Water Pipeline Leakage Detection Based on Machine Learning and Wireless Sensor Networks

Liu, Yang; Ma, Xuehui; Li, Yuting; Tie, Yong; Zhang, Yinghui; Gao, Jing

doi:10.3390/s19235086

Cited by 89 publications

(45 citation statements)

References 49 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…As examples of use of such models, [1] presents a stochastic model for failure forecasting of water mains using the collected data, and [2] provides various intelligent approaches for the prediction. Also, [32] presents a WSNbased monitoring scheme based on a leak identification method for water pipelines that relies on the combined use of Principal Components Analysis (PCA) and SVM. The highest identification accuracy achieved by this model was around 98%.…”

Section: Related Workmentioning

confidence: 99%

Leak Detection Using Flow-Induced Vibrations in Pressurized Wall-Mounted Water Pipelines

et al. 2020

View full text Add to dashboard Cite

Wireless sensor networks (WSN) provide a powerful solution to the task of monitoring the operational conditions of buried and non-buried pipes of different lengths and materials. Due to the limited energy stored in the sensor nodes, the use of low-power vibration sensors becomes the preferred choice. However, the monitoring of vibrations for leak detection in wall-mounted pipelines, and the associated complexities are not adequately dealt with in the literature. This paper offers to fill this gap by presenting a feasibility study of leak detection in wall-mounted water pipelines through vibrations measurements using low-power accelerometers. The work is divided into two steps: Firstly, a careful analysis is performed to understand the effect of various fittings such as clamps, bends, and leaks of various sizes, on the vibrations produced. Then this knowledge is used to find the best locations for placement of nodes in order to efficiently detect leaks of various sizes. This analysis revealed two important facts: (a) difficulty in detecting mediumsize leaks as their vibrations and those from the no-leak condition are very indistinguishable, (b) vibrations measured away from the leak are of a small benefit to the leak detection process. Consequently, 3 different learning models are applied, all fed with information from multiple nodes, in order to reliably detect leaks and classify their size. Comparing the performances of these models shows that the Support Vector Machine (SVM)-based model gives the best results, in that for the worst case of medium-size leaks and with the use of one sensor, the worst accuracies for leak detection and leak size classification have remarkably been improved from being respectively 51% and 36% with one sensor, to being 88% and 93%, respectively, with only a moderate increase in the number of sensors to four.

show abstract

Section: Related Workmentioning

confidence: 99%

Leak Detection Using Flow-Induced Vibrations in Pressurized Wall-Mounted Water Pipelines

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Regardless of the technique used, the amount of data generated by water distribution monitoring systems is large. As an example, experiments conducted by Liu et al showed that 100 datasets of size 5000 samples can be generated in 1200s [8]. This is equivalent to 1.5 million samples/hour for a monitoring network consisting of around 600 nodes [8].…”

Section: Introductionmentioning

confidence: 99%

“…As an example, experiments conducted by Liu et al showed that 100 datasets of size 5000 samples can be generated in 1200s [8]. This is equivalent to 1.5 million samples/hour for a monitoring network consisting of around 600 nodes [8]. To address this issue, multiple potential solutions can be adopted.…”

Section: Introductionmentioning

confidence: 99%

Water Leak Detection Survey: Challenges & Research Opportunities Using Data Fusion & Federated Learning

et al. 2021

View full text Add to dashboard Cite

With the increase in pipeline usage for fluid transportation, leak detection has become a major concern. More specifically, detecting water leaks has become a pressing challenge to both governmental and industrial stakeholders due to the financial losses it causes as well as the safety concerns associated with it. This issue is further highlighted in industrial and manufacturing environments such as the steelmaking process in which a water leak into a furnace can cause a significant explosion that would threaten both the facility and its operators. Therefore, many different water leak detection methods belonging to different types (hardware-in-the-loop-based, simulation-in-the-loop-based, or hybrid) have been proposed in the literature. However, many of these methods either are computationally complex or only suitable for particular applications. Hence, there is a need to develop innovative and novel frameworks that offer effective and efficient water leak detection mechanisms. To that end, this paper discusses two different paradigms, namely sensor data fusion and federated learning, that have the potential to further enhance water leak detection methods. Therefore, this paper first surveys the different water leak detection methods proposed in the literature along with their merits and limitations. It then describes the sensor data fusion and federated learning paradigms in more detail. Moreover, it presents different research opportunities in which these paradigms can be implemented to offer a more effective and computationally efficient water leak detection framework.

show abstract

“…Multi-sensor data fusion combines redundant data from multiple low-cost sensors to achieve a more accurate information whose quality exceeds that achieved by using a single sensor [ 36 , 38 ]. In addition, the low-power consumption requirement and the need for a WWPM to go unattended for a long period of time without any replacement of the sensor node’s battery [ 33 , 39 , 40 ], affects the choice of multi-sensor data fusion technique that can be used. Multi-sensor data fusion in WSN can either be done via a centralized, decentralized or in a distributed manner [ 37 , 41 ].…”

Section: Introductionmentioning

confidence: 99%

A Distributed Computing Solution Based on Distributed Kalman Filter for Leak Detection in WSN-Based Water Pipeline Monitoring

Nkemeni

Mieyeville

Tsafack

2020

Sensors

View full text Add to dashboard Cite

Wireless Sensor Network (WSN) applications that favor more local computations and less communication can contribute to solving the problem of high power consumption and performance issues plaguing most centralized WSN applications. In this study, we present a fully distributed solution, where leaks are detected in a water distribution network via only local collaborations between a sensor node and its close neighbors, without the need for long-distance transmissions via several hops to a centralized fusion center. A complete approach that includes the design, simulation, and physical measurements, showing how distributed computing implemented via a distributed Kalman filter improves the accuracy of leak detection and the power consumption is presented. The results from the physical implementation show that distributed data fusion increases the accuracy of leak detection while preserving WSN lifetime.

show abstract

Water Pipeline Leakage Detection Based on Machine Learning and Wireless Sensor Networks

Cited by 89 publications

References 49 publications

Leak Detection Using Flow-Induced Vibrations in Pressurized Wall-Mounted Water Pipelines

Leak Detection Using Flow-Induced Vibrations in Pressurized Wall-Mounted Water Pipelines

Water Leak Detection Survey: Challenges & Research Opportunities Using Data Fusion & Federated Learning

A Distributed Computing Solution Based on Distributed Kalman Filter for Leak Detection in WSN-Based Water Pipeline Monitoring

Contact Info

Product

Resources

About