection, reduced response times, and operational cost savings. The system achieved an F1-score of 97.3\% using a Random Forest classifier, highlighting its performance in anomaly detection. Despite minor data variability and network reliability challenges, the scalable design offers adaptability to other institutions and urban environments.

: Industrial control systems often contain sensor and actuator devices, which provide monitoring data in the form of time series, such as bridge vibrations, water distribution systems, and human physiological data. This paper proposes an anomaly detection model based on autoencoders that can consider time-series relations of the data. Moreover, the quality of the decoder output is further improved by adding a residual produced by an extra generator and discriminator. The proposed autoencoder-GAN model and detection algorithm not only improved the performance but also made the training process of GAN easier. The proposed deep learning model with the anomaly detection algorithm has been shown to achieve better results on the SWaT, BATADAL, and Rare Event Classification datasets over existing methods.

This paper proposes a scalable AI framework. The framework combines high-frequency flow and pressure data from the smart meters with a distributed edge-computing layer, where lightweight machine learning models are applied for performing initial anomaly detection in real time. Detected events then trigger an adaptive neural network, which refines the detection and estimates the leak characteristics more accurately. A pilot implementation in a mid-sized Indian city shows the capability of the system to localize leaks with a spatial resolution of better than 50 m and detect small leaks (< 2 L/min) within 30 minutes of occurrence. We use a digital copy (twin) of the distribution network, built from the available pipe layouts, nodal elevations, and historical demand patterns, as the data source for training a supervised-learning algorithm to overcome the lack of labelled leak events. By conducting a series of field trials, we provide quantitative performance results in terms of the true/false positive rates, localization error, and energy overhead of edge processing. A cost-benefit analysis is conducted that evaluates deployment scenarios at different penetration rates of smart meters and shows payback periods below three years for networks with > 60%-meter coverage. Finally, we discuss practical considerations for large-scale deployment, including integration with existing water utility management systems, scalability challenges, data privacy and security concerns, maintenance of edge-computing devices, and potential regulatory and policy implications. The results demonstrate that the proposed framework is not only technically feasible but also economically viable, providing utilities with a robust tool for improving leak detection, reducing water losses, and enhancing operational efficiency.

No abstract available

Time series anomaly detection tools (TSAD) are widely applicable across industries, such as monitoring time series data of water pipeline pressure, network traffic activities, and hardware telemetry. The primary objective is to identify anomalous segments and alert users to potential issues before any consequences. A closely related tool is time series forecasting, and some practitioners leverage it for anomaly detection. As the main objective of a forecasting model is to minimize errors, it tends to over-fit the time series, making it challenging to distinguish whether the forecasting errors occur due to model limitations or true anomalous segments. This paper introduces a method called the posterior anomaly scoring criterion, which uses deep learning time series forecasting models to estimate forecast uncertainties for TSAD. We propose replacing the forecasting model’s output layer to estimate forecast distributions and compute the probability of the posterior distribution to attain anomaly scores. These scores are processed through an automated threshold criterion to classify the anomalous segments. The experiments have demonstrated our model performs the best in four out of five datasets across seven benchmark models.

Efficient monitoring and quick feedback control are the main requirements of smart cities to guarantee the stability and safety of urban infrastructures. Real-time monitoring in order to detect anomalies can lead to the data intensive processing requires a new computing scheme to offer large-scale and low latency services. Fog architecture by extending computing to the edge of network, provides the ability to accurate and fast detection of abnormal patterns. The hierarchical fog computing architecture and the efficient hyperellipsoidal clustering algorithm presented in the previous studies have been applied in this paper to identify anomalous behaviors in water distribution grids. We created an urban water distribution grid dataset using Epanet2w simulator software by recording grid measured features as (pressure and head) for several scenarios. To evaluate the effect of applying the hierarchical anomaly detection model, we implemented the data and computing nodes at different layers by docker containers. The evaluation results proved the effectiveness of the hierarchical anomaly detection model in significant reduction of the communication latency, while preserving the detection accuracy compared to the centralized scheme.

Anomalies are defined as behavioral deviations from expected patterns and pose challenges to identify them. Anomaly detection is a fundamental activity of time series analysis. It enables informed decision-making in many control and monitoring activities, such as healthcare, water quality, seismic reflection analysis, and oil exploration. Many anomaly detection methods exist, but choosing the appropriate methods is complex due to the intrinsic nature of the time series. There is a demand for robust and adaptable anomaly detection methods. This paper introduces Refined Empirical Mode Decomposition (REMD) as a hybrid approach addressing this need, integrating Empirical Mode Decomposition (EMD) and Autoregressive Integrated Moving Average (ARIMA) models. REMD's design aims to optimize the strengths of both methods and overcome their limitations. It is evaluated against state-of-the-art methods on diverse datasets. It demonstrates superior performance, with up to three times better F1 score.

—Utilizing IoT technologies for monitoring large-scale smart facilities such as power, water and gas distribution networks has been the subject of many studies recently. The aim is to detect anomalous events in the network due to elements’ failure, bad designs, attacks or abuses of the network and alert the network operators in a timely manner. As the centralized cloud-based approaches are impractical in time-critical and real-time anomaly detection applications due to 1) high sensor-to-cloud transmission latency 2) high communication cost and 3) high energy consumption at the sensor nodes, the distributed anomaly detection methods based on Deep Neural Networks (DNN) have been applied in past studies vastly. In these methods, in order to detect anomalies in real-time, copies of the anomaly detection model are placed at the sensor nodes (rather than placing one at the cloud node) reducing the sensor-to-cloud transmissions significantly. Nevertheless, new normal samples collected at the sensor nodes still need to be transmitted to the cloud node at predefined intervals to re-train the distributed anomaly detection DNNs. In order to minimize these sensor-to-cloud transmissions during the retraining process, in this paper, two well-known lossless coding algorithms: Huffman Coding and Arithmetic Coding were studied and it was observed that the Huffman and Arithmetic Coding were able to reduce the transmission traffic up to 50% and 75% respectively using two IoT benchmark datasets of pipeline measurements. Besides, the Huffman Coding shown to be computationally feasible on resource limited sensors and resulted in up to 10% saving in energy consumption on each sensor resulting in longer network longevity. Moreover, the experimental results showed that the auto-encoder DNN could outperform the one-class SVM in the iterative distributed anomaly detection method.

There is a growing need for solutions that can improve the communication between anomaly detection algorithms and human operators. In the context of real-time monitoring of networked systems, it is crucial that new solutions do not increase the burden on an already overloaded visual channel. Sonification can be leveraged as a peripheral monitoring tool that complements current visualization systems. We conceptualized, designed, and prototyped Datascapes, a framework project that explores the potential of sound-based applications for the monitoring of cyber-attacks on AI-supported networked environments. Within Datascapes, two Design Actions were realized that applied sonification on the monitoring and detection of anomalies in (1) water distribution networks and (2) Internet networks. Two series of prototypes were implemented and evaluated in a real-world environment with eight experts in network management and cybersecurity. This paper presents experimental results on the use of sonification to disclose anomalous behavior and assess both its gravity and the location within the network. Furthermore, we define and present a design methodology and evaluation protocol that, albeit grounded in sonification for anomaly detection, can support designers in the definition, development, and validation of real-world sonification applications.

Urban water distribution networks are essential infrastructure but face persistent issues from leaks, bursts, and defects in underground pipes, resulting in significant water losses and operational inefficiencies. Traditional detection methods, including manual inspections and pressure monitoring, frequently exhibit limitations, high costs, and protracted durations. This work introduces an innovative method for identifying leaks and defects through the integration of Isolation Forest (IF), an unsupervised anomaly detection system, with big data analytics. Real-time data from sensors, flow meters, and smart meters is analyzed to detect anomalous patterns suggestive of leaks or system malfunctions. Experimental findings from a simulated urban water network indicate that the proposed method achieves a detection accuracy of over $95 \%$ while maintaining a false alarm rate under $3 \%$, improving traditional approaches. The system utilizes big data analytics to provide continuous monitoring, predictive repair, and efficient resource management, minimizing non-revenue water and assuring a stable supply. This study highlights the practical significance of integrating machine learning with big data, providing a scalable, precise, and economical framework for contemporary urban water network management.

Water Distribution Networks (WDNs) are complex, dynamic systems critical to modern society but increasingly difficult to manage due to urbanization, fluctuating demands, and resource constraints. To address these challenges, Smart Water Distribution Networks (SWDNs) utilize Internet of Things (IoT) devices and protocols like Long Range Wide Area Network (LoRaWAN) for real-time monitoring and analysis, enabling smarter and more efficient water management. This demo presents SWIM (Smart Water Interaction & Monitoring), an innovative application designed to modernize SWDNs. SWIM integrates Digital Twins (DTs), established simulation tools like EPANET, and Machine Learning (ML) to provide predictive analytics, anomaly detection, and real-time control. By employing neural networks, SWIM achieves high-accuracy hydraulic predictions with minimal input data. Built on IoTs and Low Power Wide Area Networks (LPWANs), SWIM delivers scalable, efficient, and user-friendly solutions. It aligns with the principles of Industry 5.0, demonstrating the potential to revolutionize water distribution networks and ensure their sustainability in the face of modern challenges.

This comprehensive framework for implementing smart IoT infrastructure in urban water pipeline monitoring systems integrates thousands of distributed sensors measuring critical parameters with a robust data engineering pipeline that enables real-time processing and analysis. By leveraging streaming technologies, columnar databases, and advanced analytics algorithms, the system facilitates anomaly detection, predictive maintenance, and operational optimization across extensive urban networks. The data-driven approach allows utility managers to transition from reactive to proactive infrastructure management, significantly reducing emergency repairs while improving service reliability. Through visualization dashboards and automated alert systems, stakeholders gain unprecedented visibility into network health and performance trends. The findings suggest that such smart infrastructure implementations not only enhance operational efficiency but also contribute to more sustainable and resilient urban water systems, providing valuable decision support for long-term infrastructure planning and resource allocation.

Data from network-structured applications, like sensor networks or smart grids, often reside on complex supports. Specific graph signal processing tools are needed for effective utilization. Detecting anomalous events in graph signals holds relevance across various applications, ranging from monitoring energy and water supplies to environmental surveillance. In these problems, anomalies often activate localized groups of vertices in the graph. This paper introduces the Joint Graph-Regularized Wavelet CCA (JGWCCA) approach, which combines canonical correlation analysis (CCA) with dual-tree complex wavelet packet transform (DT-CWPT) and graph regularization. JGWCCA enables time-frequency analysis of graph signals while considering the underlying graph topology. Performance validation of JGWCCA is done through numerical simulations.

Built utility infrastructures provide essential services such as water, gas, and power to communities, and their resilient operation under anomalies and spurious events is critical. In this paper, we study the deployment of heterogeneous IoT sensors in geo-distributed infrastructure networks, using stormwater as a driving usecase. These systems are responsible for drainage and flood control, but in doing so, serve as conduits that carry pollutants to receiving waters. The timely detection of such events is challenging, due to the transient/random nature of pollutants, scarce historical data, and complexity of the system. We present STEP, an integrated framework for sensor placement that leverages the network structure and topology, behavioral properties (e.g., flow rate), and community semantics such as locations of facilities (e.g., commercial spaces, residential areas, and industrial plants, etc.). We identify key metrics to capture anomaly coverage and traceability, use past pollution incidents to inform sensor deployment, and model network operations through physics-based simulations and community-scale semantics. STEP is evaluated on six real-world stormwater networks, which show the efficacy of our approach over existing methods.

Reducing Non-Revenue Water (NRW) losses in water transmission and distribution networks is a critical challenge for water utility companies. The combination of unobtrusive Internet of Things (IoT) monitoring devices and Artificial Intelligence (AI) technology is one of the most promising directions in water leak detection techniques for industrial scale infrastructure and smart cities. Currently, the complicated network topology and underground nature of transmission and distribution water pipelines pose serious limitations for the effective elimination of associated water leaks. In this paper, a realistically dimensioned IoT-enabled water transmission system provides the basis for a series of simulated leak experiments and the subsequent application of three different anomaly detection schemes. Having full control over the mechanical valve behind the simulated leaks, this test rig addresses the issue of accurate labelling in leak data and serves as testbed for the evaluation of each anomaly detection method and the comparison between them. The first anomaly detection method is the unsupervised multi-variate classification known as Isolation Forest (iForest). Second, the Support Vector Classification (SVC) approach is implemented representing supervised classification methods in the Support Vector Machine (SVM) family. Finally, a deep learning RNN-LSTM (Recurrent Neural Networks-Long Short-Term Memory) model is used in conjunction with a single threshold to signify anomalies due to high deviations between actual and forecasted values of key infield sensors. These models can detect water leaks and the results provide insights regarding both the effective applicability of sensors and Machine Learning (ML) algorithms in this context.

Monitoring the quality of water plays a vital role in the advancement towards intelligent and smart agriculture. It also facilitates the seamless transition to automated monitoring of essential components of human daily needs, as new technologies continue to be developed and adopted in both agricultural and human daily life, particularly in relation to water. Assessing water stress requires accurate measurement of water usage [1]. The objective of this study is to optimize the distribution strategy and achieve more efficient and rapid distribution, while also making better use of technology to address the urgent need for water conservation. One of the criteria for the second filtering phase, which utilized a realtime data collection system, focused on the implementation of water quality monitoring and characterization methods. Water measurement sensors and the Internet of Things (IoT) can provide immediate information on water flow and pressure rates. This information is crucial for the analysis and surveillance of water [2]. At the RCEW campus, online smart water monitoring is imperative to ensure a more efficient and sustainable use of water. By combining the Internet of Things and geographic information systems (GIS), an online smart water monitoring prototype was developed specifically for RCEW [3]. This cutting-edge technology has proven to be highly successful in providing real-time information on water parameters, saving both time and money, while enabling the continuous study of water consumption patterns. This innovative technology seamlessly integrates IoT and GIS for effective water surveillance [4,5].

The presence of impurities in water resulting from the expansion of industrial activities and the disregard of regulations aimed at protecting the environment presents considerable threats to the well-being of the general population. The establishment of reliable access to uncontaminated drinking water assumes paramount importance in the realm of disease prevention. The continuous monitoring of the quality of water in real-time serves as an indispensable measure for ensuring its safety. The objective of this article is to introduce a system that capitalizes on the Internet of Things (IoT) to automate the distribution and management of water resources within urban areas. By making use of embedded technology, this system allows for the uninterrupted surveillance and regulation of water quality. To further streamline the management process, an application compatible with the Android operating system has been developed to promptly notify users of any irregularities, thereby enabling remote control. Moreover, a web-based application has been designed to cater to the needs of iOS users, thereby broadening accessibility. This comprehensive approach not only heightens the level of user awareness but also provides them with greater control over their water supply, thereby enhancing public health outcomes.

