基于小波神经网络的一类非线性系统的故障检测

针对一类非线性系统,提出一种小波神经网络(wavelet neural network,WNN)自适应故障检测方法。WNN具有较强的泛化能力及不同的激活函数。通过设计自适应状态观测器技术,利用小波神经网络观测器良好的观测性能来观测系统的当前状态,并将其应用于一类非线性系统中,实现故障检测与诊断。利用Lyapunov直接方法从理论上证明了小波神经网络故障检测观测器的稳定性,仿真结果亦表明了该非线性系统故障检测观测器的可靠性和稳定性。

Fault Detection for a Class of Nonlinear System Based on Wavelet Neural Networks

In this paper, an adaptive fault detection method based on wavelet neural network(WNN)is proposed for a class of nonlinear systems. Compared with BP(Propagation Back)neural networks, the WNN has a more generalization ability with different activation functions. By designing adaptive state observer technique, the current state of the system is observed by using wavelet neural network observer, and it is applied to a class of nonlinear system fault detection and diagnosis. The stability of the fault detection observer for the wavelet neural network is proved by using the Lyapunov''s direct method. Finally, the reliability and stability of the fault detection observer for the nonlinear system are demonstrated by simulation.