|
3686 |
|
Wednesday, May 21, 2014, Hall A 10:50 AM Plasma Spraying 1 |
|
Optimization of an emulator based on nonlinear auto regressive model with exogenous inputs for an atmospheric plasma spray torch |
|
Taikai Liu / IRTES-LERMPS, UTBM, France Marie-Pierre Planche/ IRTES-LERMPS, France Sihao Deng/ IRTES-LERMPS, France Ghislain Montavon*/ IRTES-LERMPS, France |
|
An emulator was designed to simulate the behavior of an atmospheric mono-cathode plasma spray torch used in APS process. The in-flight particle characteristics are used as output variables in this emulator. These characteristics are measured through experience by DPV 2000, but a system error about 10% can be watched both in temperature and velocity results. Due to the unstable and random phenomena during the process, the signal acquired is very rough accompanied with random noise. For those reasons, the simulated result only has a fit about 80% in velocity and a fit about 70% in temperature. In order to improve the accuracy of this emulator, a numerical CFD simulation has been carried out to produce and collect data with high enough relevancy. Such a simulation is a key step for the emulator optimization since it allows calculating the whole plasma jet properties and gives information on in-flight particles. Before being implemented in the emulator, the numerical results have to be optimized to get the best characteristics compared with the real measured values. With these data, the emulator shows a high agreement with the experimental behavior, the difference in temperature and in velocity can be both diminished to less than 1%. |