Weighted Least Squares with Expectation-Maximization Algorithm for Burst Detection in U.K. Water Distribution Systems

In this paper, a fusion of unsupervised clustering and incremental similarity tracking of hourly water demand series is proposed. Current research using unsupervised methodologies to detect anomalous water is limited and may possess several limitations such as a large amount of dataset, the need to select an optimal cluster number, or low detection accuracy. Our proposed approach aims to address the need for a large amount of dataset by detecting anomaly through (1) clustering points that are relatively similar at each time step, (2) clustering points at each time step by the similarity in how they vary from each time step, and (3) to compare the incoming points with a reference shape for online anomalous trend detection. Secondly, through the use of Bayesian nonparametric approach such as the Dirichlet Process Mixture Model, the need to choose an optimal cluster number is eliminated and provides a subtle solution for ‘reserving’ an empty cluster for the future anomaly. Among the 165 randomly generated anomalies, the proposed approach detected a total of 159 anomalies and other anomalous trends present in the data. As the data is unlabeled, identified anomalous trends cannot be verified. However, results show great potential in using minimally unlabeled water demand data for a preliminary anomaly detection.

Section: Water Demand Data Descriptionmentioning

confidence: 89%

Section: Data Preparationmentioning

confidence: 99%

Unsupervised Bayesian Nonparametric Approach with Incremental Similarity Tracking of Unlabeled Water Demand Time Series for Anomaly Detection

Chan

Chin

2019

“…Various data-driven methods have been proposed for WDS pipe burst detection: artificial intelligence [19,20], state estimation [16,21,22], the Bayesian approach [23], classification [24,25], and SPC [15][16][17][26][27][28]. Wu and Liu [29] recently reviewed and classified data-driven approaches; please refer to them for more details of each method.…”

Section: Pipe Burst Detection Method: Western Electric Company (Wec) mentioning

confidence: 99%

Robust Meter Network for Water Distribution Pipe Burst Detection

Jung

Kim

2017

Abstract:A meter network is a set of meters installed throughout a water distribution system to measure system variables, such as the pipe flow rate and pressure. In the current hyper-connected world, meter networks are being exposed to meter failure conditions, such as malfunction of the meter's physical system and communication system failure. Therefore, a meter network's robustness should be secured for reliable provision of informative meter data. This paper introduces a multi-objective optimal meter placement model that maximizes the detection probability, minimizes false alarms, and maximizes the robustness of a meter network given a predefined number of meters. A meter network's robustness is defined as its ability to consistently provide quality data in the event of meter failure. Based on a single-meter failure simulation, a robustness indicator for the meter network is introduced and maximized as the third objective of the proposed model. The proposed model was applied to the Austin network to determine the independent placement of pipe flow and pressure meters with three or five available meters. The results showed that the proposed model is a useful tool for determining meter locations to secure high detectability and robustness.

“…Most of these methods were tested and validated on DMA (district metering areas) under real circumstances [4]. Related studies about DMA burst detection primarily involve a single-inlet flow meter [5][6][7][8][9][10][11][12][13][14]. Mounce et al (2002) presented a neural network methodology to detect bursts [5], and Mounce et al (2006) proposed static and time-delay artificial neural networks (ANNs) to detect bursts.…”

Section: Introductionmentioning

confidence: 99%

“…Besides, Ye and Fenner (2014) also proposed a methodology to use polynomial functions based on the weighted least squares method, with expectation maximization (EM) to predict normal flow or pressure values. When bursts occur in DMAs, observed data are significantly different from predictive values because predictions depend on normal historical data [10]. However, the above methods only determine whether or not burst occurs in DMA based on the predicted residual at a single time-point.…”

Section: Introductionmentioning

confidence: 99%

Real-Time Burst Detection in District Metering Areas in Water Distribution System Based on Patterns of Water Demand with Supervised Learning

Huang

Zhu

Hou

et al. 2018

This paper proposes a new method to detect bursts in District Metering Areas (DMAs) in water distribution systems. The methodology is divided into three steps. Firstly, Dynamic Time Warping was applied to study the similarity of daily water demand, extract different patterns of water demand, and remove abnormal patterns. In the second stage, according to different water demand patterns, a supervised learning algorithm was adopted for burst detection, which established a leakage identification model for each period of time, respectively, using a sliding time window. Finally, the detection process was performed by calculating the abnormal probability of flow during a certain period by the model and identifying whether a burst occurred according to the set threshold. The method was validated on a case study involving a DMA with engineered pipe-burst events. The results obtained demonstrate that the proposed method can effectively detect bursts, with a low false-alarm rate and high accuracy.