Research On Forming Prediction And Process Parameters Optimization System For Pipe Cold Bending

Pipe bending process is the pipe occurring plastic deformation in the multi-molds' co-operation,with lightweight,tough and low consumption and other advantages,which is widely used in aerospace and marine engineering and other high-tech fields.At present,the research on pipe bending results has the defects of strong experience,poor inheritance,and low efficiency.It is helpful to study the prediction of pipe bending results and optimization the process parameters,which can effectively guide the actual production of pipe bending.Many influencing factors of pipe bending process may have an impact on the forming results,which is not conducive to carry out significant analysis.In this paper,the influence factors were selected by correlation analysis,and the mandrel extension,the gap between the mandrel and the pipe,the relative pusher speed of the die and the bending time of the pipe bending were selected as the considering influencing factors.The orthogonal test design was designed by SPSS,and the corresponding wrinkle rate,flattening rate,and thinning rate were calculated by numerical simulation.The results of significance tests were analyzed by stepwise regression analysis.The training of MPSO-BP neural network is completed by a large number of training samples which were generated by numerical simulation,and the validity of the simulation is proved by comparing the prediction and numerical simulation results.According to the Topsis multi-objective optimization strategy,the process parameters of the extension and gap value are optimized and verified by the MPSO-BP neural network prediction and cell genetic algorithm.Finally,according to the above research contents,the humancomputer interaction system with the corresponding results of pipe bending prediction and optimization of process parameters is prepared based on the GUI platform of MATLAB.