Fault Diagnosis Of Satellite Attitude Control System Based On Artificial Intelligence

In recent years,aerospace technology has developed rapidly in our country,and the functions and structures of satellites are becoming more and more diversified.It is very important to guarantee the safety of these satellites.At present,the traditional artificial threshold detection can not meet the growing needs of satellite fault diagnosis.Fault diagnosis technology based on artificial intelligence has become one of the important research directions.Aiming at the attitude control system in the whole service cycle of satellite subsystem,this paper studies the fault diagnosis based on artificial intelligence methods.The main work and contributions of this paper are as follows:(1)This paper establishes a high-precision simulation model of satellite reaction wheel,in Matlab / Simulink,simulates its normal working conditions and possible fault conditions during satellite in orbit,collects data,and lays a data foundation for subsequent fault diagnosis.(2)For discrete data,a fault diagnosis model of limit learning machine based on particle swarm optimization is proposed.The method of data mining is used to solve the problem that the observer data does not change obviously when friction,slow change and other faults occur.In this paper,the improved particle swarm optimization algorithm is used to optimize the initial weight and threshold in the limit learning machine,which improves the classification performance of the limit learning machine and the accuracy of the model.(3)For time series data,a time series classification model based on long-term and shortterm memory artificial network(LSTM)is proposed to learn the long-term dependent nonlinear and linear relationship of the data.Because of the problem that the sample dimension of satellite simulation data is too large,the kernel principal component analysis method is adopted in this paper.Under the condition of ensuring the contribution rate of principal components,the sample dimension is reduced,the classification accuracy and efficiency of LSTM network are improved,and the fault diagnosis of sequential data is realized.To sum up,based on the simulation model of satellite reaction wheel,this paper puts forward the fault classification and diagnosis method,which can solve the fault diagnosis of two data types.The accuracy and effectiveness of the proposed fault diagnosis method are verified by experiments and comparative analysis of simulation data.