Smart infrastructure powered by Artificial Intelligence (AI) and Machine Learning (ML) transforms modern urban environments by optimizing energy distribution, water management, transportation, and security systems. This study presents an AI-ML-driven architecture for smart infrastructure deployment, integrating real-time and historical data from various sensors and IoT devices. The proposed system employs advanced predictive modeling techniques to enhance decision-making processes. To develop and analyze this architecture, a combination of tools, including Python, TensorFlow, Scikit-learn, and SQL-based databases, was used for data processing, model training, visualization, and system design. The dataset consists of real-time sensor readings from smart grids, water distribution networks, traffic monitoring systems, surveillance cameras, and historical records from cloud storage and edge computing environments. Key findings demonstrate that AI-driven models significantly improve resource management’s predictive accuracy, reduce data processing latency, and enhance overall system efficiency. Graphical representations, such as deployment diagrams and time-series forecasting graphs, illustrate the interactions between AI models, infrastructure components, and deployment services. The results underscore the potential of AI-ML frameworks to optimize smart infrastructure, making them more adaptive and resilient.

No abstract available

Ensuring consistent high water quality is paramount in water management planning. This paper addresses this objective by proposing an intelligent edge-cloud framework for water quality monitoring within the water distribution system (WDS). Various scenarios—cloud computing, edge computing, and hybrid edge-cloud computing—are applied to identify the most effective platform for the proposed framework. The first scenario brings the analysis closer to the data generation point (at the edge). The second and third scenarios combine both edge and cloud platforms for optimised performance. In the third scenario, sensor data are directly sent to the cloud for analysis. The proposed framework is rigorously tested across these scenarios. The results reveal that edge computing (scenario 1) outperforms cloud computing in terms of latency, throughput, and packet delivery ratio obtaining 20.33 ms, 148 Kb/s, and 97.47%, respectively. Notably, collaboration between the edge and cloud enhances the accuracy of classification models with an accuracy of up to 94.43%, this improvement was achieved while maintaining the energy consumption rate at the lowest value. In conclusion, our study demonstrates the effectiveness of the proposed intelligent edge-cloud framework in optimising water quality monitoring, and the superior performance of edge computing, coupled with collaborative edge-cloud strategies, underscores the practical viability of this approach.

No abstract available

Ensuring access to clean water in urban areas is challenging due to leaks and contamination in complex water distribution systems (WDS). Traditional leak detection methods are slow, costly, time‐consuming and not very efficient, motivating the need for advanced solutions. Therefore, this study mainly focuses on Machine learning (ML)‐based leak detection method and localization by leveraging network data. Method uses emitter coefficients to simulate water loss and evaluates ML models, including K‐nearest neighbour (KNN), Random Forest and Support Vector Machine (SVM), on two example networks EPANET Network 3 and real‐life network of National Institute of Technology Kurukshetra campus. Random Forest achieved accuracy scores of 88.13% for Network 3 and 95.84% for the NIT network, with Area Under the Curve (AUC) scores of 0.87 and 0.98, respectively. The results highlight the model's effectiveness in detecting and localizing leaks, contributing to efficient water management. Further, the advantages and limitations of the study are discussed. The future study focuses on efficient applications of these models on real‐world problem.

Due to the growing widespread use of high-frequency measurements in the water supply industry, new possibilities for precise analyses of water consumption are emerging. Studies on a global scale of the performance of water supply networks, i.e. entire water supply zones DMA (District Metered Area), are widely recognized. The implementation of precise measurements of the volume of water consumed at the local scale, even for individual buildings, allows for drawing up detailed patterns of water consumption, demonstrating the cyclicality of their consumption and the interpretation of behavioral aspects of potable water use. However, there is a research gap in the diagnosis of water supply networks in the use of the power spectra of high-frequency water consumption signals to learn about the periodicity of water consumption at the scale of a single building and to identify abnormal states as a result of a disturbed pattern of its power spectral density (PSD) scaling. The article presents the results of research on the variation in the PSD scaling of water consumption, indicated by periodograms, as a result of a short-term anomalous situation caused by the failure of the domestic water supply system. It is shown that, by means of spectral analysis, it is possible to observe anomalies of different magnitudes in the signal power spectrum. It was also confirmed that the developed method can be further supported by a continuous water consumption approach. The presented methodology can be applied to the water expert systems for a rapid diagnosis of the operational status of the water networks regarding the possible occurrence of water leakage.

Pressure management and optimal placement of pressure reducing valves (PRVs) play a critical role in minimizing leakage and maintaining operational efficiency in water distribution networks (WDNs). While evolutionary and swarm-based algorithms such as genetic algorithm (GA) and particle swarm optimization (PSO) have demonstrated promising results for PRV configuration, their efficiency is often contingent on careful parameter tuning and can be limited by convergence sensitivity or complex network scenarios. This study systematically evaluates the performance of the teaching-learning-based optimization (TLBO) algorithm – a population-based, parameter-free metaheuristic – for PRV placement and setting adjustment. Exploiting TLBO's teacher and learner phases, the approach achieves robust, repeatable leakage reduction without the need for intensive algorithm customization. Comparative results on a benchmark WDN confirm that TLBO not only minimizes the number of objective function evaluations but also consistently identifies PRV configurations that outperform or match those found by GA, PSO, and other established methods. The findings underline TLBO's practical advantages and highlight future opportunities to combine intelligent pipe selection strategies and real-time adaptive PRV control for further improvement in leakage management.

No abstract available

No abstract available

No abstract available

In the fast-moving world of information and communications technologies, one significant issue in metropolitan cities is water scarcity and the need for an intelligent water distribution system for sustainable water management. An IoT-based monitoring system can improve water distribution system management and mitigate challenges in the distribution network networks such as leakage, breakage, theft, overflow, dry running of pumps and so on. However, the increase in the number of communication and sensing devices within smart cities has evoked challenges to existing communication networks due to the increase in delay and energy consumption within the network. The work presents different strategies for efficient delay and energy offloading in IoT-integrated water distribution systems in smart cities. Different IoT-enabled communication network topology diagrams are proposed, considering the different water network design parameters, land cover patterns and wireless channels for communication. From these topologies and by considering all the relevant communication parameters, the optimum communication network architecture to continuously monitor a water distribution network in a metropolitan city in India is identified. As a case study, an IoT design and analysis model is studied for a secondary metropolitan city in India. The selected study area is in Kochi, India. Based on the site-specific model and land use and land cover pattern, delay and energy modeling of the IoT-based water distribution system is discussed. Algorithms for node categorisation and edge-to-fog allocation are discussed, and numerical analyses of delay and energy models are included. An approximation of the delay and energy of the network is calculated using these models. On the basis of these study results, and state transition diagrams, the optimum placement of fog nodes linked with edge nodes and a cloud server could be carried out. Also, by considering different scenarios, up to a 40% improvement in energy efficiency can be achieved by incorporating a greater number of states in the state transition diagram. These strategies could be utilized in implementing delay and energy-efficient IoT-enabled communication networks for site-specific applications.

Water is a vital resource for life, and its management is a key issue nowadays the water demand is increasing due to global population growth and urbanization. Present technologies for water control are currently facing interoperability problems due to the lack of support for standardization in monitoring and controlling equipment. This problem affects various processes in water management, such as water consumption, water distribution, system identification, and equipment maintenance. Under these conditions, new technologies in the water management infrastructure have been required to enable the distribution of high-quality water to users safely and cost-effectively, from the perspective of efficiently using our world’s precious water resource. The Internet of Things (IoT) Based Smart Water Management System is a cutting-edge technology that integrates sensors, communication devices, and data analytics to efficiently monitor and manage water resources. This system enables real-time monitoring of water quality, consumption, and distribution, allowing for proactive maintenance and optimization of water usage. By leveraging IoT technology, this smart water management system offers improved accuracy, reliability, and cost-effectiveness in managing water resources. Based on standards, the system proposal is a smart water management system model combining Internet of Things technology with decision support systems. This system is more efficient in distribution monitoring and control approach for a water utility to reduce water loss. This approach will help utility operators to improve water management systems, especially by exploiting emerging technologies.

The steady state of a water distribution system abides by the laws of mass and energy conservation. Hydraulic solvers, such as the one used by EPANET approach the simulation for a given topology with a Newton-Raphson algorithm. However, iterative approximation involves a matrix inversion which acts as a computational bottleneck and may significantly slow down the process. In this work, we propose to rethink the current approach for steady state estimation to leverage the recent advancements in Graphics Processing Unit (GPU) hardware. Modern GPUs enhance matrix multiplication and enable memory-efficient sparse matrix operations, allowing for massive parallelization. Such features are particularly beneficial for state estimation in infrastructure networks, which are characterized by sparse connectivity between system elements. To realize this approach and tap into the potential of GPU-enhanced parallelization, we reformulate the problem as a diffusion process on the edges of a graph. Edge-based diffusion is inherently related to conservation laws governing a water distribution system. Using a numerical approximation scheme, the diffusion leads to a state of the system that satisfies mass and energy conservation principles. Using existing benchmark water distribution systems, we show that the proposed method allows parallelizing thousands of hydraulic simulations simultaneously with very high accuracy.

There is a risk of contamination by (pathogenic) microorganisms from the outside environment into the drinking water during maintenance or pipe breaches in the drinking water distribution system (DWDS) and, consequently, the drinking water distributed to consumers may result in possible detrimental effects on public health. Traditional time-consuming microbiological testing is, therefore, performed to confirm drinking water is not microbially contaminated. This is done by culturing methods of the faecal indicators Escherichia coli, intestinal enterococci and the technical parameters coliform bacteria and heterotrophic plate counts at 22 °C (HPC22). In this study, fast methods (adenosine triphosphate (ATP), flow cytometry, enzyme activity and qPCR) were compared as an alternative for HPC22. Using dilution series and field samples, ATP (ATPtotal-lab and ATPcell-mob) and enzymatic activity (ALP-2) methods proved to be the more reliable and sensitive than flow cytometry and qPCR methods for detecting microbiological contaminations in drinking water. Significant (p < 0.05) and relatively strong correlations (R2 = 0.61-0.76) were obtained between HPC22 and both ATP methods, enzyme activity and qPCR parameters, but relations with flow cytometry were weak (R2 = 0.24 - 0.52). The samples taken after repairs or a calamity from the DWDS showed in general limited variation in the HPC22 count and were in most cases below the guidance level of 1,000 CFU/mL. We recommend that the best performing alternative methods, i.e. ATPtotal-lab and ATPcell-mob and ALP-2, should be included next to HPC22 in additional field studies to further test and compare these methods to be able to decide which fast method can replace HPC22 analysis after maintenance work in the DWDS.

In this paper, we propose an Internet of Things Enabled Water Distribution System for the Textile Dyeing Industry that uses ultrasonic sensors to autonomously monitor and adjust the water level in the tanks. The water distribution system’s manual functioning is often being replaced in industries. With the help of this method, the textile dyeing industry hopes to increase overall sustainability and efficiency while reducing human labor. Water usage in the textile sector is higher, particularly in the dye finishing process. The water supply is properly distributed by means of continuous monitoring, which allows us to keep track of the amount of water in tanks, the flow rate, and any anomalies in the distribution line. The suggested approach is implemented by means of real-time systems.

Digital Twins have emerged as a disruptive technology with great potential; they can enhance WDS by offering real-time monitoring, predictive maintenance, and optimization capabilities. This paper describes the development of a state-of-the-art DT platform for WDS, introducing advanced technologies such as the Internet of Things, Artificial Intelligence, and Machine Learning models. This paper provides insight into the architecture of the proposed platform-CAUCCES-that, informed by both historical and meteorological data, effectively deploys AI/ML models like LSTM networks, Prophet, LightGBM, and XGBoost in trying to predict water consumption patterns. Furthermore, we delve into how optimization in the maintenance of WDS can be achieved by formulating a Constraint Programming problem for scheduling, hence minimizing the operational cost efficiently with reduced environmental impacts. It also focuses on cybersecurity and protection to ensure the integrity and reliability of the DT platform. In this view, the system will contribute to improvements in decision-making capabilities, operational efficiency, and system reliability, with reassurance being drawn from the important role it can play toward sustainable management of water resources.

Water Distribution Systems (WDSs) are vital infrastructures responsible for delivering clean and safe water to various sectors. However, advancements in WDSs infrastructure, such as automated monitoring and control systems, have made these systems vulnerable to cyberattacks. These attacks can manipulate the system, causing severe disruptions like contaminating water supplies or disrupting water flow. To protect WDSs from such threats, a strong and reliable detection system is essential. This study proposes a novel approach that combines classification and regression models for anomaly detection using the BATADAL dataset. Initially, a classification model predicts whether an attack is occurring. Subsequently, a regression model evaluates water levels to determine if they are within normal ranges. If an anomaly is detected by either approach, the framework raises an attack alert. In this framework, the Extra Trees Classifier demonstrated significant performance, achieving an accuracy of 0.89 in the classification task and an average R2 score of 0.6658 in the regression task. Overall, the combined proposed framework delivered impressive results, achieving significant accuracy in detecting anomalies.

No abstract available

No abstract available

No abstract available

No abstract available

For water to be delivered to people cleanly and healthily, the tanks in which it is stored before being made available must also be reliable and clean. To prevent the risk of transmitting infectious diseases through water, it is vital to apply purification and disinfection processes to the water held in water storage tanks. For this reason, monitoring the sediments in the storage tank and their properties, such as pH, pressure, and temperature, are necessary in real-time. With SCADA systems, water flow is monitored remotely by providing central control and monitoring in drinking water distribution. Thanks to the collected data, adverse situations in the storage tanks can be automatically detected, and water distribution can be managed by intervening in the system when necessary. Additionally, water leaks can be detected, and all the water supplied to the network can be delivered to the user. In this study, the data from 4 pumping centers and 13 water storage tanks in Yozgat province were examined, and all processes from the water source to the water storage tanks and the water supply to people's use were controlled remotely with the SCADA system. While the total physical and administrative lost water rate was 64.35% and physical water loss was 27.59% before the use of SCADA, it was observed that the water losses decreased by 51% with the use of SCADA. Thanks to the system, detected sediment formations are prevented quickly, and water is delivered to the user cleanly and healthily.

This paper presents the design and application of a SCADA-IoT (Supervisory Control and Data Acquisition-Internet of Things) platform for monitoring an hydraulic system which simulates the operation of the raw water distribution network in the municipality of Tecate, Baja California (B.C.), Mexico. Based on the design and construction of an academic prototype, which represents the called Las Auras-Nopalera-Cuchumá hydraulic system, we analize the behaviour of physical variables and the integration of hardware and software components with Industry 4.0 in order to develop on-field applications using the LOGO! Web Editor software from Siemens, and cloud applications using the open source Node-RED IoT platform. Experimental results illustrate the efectiveness of the proposed prototype which retrieves level, pressure, caudal and pH real-time values as well as some signals from actuators.

Leakage in a water distribution network makes up a significant amount of loss. In certain countries, this loss represents 40% to 50% of the supplied water, whereas the global average for most systems is estimated at around 30%. Furthermore, water demand is increasing as a result of population growth, while resources are dwindling. This study used hydraulic modelling for leakage reduction through pressure control. For this purpose, a hydraulic model was created using the software EPANET for a selected water distribution network in a district-metered zone in Jeddah. The model was calibrated and validated for the district-metered zone using data obtained by SCADA. Pressure management using a pressure-reduction valve was then implemented to control the amount of leakage in water distribution systems. The preliminary findings show that pressure optimization was required as there were nodes within the network that had excessive pressure. Application of pressure management to the district metered zone showed that the reduction in water pressure from 5 bar to 2 bar resulted in an immediate reduction in water losses. The leakage volume through the district-metered area at the time of maximum pressure dropped by 10% after pressure reduction. Simultaneously, the minimum required pressure was upheld at each demand node, preventing any lack of pressure in the water distribution system. The results indicated that pressure control should be integrated with hydraulic modelling for effective leakage reduction. This research could help water-supply companies as a support tool for planning and managing leakage in water distribution systems.

