Prediction of residual stress, surface roughness, and grain refinement of 42CrMo steel subjected to shot peening by combining finite element method and artificial neural network

Shot peening (SP) is a widely used surface treatment technology of metallic materials. In order to investigate the effects of the shot velocity and the SP coverage on the surface integrity of the SPed materials, a numerical prediction framework combining the finite element method (FEM) with the artificial neural network (ANN) algorithm was proposed. A three-dimensional finite element model in conjunction with the dislocation density-based constitutive relation was developed to simulate the process of SP of 42CrMo steel. The FEM was validated by comparing the prediction results with the experimental data including the indentation profile produced by the single-shot impact and the in-depth residual stresses induced by the multiple-shot impacts. Based on the FEM simulation results, an attempt to predict the surface integrity of 42CrMo steel subjected to SP was made by taking advantage of the ANN algorithm, and the obtained results indicate that the predictions of the GA-BP-ANN algorithm (the back-propagation artificial neural network algorithm optimized by the genetic algorithm) are in good agreement with the FEM simulation results in terms of the SP-induced residual stresses, equivalent plastic strain, grain refinement, and surface roughness. This study therefore provides a new idea to predict the surface integrity of the metallic materials subjected to SP by combining the FEM simulation with ANN algorithm.