Modelling and Prediction of Responses in the IC Process using Artificial Neural Networks

In this study, an artificial neural network model has been developed to predict the characteristics of the disposable wax patterns used in the investment casting process (IC). The selected process parameters were the injection temperature, injection pressure, die temperature and injection time. The responses were the linear shrinkage and surface roughness of the wax patterns. The experiments have been performed as per Taguchi’s L18 Orthogonal array. A feed-forward back propagation neural network has been built using 13 randomly chosen experimental data and the remaining 5 data were used for testing the accuracy of the built model. It was found that the mean absolute percentage error between the actual and the predicted results of the responses was less than 15% in all cases, which shows that the model has been well built.