Water losses from distribution systems are present in all situations. Reducing them to zero is impossible. Modern technologies for operating water distribution networks are an effective means of reducing water losses. In this sense, the conducted research aimed at testing some solutions to improve the monitoring and control of pressure in a distribution network. The mathematical model of the water distribution network was developed and calibrated (Mike Urban application) and the hydraulic modeling of the network was carried out. The results were obtained through: periodic simulations with the help of the monitoring application; comparisons for flow and pressure values, data collected from field measurements, data obtained from SCADA application. These complex analyzes allow the identification, in the shortest possible time, of existing problems and areas with losses.

Supervisory Control and Data Acquisition (SCADA) systems and other Industrial Internet of Things (IIoT) systems are essential components that underlay the efficient operation of contemporary industries, such as energy, transportation, water supply, and manufacturing. But as networked systems become more interconnected, the attack surface grows, making these systems vulnerable to advanced cyber threats, with the potential of wide-spread impact upon operations or even physical destruction. Conventional, signature-based detection systems regularly miss emerging or stealthy attacks, especially in real-time OT environments. This method is capable of detection both known and zero-day attacks and does not rely on pre-defined threat signatures. A multi-tier detection structure is presented, in which edge-level anomaly detection components run at the vicinity of the sensors and actuators, and higher-level analytics at the cloud or on-site data centers perform the more advanced correlation analysis. To mitigate the constraints of operation in SCADA environments, i.e., low latency and bandwidth, the framework includes light weight AI inference models to execute on embedded IIoT devices. The data pre-processing modules clean noise and normalize sensors inputs, which result in mitigation of false positives and high detection rate. The AI models are designed using a mixture of real industrial data sets and synthetic adversarial cases emulated through a cyber-physical testbed, making it robust to different intrusion types. For more security and resilience, secure communication protocols and blockchain integrity verification for anomaly alerts are supported. Every anomaly that is detected is written in an immutable log facilitating forensics analysis and compliance reports. Furthermore, the (SIEM) integration enables detected threats to initiate automated incident response workflows, so that downtime and operational risk is reduced. Experimental results show that the proposed AI-based anomaly detection approach reduces detection latency, increases the detection rate of novel attacks and scales while deployed in distributed industrial networks. The results emphasize that AI-based AD in the context of SCADA and IIoT landscapes contributes in increasing the resilience of the critical infrastructure against cyber intrusions, introducing a proactive defense mechanism which acts as a living organism adapting to the evolution of the threat.

Access to safe and reliable water is a fundamental requirement for sustainable urban development. However, water distribution systems (WDSs), particularly in small and aging municipalities, face persistent challenges, including infrastructure degradation, population growth, and limited operational data. This study presents a comprehensive hydraulic modeling framework for the city of Farsala, Central Greece, an area characterized by significant data scarcity and outdated water network records. A novel methodology was developed that combines AutoCAD-based network digitization, GIS data integration, field surveys, and real-time SCADA telemetry to create a high-fidelity and operational hydraulic model using WaterGEMS software. The model was calibrated and validated using observed pressure and flow data, achieving high performance metrics, including a Nash–Sutcliffe efficiency (NSE) of 0.841 and R2 of 0.90. Extended period simulations (EPS) were conducted to evaluate system behavior over a 24 h cycle, revealing critical insights into pressure distribution, peak demand conditions, and leakage hotspots. The results demonstrate that even under constrained data conditions, it is possible to construct a robust and decision-supportive model, offering valuable guidance for future system upgrades, leakage control strategies, and smart infrastructure planning in similar urban environments.

Urban water security hinges on the reliable provision of safe, affordable water and the protection of people and infrastructure from contamination, losses, scarcity, and flood-related disruptions. Rapid urbanization, climate variability, aging assets, and growing cyber–physical risks expose gaps in traditional supervisory control and data acquisition (SCADA) regimes. Smart monitoring systems—continuous sensing of hydraulic and water quality parameters, edge-to-cloud analytics, and decision automation—offer near real time situational awareness to reduce non revenue water, mitigate contamination events, and optimize operations. This paper develops an integration blueprint that combines district metered areas (DMAs), multi parameter sensors (pressure, flow, acoustic leak, residual chlorine, turbidity, conductivity), remote sensing, and city platforms via interoperable standards (e.g., OGC Sensor Things) and low power wide area networks (LPWANs). We present a methodology for pilot design, data quality controls, anomaly detection, and cyber security hardening (IEC 62443/NIST 800 82), followed by a unit economics and impact framework (KPIs: NRW, response time, water quality compliance, energy per m³). A synthesis of documented deployments suggests smart monitoring can accelerate detection, reduce losses, and improve resilience when paired with governance, workforce, and data rights measures. The paper concludes with a roadmap for city utilities to scale from pilots to platform level capabilities while balancing openness, security, and affordability.

Leaks represent a major issue impacting the management and efficiency of  Drinking Water Networks (DWN) in cities worldwide. According to the Development Bank of Latin America, by 2018 the losses in DWN range from 40 to 60% in the region. In Europe, the OECD reports a wider range with few losses in cities like Amsterdam (4%) up to 37% in Naples. With this context, some regional policies have emerged like the 2020 european drinking directive “Right2Water”, that aims to encourage major suppliers (more than 50000 users), to develop tools to measure and reduce leakages by 2025. Considering this situation, we introduce here a systematic approach for leak management that combines field data, hydraulic models (HM) and machine learning.Model-based and data driven methods have been of great interest for leak location methodologies in DWN. This research will design energy-efficient and cost-efficient leak localization hotspots in the DWN. The approach is intended for sectorized DWN, equipped with a SCADA system and where a calibrated hydraulic model (e.g. EPANET) is available. This latter serves to evaluate the sensitivity of the system to leaks and identify potential points for pressure measurements in order to optimise the number of installed sensors. Given a detected leak in the network, a multiclass classifier using pressure data is developed to reduce the inspected pipe length for the leak location. The leakage localization method is implemented combining multiple individual classifiers using ensemble learning methods and a reduced number of decision variables. The methodology is tested on a real case study from a Colombian site. The method faces challenges in (a) collecting correctly labelled real leak data, and (b) modelling and calibrating hydraulic models. Those challenges are being addressed. The outcome shows that the length of pipes inspected can be reduced by one third with high performance in accuracy with few sensors required (low capital expenditures) and low computational effort (low energy and low operational expenditures). 

Instrumentation is critical for maintaining the accuracy and stability of industrial processes, particularly in measuring and controlling key variables such as flow, pressure, and temperature. However, over time, flow instruments are susceptible to issues like sensor drift, clogging, and degradation, leading to inaccurate readings and operational inefficiencies. Traditional maintenance methods, relying on periodic calibrations or reactive repairs, often fail to prevent unexpected equipment failures, resulting in costly downtime and process disruptions. This paper presents the development of a neural network machine learning-based predictive maintenance system designed to address these challenges by shifting from reactive to proactive maintenance. Historical data, including transmitter readings, system outputs, and maintenance logs, are utilized to train a machine learning model capable of identifying early warning signs of equipment malfunction, such as gradual sensor drift and anomalous flow patterns. The system analyzes data trends and delivers real-time alerts to maintenance teams, enabling timely interventions. Simulation results demonstrate the model's effectiveness, achieving a high accuracy of 91.4% and a low loss of 0.17, ensuring improved process reliability and reduced maintenance costs. This study highlights the potential of the neural network model in enhancing instrumentation reliability and operational efficiency in industrial environments.

Water leakage detection and localization in water distribution networks (WDNs) are critical for sustainable water management, economic efficiency, and infrastructure integrity. Traditional inspection and hydraulic simulation methods often fail to provide timely and accurate results, prompting growing interest in hybrid approaches that combine model-based hydraulic simulations with data-driven machine learning techniques. This study proposes a novel hybrid method that leverages simulated SCADA data to train machine learning models for leakage detection and localization. Graph Transformers are trained to predict the pressure at each junction of the WDN based on observed pressure readings and auxiliary hydraulic features such as pump flow rates. The residual signal between predicted leakage-free and leaky pressures, along with its slope, is then used for detecting and localizing abrupt and incipient leakages, respectively. A residual signal compensation mechanism and a slope-based leakage detection and localization algorithm are introduced to address complex scenarios involving concurrent abrupt and incipient leakages - conditions frequently observed in real systems but rarely considered in existing learning-based hybrid methods. The proposed method was validated using benchmark datasets from the BattLeDIM2020 challenge and compared against state-of-the-art algorithms. Results show superior economic scores (€422,660±11,553), higher true positive rates, and improved robustness under multiple-leak conditions. These findings demonstrate the potential of advanced hybrid frameworks to significantly enhance the accuracy and applicability of leakage detection and localization in complex WDNs.

SCADA systems are essential for water management in smart cities. They provide real-time monitoring, control, early detection of issues, efficiency improvements, enhanced resilience, and data-driven decision support. By leveraging SCADA technology, water utilities can optimize operations, reduce water losses, ensure a reliable supply, and contribute to sustainable water management in smart cities. Non Revenue Water (NRW) refers to water that is produced and lost before it reaches consumers or generates revenue for water service companies (WSCs). The summary highlights the importance of reducing non-potable water as it leads to increased water availability and revenues for WSCs. This is done by installing flow meters and water pressure measuring devices on the water sources or the entrances and exits of the isolated area, then calculating the amount of water feeding the area and the amount of water accounted for, and from it the amount of non-accountable water can be calculated, and work to reduce it by detecting leakage - Detecting stealth connections and installing water meters that work efficiently. The reduction of commercial losses, which includes issues like inaccurate billing or theft, can significantly improve revenue generation. Similarly, reducing physical/real losses, such as leaks or inefficient distribution systems, allows WSCs to postpone capital investments needed for water source development. To address this issue, departments such as Network Operations, Geographic Information Systems (GIS), Hydraulic Analysis, and the commercial sector must collaborate to estimate and identify the different components contributing to NRW and water loss reduction.

Manual pipeline inspections delay leak detection, causing water loss and costly repairs. This paper proposes a low-cost, energy-efficient wireless sensor system using ESP32 that monitors pressure, flow, and vibration in real time. With ESP-NOW communication, it ensures reliable data transfer, offers live dashboards, supports mesh-based expansion, and sends instant leak alerts. Experimental evaluation showed a leak detection accuracy of 96.4%, an average detection latency of 1.8 seconds, and 22% lower power consumption compared to Wi-Fi–based monitoring systems. This solution improves detection accuracy, reduces maintenance, conserves water, and is ideal for small businesses, schools, and local governments, especially in resource- limited areas.

The scarcity of water is one of the most significant problem in the worldwide is facing today in view with the growing population and climate change. A notable amount of water is being wasted only through leaks in residential areas. Leaking faucets, deteriorated pipes and excessive water pressure usually cause water leaks. The proposed system implementation of a system that detects the position of water leaks and alerts the user through Short Message Service (SMS). The system built using a water flow sensor, a microcontroller unit to interpret and evaluate the data inputs fed by the water flow sensor. This system strives to develop a real-time prototype for water leakage detection and alerting by detecting a water leak.

Leakage detection in the water distribution system not only helps to reduce water waste but also decreases the risk of drinking water pollution. To reduce reliance on hardware devices and enable real-time detection, the water utilities are transitioning towards the data-driven based approach that relies on the analysis of the flow and pressure data collected from the supervisory control and data acquisition (SCADA) system. Due to the lack of leakage data, most of these methods are unsupervised methods that rely heavily on assumptions about the distribution of anomalies; whereas, the water utility's repair records contain much valid information about the leakage and normal characteristics. To convert this information into available labels and to address the lack of leakage data, this paper proposed a new leakage detection framework to infer the pressure characteristics under normal conditions based on historical data by combining a label cleaning method-confident learning (CL)-with an unsupervised method-Gaussian mixture model (GMM)-for leak detection. The methodology is validated with synthetic and real measured data. Comparisons with four unsupervised methods demonstrate that the GMM method has superior identification of leak features from pressure data. For a real-world water distribution system in K city containing 91 pressure sensors, the average true positive rate is 0.78 and the average false positive rate is 0.11. This methodology is a promising tool to identify signs of leakage in large-scale water distribution networks (WDNs).

No abstract available

The need for effective water management systems has grown as a result of the population expansion and increasing urbanization of metropolitan areas. Water leaks in metro water pipelines, however, present a serious problem since they cause financial losses, structural damage, and water loss. As a result, creating a Smart Water Leakage Controller (SWLC) becomes essential for immediately identifying and addressing leaks. The main elements and features of a SWLC system created especially for metro water pipelines are described in this abstract. In order to continually monitor the pipeline network, the SWLC system integrates cutting-edge sensor technology, including acoustic, pressure, and flow sensors. These sensors pick up on irregularities that could be signs of leaks, like sharp reductions in pressure or very high flow rates. The sensors send real-time data to a central control unit that has sophisticated algorithms for data processing and decision-making. The SWLC system takes rapid action to minimize water loss and stop more damage after detecting a leakage occurrence. To ensure system stability, these steps can entail alerting maintenance staff, isolating the impacted pipeline portion, and modifying water flow.

An IoT-enabled water leak detection system is proposed. The proposed system uses real-time data collection and advanced machine learning techniques for continuous monitoring. It consists of two leak sensors integrated with Arduino nodes. It also uses NodeMCU to enable IoT communication and alerts. The sensors continuously monitor water flow and pressure to detect any anomalies that may indicate a leak. The system collects real-time datasets, which are then processed using a modified Long Short-Term Memory (LSTM) model for accurate leak prediction. The modified LSTM uses a new attention mechanism to increase the model's ability to focus on significant data patterns. This integration is used to increase the prediction accuracy. The loT-enabled system sends immediate alerts to users upon detecting a leak. This ensures swift response and minimizing water loss. This approach increases leak detection efficiency, resource management, and reduces manual intervention. Experimental results show that the proposed system achieves higher accuracy and best suitable for both residential and industrial water distribution networks.

Detecting leaks in water networks is a costly challenge. This article introduces a practical solution: the integration of optical network with water networks for efficient leak detection. Our approach uses a fiber-optic cable to measure vibrations, enabling accurate leak identification and localization by an intelligent algorithm. We also propose a method to access leak severity for prioritized repairs. Our solution detects even small leaks with flow rates as low as 0.027 L/s. It offers a cost-effective way to improve leak detection, enhance water management, and increase operational efficiency.

