Modeling, validation and multi-objective optimization of Al-Si/SiC-ZnO-Graphite composites: integrating neural networks-genetic algorithm with interaction effect modeling for tribological performance

While main effects of processing parameters on aluminum matrix composite tribology have been extensively studied, interactive effects of reinforcement parameters—which can dominate performance—remain inadequately understood. This research systematically investigates reinforcement content, composition, and critically, their synergistic interaction, on wear behavior of hypereutectic Al-Si composites reinforced with zinc oxide, silicon carbide, and graphite fabricated from recycled piston scrap. Three predictive models were developed: Two-Factor Interaction from Design of Experiments, and Artificial Neural Networks employing Levenberg-Marquardt and Bayesian Regularization algorithms. Digital Image Processing quantified porosity and surface roughness from microstructures and worn surfaces. Hierarchical effects analysis revealed that reinforcement interaction exhibited dominant influence on wear performance (effect magnitude 156.15), exceeding individual effects of composition (95.0) and content (94.8), providing quantitative evidence that synergistic mechanisms rather than additive contributions govern tribological behavior. All models demonstrated exceptional correlation (R² ≥ 0.999), validated through ANOVA, Kruskal-Wallis, Levene’s, Mood’s Median, and Bonferroni tests. Desirability Function Analysis and Genetic Algorithm optimization converged to identical conditions (RCMP7, 11 wt%), predicting minimum wear of 196.59 and 194.98 mg/min respectively, experimentally verified at 197 mg/min (0.2% deviation). The optimized composite achieved 97.48% wear reduction versus monolithic alloy and 61.3% weight reduction versus cast iron for cylinder liner applications, while maintaining fracture toughness (28.6 MPa·m^1/2) exceeding structural requirements. This work establishes interaction-focused optimization as essential for high-performance tribological composite development, with practical implications for automotive lightweighting and fuel efficiency improvement. Interaction effect between reinforcement composition and content dominates tribological performance over individual reinforcement parameters in aluminum-silicon ternary composites. Optimized aluminum-silicon/silicon carbide-zinc oxide-graphite composite achieves 97.48% wear reduction versus monolithic alloy under dry sliding conditions. Multi-objective optimization using Artificial Neural Network-Genetic Algorithm and 2-Factor Interaction modeling predicts wear behavior with greater than 99% correlation accuracy. Zinc oxide addition as third reinforcement provides synergistic wear resistance enhancement with 342% fracture toughness improvement to 28.6 megapascal root meter. Digital image processing validates surface roughness consistency (0.2879–0.2885 micrometers) across different reinforcement compositions under identical wear conditions.