Crack identification in plates-type structures using natural frequencies coupled with successful history-based adaptive differential evolution algorithm

Abstract

In this study, a new approach, for identification and characterization of straight cracks in plates-like structures, is presented. The finite element method using a commercial software (Abaqus)is coupled with successful history-based adaptive differential evolution algorithm (SHADE) which, ensures the minimization of the objective function based on the mean relative error, that is defined as the difference between the measured (experimental) frequencies of a plate with an unknown crack identity and numerical frequencies of a cracked plate given by the approach Shade/FEM. This method will be applied on a steel thin plate to find the identity of the crack given by length, orientation and centre coordinates. Two strategies are applied to validate the effectiveness of the proposed approach. The first one, is based on the inverse problem using natural frequencies of a plate withknown crack identity obtained by a modal simulation on Abaqus. In the second, the experimental frequencies of a cracked plate were used. With just a population size of 25 and 150 iterations, the results show a good accuracy of the proposed approach with a relative error of the objective function less than 0.8%.