摘要
针对一类满足Lipschitz条件的非线性系统,采用小波动态神经网络设计非线性观测器,以实现系统的实时鲁棒故障检测。通过小波网络逼近非线性项提高状态估计的精度,并在动态网络中加入鲁棒项增强了对不确定干扰的鲁棒性。同时采用RBF神经网络(RBFNN)对残差序列进行实时预测,实现了故障预报功能。文中证明了观测器设计的稳定性和鲁棒性,仿真示例验证了该方法的有效性。
A nonlinear robust fault detection observer based on wavelets dynamic neural network (WDNN) is proposed for a class of Lipschitz nonlinear system in this paper. The wavelets neural network is used to approximate the nonlinear term of the system to improve the precision of states estimation, and a robust term is added into the dynamic neural network to improve the robustness of uncertain disturbances. The robustness and the stable convergence of the observer are also proved. While the RBFNN is utilized to predict the residuals, scheme for real-time robust fault detection and fault prediction is established. Simulation result showed the effectiveness of the proposed approach.
出处
《火炮发射与控制学报》
北大核心
2004年第1期32-36,共5页
Journal of Gun Launch & Control
基金
国家自然科学基金(60234010)
关键词
小波动态神经网络
信息处理技术
故障检测
鲁棒性
故障预报
information processing technique
wavelets dynamic neural network
fault detection
robustness
fault prediction
