基于估计状态的非线性时滞系统的智能故障诊断

The Intelligent Fault Diagnosis of the Nonlinear Time-delay System Based on Estimation

针对一类状态不可测且模型未知的非线性时滞系统,提出了一种基于神经网络的故障诊断方法.该方法利用两个RBF神经网络分别设计了状态观测器和故障估计器用来直接估计系统的状态及故障,最后应用Lyapunov理论,论证了状态估计误差、故障估计误差、参数的收敛性以及系统的稳定性.结果表明该方法实用有效,对解决一类非线性时滞系统的故障诊断问题具有一定的实际意义.

A neural network fault diagnosis method for a class of nonlinear time-delay systems with uncertain state and unknown model is presented. This method designs the state observer and fault estimation respectively by two RBF neural networks to estimate the states and faults of the system directly. In the end, the state estimation error, fault estimation error, the convergence of the parameters and stability of the system are demonstrated by the Lyapunov theory. The results show that this method is practical and effective, and has practical significance to solve the fault diagnosis problem of a class of nonlinear systems with time delay.