To detect and locate the leakage of the pipe correctly, genetic algorithm is combined with Bayesian theory to determine the leaked pipes. Leakage detection and leakage location are carried out separately. Leakage detection is conducted based on the assumption that there is only one leaked pipe, and the simulation result demonstrates its feasibility. When the leakage detection demonstrates there is leaked pipe in the water distribution system, leakage location starts. Based on the information gathered by the manometers, leakage probability in different combinations of the virtual nodal demand can be fixed according to calculating the pressure of the monitored node, then GA is applied to search the maximum Bayesian value, the pipes with maximum Bayesian leakage possibility are believed to be leaked pipes. Optimization programme was made with combination of Matlab and Epanet, numerical simulation results demonstrate the feasibility and effectiveness of the proposed method.
Leakage Detection of Water Distribution System using Adaptive Kalman Filter
