Structural Damage Detection by Fusion of GA and PSO

Abstract

In order to solve the inverse problem on structural damage detection (SDD) in the field of structural health monitoring (SHM), a FGAPSO algorithm is proposed by a fusion of the genetic algorithm (GA) and the particle swarm optimization (PSO) in this study. For improving the simple GA with drawbacks of easy precocious and of lower computation efficiency, the real-coded GA is implemented, the chaotic logistic mapping is chosen for initializing population, the self-adaptive crossover-mutation operators and elitist strategy are employed. The GA is then mixed with the PSO algorithm for the population diversity and convergence by exchanging genes between two new populations internally and the goal of improving GA is attained at last. Further, some numerical simulations on a 13-bar planar truss structure with several damage cases have been carried out for assessing the performance of the FGAPAO. The illustrated results show that the proposed FGAPSO algorithm is better than any of conventional GA and PSO. Even for the slight damage case, it is still more feasible and effective for SDD.