Investigation of Mechanical and Corrosion Behavior of <scp>ECAP</scp> Processed <scp>AA7075</scp> Through <scp>ML</scp> , <scp>ANNW</scp> , <scp>RSM</scp> , and <scp>SA</scp> Methodologies

Abstract

ABSTRACT This study employs a multi‐perspective modeling approach combining Response Surface Methodology (RSM), Machine Learning (ML), Artificial Neural Networks (ANNW), and Simulated Annealing (SA) to optimize Equal Channel Angular Pressing (ECAP) parameters for improving the mechanical and corrosion properties of AA7075 alloy. The investigation examines microstructural evolution, mechanical, and corrosion behavior under varying die angles (90° and 120°), processing routes (A, Bc, C), and up to four passes. Significant grain refinement was achieved, with the average grain size reduced from 16.3 to 1.68 μm for route Bc after four passes at 90°. Hardness nearly doubled from 92 to 177 HV under the same conditions, with routes A and C reaching 169 and 156 HV, respectively. Tensile strength increased from 283 to 352 MPa for 4Bc at 90°, while 120° conditions showed slightly lower but still improved performance. Corrosion analysis revealed route‐dependent behaviors, with route Bc at 90° reducing the corrosion rate to 0.0298 mm/year, compared to 0.0345 mm/year for the as‐received alloy. ML‐based models achieved high predictive accuracy ( R 2 near unity), and RSM‐SA optimization closely matched experimental results. This integrated framework provides actionable insights for tailoring ECAP parameters to enhance AA7075's properties for industrial and construction applications.