Improving defects search in structure by boundary element and genetic algorithm scan method

Abstract

An inverse technique has been used to estimate the geometry for defects in elastic structures and the elastic properties of the structure from additional information such as experimental data of displacements, strain and velocity of the structure subjected to any mechanical loading or constraint. In the present paper, a newly developed inverse technique using a genetic algorithm (GA) is presented for identifying several circular and elliptical defects in a square plate subjected to a uni-directional load with various loading directions. In the present paper, the position, size and number of internal defects are unknown and are estimated by using displacement data at the outer surface of the plate. The influence of some parameters, mutation ratio and method, in GA upon the final results of inverse analysis is investigated. An inverse technique using GA leads to several candidates of solution for each loading direction owing to different sensitivities of deformation. A new method referred to as a GA scan method is presented to reconstruct defect image from solution candidates.