Burst Detection In Water Distribution Systems Based On Artificial Immune System

HaiDong Huang,Tao Tao
DOI: https://doi.org/10.1201/b14013-8
2013-01-01
Abstract:A novel burst detection method is presented in this work based on Artificial Immune System (AIS). Data corresponding to normal operation are trained through application of clonal immune algorithm (CSA) to generate a detector set which is used for monitoring the system. The state of a water distribution network is then determined in terms of Euclidean distance between a new data sample and the corresponding detector. Monte Carlo Simulation (MCS) is here applied to create artificial burst data. Then the artificial burst data are used to generate another detector sets which are used for burst identification. According to Euclidean distance between a new data sample and the corresponding detector, possible burst locations are identified. A case study with three real burst events is illustrated to evaluate the proposed method. The results obtained from the case study show that the method has the potential to be useful tool for burst detection.
What problem does this paper attempt to address?