基于滑模观测器和广义观测器的故障估计方法

Fault Estimation Approaches With Sliding Mode Observer and Descriptor Observer

针对受未知干扰影响的一类非线性系统,提出一种基于滑模观测器和广义观测器的执行器故障和传感器故障估计方法.首先通过线性变换将原系统解耦为两个降阶的子系统,其中一个子系统受执行器故障和干扰的影响,另一个含有传感器故障和干扰,进一步将后一个子系统转化为广义系统.对两类子系统分别设计滑模观测器和广义观测器,给出估计误差一致最终有界的条件,得到系统状态和未知干扰的估计值.然后,利用等效输出控制原理重构执行器故障,引入干扰补偿保证重构算法的鲁棒性,再根据广义观测器的结果获得传感器故障的估计值.最后,通过计算机仿真验证了本文方法的有效性.

In this paper, the problem of actuator fault and sensor fault estimation based on sliding mode observers and descriptor observers is addressed for a sort of nonlinear systems subject to unknown disturbances. First, the considered system is transformed into two reduced-order subsystems by linear transformation matrices. One subsystem is subject to actuator fault and the disturbance, and the other is affected by sensor fault and the disturbance, which is then further transformed into a descriptor system. A sliding mode observer and a descriptor observer are designed to process the above two subsystems, respectively. A sufficient condition is presented to ensure that estimation errors are uniformly bounded, and the estimates of system state as well as the unknown disturbance are obtained. Furthermore, by using the equivalent output injection concept, actuator fault reconstruction is achieved, whose robustness is ensured by employing the estimation of disturbance as compensation. The estimate of sensor fault is also obtained according to the output of the descriptor observer. Finally, a numerical example is employed to demonstrate the effectiveness and robustness of the proposed approach.