The Potential of the Firefly Algorithm for Damage Localization and Stiffness Identification

Abstract

The identification of those differences between the current behavior and the initial state of a structure which are indicative of the presence of damage is one of the aims of structural health monitoring. Since the last decades, considerable research advances have been conducted in the optimization field. In this paper, an objective function that minimizes the discrepancies between the analytical and the experimental modal features obtained from the measurements of the actual dynamic response of a structure is formulated. Once the stiffness parameters are set as design variables, the firefly algorithm is applied to carry out the iterations toward the global minima. Partial solutions are analyzed along different steps of the procedure and identified as local optima by calculating the new stiffness matrices and estimating the corresponding values of the objective function. Eventually, the damage detection and localization are pursued by the comparison between the stiffness matrix identified once the optimization process is finished and the starting one. This procedure is applied to a numerical example, which is representative of a generic structure meshed into finite elements where damage is introduced as a local stiffness reduction.