Investigation of Squeeze Cast Process Parameters Effects on Secondary Dendrite Arm Spacing Using Statistical Regression and Artificial Neural Network Models

Abstract

Abstract The near net shape manufacturing capability of squeeze casting process have the potential to produce high dense components with refined micro-structure. However, squeeze cast micro-structure is influenced by large number of process variables such as squeeze pressure, time delay, pressure duration, die temperature and pouring temperature. In the present work, an attempt is made to develop the model by considering aforementioned process variables. Further, significant contribution of each process parameter on the secondary dendrite arm spacing is studied by using statistical regression tool. The mathematical relationship has been developed for secondary dendrite arm spacing was used to generate the training data artificially at random and tested with the help of few test cases. It is to be noted that the test cases chosen were different from training data. Scaled conjugate gradient, Levenberg-Marquardt algorithm and regression model predictions were compared. It is interesting to note that, all models were capable to make good prediction with an average of 5 percentage deviation. Levenberg-Marquardt algorithm outperforms in terms of prediction compared to other models in the present work. The reason might be due to the nature of error surface.