Effects of Key Processing Parameters on Microstructure Features and Mechanical Properties of Die-Cast Aluminium Alloy Wheels Assisted with AI and Intelligent Prediction

Abstract

Effects of key processing parameters including ingate velocity (IV) and intensification pressure (IP), on the average/maximum sizes, volume fraction and fractal dimension of defects/porosities and grain size, and mechanical properties of die-cast Al7SiMnMg alloy wheels, were investigated by experiments and assisted with AI and intelligent prediction. Experiments show that IV considerably influences the maximum size of defects/porosities, which decreases considerably from 35.3 μm to 18μm with increasing IVs from 15 m/s to 35 m/s, afterwards remains constant. The average size and volume fraction of defects/porosities can be minimized under IVs of 35–50m/s. The grain size decreases slightly from 18 μm to 15 μm with increasing IVs from 15 m/s to 75 m/s. Microstructure features excluding fractal dimension exhibit similar reduction trends at low IPs of 300–600 bar, afterwards the effect of IP becomes not evident. IV substantially influences elongation, which reaches a high level of 10% at 35–50 m/s, but decreases remarkably to 5% at 15 m/s and 6.8% at 75m/s. IP clearly affects UTS and elongation, and UTS increases by 20MPa to 260MPa, while elongation increases absolutely from 4% to 12%, as IPs increase from 300 to 1500 bar. The yield strength is not remarkably affected by both IV and IP. Dataset of processing parameters–microstructure features–mechanical properties correlation was constructed, and XGBoost was confirmed as the most robust AI model for intelligently predicting mechanical properties within 10% error. SHAP analysis quantified underlying contributions of processing parameters to mechanical properties, and IV and IP account for 45.0% and 55.6% of the total feature impact to consequentially affect elongation.