基于小波密度函数估计的一类非线性控制系统监视方法

针对一类基于反馈线性化补偿的非线性控制系统，提出一种基于小波密度函数估计的过程监视及故障诊断策略。以正常操作范围内收集的数据，由于实施反馈线性化策略，降低了过程的非线性程度，使得线性多变量统计分析更有效。当操作工况发生变化时，利用小波变换进行非线数概率密度函数估计，比常规的密度参数估计结果具有更强的适应性。从计算机控制系统记录的大量数据中挖掘其中隐含的知识，可以对过程异常状态进行及时检测和故障诊断。通过仿真证实了该方法的可行性。

Monitoring of Nonlinear Control Systems Based on Wavelet Density Estimations

Wavelet density function estimations based monitoring approach to a class nonlinear systems controlled with the feedback linearizing strategy is suggested.Multivariable statistical analysis of data sets collected from the normal operating mode is more effective,which process nonlinearity is reduced by feedback linearizing.When the operating point changes,wavelet based density function estimation is better than the usual nonparameter estimation.It is very important that data mining and hidden knowledge discovery from these recordings on computer control systems are applied to monitor abnormity operating states and fault diagnosis.The feasibility of proposed method is illustrated by simulations.