Intelligent process modeling of robotic plasma spraying based on multi-layer artificial neural network

The implementation of multi-layer artificial neural networks (ANNs) in robotic plasma spraying was discussed and an intelligent process model was constructed to fully describe the relationships between process parameters and coating properties. Influences of plasma arc current, spray distance, robot scanning space and scanning velocity on coating properties, i. e. residual stress and porosity, were systematically studied based on the present model. Prediction can be effectively carried out after the learning of the experimental database. Theoretical analysis shows the prediction results agree well with the experiments. It is favorable to fully investigate the complex and nonlinear relationships between processing parameters and coating properties as well as to overcome the limited information indicated by the discrete variable in the processing results.