Pipe Burst Localization in Water Distribution Systems Based on Artificial Neural Network

The current paper proposes a novel methodology for near–real time burst location and sizing in water distribution systems (WDS) by means of Multi–Layer Perceptron (MLP), a class of artificial neural network (ANN). The proposed methodology can be systematized in four steps: (1) construction of the pipe–burst database, (2) problem formulation and ANN architecture definition, (3) ANN training, testing and sensitivity analyses, (4) application based on collected data. A large database needs to be constructed using 24 h pressure–head data collected or numerically generated at different sensor locations during the pipe burst occurrence. The ANN is trained and tested in a real–life network, in Portugal, using artificial data generated by hydraulic extended period simulations. The trained ANN has demonstrated to successfully locate 60–70% of the burst with an accuracy of 100 m and 98% of the burst with an accuracy of 500 m and to determine burst sizes with uncertainties lower than 2 L/s in 90% of tested cases and lower than 0.2 L/s in 70% of the cases. This approach can be used as a daily management tool of water distribution networks (WDN), as long as the ANN is trained with artificial data generated by an accurate and calibrated WDS hydraulic models and/or with reliable pressure–head data collected at different locations of the WDS during the pipe burst occurrence.

Download Full-text

Application of Artificial Neural Network Model Based on Improved PSO in Water Supply Systems

Advanced Materials Research ◽

10.4028/www.scientific.net/amr.267.609 ◽

2011 ◽

Vol 267 ◽

pp. 609-613

Author(s):

Hong Xiang Wang ◽

Wen Xian Guo

Keyword(s):

Neural Network ◽

Artificial Neural Network ◽

Distribution Systems ◽

Water Distribution ◽

Microscopic Model ◽

Macroscopic Model ◽

Control Element ◽

Calibration Data ◽

Artificial Neural ◽

Improved Pso

Parameter calibration, data collection and simulation to control element were used to improve the accuracy of microscopic model. In order to overcome the shortage of macroscopic model, theoretical and empirical equation was adopted. The artificial neural network based on PSO method was introduced to improve simulation ability of water distribution system model from microscopic model and macroscopic model. There are two hidden layers with a maximum of 64 nodes per layer in the model. The Particle Swarm Optimization (PSO) algorithm is implemented to optimize the node numbers of the hidden layers in the model. The study indicates that the artificial neural network connecting with improved PSO method is an attractive alternative to the conventional regression analysis method in modeling water distribution systems.

Download Full-text

Development of a Cyberattack Detection Model for a Water Distribution System using Water Quality and Hydraulic Criteria

Korean Society of Hazard Mitigation ◽

10.9798/kosham.2020.20.5.47 ◽

2020 ◽

Vol 20 (5) ◽

pp. 47-56

Author(s):

Kyoung Won Min ◽

Young Hwan Choi ◽

Joong Hoon Kim

Keyword(s):

Neural Network ◽

Water Quality ◽

Artificial Neural Network ◽

Distribution Systems ◽

Water Distribution ◽

Detection Performance ◽

Economic Losses ◽

Water Quality Criteria ◽

Quality Factors ◽

Artificial Neural

In recent years, Cyber-Physical Systems (CPSs) have been applied to Water Distribution Systems (WDSs) to facilitate efficient operation and maintenance. Since data are transmitted through the network in such systems, a cyberattack can disrupt the operation of WDSs, for example, by causing water supply reduction, water pollution, and economic losses. In the past few years, cyberattack detection algorithms and various cyberattack scenarios have been proposed. These studies considered either hydraulic factors, such as pipe velocity, nodal pressure, or tank level, or water quality factors. However, an algorithm which considers only one factor cannot prevent the various problems that may arise, such as water quality issues, and the hydraulic and quality factors have a correlation. Therefore, in this study, a framework was developed by considering both hydraulic and water quality factors. The proposed approach was applied to an artificial neural network model. Performance indicators were used to examine the detection performance according to the parameters of the artificial neural network. By comparing the detection performance when only hydraulic factors were considered and the performance when both hydraulic and water quality factors were considered, the effectiveness of the algorithm that consider both hydraulic and water quality factors was demonstrated. A cyberattack detection algorithm that considers both hydraulic and water quality criteria can be applicable in more realistic scenarios and contribute to the establishment of safe infrastructure for the entire process of designing and operating WDSs with CPSs.

Download Full-text