Optimization of Air Plasma Sprayed YSZ Coatings on Inconel 718 Using Taguchi Design and ANN Modeling

This study aims to optimize the air plasma spray parameters for yttria-stabilized zirconia (YSZ) thermal barrier coatings on Inconel 718 superalloy substrates. A Taguchi-based design of experiments was adopted to evaluate the influence of key process parameters, namely, power, stand-off distance, and powder feed rate, on the coating performance. NiCr (80/20) was used as the bond coat. Hardness, porosity, and machining-induced delamination factor (MIDF) factor were selected as the output responses, and signal-to-noise (S/N) ratios were used for analysis. The results revealed that the stand-off distance had the most significant influence on both porosity and hardness, whereas the powder feed rate played a major role in MIDF. The optimal conditions for minimizing porosity and MIDF and maximizing hardness were identified as power level 3 (30 kW), stand-off distance level 2 (200 mm), and powder feed rate level 2 (40 g/min). The findings were supported by ANOVA and S/N ratio analysis, validating the reliability and effectiveness of the optimized parameters for improving the coating quality. An ANN model was developed to predict the output response with respect to the variation in the input factors. A feed-forward backpropagation network with a 3-3-3 architecture was adopted to supervise the learning. Levenberg–Marquardt was used to improve the predictability, and the developed model had better predictability, with R 2 values of 99.4.