Optimization of a Pipelines Leak Detection Method Based on Inverse Transient Analysis Using a Genetic Algorithm

To date, there are few studies that have investigated the association between genetic algorithms and source localization problems. In this paper, we present a new optimization approach for leak detection in water pipelines while using advanced signal processing and statistical methods. This study is based on the work of Wang and Ghidaoui (Mech Syst Signal Process 107:529–548, 2018). The leak problem is formulated as an inverse source localization problem solved with an exhaustive search algorithm. It should be noted that with this type of application the computation time is a major drawback since the search space expands with the length of the pipeline. Thus, adopting a genetic algorithm (GA) may be necessary. The main goal of our approach based on GA is to reduce computation time and locate close leaks. Numerical results demonstrate that the proposed GA significantly outperforms exhaustive search in terms of execution time, while still maintaining excellent leakage estimation. Finally, a hybrid GA with the MFP (matched field processing) method has been proposed in order to limit the search space, which led to much more interesting results from a practical point of view.