Applicability of a Fuzzy Genetic System for Crack Diagnosis in Timoshenko Beams

Abstract

The suitability of a new fuzzy genetic fused system is investigated in this study for damage identification in Timoshenko beam-type structures. Damage scenarios are created by introducing cracks of varying depth and locations. Both single and multiple crack cases are considered. The inverse problem of determining crack location and extent is carried out using the hybrid system. The hybridization is done in order to use the real power of fuzzy logic approach where tuning of the membership functions is automated by utilizing a genetic algorithm. Both global parameters (natural frequencies) and local parameters (maxima of rotational mode shape deviations) are used as input for the integrated system. The efficiency of developed hybrid models is evaluated through several error statistics. The error metrics show that the estimates of the hybrid models are in good agreement with theoretical results. The errors also show that the optimal models perform rather accurately in a noise ratio expected in practice. Consequently, the fused system can be used as nondestructive procedure for health monitoring of beam-type structures.