Improvement of the Computational Efficiency of Metaheuristic Algorithms for the Crack Detection of Cantilever Beams Using Hybrid Methods

This study examines the capability of various optimization algorithms and proposes novel hybrid algorithms for more precise prediction of open-edge cracks in cantilever beams. The natural frequencies of the beam with a crack are obtained by modal analysis and experimentally validated by impact testing. The performance of Harris hawk optimization (HHO), electrostatic discharge algorithm (ESDA), pathfinder algorithm (PFA) and Henry gas solubility optimization (HGSO) algorithms from the literature is evaluated to determine the location and depth of an open-edge crack for an Euler-Bernoulli beam. Then, hybrid algorithms (HHO-NM, ESDA-NM and PF-NM) are proposed to improve the results of the aforementioned algorithms. Simulation results show that the proposed hybrid algorithms yield much more precise results with fewer function evaluations than the previously introduced algorithms and, therefore, have superior crack detection capability. Statistical post hoc analysis shows that the proposed hybrid algorithm can be considered a high-performance algorithm, which can significantly improve the efficiency of crack detection applications.