This study focuses on data-driven approaches for burst detection and classifies them into three categories: classification method, prediction-classification method and statistical method. The performance of these methods is discussed. By analysing uncertainty in burst detection, this paper revealed that non-stationary monitoring data and limitations present in these methods challenge the reliability of detection results. Data pre-processing and probabilistic solutions to deal with the uncertainty are summarised. From these findings and discussions, this paper concludes and recommends that: a) data-driven approaches are promising in real-life burst detection and reducing false alarms is an important issue; b) more comprehensive performance evaluation might be necessary, in particular regarding detectable burst size; c) further research on new methods employing multivariate analysis and a new category based on clustering analysis would be beneficial to tackle uncertainty; d) more focus on the use of...

A review of data-driven approaches for burst detection in water distribution systems

This paper presents a new method for identifying leakage zones of water distribution systems. A large water network is first divided into a number of zones. The zone number is used as the category label of the multiclass support vector machine (M-SVM), which is trained with the data set generated by simulation of the possible leakages using a hydraulic model. The trained M-SVM is used as the leakage zone identification model and applied to determine the likely leakage zones with the observed field data. Two case studies are presented in this paper to demonstrate the effectiveness of the method. The results indicate that this method has many unique advantages in solving the nonlinear and high-dimensional pattern recognition problem with a small sample data set. Together with the method of pressure-dependent leakage detection (PDLD), the proposed approach enables engineers to improve the effectiveness and efficiency of leakage detection for large water distribution systems.

Leakage Zone Identification in Large-Scale Water Distribution Systems Using Multiclass Support Vector Machines

Water is a precious resource that should be managed carefully. However, due to leakages in water distributed networks (WDNs), a large amount of water is lost each year that suggests the need for reliable and robust leak detection and localization system. This paper attempts to review the current technologies for leakage detection in WDN as well as several proposed intelligent methodologies (such as support vector machine, neural network, and convolution neural network) over the past few years. The current methodologies and their limitations are discussed. Uncertainties involved in the implementation of WDN leakage detection are also discussed, and several suggestions to overcome such uncertainties are provided for future implementations.

Review of Current Technologies and Proposed Intelligent Methodologies for Water Distributed Network Leakage Detection

Burst detection in district metering areas using a data driven clustering algorithm.

Transient wave-based methods for anomaly detection in fluid pipes: A review

The problem of bursts and leakages in water distribution systems has received significantly increased attention over the past two decades. As they represent both an environmental and an economical issue, how to reduce water loss through bursts and leakages is a challenging task for water utilities. Consequently, various techniques have been developed to detect the location and size of leakages. The methods for bursts (or leaks) detection and location can be broadly divided into two main categories, one based on hardware and the other based on software. Hardware-based methods include (i) acoustic detection methods such as listening rods, leak correlators, leak noise loggers and (ii) non-acoustic detection methods such as gas injection, ground penetrating radar technology and infrared photography. Software-based methods make use of the data collected by real-time pressure and/or flow sensors and several artificial intelligence techniques and statistical data analysis tools, including (i) methods based on numerical modeling methods, such as inverse transient analysis, time domain analysis and frequency domain analysis, and (ii) some non-numerical modeling methods, such as artificial neural networks, Bayesian inference systems, the Golden section method, and Kalman filtering. In this article, the authors describe the methods for pipe network burst location and detection, summarize the features of each method, and propose a suggestion for future work.

A review of methods for burst/leakage detection and location in water distribution systems

A new method using artificial immune network is presented to identify pipe burst in water-distribution systems. Burst detection is considered as the problem of pattern recognition in the proposed method. An artificial database that includes information on burst events (BEs) is first established. Using the clonal selection algorithm, the artificial immune network is constructed based on the principle of immune system. The burst location is finally identified using the nearest neighbor method. Three offline case studies are illustrated in detail to evaluate the current method. A total of five possible burst locations are identified from 34 nodes in Case Study 1, whereas four possible burst locations are identified from 77 nodes in Case Study 2. The results derived from the first two case studies show that the method can identify the possible burst areas, including the true burst location, using model-simulated results. The data derived from real BEs in Case Study 3 are used to evaluate the proposed ...

Burst Detection Using an Artificial Immune Network in Water-Distribution Systems

A new methodology was proposed for contamination source identification using information provided by consumer complaints from a probabilistic view. Due to the high uncertainties of information derived from users, the objective of the proposed methodology doesn’t aim to capture a unique solution, but to minimize the number of possible contamination sources. In the proposed methodology, all the possible pollution nodes are identified through the CSA methodology firstly. And then based on the principle of total probability formula, the probability of each possible contamination node is obtained through a series of calculation. According to magnitude of the probability, the number of possible pollution nodes is minimized. The effectiveness and feasibility of the methodology is demonstrated through an application to a real case of ZJ City. Four scenarios were designed to investigate the influence of different uncertainties on the results in this case. The results show that pollutant concentration, injection duration, the number of consumer complaints nodes used for calculation and the prior probability with which consumers would complaint have no particular effect on the identification of contamination source. Three nodes were selected as the most possible pollution sources in water pipe network of ZJ City which includes more than 3 000 nodes. The results show the potential of the proposed method to identify contamination source through consumer complaints.

Identification of contamination source in water distribution network based on consumer complaints

An improved method is proposed to obtain the optimal placement of pressure monitoring station which aims to detect leakage or bursts in pipe networks. Self-organizing map network (SOM) is used as a tool for clustering due to its strong capability to classify the inputs, without knowing the number of cluster in advance. In addition, the effective detected domain initially used in the placement of flow monitoring stations is introduced herein to make sure that pressure monitoring points can detect bursts effectively. A real world network including 77 nodes and 108 pipes is used to demonstrate the effectiveness of the proposed method. 9 nodes are chosen as the locations of pressure monitoring points through comparative analysis. The results show that it can be easy and quick to obtain the optimal placement of pressure monitoring stations through application of the proposed method.

Optimal pressure meters placement for bursts detection based on SOM

The invention relates to a relevant control technology for tube bursting accidents of a water supply tube network and provides a tube bursting positioning method based on an artificial immunity system. The artificial immunity system is constructed so as to recognize and position a possible tube bursting position, so that the possible tube bursting occurrence area is greatly reduced, and finally the time for accurately positioning the tube bursting position is reduced. The tube bursting positioning method comprises the following steps: firstly, acquiring adequate tube bursting operation data by virtue of simulating the tube bursting of the water supply tube network; then according to the correlation theory of the artificial immunity system, training the operation data of burst tubes and constructing one artificial immunity system; finally recognizing the operation data of new tube bursting accidents according to the artificial immunity system and finally determining possible tube bursting occurrence positions according to a nearest neighbor classification method. Compared with a same-type tube bursting detection technology, the tube bursting positioning method provided by the invention can be used for achieving higher computation efficiency and obtaining more accurate possible tube bursting region.

Haidong Huang

Papers

A review of methods for burst/leakage detection and location in water distribution systems

Burst Detection Using an Artificial Immune Network in Water-Distribution Systems

Identification of contamination source in water distribution network based on consumer complaints

Optimal pressure meters placement for bursts detection based on SOM

Tube bursting positioning method based on artificial immunity system