基于小波Elman神经网络的活塞环渗氮硬化质量预测控制

针对活塞环渗氮硬化工序建模困难的情况，通过主成分分析法（PCA）提取氮化工序特征参数，降低了质量模型输入样本维数，建立了基于小波Elman神经网络的活塞环制造关键工序质量预测模型，实现了工序过程质量波动趋势的预测，为后续的工艺优化和质量改进奠定基础。结果表明，该方法可以有效地改进渗氮硬化工序的质量控制，质量预测模型对输出质量特征值的预测准确率达到89%，具有比标准Elman网络更好的预测精度和收敛速度.

Quality Prediction and Control for Nitride Hardening of Piston Rings based on Elman Neural Network and Wavelet Transform

Aiming at the problem of difficult modeling for nitride hardening of piston rings, the features of nitride process are extracted using principal component analysis method and the input samples’ dimension of quality model are reduced, then quality prediction model of key process for piston rings’ manufacturing is built based on wavelet Elman neural network. It can not only effectively predict the process quality fluctuation, but also lay the foundation for further technics optimization and quality improvement. The experimental results show this method can obviously improve the quality control of nitride hardening, and the quality characteristic value can be predicted accurately more than 85 percent. Particularly, the wavelet Elman neural network converges more quickly , and its accuracy is higher than the normal Elman neural network.