Traditional approaches to designing process parameters for new alloys often rely on inefficient trial-and-error methods. In this study, elemental physicochemical parameters were integrated with feature selection to develop a machine learning prediction model with robust predictive capability. Through Pareto analysis, the optimal extrusion parameters for high-strength Al-Mg-Si alloys were determined as an extrusion ratio (EXR) of 40, temperature (EXT) of 540 °C, and speed (EXS) of 1.1 mm/s. Under these conditions, the ultimate tensile strength (UTS), yield strength (YS), and elongation (EL) were 402.4 MPa, 385.5 MPa, and 10.3 %, respectively. Additionally, significant nonlinear interactions between process parameters and mechanical properties were revealed through Shapley additive explanations (SHAP) analysis. Through electron backscatter diffraction (EBSD) and transmission electron microscopy (TEM), the strength-ductility mechanism was attributed to the competing effects of grain boundaries, dislocations, textures, and precipitates. This data-driven strategy provides a robust methodology for optimizing and designing alloy processing parameters.
扫码关注我们
求助内容:
应助结果提醒方式:
