Extraction of kaolin and tribo informative analysis of the Al-kaolin composite through machine learning approaches

Microwave fabrication of aluminium composites has emerged as a novel and trending technique in the current industrial landscape due to its efficiency and energy-saving potential. In this study, Al-kaolin composites were fabricated using microwave energy techniques, focusing on predictive modelling of the microwave-assisted Al-kaolin composite's wear rate and coefficient of friction (COF). The fabricated composites were evaluated for hardness, wear rate, and coefficient of friction (COF) under varying parameters. It was observed that 4 wt% kaolin is the optimal reinforcement fraction, resulting in a 34% improvement in tensile strength, while hardness showed a consistent increase up to 4 wt% Kaolin, reaching a maximum value of 96 RHC. Additionally, wear rate and COF exhibited a decreasing trend with increasing kaolin content, indicating enhanced tribological performance. The lowest wear rate of 3.2 × 10⁻⁴ mm³/Nm and COF of 0.42 were observed for the 4 wt% Kaolin composite, demonstrating improved wear resistance. To further understand and predict the behaviour of the composites, a systematic dataset was collected, and various machine learning (ML) models were trained and tested for predictive modelling of wear rate and COF. Among the trained models, XGBoost demonstrated the highest predictive accuracy, achieving 94.33% for wear rate and 94.62% for COF. A feature importance analysis revealed that the standard of distance (Sod) was the most influential parameter affecting these outputs.