Application of Neural Network and FEM to Optimize Load Path of T-shape Tube Hydro Forming

During tube hydroforming (THF) process, failure modes such as buckling, necking and bursting may occur because axial feeding and internal pressure are imposed simultaneously. As load path has great influence on THF, prediction of properties of final product is difficult and a time consuming work. In this study, Neural Network algorithm and ANSYS LS-DYNA and ANSYS Program Language Design (APDL), was used to predict final product properties of T- shapebranch workpiece. FE model and Neural Network were verified using experimental result for a determined load path. Finally direct search method has been used to obtain optimum load path for higher formability.

کلیدواژه ها:

Tube hydroforming ، Bursting failure ، Load path ، Neural Network ، Direct search pattern