Hybrid CCD-ANFIS-GWO Framework for Multi-Objective Optimization of Anthill Clay-Sand Moulding Parameters and Grey Cast Iron Casting Quality

Stringent government and environmental regulations on sand quarrying pose scarcity challenges to meet 70% of global casting production in foundries. Anthill clay (a binder and partial replacement for silica sand) is used to prepare moulds, affecting collapsibility (CP) and mould hardness (MH). The CP and MH are influenced by the anthill-to-sand ratio, mulling time, number of strokes, coal dust, and water. Central composite design (CCD) experimental plans are used to study variables and to analyze process insights (main effects, interaction effects, and the establishment of an empirical equation) for CP and MH. All factors significantly contribute to both responses. The number of strokes, anthill-to-sand ratio, and mulling time show nonlinear relationships with CP and MH. The CP and MH show a direct relationship, with a coefficient of determination of 98.5%. The mean absolute percent errors in predicting the 10 random sets of experimental cases for MH and CP are 2.91% and 3.2% for CCD, and 2.11% and 2.97% for the Adaptive Neuro-Fuzzy Inference System (ANFIS), respectively. The computed highest overall desirability value of 0.8588 corresponds to Case 2 (maximum weight fraction of 0.9 for CP and minimum weight fraction of 0.1 for MH), which is recommended as the optimal condition determined by the Grey Wolf Optimizer (GWO). Optimized anthill clay sand moulding conditions increase the tensile strength by 15.02%, impact strength by 15.56%, and casting density by 3.8% in FG200 grey cast iron castings (automobile engine mounting bush) compared to sub-optimal sand-moulding conditions.