Implementation of modal flexibility variation method and genetically trained ANNs in fault identification

The main objective of the present study is to develop a new two-phase procedure in order to localize the faults and corresponding severity in thin plate structures. Initially, the variation of modal flexibility and load-deflection differential equation of plate in conjunction with the invariant expression for the sum of transverse load are employed to formulate the damage indicator. Then an Artificial Neural Network (ANN) techniques and genetic algorithm are implemented to determine the corresponding damage severity. Genetic algorithm (GA) is used to automate the parameter selection process in artificial neural networks and eliminate the context dependent notion of the ANNs. The feasibility of the present Modal Flexibility Variation method (MFV) is verified through some numerical simulation and experimental tests on a steel plate. The results show that the performance of the proposed algorithm is quite encouraging and the maximum differences are less than three percent.