静电陀螺故障诊断的探讨

Exploration on Failure Diagnosis of Electrostatic Suspended Gyro

基于最优估计概念和ＡＩＣ准则，对静电陀螺随机漂移数据用递推算法建立时变系数ＡＲ模型，进而对时变系数和其规范化标准差，以及模型残差的方差进行讨论，根据系数和规范化标准差变化情况，判断陀螺是否存在奇异信号或故障。

Based on the idea of optimum estimation and AIC criterion, an AR model with time varying coefficients is established for random drift rate of electrostatic suspended gyro (ESG) by recursive algorithm. Further more, the time varying coefficients, their canonical standard errors and models residual variance are discussed. According to the variant condition of the coefficients and their canonical standard errors we distinguish that whether or not there is singular signal (or failure).