A VIBRATION BASED INVERSE HYBRID INTELLIGENT METHOD FOR STRUCTURAL HEALTH MONITORING

In this paper, a novel identification algorithm (hybrid intelligent system) using inverse analysis of the vibration response of a cracked cantilever beam has been proposed. The crack identification algorithm utilizes the vibration signatures of the cracked beam derived from finite element and theoretical analysis. The hybrid controller is designed to predict the crack locations and their severities by integrating the capabilities of fuzzy logic and neural network technique. The measured modal parameters along with the outputs from the fuzzy controller are the inputs to the neural segment of the hybrid system, where as final relative crack depths and final relative crack locations are the output parameters. The derived vibration parameters are used to establish series of fuzzy rules and training patterns for the fuzzy and neural controller. Finally, the reliability of the proposed crack identification algorithm is established by comparing the results obtained from the experimental analysis.