PROBABILISTIC OPTIMIZATION APPROACH FOR DAMAGE IDENTIFICATION USING FREQUENCY RESPONSE

This paper presents a novel probabilistic optimization approach to identify damage characteristics by using the frequency response function (FRF). The proposed approach has been developed to predict the probability of damage existence and to further identify salient details about damage location and damage severity in a probabilistic manner. The optimization problem has been developed as a function of measured and simulated frequency responses and is formulated in a multi-stage sequence to detect the probability of damage parameters including crack depth and crack location while minimizing uncertainties in the analysis outcomes. To demonstrate the proposed approach, a simply supported beam has been modeled with an open edge crack and characterized by using Linear Elastic Fracture Mechanics (LEFM). Several frequency responses obtained from the structure have been incorporated with different levels of noise to evaluate the robustness of the proposed algorithm. The algorithm has been tested through multiple simulations with various damage characteristics and different levels of noise. In all cases, the proposed algorithm has successfully predicted the presence of damage with a relatively high probability. Evaluation of the results demonstrates that the probabilistic optimization approach provides significant advantages over conventional deterministic methods for damage detection in structural health monitoring.