Prediction and Optimization of Dimensional Shrinkage Variations in Injection Molded Parts Using Forward and Reverse Mapping of Artificial Neural Networks

Abstract

The most significant process parameters affecting dimensional shrinkage in transverse and longitudinal directions of molded parts in Plastic Injection Molding (PIM) process are injection velocity, mold temperature, melt temperature and packing pressure. In the present work, ANN model was developed for forward and reverse mapping prediction. In forward mapping PIM process parameters are expressed as the input parameters to predict dimensional shrinkage, whereas in reverse mapping, attempts were made to predict an appropriate set of process parameters required for arriving at the required dimensional shrinkage. The trained network with one thousand input-output data randomly generated from regression equations reported by earlier researchers resulted in minimum mean squared error. The performance of developed model was compared with experimental values for ten different test cases. The results show that ANN model with both forward and reverse mapping is capable of prediction with an error level of less than ten percent.