Experimental research and optimization based on response surface methodology on machining characteristics of cast Al-7Si-0.6 Mg alloy: Effects of cutting parameters and heat treatment

In this study, structural, mechanical and machining features of Al-7Si-0.6 Mg alloy manufactured by permanent mold casting (PMC) in both as-cast (AC) and T6 heat-treated (HT) conditions were investigated. Structural and mechanical properties were determined using conventional methods. Milling experiments were conducted with titanium aluminum nitride (TiAlN) coated carbide end mills in CNC milling machine under conditions of three different cutting speeds (V: 50, 80 and 110 m/min), feed rate (f: 0.08; 0.16 and 0.24 mm/rev) and constant depth of cut (DoC) (1.5 mm). It was seen that structure of AC Al-7Si-0.6 Mg comprised of aluminum-rich alpha-Al, coral-like eutectic Al-Si, primary Si, Mg2Si, Fe-rich plate, acicular beta-Fe (beta-Al5FeSi) and script-like pi-Fe (pi-AlSiMgFe) intermetallic phases. Eutectic Al-Si, primary Si, beta-Fe and pi-Fe phases in the structure of the alloy in the HT condition were partially dissolved and turned into an agglomerated and spherical state. While the hardness (HB), yield (YS) and tensile strength (TS) of the alloy raised with HT, the elongation to fracture (EF) reduced. While cutting force (CF), surface roughness (SR), built-up layer (BUL) and built-up edge (BUE) decreased with increasing V, they increased with increasing f in the milling of both cast and HT alloys. Adhered layers and feed marks were consisted of the cutted surfaces of the alloys. While the most adhered layer was observed in AC alloys at a low V (50 m/min) and high f (0.24 mm/rev) combination, the least adhered layer formation was observed at the highest V and the lowest f value in HT alloys. Optimum parameters with Response Surface Methodology (RSM) were determined as V: 125 m/min and f: 0.04 mm/rev in AC and HT cases. The statistical significance of V and f on CF and SR was revealed by analysis of variance (ANOVA), while mathematical models were improved to predict CF and SR.