Modeling of wear process parameters of in-situ RZ5-10wt%TiC Composite using artificial neural network

Abstract

In this paper modeling of wear process parameters of in-situ RZ5-10wt%TiC metal matrix composite investigated. The RZ5-10wt%TiC composite has fabricated by argon assisted self-propagating high-temperature method. It is used in automotive, aerospace application to protect parts against wear. The pin on disc machine has used for the wear test. It has performed by varying input as the applied load and sliding distance. The coefficient of friction and weight loss considered as the output of the test. The results showed improvement in wear resistance of the RZ5-10 wt% TiC composite compares to the unreinforced alloy. Further, modelling of wear resistance of RZ5-10 wt% TiC composites were done using the back propagation neural network with Broyden–Fletcher–Goldfarb–Shanno(BFGS)algorithm. The ANN model trained using experimental data. After the completion of training process test data were used to check the system accuracy. It found success in the prediction of wear process parameters of RZ5-10wt%TiC composite. The results have shown that the values of wear process parameters obtained from ANN are very close to the experimental values. Also, it is useful compare to time-consuming experimental processes.