A hybrid back-propagation neural network and intelligent algorithm combined algorithm for optimizing microcellular foaming injection molding process parameters

Abstract

The microcellular foaming injection molding process can effectively reduce the product weight and facilitate the realization of automobile lightweight, which is greatly affected by process parameters. Warpage of microcellular foaming material will increase, affecting the dimensional stability of the product, unless the process parameters are effectively controlled by process parameters. Warpage of microcellular foaming material has been based on key process variables including mold temperature, melt temperature, coolant temperature, coolant Reynolds number, v/p switch over, initial foaming volume, initial bubble radius and initial gas concentration. In order to reduce warpage of microcellular foaming material, the input and output data obtained from the Finite Element (FE) simulations are used to train BP neural network, GABP neural network and PSOBP neural network as prediction model for the warpage. Comparing the performance of the three methods in prediction error and training time, PSOBP is considered as the best prediction model for the warpage of microcellular foaming material. It has been proved that the prediction model has the ability to predict the warpage of the plastic within an error range of 1 %. The prediction model can be optimized by genetic algorithm to find the best combination of process parameters. The optimized warpage value is 0.7038 mm, which effectively reduces warpage of microcellular foaming material. Finally, finite element simulation and physical tests are carried out to verify the accuracy of the method to optimize microcellular foaming injection molding process parameters.