MODELING OF PRESSURE DIE CASTING PROCESS: AN ARTIFICIAL INTELLIGENCE APPROACH

In the present work, both forward and reverse modeling is carried out for the high pressure die casting process by utilizing back-propagation neural network (BPNN) algorithm. The pressure die casting process is considered as an input-output model with the fast shot velocity, intensification pressure, phase change over point and holding time as the input parameters, whereas surface roughness, hardness and porosity as the output of the system. Batch mode of training had been provided to the networks with the help of one thousand input-output training data. These training data were generated artificially from the regression equations, which were obtained earlier by the same authors. The regression equations used in the present work were obtained by applying design of experiments and response surface methodology techniques. The performance of BPNN in forward and reverse modeling has been tested with the help of test cases. Further, the performance of BPNN in forward modeling was compared with statistical regression models. The results showed that the BPNN approach is able to carry out both the forward as well as reverse mappings effectively and can be used in the foundries.