Prediction of Stress Distribution in Plastic Deformation of Bending Plate Member Using Machine Learning

Abstract

We establish procedure to predict stress distribution by machine learning focusing on the plastic deformation of bending plate in press forming. Since the press forming is nonlinear elasto-plastic behavior, implicit or explicit methods generally use incremental analysis. On the other hand, many data are indispensable for building the machine learning model. Therefore, we adopt the one-step method, which can significantly reduce the calculation cost of the finite element analysis. We validate prediction of stress distribution due to plastic deformation in the plate by machine learning by comparing the finite element analysis results.