Reducing shrinkage in injection moldings via the Taguchi, ANOVA and neural network methods

Abstract

Plastic injection molding is suitable for mass production articles since complex geometries can be obtained in a single production step. However, the difficulty in setting optimal process conditions may cause defects in parts, such as shrinkage. In this study, optimal injection molding conditions for minimum shrinkage were determined by the Taguchi, experimental design and the analysis of variance (ANOVA) methods. Polypropylene (PP) and polystyrene (PS) were injected in rectangular-shaped specimens under various processing parameters: melt temperature, injection pressure, packing pressure and packing time. SIN ratios were utilized for determining the optimal set of parameters. According to the results, 260 degrees C of melt temperature, 60 MPa of injection pressure, 50 MPa of packing pressure and 15 s of packing time gave minimum shrinkage of 0.937% for PP and 1.224% for PS. Statically the most significant parameters were found to be as packing pressure and melt temperature for the PP and PS moldings, respectively. Injection pressure had the least effect on the shrinkage of either material. After the degree of significance of the studied process parameters was determined, the neural network (NN) model was generated and was shown to be an efficient predictive tool for shrinkage. (C) 2009 Elsevier Ltd. All rights reserved.