Pipeline leakage, which leads to water wastage, financial losses, and contamination, is a significant challenge in urban water supply networks. Leak detection and prediction is urgent to secure the safety of the water supply system. Relaying on deep learning artificial neural networks and a specific optimization algorithm, an intelligential detection approach in identifying the pipeline leaks is proposed. A hydraulic model is initially constructed on the simplified Net2 benchmark pipe network. The District Metering Area (DMA) algorithm and the Cuckoo Search (CS) algorithm are integrated as the DMA-CS algorithm, which is employed for the hydraulic model optimization. Attributing to the suspected leak area identification and the exact leak location, the DMA-CS algorithm possess higher accuracy for pipeline leakage (97.43%) than that of the DMA algorithm (92.67%). The identification pattern of leakage nodes is correlated to the maximum number of leakage points set with the participation of the DMA-CS algorithm, which provide a more accurate pathway for identifying and predicting the specific pipeline leaks.

Pipe leakage is an inevitable phenomenon in water distribution networks (WDNs), leading to energy waste and economic damage. Leakage events can be reflected quickly by pressure values, and the deployment of pressure sensors is significant for minimizing the leakage ratio of WDNs. Concerning the restriction of realistic factors, including project budgets, available sensor installation locations, and sensor fault uncertainties, a practical methodology is proposed in this paper to optimize pressure sensor deployment for leak identification in terms of these realistic issues. Two indexes are utilized to evaluate the leak identification ability, that is, detection coverage rate (DCR) and total detection sensitivity (TDS), and the principle is to determine priority to ensure an optimal DCR and retain the largest TDS with an identical DCR. Leakage events are generated by a model simulation and the essential sensors for maintaining the DCR are obtained by subtraction. In the event of a surplus budget, and if we suppose the partial sensors have failed, then we can determine the supplementary sensors that can best complement the lost leak identification ability. Moreover, a typical WDN Net3 is employed to show the specific process, and the result shows that the methodology is largely appropriate for real projects.

Considering the problem of difficulty in transmission and storage due to a large amount of data in the water-supply network monitoring system based on a wireless sensor network (WSN), we propose a sparse representation of the water-supply network monitoring data by using compressed sensing (CS) method. At the same time, aiming at the problem of low leakage identification accuracy caused by information loss under compressed sensing, we propose a leak identification method for a water-supply pipe network based on compressed sensing and deep residual neural network (ResNet). Firstly, under the condition that the observation matrix ensures the integrity of signal information, the compressed sensing theory is used to compress and observe leakage signals to obtain observation data, to reduce the redundant information and volume of the data. At the same time, the observation data is preprocessed to realize the transformation of a one-dimensional signal to a two-dimensional matrix. Then the residual neural network is trained by using the two-dimensional data to realize the automatic, efficient, and accurate leak identification under different leakage apertures. Experimental results show that the proposed method can obtain relatively high accuracy and greatly reduce the training time of ResNet by using compressed data. When the Compression rate (CR) is 70% and the observation matrix is a Gaussian random matrix, the average accuracy is 96.67% and the training time is only 50% compared to uncompressed data. The research work provides a new intelligent leak identification under different leak apertures using WSN and has important application prospects in saving water resources.

The Water distribution networks (WDN) experience substantial water loss due to leakage in pipelines caused by pipe aging or pipe bursts. The Machine Learning (ML) and Deep Learning (DL) based acoustic method which depends on signal information and limited diversity of available data. However, the existing methods fail to detect leak points when sudden change in the input data. To overcome this limitation, a Transductive- Long Short-Term Memory (LSTM) is proposed for leakage detection and identification in a water supply network. The dataset utilized for detection was collected from the residential area are fed as input to the proposed model. After that, the proposed model considers neighbor data points and a new sequence of data was formed for identifying the unknown leakage points in the pipeline. For tuning the parameters Adam optimizer is utilized. The experimental results of the proposed T-LSTM are evaluated by the performance metrics that attained $\mathbf{9 8. 3 \%}$ accuracy, $98.2 \%$ precision, $98.2 \%$ recall, and $\mathbf{9 8. 1 \%}$ F1-score which is greater than the existing methods like Resnet, Change Detection Test (CDT), Artificial Intelligence-Deep Neural Network (AI-DNN), Ensemble ML models, Convolutional Neural Network (CNN)SqueezeNet.

Water is one among the important source around the world the available of source compared to our requirement is very minimum. The major water loss occurred in water distribution is the pipeline leakage during the transportation. The identification and location of leaks is an issued faced by the people. This issue should be sort ought to improve the distribution system by managing the minimal loss. The pipe leaks may be occurred plenty of reasons like aging effect of pipelines, and long span of corrosions such reasons are caused for water leaks it leads the water losses. Hence finding the leakages and troubleshooting the faulty pipeline is very critical during water Transportation. So that we need a solution to obtain the exact location of leak when damage occurs in the pipeline. The leak detection can be obtained using fluid dynamics, kinematics based on the fluid rate with the help of flow meter. Here Tiva C series used as controller, Prototype has been developed to find leakage in underground pipelines has been implemented and Final results of proposing methodologies is to find out the location of pipe leaks with the maximum distance of 20 meters. This leads to improve the leak detection as time limit.

A novel neural network model able to identify the nature of leaks in a live urban underground water pipe grid is proposed in this article. Traditional leak detection methods are often limited to detecting the existence or otherwise of a possible leak in the signals, yet do not provide information about the actual nature of the leak. This is of utmost importance for utilities to be able to rank and schedule their repair and rehabilitation efforts. The proposed model, termed multiclass time–frequency convolutional neural network (M-TFCNN), uses a multimodel categorization approach, enabling the identification of various types of leak-related events, namely, pipe breaks, hydrant leaks, service valve leaks, private leaks, and others. The proposed neural network architecture consists of a feature extraction module, a convolutional base, and a classification head, which has been meticulously engineered through numerous experiments and iterations. The model was trained on a large-scale dataset of vibroacoustic signals collected from an operating water main network in a major Australian city. The data encompass a substantive signal collection of validated representations of different types of leaks gathered over a span of three and a half years. The model’s performance is evaluated using traditional metrics such as accuracy, precision, recall, and ${F}1$ -score. The presented results show that the proposed model surpasses conventional leak detection methods, accurately identifying different types of water leaks and achieving accuracies of up to 98%. Overall, the proposed neural network model represents a notable practical step forward in the field of water leak detection by subcategorizing leaks and has the potential to revolutionize the way industry practitioners manage larger water infrastructure.

Smart meters such as advanced metering infrastructure (AMI) can significantly improve identifying realistic sized leaks in water distribution networks (WDNs). However, to date, detection/localization methods for AMI systems are extremely limited. In this study, to examine the benefits of using AMIs for leak detection within distribution network, a three-dimensional (3D) convolutional neural network (CNN) deep learning (DL) model is proposed that can account for temporally and spatially distributed information of pressures. The 3D CNN is tested for a real WDN in Austin using the realistic sized leaks (e.g., 3 L/s for 150-mm pipes) that are generated from hydraulic simulations. The model's performance is evaluated using detection probability, false alarm rate, and localization pipe distance metrics. In addition, the strength of using DL for leak identification is examined by comparing the CNN results with those from an optimization-based model. The 3D CNN performed better than the optimization model indicating that DL has advantages over conventional tools such as optimization methods. However, its adaptability may limit its use in some cases. Since DL can be significantly impacted by hydraulic simulation model, a way to handle modelling error must be determined. In addition, as network changes occur, retraining is required that may be time consuming and have difficulty with the number of failure conditions.

Leak detection is important to enable automatic and early identification of leakages in water distribution systems, which may prevent water wastage, reduce the environmental impact of leakages, and also avoid structural damages to pipe networks. While there are different sensors, hardware and software-based methods for leak detection, this field still faces issues, especially regarding data scarcity and imbalance. In this paper, a comparison of different Machine Learning (ML) models for leak detection and leak type classification has been carried out, addressing data augmentation and balancing issues, and providing a comparison on the performance of pressure and accelerometric data for leak monitoring with ML. Compared to the state-of-the-art, mainly focusing on leak detection, the focus has also been on leak type identification, which may be useful for diagnosing pipe damage. Results indicate that accelerometer data is more consistent across tasks and performs better than pressure data for multi-class leak classification. The best-performing model is an Ensemble Bagged Trees classifier using accelerometer data and features extracted from 5 s long windows, and it achieved 91.7% accuracy in multi-class leak classification. While the proposed approach still faces issues in identifying longitudinal cracks, it provides a solid base for leak classification with ML and it identifies the accelerometer as a viable, non-invasive solution for leak monitoring.

Nowadays, reducing the loss of water resources is of primary concern in many countries worldwide. Different methodologies exist for the identification and localisation of leaks in water distribution networks, which can be applied simultaneously at a different scale, to improve success rate. Acoustic correlation has been employed for many years to localise such leakages, based on the post-processing of vibroacoustic signals acquired by two sensors placed on both sides of the pipe where the leakage is supposed to be located. The practical application of this methodology still presents some significant issues. The implementation of the noise correlation methodology was investigated through an experimental test rig by simulating leakages in different conditions of pressure and flow and investigating the effects of background noise. Hydrophonic measurements and recordings were performed, identifying the frequency range affected by the leak noise, and the signal to assess the accuracy of the noise correlation method. A pre-filtering of the signals to reduce the effect of background noise masking was also investigated. The accuracy of acoustic correlation in localising the leakage was assessed by computing the propagation velocity, which depends on the vibroacoustic fluid-structure interaction, through approximated analytical equations

A great deal of water could be saved by implementing automatic leak detection all over the world. In developed nations, pipe induced vibrations or sounds are used in acoustic devices for detecting leaks while in developing countries it’s majorly through visible physicality on floors usually resulting to wastage of a lot of water as they are not notified until something goes wrong there. Pipe leaks due to various reasons such as worn-out, natural disasters and improper installation can be rectified with the solution that detects and locates damages within the earliest times possible. A affordable way out would be having an embedded system with flow sensors at several points along the pipeline being used as means of identifying leaking waters at minimum cost. Detected faults could be sent wirelessly to appropriate departments via cloud-based information storage. This concept also involves incorporating AR technology into our designed system since underground sewage systems are not easy to locate in most cases. The AR interface displays the underground sewage pipeline blueprint, allowing users to visualize pipe locations without technical expertise. Moreover, the system identifies leaks and potential pipeline issues within the AR environment, providing real-time insights for prompt resolution. Efficient water resource management is a key global goal due to increasing water demand. Automation in such systems reduces errors, enhances efficiency, and bridges the supply-demand gap.

Underground water pipe leaks cause significant water and economic losses and are hard to detect due to their buried nature. Long-wavelength PolSAR signals offer new opportunities for leak detection with their high penetration and groundwater fraction recording capabilities. Using SAOCOM data, this study extracts urban surface scattered energy through Singh's seven-component polarization decomposition and trains Random Forest, Multilayer Perceptron, and XGBoost models for leak prediction. The integrated model achieves $81.20 \%$ accuracy through a voting mechanism. DBSCAN density clustering optimizes the results, reducing potential leaks to 1,265, with all real leaks within a 150 -meter buffer zone. This study highlights the potential of combining PolSAR data with machine learning for urban water main leak detection.

The detection of leakages in Water Distribution Networks (WDNs) is usually challenging and identifying their locations may take a long time. Current water leak detection methods such as model-based and measurement-based approaches face significant limitations that impact response times, resource requirements, accuracy, and location identification. This paper presents a method for determining locations in the WDNs that are vulnerable to leakage by combining six leakage-conditioning factors using logistic regression and vulnerability analysis. The proposed model considered three fixed physical factors (pipe length per junction, number of fittings per length, and pipe friction factor) and three varying operational aspects (drop in pressure, decrease in flow, and variations in chlorine levels). The model performance was validated using 13 district metered areas (DMAs) of the Sharjah Electricity and Water Authority (SEWA) WDN using ArcGIS. Each of the six conditioning factors was assigned a weight that reflects its contribution to leakage in the WDNs based on the Analytic Hierarchy Process (AHP) method. The highest weight was set to 0.25 for both pressure and flow, while 0.2 and 0.14 were set for the chlorine and number of fittings per length, respectively. The minimum weight was set to 0.08 for both length per junction and friction factor. When the model runs, it produces vulnerability to leakage maps, which indicate the DMAs’ vulnerability classes ranging from very high to very low. Real-world data and different scenarios were used to validate the method, and the areas vulnerable to leakage were successfully identified based on fixed physical and varying operational factors. This vulnerability map will provide a comprehensive understanding of the risks facing a system and help stakeholders develop and implement strategies to mitigate the leakage. Therefore, water utility companies can employ this method for corrective maintenance activities and daily operations. The proposed approach can offer a valuable tool for reducing water production costs and increasing the efficiency of WDN.

The fire protection pipeline network is a critical infrastructure for ensuring fire safety, and its reliability directly influences the effectiveness of fire suppression. These networks are susceptible to corrosion, aging, or cracking-induced leakage. Traditional single-point leakage models struggle to accurately capture the complex characteristics of multi-point leak scenarios, which often result in greater water loss and more significant pressure drops. Therefore, it is essential to develop specialized methods for the effective identification and accurate localization of multi-point leaks, which are commonly represented by dual-point leaks. This study proposes a dual-point leak localization method based on a Bayesian Optimization and Light Gradient Boosting Machine (BO-LightGBM) model for fire protection pipe networks. This approach integrates experimental data with numerical simulations to create a dataset of dual-leak scenarios under varying operating conditions. The Bayesian optimization algorithm is used to automatically determine the optimal hyperparameters for the LightGBM model, which is then trained as the final BO-LightGBM localization model. Testing on simulated datasets shows that the proposed model achieves 96.11% accuracy. The BO-LightGBM model demonstrates superior performance compared to other localization models developed using mainstream machine learning algorithms. In conclusion, the BO-LightGBM method presented in this study effectively detects and localizes dual-point leaks in fire protection pipeline networks. It significantly reduces manual monitoring requirements and addresses limitations such as poor real-time performance and data imbalance in existing methods. This technology provides strong support for advancing intelligent fire protection systems.

Water leak detection is a critical component in the protection and efficient management of distribution networks. Undetected leaks can result in significant structural damage, costly repairs, business interruptions, and loss of valuable assets. The system must achieve high sensitivity to detect faint leak sounds in diverse pipe materials (metal $/$ non-metallic), incorporate noise suppression to isolate leaks from ambient interference (e.g., traffic or machinery), enable automation for real-time detection and classification without expert intervention, and ensure scalability to adapt to urban and rural water distribution networks. The system faces background noise contamination from fluid dynamics, environmental sources, and pipe vibrations; limited accuracy in non-metallic pipes due to rapid acoustic signal attenuation; manual dependency in traditional methods, resulting in delays and higher costs; and false alarms caused by overlapping frequencies between leaks and ambient noise. Early leak identification prevents catastrophic pipe bursts, reduces repair costs by up to 70 %, and supports sustainability goals by conserving water resources. This paper addresses critical challenges in water leak detection by integrating ground microphone technology with advanced signal processing and machine learning, offering a novel solution for automated leak classification, and using an ultra-low noise amplifier for noise reduction. It consists of four stages such as signal acquisition, signal division, signal amplification, and automation. By employing an ultra-low noise amplifier and filtering, the system achieves $\mathbf{4 0 - 6 0 ~ d B}$ noise suppression, effectively isolating leaks from ambient interference such as traffic or pipe vibrations. This enhancement enables reliable detection even in non-metallic pipes, where traditional methods struggle due to rapid acoustic signal attenuation. For feature extraction, Support Vector Machines (SVM) are utilized to identify discriminative spectral and temporal patterns, enhancing the separability of leak signatures from noise. It achieve $\mathbf{9 2 \%}$ precision in leak classification, outperforming thresholdbased systems by accurately distinguishing between small and large leaks.

