An artificial neural network for predicting the friction coefficient of deposited Cr1−xAlxC films

This paper applies a generalized regression neural network (GRNN) for predicting the friction coefficient of deposited Cr1−xAlxC films on high-speed steel substrates via direct current magnetron sputtering systems. The Cr1−xAlxC films exhibited some excellent characteristics, such as low friction coefficient, high hardness, and large contact angle. In this study, a GRNN model is applied for predicting the friction coefficient of Cr1−xAlxC films on high-speed steel substrates instead of complex practical experiments. The results exhibit good prediction accuracy of friction coefficient since about ±0.97% average errors and show the feasibility of the prediction model. Compared to the conventional back propagation model, the GRNN model is more suitable to predict the friction coefficient of Cr1−xAlxC films.

Figure optionsDownload as PowerPoint slideHighlights
► A neural network was applied to predict the friction coefficient of CrAlC films.
► The applied artificial neural networks can reduce the complex actual experiments.
► General and modified networks can both predict the estimated output values.
► The modified network model has good prediction accuracy of friction coefficient.
► Compared to conventional network, the applied model in this case is more suitable.