Optimized deep learning regression assisted wear rate analysis of Cu–AlCoCrCuFe HEA composite

Abstract

At present, composites have brought materials to a new era with superior characteristics. The proper selection of materials with optimal processing conditions is one of the prime factors in determining the efficiency of composites. This article introduces a new composite by fabricating a copper matrix with reinforced AlCoCrCuFe High Entropy Alloy (HEA). The ultimate theme of this research is to forecast the wear rate of the Cu–AlCoCrCuFe HEA composite. Using the pin on the disk, the wear test is carried out on varying combinations of parameters. Then Taguchi analysis is carried out to get the precise fit parameters. The wear rate mainly depends on the percentage of HEA, sliding distance, sliding velocity, and the load applied. Correlation analysis is performed to determine the exact parameter determining the optimal wear rate and the coefficient of friction. After feature extraction, the parameters are optimized, and the neural network regression is given the optimized parameters. The network has been trained, and predictions are made using it. The model successfully predicts the wear rate, as evidenced by the declining RMSE of 0.28 and rising R2 values up to 92%. The analysis shows a significant decline in the wear rate with HEA additions. In addition, the wear rate increases with a rise in load, sliding velocity, and sliding distance.