基于广义未知输入观测器的鲁棒故障估计方法

当干扰存在时，有效地估计故障且放松故障的限制条件需要进一步的研究，为此针对含未知干扰的非线性连续系统的鲁棒故障估计问题提出一种广义未知输入观测器方法。首先，将执行器故障向量和传感器故障向量与原系统状态向量组成广义系统，放松对故障类型的限制，对此广义系统设计未知输入观测器解耦干扰，保证鲁棒性的同时估计出状态变量、执行器故障及其一阶微分和传感器故障。然后通过解线性矩阵不等式（LMI）给出估计误差渐近收敛的条件。最后，在MATLAB 的simulink平台上用三叶片水平轴风力模型仿真验证本文观测器的故障估计有效性鲁棒性。

Robust fault estimation approach based on generalized unknown input observer

Estimating faults effectively and relaxing the constraints of faults when disturbance exists need further study, this paper proposed a generalized unknown input observer method aiming at nonlinear systems with unknown disturbance. First, actuator fault vector and sensor fault vector together with states of original system formed a new generalized system, relaxing the constraints of faults. Designing unknown input observer to decoupling disturbance, the observer can estimate states and actuator fault and its first order differential vector and sensor fault. Second, The conditions of asymmetric convergence of the estimation error is given by solving the linear matrix inequality(LMI). Finally, a three-blade horizontal-axis turbine model demonstrated the effectiveness and robustness of proposed approach by MATLAB simulations.