基于奇异值分解的冗余惯导系统故障诊断  

Fault Detection for Redundant Inertial Navigation System Based on SVD Method

作　　者：贾鹏[1] 张洪钺[1] 

机构地区：[1]北京航空航天大学自动化科学与电气工程学院,北京100083

出　　处：《宇航学报》2006年第5期1076-1080,共5页Journal of Astronautics

基　　金：国家自然科学基金(60234010)

摘　　要：为了验证奇异值分解法(SVD)在冗余惯性组件系统故障诊断中的实用性,进行了一系列的研究。以沿正十二面体对称侧面配置的六陀螺冗余惯性组件为平台,对2种基于奇异值分解的故障诊断方法进行了相应的实验,分析了不同类型故障下的故障诊断效果及其原因。实验结果证明了方法在实际应用中的有效性,2种方法均可以正确的诊断出故障。奇异值分解法中的方法2不仅可以检测一个陀螺的故障,还可以检测2个陀螺同时发生故障的情况,有更广泛的实用价值。The single valued decomposed （SVD） method applied in FDI of redundant inertial navigation system was mainly researched.2 FDI methods were discussed based on SVD; in which 2^nd method can detect the situation of 2 faults occur simultaneously, Based on the six-gyroscope redundant inertial measurement unit, different performances of these methods were evaluated; the possible causes which affect their fault detection capability were analyzed. The result proves that the SVD method is effective in practice.

关 键 词：奇异值分解法 故障诊断 冗余惯性组件 

分 类 号：V241.6[航空宇航科学与技术—飞行器设计]