基于神经网络观测器的卫星姿态控制系统陀螺故障诊断

Gyroscopic fault diagnosis method for satellite attitude control system based on neural network observer

摘要:

针对基于解析模型的卫星姿态控制系统陀螺故障诊断方法存在设计复杂、参数求解困难的问题,提出一种基于神经网络观测器的陀螺故障诊断方法。由系统内的冗余关系导出故障诊断逻辑,实现对陀螺故障的检测和隔离；同时利用先验模型知识和神经网络的非线性建模特性对陀螺故障进行估计。仿真结果表明,该方法能够实现对陀螺故障的检测、隔离和估计。

Abstract:

In order to solve the problem of complicated design and difficult parameter solving for the fault diagnosis method of satellite attitude control system based on analytic model, a new method of gyro fault diagnosis based on neural network observer is proposed. The fault diagnosis logic is derived from the redundancy relationship in the system to detect and isolate the fault of the gyroscope. At the same time, the fault of the gyroscope is estimated by using the prior model knowledge and the nonlinear modeling characteristics of the neural network. Simulation results show that the method can detect, isolate and estimate gyro faults.