Open-Cell AlSn6Cu-SiC Composites: Fabrication, Dry-Sliding Wear Behavior, and Machine Learning Methods for Wear Prediction

Abstract

Open-cell AMMCs are high-strength and lightweight materials with applications in different types of industries. However, one of the main goals in using these materials is to enhance their tribological behavior, which improves their durability and performance under frictional conditions. This study presents an approach for fabricating and predicting the wear behavior of open-cell AlSn6Cu-SiC composites, which are a type of porous AMMCs with improved tribological properties. The composites were fabricated using liquid-state processing, and their tribological properties are investigated by the pin-on-disk method under different loads (50 N and 100 N) and with dry-sliding friction. The microstructure and phase composition of the composites were investigated by scanning electron microscopy, energy-dispersive X-ray spectroscopy, and X-ray diffraction. The mass wear and coefficient of friction (COF) of the materials were measured as quantitative indicators of their tribological behavior. The results showed that the open-cell AlSn6Cu-SiC composite had an enhanced tribological behavior compared to the open-cell AlSn6Cu material in terms of mass wear (38% decrease at 50 N and 31% decrease at 100 N) while maintaining the COF at the same level. The COF of the composites was predicted by six different machine learning methods based on the experimental data. The performance of these models was evaluated by various metrics (R2, MSE, RMSE, and MAE) on the validation and test sets. Based on the results, the open-cell AlSn6Cu-SiC composite outperformed the open-cell AlSn6Cu material in terms of mass loss under different loads with similar COF values. The ML models that were used can predict the COF accurately and reliably based on features, but they are affected by data quality and quantity, overfitting or underfitting, and load change.