基于自适应模糊神经网络的非线性系统鲁棒故障诊断

Robust Fault Diagnosis Based on Adaptive Fuzzy Neural Network for Nonlinear Systems

针对一类非线性闭环稳定系统,在考虑噪声干扰和系统参数不确定误差条件下,用自适应模糊系统进行非线性补偿,提出了一种基于自适应模糊状态观测器实现状态观测和故障检测,再由 RBF 神经网络故障逼近器实现故障在线跟踪的鲁棒故障诊断方法。为确保观测器和诊断系统的鲁棒性,给出了闭环系统在有界噪声干扰和系统不确定误差下的稳定性定理,并进行了证明。

Aiming at a kind of nonlinear closed-loop stable systems,a robust fault diagnosis method based on adaptive-fuzzy state observer and neural network learning structure is presented under external disturbance and internal system parameter error,the adaptive-fuzzy systems is adopted to repair nonlin- ear parts and RBF neural network is designed to track and diagnose faults.To guarantee the robustness of observers and diagnosis systems under disturbance and system parameter error,stability theorem on detection and diagnosis is proposed and proved as well.