Water pipe leakage poses a significant challenge for water distribution authorities worldwide, as pinpointing the exact source of the problem becomes a daunting task. This issue can arise due to various factors, including pipeline deterioration due to age or ongoing urban construction projects like those in Dar es Salaam. Consequently, these distribution authorities encounter immense difficulties identifying the cause of the leak and implementing appropriate measures. Hence, the objective of this endeavor was to produce an Internet of Things based framework to find water leaks, monitor water quality, and track water levels. The prototype consisted of two sensors embedded at the source and destination points, enabling the measurement of water flow rate. The findings revealed that comparing the volume of water generated at the starting point with that at the endpoint facilitated the identification of leaks. Considering a more comprehensive approach to distance calculation could yield intriguing insights, promoting further research into IoT monitoring systems.

: This paper promotes water distribution networks’ (WDNs) sustainability and efficiency by integrating intelligent data analysis with ground-penetrating radar (GPR) to better interpret GPR images for detecting water leaks, favouring their asset assessment. This work uses GPR data from a laboratory setting to investigates the effects of various parameters on image interpretability across pipes. This methodology aims to advance the automation of leak and pipe identification, improving data interpretation and reducing dependency on human experts for leakage detection purposes. The findings suggest the possibility of uncovering new features enhancing GPR image interpretability, presented in 3D models.

Numerous cities developed in the 20th century have a Water Distribution Network(WDN) with decade-old pipes. Aging pipelines can leak with lower the water revenue ratio, thus resulting in waste and economic losses. In this study, new methods for localizing leakage in WDNs are proposed and tested. Two meta-heuristic methods based on Harmony Search(HS) and Genetic Algorithm(GA) are developed to detect leakage of WDNs through an evaluation of emitter coefficients. The old town of G-Town in South Korea has a 50-year-old WDN and a low water revenue ratio. High quality field measurements in G-Town allowed detailed testing of optimization methods based on emitter coefficients for leakage detection. As a result, both GA and HS yielded comparable outputs. The HS method performed with slightly better accuracy in minimizing the objective function than GA after about 1,000 iterations. Leakage detection becomes fairly accurate after approximately 30 minutes of optimization and pipe networks with more than 100 pipes and 50 nodes require more iterations and calculation time.

This study presents an advanced approach to Ground Penetrating Radar (GPR) for characterising buried pipelines and detecting water leakages. GPR is a non-destructive technique used to detect subsurface dielectric objects such as pipelines. The research combines theoretical modelling, experimental validation, and numerical simulation across various pipe materials (metal, fiberglass, PVC) and soil types (clay, sand, and silt). Hyperbolic signatures and power reflectivity in radargrams were analysed to estimate pipe radius, with an accuracy of 75.7%. Methods for feature recognition, anomaly detection, and signal enhancement were applied to improve the identification of buried utilities and leak zones. The study incorporates four comparative approaches: attenuation characteristic analysis, time-frequency spectral analysis, and Finite-Difference Time-Domain (FDTD) simulation. Corroded pipes are effectively detected through attenuation analysis, while cracked pipes are identified using time-frequency spectral methods. FDTD simulations reveal that sand-based models are computationally less intensive than those involving finer soils like clay or silt. Signal filtering techniques, including fitted velocity modelling and hyperbola extraction, improve the clarity and reliability of radargrams. In conclusion, this integrated GPR-based approach significantly advances underground pipeline diagnostics, offering a powerful tool for infrastructure maintenance, leak detection, and utility mapping in urban environments.

In data assimilation, observations are fused with simulations to obtain an accurate estimate of the state and parameters for a given physical system. Combining data with a model, however, while accurately estimating uncertainty, is computationally expensive and infeasible to run in real-time for complex systems. Here, we present a novel particle filter methodology, the Deep Latent Space Particle filter or D-LSPF, that uses neural network-based surrogate models to overcome this computational challenge. The D-LSPF enables filtering in the low-dimensional latent space obtained using Wasserstein AEs with modified vision transformer layers for dimensionality reduction and transformers for parameterized latent space time stepping. As we demonstrate on three test cases, including leak localization in multi-phase pipe flow and seabed identification for fully nonlinear water waves, the D-LSPF runs orders of magnitude faster than a high-fidelity particle filter and 3-5 times faster than alternative methods while being up to an order of magnitude more accurate. The D-LSPF thus enables real-time data assimilation with uncertainty quantification for the test cases demonstrated in this paper.

Leaks in pipelines are a very important issue and should not be underestimated in terms of handling and resolution. If a leak in a pipe is not addressed immediately, it can cause adverse effects, such as a decrease in water flow efficiency, and in some conditions, it can even stop the flow of water completely. Therefore, a water pipe leak detection system is needed to be implemented in homes. In this research, a prototype of a water pipe leak detection system consisting of a 4-branch and a T-branch was built to simulate piping conditions in users' homes. This system uses a water flow meter sensor as a leak detection sensor by comparing the difference in water discharge readings between sensors. me-ter water flow sensor is a sensor that is very suitable for this task. The water flow meter sensor works on the principle of hall effect and can read the water flow rate precisely and accurately with a reading accuracy of 9.9%. The system is IoT connected to blynk for easy monitoring and maintenance by the user. From 10 experiments conducted, the system successfully detected leaks with 100% success rate. The delay that occurs when sending data from the microcontroller to blynk is 67 ms.

Water pipe underground leakage detection was very important reduces water loss in the plumbing system. The occurrence of leakage in underground water pipelines represents a considerable challenge to the effective conservation of water resources and the maintenance of associated infrastructure. This study proposes a novel detection system based on soil moisture monitoring and fuzzy logic to accurately identify and classify leak severity in subterranean water pipelines. Soil moisture sensors situated in close proximity to the pipeline are employed to ascertain the real-time moisture levels, which are then processed by a fuzzy logic algorithm to determine the leak condition with high accuracy. The system's efficacy was validated through a prototype that achieved an average error rate of 6.64%, thereby demonstrating reliable performance in simulated leakage scenarios. This methodology offers a cost-effective, scalable solution to detect pipeline leaks and minimize water loss, with promising applications in urban and rural water distribution systems

No abstract available

No abstract available

No abstract available

No abstract available

No abstract available

No abstract available

No abstract available

No abstract available

No abstract available

Detecting and mapping subsurface utilities in urban areas is crucial for identifying defects or damages in drinking and sewage pipes that can cause leaks. These leaks make it difficult to accurately characterize the pipes due to changes in their reflective properties. This study focused on detecting leaks originating from underground pipes and distinguishing between these various types of pipes. It also aimed to create a visual fingerprint model that displays the reflection characteristics of these pipes during different leak conditions, enabling efficient maintenance and handling procedures on the pipes. To achieve this, a finite-difference time-domain (FDTD) method was used to simulate two types of pipe materials with and without leak areas to construct different scenarios. Additionally, a ground-penetrating radar (GPR) field survey was conducted using a 600 MHz antenna in a part of the El Hammam area on Egypt’s northwest coast. The simulated images produced with numerical modeling were compared with the radar profiles obtained using GPR at particular locations. The numerical simulations and radar profiles demonstrated the noticeable influence of water leaks from the different pipes, wherein the reflection of saturated soil waves was interrupted due to the presence of saturated soil. Envelope and migration techniques were employed in a new application to accurately distinguish between different pipe types, specifically focusing on leak areas. The strong correlation between the real radar profile and the specific signal of a water pipe leak in the simulated models suggests that GPR is a reliable non-destructive geophysical method for detecting water pipe leaks and distinguishing between the different pipe materials in various field conditions. The simulated models, which serve as image-matching fingerprints to identify and map water pipe leaks, help us to comprehend reality better.

The design of leak detection systems that occur on water pipes is a priority area of applied research that has an economic and health impact on the future of any nation. The various control systems and tools that currently exist throughout the world are designed to ensure permanent and effective monitoring of natural resources that have become rare and precious. The determining factor for the choice of a good detector is the cost in the first place, flexibility, and speed of processing. In this work, the basic idea is to simultaneously acquire from a new inexpensive electronic device two signals from two pressure transmitters installed on a prototype pipe carried out at the laboratory. These signals are usually immersed in noise. For this, denoising by an appropriate digital filter is indispensable. In our case, the Savitzky-Golay filter (S.G) presents its efficiency. The denoising performances are obtained from the calculation of SNR. The denoised signals are analysed to confirm the presence of the leak in the case of its existence. Mathematical equations are applied to determine the exact position of the leak with regards to one of the sensors. Validation tests are required to determine the position of the leak when the difference time between the signals is known.

No abstract available

With drinking water resources rapidly depleting with time, water conservation efforts have received special emphasis, especially in arid regions like California. One of the major sources of unused water expenditure is inconspicuous leaks in underground water distribution networks (WDN), making it highly essential to quickly detect and localize them. The leak detection and localization problem has been widely studied for a straight pipeline system, however, estimating the leak location in a pipe network remains largely unexplored. In this study, we measure the acoustic pressure signals inside a pipe network at multiple locations using state-of-the-art hydrophone-enabled devices. To localize the leaks in pipe networks, we propose maximum likelihood estimation, which has previously shown high efficacy in localizing mobile devices in a cellular network. In this approach, the cross-correlation of the filtered signals from different sensor pairs yields multiple time delays corresponding to multiple acoustic paths traversed by the leak noise in the pipe network, which is more difficult to solve compared to a straight pipe system. The leak location is then identified by maximizing a conditional probability distribution function of the distance between the sensor and the leak location.

In water supply pipe leakage detection, a single type of sensor data makes it difficult to describe the pipeline's operational status comprehensively, resulting in a high false detection rate when determining the extent of leakage. Therefore, this paper proposes a leakage detection method based on the Improved Bi-sensor Gramian Angular Field (IBGAF) and Dual-channel Multiscale Feature Fusion Network (DMFFN). The method can fully utilize the complementary information of the two sensors and improve the reliability of leak detection. In IBGAF, the two types of sensor data are mapped as points in a spherical coordinate system. The inverse cosines of their normalized amplitudes are set as the polar and azimuthal angles, respectively. Based on the positional correlation of the points, three IBGAF matrices are computed to encode the R, G, and B color channels of the image for signal fusion. The DMFFN adopts a multi-scale convolutional architecture combined with the attention mechanism, which can fully mine the complementary information of the fused signals. The designed 2-D convolutional parallel feature fusion mechanism performs adaptive fusion of multi-scale information, effectively reducing network complexity. The proposed method is compared with the single-class signal detection-based method, and the experimental results demonstrate that the method can significantly enhance detection accuracy. Compared with other detection methods, the method has the best performance and can accurately recognize four leakage conditions of pipelines.

The increasing scarcity of potable water resources, exacerbated by climate change and aging urban infrastructure, has intensified the demand for real-time monitoring and detection of water losses in distribution networks. Traditional leak detection methods often rely on single-point analysis and lack detailed information on the hardware and software architecture of proposed systems, limiting their scalability and applicability. In this study, a wireless sensor network (WSN) based on FSR sensors was designed and implemented, with a gateway forwarding data to a cloud platform for remote monitoring, to address these limitations. The proposed system integrates both hardware and software components, providing secure wireless communication and real-time monitoring of pressure variations along the pipeline. Experimental validation was conducted using a laboratory-scale prototype pipeline where multiple simultaneous leakage scenarios were created. Results demonstrated that the developed device reliably detected and differentiated leaks at varying flow rates and multiple points, with sensors closest to leakage sites exhibiting rapid pressure drops, while downstream sensors showed delayed but converging responses. These findings not only validate the effectiveness of FSR-based sensing for multi-leak detection but also highlight the advantages of combining IoT and WSN architectures for scalable and low-cost monitoring solutions.

Generally, municipal water supply companies use manual collection and laboratory analysis for water quality testing. However, these methods have limitations such as lack of real-time information, inability to sample the entire water supply, and high costs. Therefore, continuous, real-time water quality monitoring is crucial for public health protection and ensuring that the whole water supply network is monitored. This paper proposes an Internet of Things (IoT) platform for the measurement of consumption and the quality of drinking water in rural or semi-rural environments. Data collected through temperature, flow, potential of hydrogen (pH), turbidity, and Oxidation Reduction Potential (ORP) sensors is exchanged with a database through a long-range wireless communication protocol. Two mobile applications and one desktop application were also developed, with the purpose of being used by simple users, technicians, and network administrators respectively. The presented implementation process includes the design of the hardware surrounding the ESP32 microcontroller and its mounted peripherals, as well as the software run by the microcontroller and the mobile devices. A prototype system was built and tested under controlled conditions, successfully recognizing an increase in water turbidity and its unsuitability when contaminated with different agents. This method may prove to be a financially advantageous solution for rural, semi-rural, and even urban environments when used with groups of data collection nodes, helping significantly in the upkeep and surveillance of the water supply network. The IoT platform is equipped with sensors that measure water consumption as well as turbidity, temperature, and pH. Tested under controlled conditions, detecting different signs of water contamination. A low-cost option for continuous water monitoring, particularly in resource-limited environments. The IoT platform is equipped with sensors that measure water consumption as well as turbidity, temperature, and pH. Tested under controlled conditions, detecting different signs of water contamination. A low-cost option for continuous water monitoring, particularly in resource-limited environments.

In wireless sensor network-based water pipeline monitoring (WWPM) systems, a vital requirement emerges: the achievement of low energy consumption. This primary goal arises from the fundamental necessity to ensure the sustained operability of sensor nodes over extended durations, all without the need for frequent battery replacement. Given that sensor nodes in such applications are typically battery-powered and often physically inaccessible, maximizing energy efficiency by minimizing unnecessary energy consumption is of vital importance. This paper presents an experimental study that investigates the impact of a hybrid technique, incorporating distributed computing, hierarchical sensing, and duty cycling, on the energy consumption of a sensor node in prolonging the lifespan of a WWPM system. A custom sensor node is designed using the ESP32 MCU and nRF24L01+ transceiver. Hierarchical sensing is implemented through the use of LSM9DS1 and ADXL344 accelerometers, distributed computing through the implementation of a distributed Kalman filter, and duty cycling through the implementation of interrupt-enabled sleep/wakeup functionality. The experimental results reveal that combining distributed computing, hierarchical sensing and duty cycling reduces energy consumption by a factor of eight compared to the lone implementation of distributed computing.

Pipe bursts in water distribution networks (WDNs) not only cause significant water wastage but also pose risks of secondary disasters, such as road collapses. In most cases, these failures often result in a simultaneous decrease in values across multiple pressure sensors, indicating strong correlations in pressure data from the same event. This correlation suggests that leveraging the cooperative interaction between multiple sensors can significantly enhance burst detection sensitivity and expand coverage—an advantage largely overlooked in current methods. To address this gap, this study introduces a sensor cooperation gain system that constructs virtual sensors based on pressure data correlations and applies value fusion principles to form a multi‐sensor, cooperation‐based burst detection model, thus enabling more accurate system‐wide decision‐making. Case studies demonstrate that, compared to the existing method, the proposed gain fusion approach improves monitoring coverage by 25.34% and reduces the minimum coverage burst flow rate by 29.69 m3/hr.

