Influence of titanium carbide and scandium particulates dispersion on Al–Zn alloy by ultrasonic assisted stir casting: enhancing characterisation, tensile and tribological properties

In this research, the optimised process parameters for Ultrasonic Assisted Stir Casting (UAS) of AA7075/TiC/Sc Metal Matrix Composites were determined. The input variables considered were Sc content (0.25, 0.5, 0.75 wt.%), melting temperature (700, 725, 750°C), stirring speed (200, 225, 250 rpm), and stirring duration (5, 10, 15 min). Tensile strength and wear loss were chosen as response parameters. Grey Relational Analysis (GRA) and Deep Q-Network (DQN) machine learning were employed for optimisation, while XRD, OM, and SEM analyses were used for metallurgical, worn surface, and fracture characterisation. GRA results indicated that 0.5 wt.% Sc, 750°C, 200 rpm, and 10 min were optimal, with SC6 achieving the highest GRG at 0.912. DQN confirmed SC6 as optimal, attaining the maximum reward (1.0) with superior tensile strength (299.5 MPa, norm 1.0) and low wear (38 mg, norm 0.222). DQN demonstrated better predictive capability than GRA, with higher consistency in performance. XRD analysis of AA7075/0.6 wt.% Sc/12 wt.% TiC composites showed primary and secondary phases; sharp Al peaks confirmed matrix dominance, Al₃Sc precipitates enhanced grain refinement and strength, while Mg₂Zn and TiC provided precipitation strengthening, hardness, and wear resistance.