Effects of Fe Content and Natural Ageing on Microstructure and Mechanical Properties of Recyclable Al5.5Mg2Si Die-Cast Alloys Assisted with Machine Learning

Abstract

Effects of Fe content and natural ageing (NA) on microstructure and mechanical properties of Al5.5Mg2SixFe (x = 0.12, 0.3, and 0.6, in wt%) heat treatment-free die-cast alloys are investigated in assistance with machine learning (ML). The main intermetallics in the alloys include β-Mg2Si and α-Al15(Fe,Mn)3Si2, while β-Al13(Fe,Mn)4Si0.25 with the number density of (0.15 ± 0.02) × 10^−2 μm^−2 appears in the Al5.5Mg2Si0.6Fe alloy. As Fe increases from 0.12% to 0.6%, the as-cast yield strength (YS) and elongation (El) reduce from 163.6 ± 2.6 MPa to 160.6 ± 2.5 MPa and 13.09 ± 1.16% to 10.58 ± 1.27%, respectively, which is attributed to the β-Al13(Fe,Mn)4Si0.25 and the increased number density of α-Al15(Fe,Mn)3Si2 from (1.26 ± 0.21) × 10^−2 to (6.32 ± 0.72) × 10^−2 μm^−2. After NA, the alloys show increased YS and decreased El. The Al5.5Mg2Si0.12Fe alloy exhibits considerable NA strengthening with YS increasing by 9.2 ± 5.1 MPa and El decreasing by 2.33 ± 3.21% after NA for 30 days, which is due to the nanoscale β″ precipitates. The quantitative relationship between Fe content, NA time, and tensile properties is established by the Random Forest ML model, i.e., YS(MPa) = 32485.5 − 32324.7 × exp(−0.5 × ((t(day) − 3.9)/2083.9)^2 − 0.5 × ((Fe(wt%) − 0.48)/18.87)^2) and El(%) = 827.5 − 1529.9 × exp(−0.5 × ((t(day) − 517.61)/3777.2)^2 − 0.5 × ((Fe(wt%) − 269.65)/242.01)^2). A high-performance and recyclable Al5.5Mg2Si0.56Fe die-cast alloy is predicted with the YS and El of 163.75 ± 3.4 MPa and 10.05 ± 0.21% after NA for 30 days. This study provides insights for intelligently developing high-performance and recyclable die-cast alloys.