Predictive modelling of surface roughness and residual stress induced by milling of hot forged and heat treated AA7075

This study investigates the influence of cutting parameters on residual stress and surface roughness during the milling of hot-forged and T6 heat-treated AA7075 components. Using Taguchi L9 and full-factorial experimental designs and regression modelling, the research highlights important relationships between cutting parameters (cutting speed, feed rate, and depth of cut), residual stress and surface roughness. Higher cutting speeds (350 m/min) and lower feed rates (0.1 mm/tooth) significantly minimized residual stresses, with hoop stress values decreasing from 108.7 MPa at lower speeds (150 m/min) to approximately 73.4 MPa at higher speeds, and axial stress values ranging from 45.9 MPa to 88.5 MPa. Surface roughness (Ra) was most influenced by feed rate, with measurement values varying between 0.25 μm and 0.92 μm. Support Vector Regression (SVR) demonstrated better accuracy for predicting residual stress (MAPE: 11.5%) and surface roughness (MAPE: 7%), outperforming Lasso and Ridge regression models. These findings provide a consistent framework for optimizing cutting parameters and enhancing residual stress and surface roughness in AA7075 machining processes, offering practical implications for improving component performance and manufacturing efficiency.