Potential of Two Metaheuristic Optimization Tools for Damage Localization in Civil Structures

Abstract

In structural health monitoring, the presence of damage is detected and localized by outlining the differences between the initial state and current behavior of a given structure. The problem is often formulated as an optimization problem. In this paper, a highly nonlinear objective function that minimizes the discrepancies between the analytical and experimental features of a structure is introduced. Within a finite-element discretization, some stiffness parameters are chosen as reference variables. Two metaheuristic tools, the artificial bee colony (ABC) algorithm and the firefly algorithm (FA), are applied to proceed the iterations toward the global minima of the objective function. By comparing the identified and analytical stiffness matrices, the damage detection and localization are performed. These methods are applied to a steel structure. The efficiency of the two tools is compared.