Hybrid RSM-Fuzzy Logic framework for multi-objective optimization and predictive modeling of HVOF-sprayed WC-12Co coatings on SUS 400 stainless steel

• Developed a hybrid RSM-Fuzzy Logic framework for multi-objective optimization of HVOF-sprayed WC-12Co coatings on SUS 400 stainless steel. • Achieved high predictive accuracy for surface roughness, hardness, porosity, and coating thickness, with R² values exceeding 0.98 for all responses. • Identified optimal HVOF process conditions leading to dense, uniform, and well-bonded coating microstructures. • Demonstrated strong agreement between experimental results and model predictions, with MAPE values below 1.2% across all responses. • Confirmed the hybrid RSM-Fuzzy Logic approach as a robust and interpretable tool for predictive modeling and process optimization in HVOF spraying. High-Velocity Oxy-Fuel (HVOF) thermal spraying is widely employed to produce dense and wear-resistant coatings for high-performance engineering components; however, achieving consistent coating quality at the industrial scale remains challenging due to complex nonlinear interactions among process parameters. This study proposes a hybrid predictive and optimization framework integrating Response Surface Methodology (RSM) with Fuzzy Logic to enhance the mechanical and structural performance of WC-12Co coatings deposited on SUS 400 stainless steel. A Central Composite Design (CCD) comprising 30 experimental runs was employed to systematically investigate the effects of oxygen flow rate, propane flow rate, powder feed rate, and spray distance on four critical coating responses: surface roughness (Ra), hardness, porosity, and coating thickness. Analysis of variance (ANOVA) identified powder feed rate as the most influential parameter governing coating porosity (p = 0.0395), while the developed RSM models successfully captured both main and interaction effects within the investigated design space. To further improve predictive reliability, a Mamdani-type Fuzzy Inference System (FIS) was developed, achieving mean absolute percentage errors below 1.5% and providing transparent, rule-based interpretability suitable for industrial decision-making. Under optimized operating conditions, the hybrid framework yielded a coating hardness of approximately 1296.2 HV, low porosity of about 1.1%, uniform coating thickness around 265.5 µm, and surface roughness of approximately 1.71 µm. Overall, the proposed RSM-Fuzzy Logic hybrid framework demonstrates superior predictive accuracy and robust multi-objective optimization capability compared with conventional regression-based approaches, offering an effective decision-support tool for achieving consistent coating quality in industrial HVOF applications.