Evaluating microstructural and machine learning predictive models for friction drilling of sustainable snail shell reinforced aluminium matrix composites

For lightweight automotive applications, friction drilling is a choice candidate for ecofriendly drilling of aluminium matrix composites (AMCs) with green snail shell reinforcement. The present work investigates the effects of significant process variables such as spindle speed, feed rate, workpiece thickness, and drill diameter on bushing length, bushing thickness, and roundness. Higher spindle speeds and lower feed rates enhance bushing length due to controlled material flow and heat generation. However, higher spindle speeds and inappropriate feed rates resulted in decreased roundness. Microstructural examination reveals different zones across the bush formation during the friction drilling process. The head petal and tail petal regions have coarse grains in the range of 20 to 40 μm due to disrupted material flow. Surface topography of all bushing zones reflects the interconnected mechanical and thermal impacts which occur during friction drilling. Surface quality is on higher side at areas with optimal heat generation and material movement such as Upper Critical Region, and lower side at areas with excessive or insufficient deformation such as head and tail Petal Regions. The surface topography analysis at bushing zone reflects the coupled thermal and mechanical effects during friction drilling. Excessive deformation plays a crucial role in the surface quality. Poor surface quality can be observed at Head and Tail Petal Regions due to the excessive deformations. Dynamic recrystallisation produced fine grains in the order of 5 μm in the Upper Critical Region with ideal temperature and mechanical conditions. Random Forest (RF), Multilayer Perceptron (MLP), Gaussian Process Regression (GPR) and Support Vector Machine (SVM) models were employed for the prediction of distinct output responses.