Quality Prediction And Intelligent Adjustment Based On Pattern Recognition In Plasma Spray Tooling

Rapid tooling (RT) has become a key technology in the design and development of new products for its low manufacturing cost and rapid response to the market demand. As for its few limits of the complex pattern's size and spraying materials and high formation quality, rapid plasma spray tooling (RPST) has received widespread attention. However, the key problem in RPST process is how to keep formability and final quality of coatings to avoid coating wrap and fissure due to the asymmetry of coatings thickness and temperature distribution. In the process of plasma spraying, different technological parameters could lead to different coating qualities. So it is especially necessary and important to predict the coating quality and intelligently adjust the main processing parameters.The main processing parameters were selected on the help of the experiment result of the robotic spray path in our laboratory, and deeply experimental studies were carried out to investigate the influence of different parameters on the quality of sprayed coatings during the RPST process.In this paper, the author firstly discussed theory, procedure and major arithmetic. Then the diagnostic system of coatings quality based on pattern recognition was developed. The system consists of clustering analysis of sprayed coating under different process conditions using different reflection methods and comparison of different resulting conclusions, development of a prediction model of spray coating to predict coating quality on the basis of reflection method, investigation of influence grade of the main processing parameters on the quality of sprayed coatings by feature extracting, analysis of the minimum training sample points on purpose of getting fine clustering effect, optimizing of processing parameters in plasma spraying and determination of the optimum parameter range using evolutionary method and mapping reversion.Finally, experimental study was carried out to test and verify the model mentioned above, results proved that the diagnostics system of coatings quality based on pattern recognition can gain good reliability and practicability, and it has great significance for planning of reasonable spraying processes.