Optimization of robot plasma coating efficiency using genetic algorithm and neural networks / S.Prabhu and B.K.Vinayagam

This work describes the Taguchi analysis coupled with Artificial Neural network and Genetic algorithm to optimize the robot deposition parameters used for plasma coating on titanium aluminum alloy material. L27 orthogonal array have been used for coating the work piece using robot. The Arc current (Amp), Arc voltage (volt), powder feed rate(mm/sec), substrate Surface Roughness (μm), Spray gun distance (mm) and TiO2 content in feedstock (%) have been considered as input parameters and coating efficiency is considered as output parameters. Using feed forward Artificial Neural Networks (ANNs) trained the experimental values with the Levenberg–Marquardt algorithm, the most influential of the factors were determined. Regression analysis are used to predict the robot coating efficiency and ANOVA analysis are used to contribute the individual process parameter on robot deposition coating efficiency. The developed mathematical model was further analyzed with Genetic algorithm to find out the optimum conditions leading to the maximum coating efficiency.