Quality Modeling and Prediction of Nitride Hardening for Piston Rings Based on Elman Neural Network

Quality prediction and control methods are crucial to obtain safe and reliable operation in motor drive systems. In view of complex factors and strong relativity of the nitride hardening process for piston rings, the principal component analysis is used to eliminate minor factors and to extract major factors. Then the quality prediction model based on an improved Elman neural network is successfully applied in the nitride process, which inspects any odd change and predict the process characteristics. Finally, Extended Kalman Filter is used to improve the training speed, which can effectively prevent the neural network to trap in local minimum and to improve the approximation accuracy. The simulation and experiment results show the quality model can effectively predict the characteristic values of process quality, and it also can identify abnormal change pattern and enhance process control accuracy.