Microstructure Study of Friction Stir Processed Hypereutectic Al-20Si Alloy and Analysis of the Wear Behaviour using Machine Learning Algorithms

Al–20Si alloy was subjected to friction stir processing to find its effect on the microstructure and wear behaviour. The microstructures of as-cast and friction stir processed (FSP) alloy were studied using optical microscopy and field emission scanning electron microscopy. The microstructure analysis showed significant refinement of Si particles in Al–20Si alloy by FSP. Similarly, for the experimental study of the wear behaviour, three different parameters: sliding velocity, normal load, and sliding distance were considered. In this study, five different machine learning (ML) algorithms were used for the prediction of wear rate. The hyper parameter tuning of each model was carried out for accurate comparisons. The models were then evaluated on the basis of different statistical metrics to find the superior model. Random Forest model showed the highest prediction accuracy (R2 = 0.8846) and was considered for comparing the wear rates with experimental values.