Analysis of fracture mechanics via finite elements and artificial intelligence

In this thesis, a finite element analysis has been carried out for accurate calculation and evaluation of the Stress Intensity Factor (SIF) in Linear Elastic Fracture Mechanics (LEFM) and J-integral in elasto-plastic Fracture Mechanics (EPFM). The direct relations between the parameters and some guide lines for future works have been investigated. This study also presents the influence of mesh refinement on the SIF and J-integral values of Mixed Mode (Mode I and Mode II) loading. Furthermore, the prediction of crack path has been performed numerically using maximum circumferential stress criterion. Finally, explicit formulations have been obtained in the evaluation of SIF and J-integral parameters using Neural Networks (NN?s) and Genetic Programming (GP). Keywords: Stress Intensity Factor, J-integral, Crack Path Prediction, Displacement Extrapolation Method, Finite Element Method, Artificial Intelligence, Neural Networks and Genetic Programming.