摘要
针对缺乏故障样本的卫星姿态控制系统(ACS)故障诊断问题,研究一种基于迁移成分分析(Transfer Component Analysis, TCA)的卫星姿态控制系统故障定位方法。以设计阶段的卫星模型作为标称模型,以在轨卫星作为实际模型;利用标称模型和实际模型的健康数据拟合ACS系统模型,生成残差信号,提取故障特征。以标称模型残差特征、实际卫星残差特征分别作为源域、目标域数据,利用TCA方法进行特征变换,以减小源域和目标域的残差特征分布差异,从而解决在轨卫星的故障诊断问题。基于三轴气浮台的半物理仿真试验表明,在无故障样本情况下,本文所设计的引入TCA的卫星ACS系统故障定位方法明显提高了故障部件定位的准确性。
In this paper, a Transfer Component Analysis(TCA) based on satellite attitude control system fault diagnosis method is proposed for satellite attitude control system(ACS) in absence of fault samples. The designed satellite model and the satellite in orbit are taken as the nominal model and the actual model, respectively. The health data of the nominal model and the actual model are used to fit the ACS system models which are used to generate residual signals and then extract fault features. The residual signals of the nominal model and the actual satellite are used as the source domain and target domain data respectively. TCA is applied for reducing the distribution difference of residual features between the source domain and the target domain, such that the fault diagnosis problem is solved for the satellite in orbit. A semi-physical simulation experiment based on a micro triaxial air bearing table simulation system shows that the proposed fault location method can significantly improve the accuracy of fault component location in absence of fault samples.
作者
何漫丽
程月华
韩笑冬
王泽
叶正宇
He Manli;Cheng Yuehua;Han Xiaodong;Wang Ze;Ye Zhengyu(College of Astronautics,Nanjing University of Aeronautics and Astronautics,Nanjing 211100,China;College of Automation Engineering,Nanjing University of Aeronautics and Astronautics,Nanjing 211100,China;China Academy of Space Technology,Beijing 100094,China)
出处
《航天控制》
CSCD
北大核心
2022年第2期28-34,共7页
Aerospace Control
基金
国家自然科学基金面上项目(61972398)。
