Anomaly detection of spacecraft attitude control system based on principal component analysis

Abstract

Based on the high fault rate of the attitude control system in the process of spacecraft running, this paper analyzes the variation tendency, characteristics and regularities of different types of data generated by the attitude control system. In this paper, we construct the principal component analysis (PCA) model to express the correlations among the variables using the data collected with normal operation. Then we propose the algorithm of PCA-based anomaly detection of the attitude control system and judge the state of the system by monitoring multi-variable statistics which can implement the performance monitoring of the orbiting spacecraft. The experimental results indicate the feasibility and validity of the method to detect the abnormality of the spacecraft attitude control system based on multi-variable statistics of the PCA model.