Enhancing Casting Strength through Optimization of Porosity Formation in Pressure Die Casted AlSi9Cu3 Alloy: Modeling and Optimization using Prioritized Machine Learning–RSM Approach
摘要
High-pressure die casting (HPDC) has become a dominant method in industries because it enables the rapid production of complex parts with excellent surface finish and tight dimensional control. This study aims to enhance the strength of the cast component for an AlSi9Cu3 alloy casting by minimizing porosity and shrinkage. Key HPDC variables such as first-phase and second-phase injection speeds, die and furnace temperature, and accumulation pressure were designed by using design of experiment method L27 orthogonal array with porosity as primary outcome. Experimental runs were optimized and compared using a prioritized k-means clustering approach to establish relationships. Moreover, the best cluster outcomes were further processed for strength-based analysis to achieve the best outcome for the desired product. First speed phase has a stronger relationship with porosity as it directly controls metal flow, filling quality, and solidification, governing the turbulence and filling. The entropy weightage indicates first-phase velocity (85.88%) dominates porosity control due to its strong influence on melt flow and turbulence, while second-phase velocity (5.68%) and pressure (6.11%) have moderate effects on densification. K-means clustering identified Sample 8 as the best performer leading to reduced porosity and superior mechanical strength followed by samples 22, 13, 16, and 19 also clustered near the centroid, showing strong performance. Strength analysis of Sample 8 further showed highest maximum force (before break) of 26.63 kN and elongation of 1.15 mm results. Optimized HPDC settings significantly reduced porosity levels, resulting in improved strength and break load capacity of the selected samples. This study will contribute to advancing die casting techniques by promoting higher quality, reduced defects, and improved overall manufacturing efficiency.