Application of RBF Neural Network and Finite Element Analysis to Solve the Inverse Problem of Defect Identification

This paper presents an approach which is based on the use of radial basis function (RBF) neural network and finite element analysis to solve the inverse problem of defect identification. The approach is used to identify unknown defects in metallic walls. The methodology used in this study consists in the simulation of a large number of defects in a metallic wall, using the finite element method (FEM). Both variations in with and height of the defects are considered. Then the obtained results are used to generate a set of vectors for the training of a RBF neural network. Finally, the obtained neural networks are used to identify a group of new defects, simulated by the FEM, but not belonging to the original dataset. Performance of the RBF network was also compared with the most commonly used multilayer perceptron (MLP) network model. The reached results demonstrate the efficiency of the proposed approach, and that RBF network performs better than MLP network model