This study introduces the build out and implementation of a wireless sensor network that was used to monitor dynamic pressure of water flowing inside a pipe against leak. At 5 chosen points along a prototype water pipeline, pressure sensors (HK1100C) were used to collect pressure data and its voltage data equivalent. Five locally designed and constructed field sensor nodes made up of pressure sensors, Ardiuno boards containing an 8-bit ATmega328P microcontrollers, GSM SIM800L modules, and power sources were used to process and analyze the pressure data such that when the pressure data falls below 9psi, 9.5psi, 10psi, 10.5psi, and 11psi respectively at the 5 chosen points, the information is processed by the Ardiuno boards containing an 8-bit Atmega328P microcontrollers and then passed on to the GSM SIM800L modules of respective field sensor nodes which then transmit these information as radio waves to the master sensor node of the network that has an LCD JHD162 module and a buzzer.. The received radio waves were converted back into electrical signals by the transceiver contained in the GSM SIM800L module of the master sensor node and used to power the LCD and buzzer to indicate ‘low pressure’ and sound an alarm simultaneously. The location of the pressure drop is identified and the cause rectified by trained personnel. SIM cards of chosen GSM networks were inserted into the various GSM SIM800L modules contained in the sensor nodes (fields and master nodes) for communication. Also, an experiential evaluation of the relationship between leaks and pressure variations was carried out at one of the chosen 5 points along the prototype water pipeline.

Traditional laboratory-based water quality monitoring and testing approaches are soon to be outdated, mainly because of the need for real-time feedback and immediate responses to emergencies. The more recent wireless sensor network (WSN)-based techniques are evolving to alleviate the problems of monitoring, coverage, and energy management, among others. The inclusion of the Internet of Things (IoT) in WSN techniques can further lead to their improvement in delivering, in real time, effective and efficient water-monitoring systems, reaping from the benefits of IoT wireless systems. However, they still suffer from the inability to deliver accurate real-time data, a lack of reconfigurability, the need to be deployed in ad hoc harsh environments, and their limited acceptability within industry. Electronic sensors are required for them to be effectively incorporated into the IoT WSN water-quality-monitoring system. Very few electronic sensors exist for parameter measurement. This necessitates the incorporation of artificial intelligence (AI) sensory techniques for smart water-quality-monitoring systems for indicators without actual electronic sensors by relating with available sensor data. This approach is in its infancy and is still not yet accepted nor standardized by the industry. This work presents a smart water-quality-monitoring framework featuring an intelligent IoT WSN monitoring system. The system uses AI sensors for indicators without electronic sensors, as the design of electronic sensors is lagging behind monitoring systems. In particular, machine learning algorithms are used to predict E. coli concentrations in water. Six different machine learning models (ridge regression, random forest regressor, stochastic gradient boosting, support vector machine, k-nearest neighbors, and AdaBoost regressor) are used on a sourced dataset. From the results, the best-performing model on average during testing was the AdaBoost regressor (a MAE¯ of 14.37 counts/100 mL), and the worst-performing model was stochastic gradient boosting (a MAE¯ of 42.27 counts/100 mL). The development and application of such a system is not trivial. The best-performing water parameter set (Set A) contained pH, conductivity, chloride, turbidity, nitrates, and chlorophyll.

The problem of the need for clean water is very important. The several diseases triggered by poor water quality reach more than 200 cases each year and cause more than 5 million deaths worldwide. Thus, monitoring water quality becomes important for the availability of safe and clean water. Wireless sensor networks have become a promising alternative to adopted for supplementing the conventional monitoring process. This network allows measurements from the remote location in real-time and with little human intervention. Wireless sensor network topology performance will support the stability of real-time data transmission. The difference in network topology between each router-node station affects the disruption of data distribution. Quality of Services (QoS) measurement is based on wireless sensor connectivity in transmit sensor data from several parameters, including temperature, total dissolved solids, and pH in the node station to the website service. The delay in transmitting data affects the performance of the measurement of the water pollution monitoring system.

The development of accurate corrosion models from corrosion data is a key requirement for well informed, preemptive actions towards extending the life of water infrastructure systems and reducing maintenance costs. This work presents the development of a custom proof-of-concept system implementation performing water corrosion monitoring using custom integrated corrosion sensing technologies powered by energy harvested through an in-line turbine generator. Implementing a star topology network of wireless corrosion sensor nodes distributed through water infrastructure systems offers an innovative solution for real-time and on-line monitoring of corrosion in water infrastructure systems. The wireless sensor network includes a wireless access point that provides an interface for retrieval of the corrosion data and a nexus for communications. A low-power bidirectional data transfer communications protocol using a form of time-division multiplex access is designed and implemented towards the optimisation of energy efficiency in the network. In an emulated pipeline field setup, the energy harvested and stored provides 125.80 mWh at 4.89 V for an in-pipe flow rate of 13.25 l/min, a typical value for municipal potable water supply in South Africa. The local node ER sensor can of measure changes in resistance values to an accuracy of 1% and will be usable in-system for a year. The access point implements EIS and LPR sensing. The extracted LPR value predicts a corrosion rate is 63.0 um/year in potable water.

Nowadays, air and water quality are a big problem in the cities because they affect human health deeply. To determine air and water quality, the monitoring system containing a wireless sensor node having several sensors has been developed with the internet of things (IoT) technology in this study. It regularly monitors various parameters such as temperature, humidity, bar pressure, wind direction, wind speed max, wind speed average, rain fall one hour, rain fall one day, $PM_{2.5}$ and PM10 for air quality monitoring and consumed water, waste water, incoming water, pH sensor, conductivity sensor, turbidity sensor, dissolved oxygen and temperature sensors for water quality monitoring. To improve smart city environments, air and water quality monitoring has to be common, present everywhere, and rapidly responsive. The monitoring system based on IoT is able to track air and water pollution in real time and transmit the information fast through a wide area network. This system can be integrated with innovative approaches like smart city. Therefore, the demand for a real time monitoring system will increase day by day.

Real-time, high-resolution monitoring of chemically diverse water pollutants remains a critical challenge for smart water management. Here, we report a fully integrated, multi-modal nano-sensor array, combining graphene field-effect transistors, Ag/Au-nanostar surface-enhanced Raman spectroscopy substrates, and CdSe/ZnS quantum dot fluorescence, coupled to an edge-deployable CNN-LSTM architecture that fuses raw electrochemical, vibrational, and photoluminescent signals without manual feature engineering. The 45 mm × 20 mm microfluidic manifold enables continuous flow-through sampling, while 8-bit-quantised inference executes in 31 ms at <12 W. Laboratory calibration over 28,000 samples achieved limits of detection of 12 ppt (Pb2+), 17 pM (atrazine) and 87 ng L−1 (nanoplastics), with R2 ≥ 0.93 and a mean absolute percentage error <6%. A 24 h deployment in the Cherwell River reproduced natural concentration fluctuations with field R2 ≥ 0.92. SHAP and Grad-CAM analyses reveal that the network bases its predictions on Dirac-point shifts, characteristic Raman bands, and early-time fluorescence-quenching kinetics, providing mechanistic interpretability. The platform therefore offers a scalable route to smart water grids, point-of-use drinking water sentinels, and rapid environmental incident response. Future work will address sensor drift through antifouling coatings, enhance cross-site generalisation via federated learning, and create physics-informed digital twins for self-calibrating global monitoring networks.

Water distribution sector is currently facing operation challenges, especially in urban areas, as a result of climate change and high rate of urbanization experienced in these areas. These two aforementioned factors pose a threat of water shortage in these cities and consequently affects the water distribution sector as there will be interruption in water supply. To mitigate this, efficient management of the entire distribution process and automated monitoring and control systems are needed. This paper showcases such an automated water distribution system that is responsible for providing clean water to urban residents. The system uses an IoT (Internet of Things) approach for constant and real-time monitoring of water supply from the Base Station down to the water tanks of each block of house. Water levels and water quality are continuously monitored in the end-user’s water tanks, and the tanks are automatically refilled based on user requirements and Base Station planning. Both end users and water board managers can follow the water flow information in a user-friendly web application.

No abstract available

The need for safe drinking water in refugee camps has underscored the importance of deploying sustainable, low-cost water quality monitoring solutions. This study explores the design, implementation, and performance of a Low-Cost IoT Sensor Network (LCISN) tailored for realtime water quality monitoring in humanitarian settings. The proposed system integrates low-power microcontrollers, opensource software platforms, and a suite of sensors (pH, turbidity, temperature, and electrical conductivity) to collect and transmit water quality data over wireless mesh networks. Data are aggregated on cloud-based dashboards to enable rapid detection of contamination and facilitate timely intervention by camp management. The system emphasizes energy efficiency through solar-powered nodes and intermittent data transmission protocols, enabling extended operation in infrastructure-limited environments. The solution provides a transformative approach for water safety in crisis zones, empowering communities with actionable environmental intelligence. This research contributes to the advancement of humanitarian engineering in water quality management for displaced populations.

we are now moving into an era where the Iota and sensor-based technologies can monitor many parameters, including water quality. Since fresh, potable water is becoming increasingly scarce in many parts of the globe, maintaining a clean and sustainable source for consumption remains one of our biggest challenges today. This challenge can be addressed by leveraging IoT and sensor-based systems that accumulate data from several devices plus study it toward make insight designed meant for better executive. Most available IoT and sensor-based water quality monitoring systems use a network of devices, generally called sensors or data loggers that are connected to take in the multiple variables related to parameters like pH levels, dissolved oxygen placeholders d, and pollutants. This data is subsequently integrated with scientific algorithms and machine learning, which helps to determine water quality fluctuations in near real-time. These systems can be combined with geographic information systems, illuminating the data and making it mappable so that problem areas will likely emerge. IoT and sensor-based water quality monitoring are advantageous against conventional methods in terms of better accuracy, increased efficiency, cost-effectiveness to monitor long term, and the ability to demonstrate effective real-time information about water fettle. Information like this can be critical for managing water resources, as it can identify potential problems such as contamination or water running out. It optimizes these technologies to monitor remote locations, decreasing human labor requirements and enabling real-time decision-making in sustainable water resource management.

The development of sophisticated monitoring systems that can do thorough and real-time assessments has been spurred by growing worries about the quality of water. In this study, we suggest a unique method for dynamically monitoring the quality of water by combining machine learning techniques with an Internet of Things (IoT) sensor network. With carefully placed IoT sensors inside water bodies or distribution networks, the system is intended to continually gather multiple parameter data, such as pH, turbidity, temperature, and dissolved oxygen. Modern machine learning algorithms housed on cloud infrastructure are used to process and analyze the gathered data. Our method seeks to identify abnormalities, forecast changes in water quality, and offer current information on the state of water resources. Machine learning models are trained on past data in order to detect trends, spot departures from the norm, and make it easier to make proactive decisions in reaction to changes or possible pollutants. We outline the design of our Internet of Things (IoT) sensor network, how cloud computing is integrated for data processing, and how machine learning algorithms are put into practice for predictive analytics. We also go over the system’s flexibility to changing environmental circumstances, scalability, and possible uses in environmental protection and water resource management.

No abstract available

The rapid detection and localisation of pipe burst incidents in water pipe systems are crucial for timely repair, minimisation of leakage, and assurance of water supply security. However, previous researches have insufficiently emphasised the identification of pressure characteristics during pipe burst, failing to accurate detection and localisation of pipe burst. To address this problem, an experimental system was designed to elucidate the complete dynamic process of pipe burst pressure fluctuations. Specifically, the continuous transition from steady state to transient state and back to steady state in the hydraulic state of the system after a pipe burst event, was observed both in time and space dimensions. The effects of initial pressure and flow rate at steady state, location and size of pipe burst on pressure characteristics were carefully investigated. Subsequently, a pipe burst model considering unsteady friction effect was developed and validated through experiments. Based on the above study, analytical solution regarding pressure drop during burst events and steady-state pressure at the burst point after burst events were derived to quantify the influence laws of different parameters and validated by the pipe burst model. The obtained results are characterised by the dominance analysis so as to explore the importance ranking of different factors to the pressure drop and steady-state burst pressure in the pipeline system. The analysis solution of pressure drop indicates the following ranking of factors from high to low is: pipe/burst area ratio, initial steady-state pressure at the burst point, wave speed; while the ranking of influencing factors for steady-state burst pressure from high to low is: the boundary pressures, the location of the pipe burst, the initial steady state flow rate, and the burst outflow coefficient. The research findings can provide theoretical and practical guidance for the rapid diagnosis and quantitative analysis of burst pipes.

Urban water systems worldwide are confronted with the dual challenges of dwindling water resources and deteriorating infrastructure, emphasising the critical need to minimise water losses from leakage. Conventional methods for leak and burst detection often prove inadequate, leading to prolonged leak durations and heightened maintenance costs. This study investigates the efficacy of logic- and machine learning-based approaches in early leak detection and precise location identification within water distribution networks. By integrating hardware and software technologies, including sensor technology, data analysis, and study on the logic-based and machine learning algorithms, innovative solutions are proposed to optimise water distribution efficiency and minimise losses. In this research, we focus on a case study area in the Sunbury region of Victoria, Australia, evaluating a pumping main equipped with Supervisory Control and Data Acquisition (SCADA) sensor technology. We extract hydraulic characteristics from SCADA data and develop logic-based algorithms for leak and burst detection, alongside state-of-the-art machine learning techniques. These methodologies are applied to historical data initially and will be subsequently extended to live data, enabling the real-time detection of leaks and bursts. The findings underscore the complementary nature of logic-based and machine learning approaches. While logic-based algorithms excel in capturing straightforward anomalies based on predefined conditions, they may struggle with complex or evolving patterns. Machine learning algorithms enhance detection by learning from historical data, adapting to changing conditions, and capturing intricate patterns and outliers. The comparative analysis of machine learning models highlights the superiority of the local outlier factor (LOF) in anomaly detection, leading to its selection as the final model. Furthermore, a web-based platform has been developed for leak and burst detection using a selected machine learning model. The success of machine learning models over traditional logic-based approaches underscores the effectiveness of data-driven, probabilistic methods in handling complex data patterns and variations. Leveraging statistical and probabilistic techniques, machine learning models offer adaptability and superior performance in scenarios with intricate or dynamic relationships between variables. The findings demonstrate that the proposed methodology can significantly enhance the early detection of leaks and bursts, thereby minimising water loss and associated economic costs. The implications of this study are profound for the scientific community and stakeholders, as it provides a scalable and efficient solution for water pipeline monitoring. Implementing this approach can lead to more proactive maintenance strategies, ultimately contributing to the sustainability and resilience of urban water infrastructure systems.

