سیستم هوشمند برای پیش بینی و کنترل : استفاده در فرآیند پلاسما اسپری

Parametric drifts and fluctuations occur during plasma spraying. These drifts and fluctuations originate primarily from electrode wear and intrinsic plasma jet instabilities. One challenge is to control the manufacturing process by identifying the parameter interdependencies, correlations and individual effects on the in-flight particle characteristics. Such control is needed through methods that (i) consider the interdependencies that influence process variability and that also (ii) quantify the processing parameter-process response relationships. Due to the large amplitudes of the drifts and fluctuations, the strategy to adopt would depend on the required corrections to apply to the in-flight particle characteristics. Artificial intelligence is a pertinent tool to reach this objective. The system is flexible in order to permit a full control based on pre-defined rules aiming at maintaining at constant values in-flight particle characteristics (average surface temperature and velocity) by adjusting in real time the arc current intensity, the total plasma gas flow and the hydrogen content whatever the fluctuations.

The performance of the industrial equipments and the production rate depend on the capability of some parts to resist wear. Thus, to improve these performances and to confer specific characteristics to the part, the surface treatment methods such as atmospheric plasma spraying (APS) are used. This process permits to create a coating on a surface in order to confer singular characteristics to protect it from phenomena such as wear, corrosion, erosion, etc. (Normand, Fervel, Coddet, & Nikitine, 2000). APS consists in injecting in a viscous enthalpy plasma jet (animated by a momentum) powder particles whose average size ranges from 10 to 100 μm. These particles are melted and simultaneously accelerated towards the surface of the part to be covered. They form, after impact, spreading and solidification, lamellae of a few tens to hundreds micrometers in diameter and a few micrometers thick. The coating results from the stacking of these lamellae (Fauchais & Vardelle, 2000). Fig. 1 summarizes the principle. APS process is characterized by several parametric drifts and fluctuations at different characteristic times (100–500 μs). These phenomena originate especially from the electrode wear (in tens of hours) and intrinsic plasma jet instabilities.The objective of this work was to develop an expert system which can adjust the operating process parameters as a function of the measured in-flight particle characteristics to elaborate a coating. Due to the large amplitudes of these drifts and fluctuations, the strategy to adopt would depend on the required corrections to apply to the in-flight particle characteristics. Artificial intelligence (AI) based on artificial neural networks (ANN) concerning the prediction of the parameters to be reached (which adjustment must be proceeded to tend towards the target value?) and on fuzzy logic (FL) concerning the strategy to adopt to control the process parameters (on which parameters to act and with which amplitude?) (Kanta et al., 2006 and Kanta et al., 2008). The system is flexible in order to permit a full control based on pre-defined rules aiming at maintaining at constant values in-flight particle characteristics (surface temperature T and velocity V) by adjusting arc current intensity (I), total plasma gas flow (H2 + Ar) and hydrogen content (H2/Ar) whatever the fluctuations. The methodology and the key steps of the process were validated experimentally in the specific case of alumina–titania coating.

The APS process is characterized by several parametric drifts and fluctuations at different characteristic times due, especially, to the electrode erosion and intrinsic plasma jet instabilities. The expert system developed allows full process control on the basis of pre-defined rules. The rules control the “real time” adjustment of the power parameters so that constant values for the in-flight particle characteristics can be maintained. The knowledge base stores data on the effect of power parameters on in-flight particle velocity and temperature. Thus, the required correction, which depends on measured magnitudes of drifts and fluctuations, can be implemented. The regulator maintains constant velocity and temperature of the in-flight particles by adjusting the process parameters: arc current intensity, total plasma gas flow and the hydrogen percentage in the plasma gas. This regulation system precisely permitted to take into accounts the instabilities and intrinsic fluctuations inherent to the process.