A method combining optimization algorithm and inverse-deformation design for improving the injection quality of box-shaped parts

Volume shrinkage and warpage deformation are very critical quality indicators in the plastic injection molding (PIM) of box-shaped thin-walled plastics. These two performance indexes are greatly affected by the molding parameters. Therefore, in this paper, six optimization algorithms and inverse-deformation designs (IDD) are used to reduce volume shrinkage and warpage deformation. Firstly, six important molding parameters, namely filling time (A), plasticity temperature (B), mold temperature (C), holding time (D), maximum holding pressure (E) and cooling time (F), are determined, and the L 25 (5 6 ) orthogonal experimental design (OED) is established. Taguchi grey correlation (TGC) theory analysis is used to determine the optimal combination of molding parameters. Secondly, different combinations of Box-Behnken design (BBD) response surface method, BP neural network (BPNN) training, and NSGA-II genetic algorithms are used to generate four combined optimization algorithms, in order to perform multi-objective optimization of the six molding parameters. The result shows that the effectiveness of four optimization analyses are ranked as follows: BPNN-BBD-NSGA-II > BPNN-BBD > BBD-NSGA-II > BBD. The BPNN-BBD-NSGA-II method holds the best prediction results. Finally, a global optimization platform based on NX/Moldex3D is established considering the IDD theory to simulate the molding process. Optical scanning instruments are used to examine the molding quality. The result proves that the warpage deformation in box-shaped thin-walled injection-molded products is almost completely eliminated and a high molding quality can be achieved. This research is favorable for designing the molding process and guiding the molding of box-shaped thin-walled injection-molded products.