A genetic algorithm with tabu search method for 3 dimensional detect profile reconstruction using MFL measurement

In nondestructive testing, magnetic flux leakage (MFL) inspection is extensively employed for the inversion of pipeline defects. Exact reconstruction of the defect plane with measurements is a pressing issue in the field of MFL detection. This article proposes a method that incorporates a layer-by-layer genetic algorithm with tabu search (GA-TS). During the iterative process, our objective is to utilize a tabu search based on evolutionary algorithms to avoid local optimal solutions, obtain global optimal solutions, and reconstruct defects with enhanced accuracy. In addition to enhancing the accuracy of pipeline defect categorization, our approach has two significant limitations: its limited global search capability and slow convergence rate. Finally, the method is evaluated via experiments in which MFL signals obtained from an experimental platform and simulation signals are used. Practical results and a thorough comparative study of alternative approaches confirm the model's superiority.