New approach for thermal spray process control with an artificial neural network in aero-engines

Thermal spray coatings on basic materials of aero-engine components arc indispensable for prolonging life and improving efficiency. Thermal spray processes are extremely complex; therefore it is impossible to represent them accurately enough in a traditional model, hence some reliable simulation. In this research work it will be shown how such processes can be represented with an artificial neural network that allows the thermal spray process to be controlled automatically. Data from plasma spray processes on turbo-compressor casings is used to train and verify the neural network. The results of the first prototype for a closed loop arc very promising.