Research On Shape Diagnosis Of Plasma Spraying Jet Based On Pattern Recognition

In Rapid Plasma Spray Tooling (RPST) processing,there are many kinds of parameters, such as plasma jet temperature field and configuration,structure of spraying gun,power,gas character and flow rate,powder flow rate,particle velocity,distance and angle between gun and the substrate,and so on,all which will affect the quality of mould.Among this parameters,the effect of shape on the quality of plasma spraying jet is most obviously.The plasma jet shape is the most intuitionist information of the plasma jet we obtained at the beginning of the process,it can reflects the stability of various parameters.So it is very urgent to found a applied and reliable plasma spraying diagnostic system to measure the plasma jet shape in order to get a kind of reasonable plasma spray process and the coating quality needed by RPST.Based on image processing technology, the aspect ratio and roundness of plasma jet have been first proposed for presenting shape characteristics of plasma jet in this paper.On the basis of pattern recognition technology,shape characteristics of spray jet under different spraying conditions have been picked-up, and principles of quality diagnosing of plasma jet and coating quality predicting have been established through analyzing the relationship between the spraying parameters and shape characteristics of spray jet. The shapes of plasma jet under different conditions were collected by CCD in the experiments,and clear spray jet contour edge has been picked-up by using sobel operator, and the aspect ratio and roundness of plasma jet have been obtained through analyzing the special point of spray jet contour. The relation between technical parameters and shape characteristics of plasma jet is analyzed through the method of characteristic distilling of plasma jet, and was validated by the measurement experiments of In-flight particle behaviors. The result shows as follows:with the input voltage increasing,the eigenvalue of aspect ratio of plasma jet is also increasing, plasma jet becomes more narrow and longer,flow turbulence becomes lower. In order to set up the principle of the quality diagnosing of plasma jet and realize coating quality predicting, 25 groups of experimental data in plasma spraying experiment have been chosen to implement swatch learning. Microstructure of the coatings has been observed by SX-40 SEM,and hardness of the coatings has been measured by HXD-1000A Vickers hardness tester under the corresponding condition.According to the swatch learning result, the method of classifying the quality of plasma jet has been proposed based on the aspect ratio and roundness of plasma jet, and principles of quality predicting of spraying coating have also been established according to the different classifications of plasma jet. This research is helpful for the establishment of intelligent diagnosis system of plasma spraying process.