Utility companies lose approximately 35 liters of water for every 100 produced due to incorrect, illegal connections and the poor condition of pipes. This study develops an intelligent model to detect leaks using the Kalman filter, BiLSTM neural networks, and the Monte Carlo Dropout algorithm. Using data from the Empresa de Acueductos y Alcantarillados de Bogotá (EAAB), Colombia, autocorrelation analysis, PCA, cluster analysis, ADF and Durbin-Watson tests, Hurst exponent, spectral analysis, and wavelet transform were performed. Then, Kalman filtering techniques were applied, and a BiLSTM architecture controlled with Monte Carlo dropout was implemented. The results showed an accuracy of 87.48% in training and 80.48% in validation. Temporal analysis revealed a stationary behavior in the flow series, and the decrease in spectral intensity around 0.25 Hz was related to pressure perturbations caused by leaks. A detailed evaluation of pressure and flow signals identified leak patterns with high precision, demonstrating the effectiveness of the wavelet spectrogram in detecting energy disturbances. The novelty of the study lies in the integration of advanced artificial intelligence and combinatorial optimization techniques to improve water resource management, allowing early and accurate detection of leaks, significantly improving compared to traditional methods. Doi: 10.28991/CEJ-2024-010-07-01 Full Text: PDF

Leaks occur at different rates in water distribution systems (WDSs). Network characteristics, high pressure, environmental factors and operational factors are effective on leaks. Field detection and monitoring activities should be implemented to reduce the volume of leaks resulting from faults in the WDS. The aim of this study is to create and calibrate the district metered area (DMA) based hydraulic model to understand the network behavior and monitor the hydraulic components. The hydraulic model is based on consumption data, network topology, characteristics and pipe roughness information. Calibration should be performed by comparing the pressures obtained from the model with the pressures measured in the field in order to apply the model in leakage management. Incomplete or incorrect network information may cause the difference between these two pressures to be large. In particular, basic data such as incomplete creation of the network topology, incomplete or incorrect acquisition of roughness and consumption information are effective in not providing model calibration. In the calibrated hydraulic model, it is possible to detect and prevent potential leaks by monitoring pressure changes at the nodes. It is thought that the results obtained in this study will constitute a reference in leakage management and hydraulic analysis.

This research article presents a data-driven approach for detecting bursts in water distribution networks (WDNs). The framework uses spatiotemporal information from monitoring pressure and unsupervised learning model. This approach employs three stages: (1) benchmark dataset acquisition, (2) spatiotemporal information analysis, and (3) burst detection model construction. First, the benchmark datasets were the normal dataset initially obtained by the clustering algorithm. Second, spatiotemporal information features are extracted from multimoment time windows from multiple sensors, including the distance and shape features. Third, burst detection was performed based on the isolation forest technique. A WDN is used to evaluate the performance of the method. Results show that the method can effectively detect the burst.

Pressure and flow estimation in Water Distribution Networks (WDN) allows water management companies to optimize their control operations. For many years, mathematical simulation tools have been the most common approach to reconstructing an estimate of the WDN hydraulics. However, pure physics-based simulations involve several challenges, e.g. partially observable data, high uncertainty, and extensive manual configuration. Thus, data-driven approaches have gained traction to overcome such limitations. In this work, we combine physics-based modeling and Graph Neural Networks (GNN), a data-driven approach, to address the pressure estimation problem. First, we propose a new data generation method using a mathematical simulation but not considering temporal patterns and including some control parameters that remain untouched in previous works; this contributes to a more diverse training data. Second, our training strategy relies on random sensor placement making our GNN-based estimation model robust to unexpected sensor location changes. Third, a realistic evaluation protocol considers real temporal patterns and additionally injects the uncertainties intrinsic to real-world scenarios. Finally, a multi-graph pre-training strategy allows the model to be reused for pressure estimation in unseen target WDNs. Our GNN-based model estimates the pressure of a large-scale WDN in The Netherlands with a MAE of 1.94mH$_2$O and a MAPE of 7%, surpassing the performance of previous studies. Likewise, it outperformed previous approaches on other WDN benchmarks, showing a reduction of absolute error up to approximately 52% in the best cases.

This paper studies the problem of optimal placement of water quality (WQ) sensors in water distribution networks (WDNs), with a focus on chlorine transport, decay, and reaction models. Such models are traditionally used as suitable proxies for WQ. The literature on this topic is inveterate, but has a key limitation: it utilizes simplified single-species decay and reaction models that do not capture WQ transients for nonlinear, multi-species interactions. This results in sensor placements (SP) that do not account for nonlinear WQ dynamics. Furthermore, as WQ simulations are parameterized by hydraulic profiles and demand patterns, the placement of sensors are often hydraulics-dependent. This study produces a greedy algorithm that addresses the two aforementioned limitations. The algorithm is grounded in nonlinear dynamic systems and observability theory, and yields SPs that are submodular and robust to hydraulic changes. Case studies on benchmark water networks are provided. The key findings provide practical recommendations for WDN operators.

Leakages in drinking water distribution networks pose significant challenges to water utilities, leading to infrastructure failure, operational disruptions, environmental hazards, property damage, and economic losses. The timely identification and accurate localisation of such leakages is paramount for utilities to mitigate these unwanted effects. However, implementation of algorithms for leakage detection is limited in practice by requirements of either hydraulic models or large amounts of training data. Physics-informed machine learning can utilise hydraulic information thereby circumventing both limitations. In this work, we present a physics-informed machine learning algorithm that analyses pressure data and therefrom estimates unknown irregular water demands via a fully connected neural network, ultimately leveraging the Bernoulli equation and effectively linearising the leakage detection problem. Our algorithm is tested on data from the L-Town benchmark network, and results indicate a good capability for estimating most irregular demands, with R2 larger than 0.8. Identification results for leakages under the presence of irregular demands could be improved by a factor of 5.3 for abrupt leaks and a factor of 3.0 for incipient leaks when compared the results disregarding irregular demands.

The Alqueva Multi-Purpose Project (EFMA) is a massive abduction and storage infrastructure system in the Alentejo, which has a water quality monitoring network with almost thousands of water quality stations distributed across three subsystems: Alqueva, Pedrogão, and Ardila. Identification of pollution sources in complex infrastructure systems, such as the EFMA, requires recognition of water flow direction and delimitation of areas being drained to specific sampling points. The transfer channels in the EFMA infrastructure artificially connect several water bodies that do not share drainage basins, which further complicates the interpretation of water quality data because the water does not flow exclusively downstream and is not restricted to specific basins. The existing user-friendly GIS tools do not facilitate the exploration and visualisation of water quality data in spatial-temporal dimensions, such as defining temporal relationships between monitoring campaigns, nor do they allow the establishment of topological and hydrological relationships between different sampling points. This thesis work proposes a framework capable of aggregating many types of information in a GIS environment, visualising large water quality-related datasets and, a graph data model to integrate and relate water quality between monitoring stations and land use. The graph model allows to exploit the relationship between water quality in a watercourse and reservoirs associated with infrastructures. The graph data model and the developed framework demonstrated encouraging results and has proven to be preferred when compared to relational databases.

Due to the global water crisis there is a strong need for real-time water quality monitoring with high temporal and spatial resolution. This paper presents an economical multiparameter water quality monitoring system for continuous monitoring of fresh waters. It is based on a sensor node that integrates turbidity, temperature, and conductivity sensors, a miniature eighteen-channel spectrophotometer, and a sensor for the detection of thermotolerant coliforms, which is a major novelty of the system. Due to the influence of water impurities on the measurement of thermotolerant coliforms, a heuristic method has been developed to mitigate this effect. Moreover, the sensor is low power and with an integrated Long Range Wide Area Network module, it comprises a system that is wireless sensor network ready and can send data to a dedicated server. In addition, the system is submersible, capable of long-term field operation, and significantly cheaper in comparison to existing solutions. The purpose of the system is to give early warning of incidental pollution situations, thus enabling authorities to take action regarding further prevention of such occasions.

A state-space representation of water quality dynamics describing disinfectant (e.g., chlorine) transport dynamics in drinking water distribution networks has been recently proposed. Such representation is a byproduct of space- and time-discretization of the PDE modeling transport dynamics. This results in a large state-space dimension even for small networks with tens of nodes. Although such a state-space model provides a model-driven approach to predict water quality dynamics, incorporating it into model-based control algorithms or state estimators for large networks is challenging and at times intractable. To that end, this paper investigates model order reduction (MOR) methods for water quality dynamics with the objective of performing post-reduction feedback control. The presented investigation focuses on reducing state-dimension by orders of magnitude, the stability of the MOR methods, and the application of these methods to model predictive control.

Traditional control and monitoring of water quality in drinking water distribution networks (WDN) rely on mostly model- or toolbox-driven approaches, where the network topology and parameters are assumed to be known. In contrast, system identification (SysID) algorithms for generic dynamic system models seek to approximate such models using only input-output data without relying on network parameters. The objective of this paper is to investigate SysID algorithms for water quality model approximation. This research problem is challenging due to (i) complex water quality and reaction dynamics and (ii) the mismatch between the requirements of SysID algorithms and the properties of water quality dynamics. In this paper, we present the first attempt to identify water quality models in WDNs using only input-output experimental data and classical SysID methods without knowing any WDN parameters. Properties of water quality models are introduced, the ensuing challenges caused by these properties when identifying water quality models are discussed, and remedial solutions are given. Through case studies, we demonstrate the applicability of SysID algorithms, show the corresponding performance in terms of accuracy and computational time, and explore the possible factors impacting water quality model identification.

The quality of water is key for the quality of agrifood sector. Water is used in agriculture for fertigation, for animal husbandry, and in the agrifood processing industry. In the context of the progressive digitalization of this sector, the automatic assessment of the quality of water is thus becoming an important asset. In this work, we present the integration of Ultraviolet-Visible (UV-Vis) spectroscopy with Machine Learning in the context of water quality assessment aiming at ensuring water safety and the compliance of water regulation. Furthermore, we emphasize the importance of model interpretability by employing SHapley Additive exPlanations (SHAP) to understand the contribution of absorbance at different wavelengths to the predictions. Our approach demonstrates the potential for rapid, accurate, and interpretable assessment of key water quality parameters.

Water quality monitoring is a core component of ecological environmental protection. However, due to sensor failure or other inevitable factors, data missing often exists in long-term monitoring, posing great challenges in water quality analysis. This paper proposes a Neural Tucker Convolutional Network (NTCN) model for water quality data imputation, which features the following key components: a) Encode different mode entities into respective embedding vectors, and construct a Tucker interaction tensor by outer product operations to capture the complex mode-wise feature interactions; b) Use 3D convolution to extract fine-grained spatiotemporal features from the interaction tensor. Experiments on three real-world water quality datasets show that the proposed NTCN model outperforms several state-of-the-art imputation models in terms of accuracy.

With the advent of integrated sensor technology (smart flow meters and pressure sensors), various new numerical algorithms for leak localization (a core element of water distribution system operation) have been developed. However, there is a lack of theory regarding the limitations of leak localization. In this work, we contribute to the development of such a theory by introducing an example water network structure with parallel pipes that is tractable for analytical treatment. We define the leak localization problem for this structure and show how many sensors and what conditions are needed for the well-posedness of the problem. We present a formula for the leak position as a function of measurements from these sensors. However, we also highlight the risk of finding false but plausible leak positions in the multiple pipes. We try to answer the questions of how and when the leaking pipe can be isolated. In particular, we show that nonlinearities in the pipes' head loss functions are essential for the well-posedness of the isolation problem. We propose procedures to get around the pitfall of multiple plausible leak positions.

Water Distribution Networks (WDNs), critical to public well-being and economic stability, face challenges such as pipe blockages and background leakages, exacerbated by operational constraints such as data non-stationarity and limited labeled data. This paper proposes an unsupervised, online learning framework that aims to detect two types of faults in WDNs: pipe blockages, modeled as collective anomalies, and background leakages, modeled as concept drift. Our approach combines a Long Short-Term Memory Variational Autoencoder (LSTM-VAE) with a dual drift detection mechanism, enabling robust detection and adaptation under non-stationary conditions. Its lightweight, memory-efficient design enables real-time, edge-level monitoring. Experiments on two realistic WDNs show that the proposed approach consistently outperforms strong baselines in detecting anomalies and adapting to recurrent drift, demonstrating its effectiveness in unsupervised event detection for dynamic WDN environments.

Most reported research for monitoring health of pipelines using ultrasonic guided waves (GW) typically utilize bulky piezoelectric transducer rings and laboratory-grade ultrasonic non-destructive testing (NDT) equipment. Consequently, the translation of these approaches from laboratory settings to field-deployable systems for real-time structural health monitoring (SHM) becomes challenging. In this work, we present an innovative algorithm for damage identification and localization in pipes, implemented on a compact FPGA-based smart GW-SHM system. The custom-designed board, featuring a Xilinx Artix-7 FPGA and front-end electronics, is capable of actuating the PZT thickness shear mode transducers, data acquisition and recording from PZT sensors and generating a damage index (DI) map for localizing the damage on the structure. The algorithm is a variation of the common source method adapted for cylindrical geometry. The utility of the algorithm is demonstrated for detection and localization of defects such as notch and mass loading on a steel pipe, through extensive finite element (FE) method simulations. Experimental results obtained using a C-clamp for applying mass loading on the pipe show good agreement with the FE simulations. The localization error values for experimental data analyzed using C code on a processor implemented on the FPGA are consistent with algorithm results generated on a computer running MATLAB code. The system presented in this study is suitable for a wide range of GW-SHM applications, especially in cost-sensitive scenarios that benefit from on-node signal processing over cloud-based solutions.

Water networks are used in numerous applications, all of which have the essential task of real-time monitoring and control of water states. A framework for the generation of efficient models of water networks suitable for real-time monitoring and control purposes is proposed. The proposed models preserve the distributed parameter character of the connected local elements. Hence, the spatial resolution of the property under consideration is recovered. The real-time feasibility of the network model is ensured by means of reduced order modeling of the models constituting components. A novel model order reduction procedure that preserves the model parametric dependency is introduced. The proposed concept is evaluated with the water temperature as the property under consideration. The formulated model is applied for the prediction of the water temperature within an experimental test bench of a 60 meter exemplary circulation network at the company VIEGA. A reduced order model (ROM) with a 50 mm spatial resolution, i.e. 1200 discretization points, is constructed and utilized as the identification model for a single path of the test bench. Afterwards, the ROM is evaluated in a Hardware in the Loop experiment for the prediction of the downstream temperature showing high prediction accuracy with mean relative error below 3.5 \%. The ROM single step computation time did not exceed 2 msec highlighting the real-time potential of the method. Moreover, full network model validation experiments featuring both diffusion and transport dominated parts were conducted. The network model is able to predict the temperature evolution, flow rate, and pressure accurately at the different paths of the network with mean relative errors below 4 \%, 2 \%, and 2 \%, respectively.

Water management is one of the most critical aspects of our society, together with population increase and climate change. Water scarcity requires a better characterization and monitoring of Water Distribution Networks (WDNs). This paper presents a novel framework for monitoring Water Distribution Networks (WDNs) by integrating physics-informed modeling of the nonlinear interactions between pressure and flow data with Topological Signal Processing (TSP) techniques. We represent pressure and flow data as signals defined over a second-order cell complex, enabling accurate estimation of water pressures and flows throughout the entire network from sparse sensor measurements. By formalizing hydraulic conservation laws through the TSP framework, we provide a comprehensive representation of nodal pressures and edge flows that incorporate higher-order interactions captured through the formalism of cell complexes. This provides a principled way to decompose the water flows in WDNs in three orthogonal signal components (irrotational, solenoidal and harmonic). The spectral representations of these components inherently reflect the conservation laws governing the water pressures and flows. Sparse representation in the spectral domain enable topology-based sampling and reconstruction of nodal pressures and water flows from sparse measurements. Our results demonstrate that employing cell complex-based signal representations enhances the accuracy of edge signal reconstruction, due to proper modeling of both conservative and non-conservative flows along the polygonal cells.

