Surrogate-Assisted Inverse Transient Analysis (SAITA) for Leakage Detection in Pressurized Piping Systems

The inverse transient analysis (ITA) is a well-known leak detection method for pressurized piping systems. In this method, unknown parameters of the number, location, and size of leakages are obtained through solving a nonlinear programming problem using an optimization technique. Both transient analysis and optimization are time consuming and make the whole process of ITA computationally inefficient. In this study to increase the efficiency of the ITA, the methods of Surrogate Assisted Genetic Algorithm-Local Search (SAGA-LS), SAGA-Pre Selection, and SAGA-Clustering Technique are investigated. The methods are applied to a benchmark example and the results are compared to the standard ITA. The results indicate that the ITA-SAGA-LS has the best performance with 92% increase in the accuracy of the calculated leakage area as compared to the standard ITA.