摘要
研究利用ＷａｖｅＡＲＸ神经网络构造间歇过程的非参数化逼近模型，在分析其逼近偏差的基础上提出一种非线性次优滤波器设计方法，用于间歇过程的工况监测。实例仿真研究证实了该方法的可行性和有效性。
A means to approximate the real batch process based on WaveARX neural network was studied. After a nonparametric approximate model was constructed, a nonlinear suboptimal filter was designed to generate the residuals for fault detection based on the analysis of the upper bound approximate error. The application result has proved the feasibility and effectiveness of the method.
出处
《控制与决策》
EI
CSCD
北大核心
1998年第2期151-155,共5页
Control and Decision
基金
国家自然科学基金
关键词
神经网络
间歇过程
工况监测
离散系统
wavelet transform, neural network, suboptimal filter, batch process, process monitoring