The optimization of water distribution systems (WDSs) is vital to minimize energy costs required for their operations. A principal approach taken by researchers is identifying an optimal scheme for water pump controls through examining computational simulations of WDSs. However, due to a large number of possible control combinations and the complexity of WDS simulations, it remains non-trivial to identify the best pump controls by reviewing the simulation results. To address this problem, we design a visual analytics system that helps understand relationships between simulation inputs and outputs towards better optimization. Our system incorporates interpretable machine learning as well as multiple linked visualizations to capture essential input-output relationships from complex WDS simulations. We demonstrate our system's effectiveness through a practical case study and evaluate its usability through expert reviews. Our results show that our system can lessen the burden of analysis and assist in determining optimal operating schemes.

Knowing the pressure at all times in each node of a water distribution system (WDS) facilitates safe and efficient operation. Yet, complete measurement data cannot be collected due to the limited number of instruments in a real-life WDS. The data-driven methodology of reconstructing all the nodal pressures by observing only a limited number of nodes is presented in the paper. The reconstruction method is based on K-localized spectral graph filters, wherewith graph convolution on water networks is possible. The effect of the number of layers, layer depth and the degree of the Chebyshev-polynomial applied in the kernel is discussed taking into account the peculiarities of the application. In addition, a weighting method is shown, wherewith information on friction loss can be embed into the spectral graph filters through the adjacency matrix. The performance of the proposed model is presented on 3 WDSs at different number of nodes observed compared to the total number of nodes. The weighted connections prove no benefit over the binary connections, but the proposed model reconstructs the nodal pressure with at most 5% relative error on average at an observation ratio of 5% at least. The results are achieved with shallow graph neural networks by following the considerations discussed in the paper.

In this document, the supervisory control and data acquisition (SCADA) and phasor measurement unit (PMU) measurement chain modeling will be studied, where the measurement error sources of each component in the SCADA and PMU measurement chains and the reasons leading to measurement errors exhibiting non-zero-mean, non-Gaussian, and time-varying statistical characteristic are summarized and analyzed. This document provides a few equations, figures, and discussions about the details of the SCADA and PMU measurement error chain modeling, which are intended to facilitate the understanding of how the measurement errors are designed for each component in the SCADA and PMU measurement chains. The measurement chain models described here are also used for synthesizing measurement errors with realistic characteristics in simulation cases to test the developed algorithms or methodologies.

Calibration is a critical process for reducing uncertainty in Water Distribution Network Hydraulic Models (WDN HM). However, features of certain WDNs, such as oversized pipelines, lead to shallow pressure gradients under normal daily conditions, posing a challenge for effective calibration. This study proposes a calibration methodology using short hydrant trials conducted at night, which increase the pressure gradient in the WDN. The data is resampled to align with hourly consumption patterns. In a unique real-world case study of a WDN zone, we demonstrate the statistically significant superiority of our method compared to calibration based on daily usage. The experimental methodology, inspired by a machine learning cross-validation framework, utilises two state-of-the-art calibration algorithms, achieving a reduction in absolute error of up to 45% in the best scenario.

Leakage in water systems results in significant daily water losses, degrading service quality, increasing costs, and aggravating environmental problems. Most leak localization methods rely solely on pressure data, missing valuable information from other sensor types. This article proposes a hydraulic state estimation methodology based on a dual Unscented Kalman Filter (UKF) approach, which enhances the estimation of both nodal hydraulic heads, critical in localization tasks, and pipe flows, useful for operational purposes. The approach enables the fusion of different sensor types, such as pressure, flow and demand meters. The strategy is evaluated in well-known open source case studies, namely Modena and L-TOWN, showing improvements over other state-of-the-art estimation approaches in terms of interpolation accuracy, as well as more precise leak localization performance in L-TOWN.

Water is a critical resource that must be managed efficiently. However, a substantial amount of water is lost each year due to leaks in Water Distribution Networks (WDNs). This underscores the need for reliable and effective leak detection and localization systems. In recent years, various solutions have been proposed, with data-driven approaches gaining increasing attention due to their superior performance. In this paper, we propose a new method for leak detection. The method is based on water pressure measurements acquired at a series of nodes of a WDN. Our technique is a fully data-driven solution that makes only use of the knowledge of the WDN topology, and a series of pressure data acquisitions obtained in absence of leaks. The proposed solution is based on an feature extractor and a one-class Support Vector Machines (SVM) trained on no-leak data, so that leaks are detected as anomalies. The results achieved on a simulate dataset using the Modena WDN demonstrate that the proposed solution outperforms recent methods for leak detection.

Detecting and localizing leaks in water distribution network systems is an important topic with direct environmental, economic, and social impact. Our paper is the first to explore the use of factor graph optimization techniques for leak localization in water distribution networks, enabling us to perform sensor fusion between pressure and demand sensor readings and to estimate the network's temporal and structural state evolution across all network nodes. The methodology introduces specific water network factors and proposes a new architecture composed of two factor graphs: a leak-free state estimation factor graph and a leak localization factor graph. When a new sensor reading is obtained, unlike Kalman and other interpolation-based methods, which estimate only the current network state, factor graphs update both current and past states. Results on Modena, L-TOWN and synthetic networks show that factor graphs are much faster than nonlinear Kalman-based alternatives such as the UKF, while also providing improvements in localization compared to state-of-the-art estimation-localization approaches. Implementation and benchmarks are available at https://github.com/pirofti/FGLL.

Water Distribution Systems(WDS) are critical infrastructures that deliver potable water to residential areas. Incidents to pipelines cause water loss and contamination in pipelines. Hence, water quality monitoring is one of the requirements for utility managers to ensure the health of water. However, it is challenging to access all parts of the WDS since they are long and comprise pipes with different configurations and sizes. In this paper, we propose a size-adaptable and modular in-pipe robot so-called SmartCrawler that works based on wheel wall press mechanism. We develop a two-phase motion control algorithm that enables reliable motion in straight and non-straight configurations of in-service pipelines. The controller in phase 1 stabilizes the robot in the straight paths and tracks the desired velocity with high-level linear quadratic regulator (LQR) and low-level proportional-integral-derivative (PID) based controllers. The controller in phase 2 with a designed error-check submodule and velocity controller, enables the robot to steer to the desired directions at non-straight configurations. The performance of the two-phase controller is evaluated with experimental and simulation results. Wireless underground communication is a challenging task for underground applications. We propose a bi-directional wireless sensor module based on active radio frequency identification (RFID) that works in 434MHz carrier frequency and evaluate its performance with experimental results. At the end of this work, we design the printed circuit board (PCB) for the SmartCrawler that embeds the electronic components in a confined and sealed environment. The simulation and the experimental results prove the proposed robotic system can be used for in pipe missions where wireless communication is needed to communicate with the robot during operation.

Timely leak detection in water distribution networks is critical for conserving resources and maintaining operational efficiency. Although Graph Neural Networks (GNNs) excel at capturing spatial-temporal dependencies in sensor data, their black-box nature and the limited work on graph-based explainable models for water networks hinder practical adoption. We propose an explainable GNN framework that integrates mutual information to identify critical network regions and fuzzy logic to provide clear, rule-based explanations for node classification tasks. After benchmarking several GNN architectures, we selected the generalized graph convolution network (GENConv) for its superior performance and developed a fuzzy-enhanced variant that offers intuitive explanations for classified leak locations. Our fuzzy graph neural network (FGENConv) achieved Graph F1 scores of 0.889 for detection and 0.814 for localization, slightly below the crisp GENConv 0.938 and 0.858, respectively. Yet it compensates by providing spatially localized, fuzzy rule-based explanations. By striking the right balance between precision and explainability, the proposed fuzzy network could enable hydraulic engineers to validate predicted leak locations, conserve human resources, and optimize maintenance strategies. The code is available at github.com/pasqualedem/GNNLeakDetection.

This paper presents a battery-free and gateway-free water leak detection system capable of direct communication over LTE-M (Cat-M1). The system operates solely on energy harvested through a hydroelectric mechanism driven by an electrochemical sensor, thereby removing the need for conventional batteries. To address the stringent startup and operational power demands of LTE-M transceivers, the architecture incorporates a compartmentalized sensing module and a dedicated power management subsystem, comprising a boost converter, supercapacitor based energy storage, and a hysteresis controlled load isolation circuit. This design enables autonomous, direct to cloud data transmission without reliance on local networking infrastructure. Experimental results demonstrate consistent LTE-M beacon transmissions triggered by water induced energy generation, underscoring the system's potential for sustainable, maintenance free, and globally scalable IoT leak detection applications in smart infrastructure.

This work presents a self powered water leak sensor that eliminates both batteries and local gateways. The design integrates a dual compartment electrochemical harvester, a low input boost converter with supercapacitor storage, and a comparator gated LTE-M radio built on the Nordic Thingy:91 platform. Laboratory tests confirm that the system can be awakened from a dormant state in the presence of water, harvest sufficient energy, and issue repeated cloud beacons using the water exposure as the power source. Beyond conventional LTE-M deployments, the system's compatibility with 3GPP standard cellular protocols paves the way for future connectivity via non terrestrial 5G networks, enabling coverage in infrastructure scarce regions.

This paper presents a battery-less, self-powered water leak detection system that utilizes LoRa communication for long-range, real-time monitoring. The system harvests hydroelectric energy through a layered stack of conductive nanomaterials and metals, achieving a peak short-circuit current of over 500 mA and 1.65 V open-circuit voltage upon exposure to water. To address LoRa's higher power demands, an energy management subsystem -- comprising a DC-DC boost converter and a 100 mF supercapacitor -- ensures stable power delivery for the LLCC68 LoRa module. Experimental results demonstrate the system's ability to detect leaks as shallow as 0.5 mm, activate within 50 seconds across varying water depths, and transmit data reliably over LoRaWAN. This solution eliminates battery dependency, offering a scalable, maintenance-free approach for industrial, commercial, and residential applications, while advancing sustainable IoT infrastructure.

Water distribution systems (WDS) carry potable water with millions of miles of pipelines and deliver purified water to residential areas. The incidents in the WDS cause leak and water loss, which imposes pressure gradient and public health crisis. Hence, utility managers need to assess the condition of pipelines periodically and localize the leak location (in case it is reported). In our previous works, we designed and developed a size-adaptable modular in-pipe robot [1] and controlled its motion in in-service WDS. However, due to the linearization of the dynamical equations of the robot, the stabilizer controller which is a linear quadratic regulator (LQR) cannot stabilize the large deviations of the stabilizing states due to the presence of obstacles that fails the robot during operation. To this aim, we design a self-rescue mechanism for the robot in which three auxiliary gear-motors retract and extend the arm modules with the designed controller towards a reliable motion in the negotiation of large obstacles and non-straight configurations. Simulation results show that the proposed mechanism along with the motion controller enables the robot to have an improved motion in pipelines.

In this article, we propose an approach to leak localisation in a complex water delivery grid with the use of data from physical simulation (e.g. EPANET software). This task is usually achieved by a network of multiple water pressure sensors and analysis of the so-called sensitivity matrix of pressure differences between the network's simulated data and actual data of the network affected by the leak. However, most algorithms using this approach require a significant number of pressure sensors -- a condition that is not easy to fulfil in the case of many less equipped networks. Therefore, we answer the question of whether leak localisation is possible by utilising very few sensors but having the ability to relocate one of them. Our algorithm is based on physical simulations (EPANET software) and an iterative scheme for mobile sensor relocation. The experiments show that the proposed system can equalise the low number of sensors with adjustments made for their positioning, giving a very good approximation of leak's position both in simulated cases and real-life example taken from BattLeDIM competition L-Town data.

In this paper, we present a nodal hydraulic head estimation methodology for water distribution networks (WDN) based on an Unscented Kalman Filter (UKF) scheme with application to leak localization. The UKF refines an initial estimation of the hydraulic state by considering the prediction model, as well as available pressure and demand measurements. To this end, it provides customized prediction and data assimilation steps. Additionally, the method is enhanced by dynamically updating the prediction function weight matrices. Performance testing on the Modena benchmark under realistic conditions demonstrates the method's effectiveness in enhancing state estimation and data-driven leak localization.

This article presents a leak localization methodology based on state estimation and learning. The first is handled by an interpolation scheme, whereas dictionary learning is considered for the second stage. The novel proposed interpolation technique exploits the physics of the interconnections between hydraulic heads of neighboring nodes in water distribution networks. Additionally, residuals are directly interpolated instead of hydraulic head values. The results of applying the proposed method to a well-known case study (Modena) demonstrated the improvements of the new interpolation method with respect to a state-of-the-art approach, both in terms of interpolation error (considering state and residual estimation) and posterior localization.

Disinfection byproducts are contaminants that can cause long-term effects on human health, occurring in chlorinated drinking water when the disinfectant interacts with natural organic matter. Their formation is affected by many environmental parameters, making it difficult to monitor and detect disinfection byproducts before they reach households. Due to the large variety of disinfection byproduct compounds that can be formed in water distribution networks, plus the constrained number of sensors that can be deployed throughout a system to monitor these contaminants, it is of outmost importance to place sensory equipment efficiently and optimally. In this paper, we present DBPFinder, a simulation software that assists in the strategic sensor placement for detecting disinfection byproducts, tested at a real-world water distribution network in Coimbra, Portugal. This simulator addresses multiple performance objectives at once in order to provide optimal solution placement recommendations to water utility operators based on their needs. A number of different experiments performed indicate its correctness, relevance, efficiency and scalability.

Optimal sensor placement is essential for state estimation and effective network monitoring. As known in the literature, this problem becomes particularly challenging in large-scale undirected or bidirected cyclic networks with parametric uncertainties, such as water distribution networks (WDNs), where pipe resistance and demand patterns are often unknown. Motivated by the challenges of cycles, parametric uncertainties, and scalability, this paper proposes a sensor placement algorithm that guarantees structural observability for cyclic and acyclic networks with parametric uncertainties. By leveraging a graph-based strategy, the proposed method efficiently addresses the computational complexities of large-scale networks. To demonstrate the algorithm's effectiveness, we apply it to several EPANET benchmark WDNs. Most notably, the developed algorithm solves the sensor placement problem with guaranteed structured observability for the L-town WDN with 1694 nodes and 124 cycles in under 0.1 seconds.

The area of fault detection is becoming more interesting since there have been many unique designs to detect or even compensate the faults, either from sensor or actuator. This paper applies the hydraulic system with interconnected tanks by implementing a leakage on one of the three tanks. The mathematical model along with the details of stability properties are highly discussed in this paper by imposing the Lyapunov and boundedness stability. The theory of fault detection with certain threshold after the occurrence of the fault corresponding to the state estimation error is mathematically presented ended by simulation. Moreover, the system compares the effectiveness of the proposed observer using Luenberger observer, adaptive-scaling Kalman and consensus filtering. The results for some different initial condition guarantee the detection of the fault for some time $t_d > t_f$