Damage Detection of Truss Structures using an Improved Charged System Search Algorithm

In this paper, detection and assessment of structural damage using the changes in a structure's natural frequencies is addressed as an optimization problem. The damaged element(s) and the percentage of damage are considered as the problem variables. The objective is to set these variables such that the natural frequencies of the model correspond to the experimentally measured frequencies of the actual damaged structure. This is a problem with several global optimal solutions each representing a probable state of damage. Obviously, unlike many other optimization problems, it is not sufficient to find one of these optimal solutions; it is important to find all such possible states and to compare them. On the other hand, meta-heuristic optimization algorithms tend to converge to a single solution in each run. These algorithms do not generally use any optimality criterion and evaluate the quality of solutions only by direct comparison. Experimental results show that some of the optimal solutions have a greater probability of being found by the algorithms than others. In fact the algorithm agents neglect some of the promising regions of the search space to the benefit of some others. In this paper, the charged system search algorithm is improved and utilized to tackle the problem of finding as many global optimal solutions as possible in a single run.