卫星姿态控制系统执行器微小故障检测方法

针对卫星姿态控制系统执行器微小故障检测问题，提出一种基于神经网络干扰观测器的微小故障检测方法。该方法利用卫星姿态控制系统内的冗余关系，分别构建陀螺干扰观测器和干扰力矩观测器，对系统内的测量误差、扰动等进行估计，并对故障检测观测器进行扰动补偿，提高对执行器微小故障的检测能力。仿真结果表明，与基于解析模型的方法相比，该方法能够较精确地对解析模型的误差进行补偿，明显降低了检测阈值，实现了对扰动掩盖下的微小执行器故障检测。 

Small fault detection method for actuator of satellite attitude control system

Aimed at the problem of small fault detection for actuator of satellite attitude control system, a small fault detection method based on neural network disturbance observer is proposed. By using the redundant relationship in the satellite attitude control system, a gyro disturbance observer and disturbance torque observer are constructed to estimate the measurement error and disturbance torque in the system, and the fault detection observer is compensated for disturbances to improve actuator's small fault detection capability. The simulation results show that, compared with the analytical model based method, this method can compensate the error of the analytical model more accurately, reduce the detection threshold significantly, and realize the small fault detection for actuator under disturbance concealment.