A Hybrid Search Technique for Inverse Transient Analysis in Water Distribution Systems

Leak detection and calibration of water distribution system (WDS) hydraulic models for pipe internal roughnesses and other parameters is, undoubtedly, an important issue. In this paper, an inverse transient model capable of detecting leaks in pipe networks is presented. First, a brief review of existing leak detection approaches is given. The review is followed by the presentation of a inverse transient model which is formulated as an optimisation problem, with the objective function in the form of weighted least squares and a set of relevant constraints. After identifying advantages and disadvantages of both, Genetic Algorithms (GA) and the Levenberg-Marquardt (LM) search method, a hybrid genetic algorithm was developed to solve the inverse transient problem. Finally, a summary is made and relevant conclusions are drawn.