Fault Prognosis Approach for Satellite Attitude Control System Based on T-S Model and Time Series Analysis

A new method based on T-S fuzzy model and time series analysis is proposed for predicting faults in satellite attitude control system. Firstly, satellite attitude control system with nonlinearity and uncertainty is modeled. The residual errors can be obtained by fuzzy parity equation and they are only sensitive to the output of specific actuators (or sensors). Secondly, the time series of the output errors are used to build the autoregressive (AR) model. Therefore, the faults in the satellite attitude control system are predicted by using the AR model. Finally, the results of fault prognosis are given by fault probability. The confidence factor is determined which shows the confidence level of the fault prognosis.