Performance evaluation of modified genetic and swarm-based optimization algorithms in damage identification problem

摘要

An experimental verification of a damage detection process using novel optimization techniques such as modifiednreal coded genetic algorithms and swarm-based algorithms is presented. Here, the objective function is defined asnthe sum of differences of the modal frequencies between intact and stiffness damaged state, which has to benminimized to identify the damage location and its severity in the process of model updating. In addition to thenstructural or damage variables such as the mass or stiffness of the numerical model, the profiles of modal frequencynshifts are also damage-sensitive features. The iterative process that uses the proposed population-based optimizationnalgorithms successfully identifies the local mass change of a test structure by updating the damage variables tonfit the modal data of test structures such as a cantilevered beam and multibay truss frame. Copyright © 2011 JohnnWiley & Sons, Ltd.