Efficient Cultural Algorithm for Structural Damage Detection Problem Based on Modal Data

Abstract

This paper aims to present an efficient cultural algorithm to solve the target optimization problem of vibration-based damage detection. The modal flexibility error residual is employed as objective function to be minimized. Cultural algorithms are inspired from the cultural evolutionary process in nature and use social intelligence to solve optimization problems. A cultural algorithm is composed of a belief space which consists of different knowledge sources, a population space and a set of communication protocols that enables interaction of these two spaces. Cultural algorithm offers powerful tools to solve various optimization problems as a result of its robustness as well as computation effectiveness. In this work, cultural algorithm is applied to generate solutions using three knowledge sources namely situational knowledge, normative knowledge, and domain knowledge. The core idea of using domain knowledge source is to speed up the convergence of the algorithm and thus, reducing its computational cost. The performance of the proposed algorithm is demonstrated through a numerical example, with different damage scenarios and noise levels. Comparison of the proposed algorithm with other basic and state-of-the-art algorithms reveals its superiority in accurately detecting the sites and the extents of structure damages in spite of contaminated vibration data by noise.