Fault Detection for Satellite Reaction Wheel Based on Zonotopic Set-Membership Estimation

Abstract

A fault detection method is proposed based on zonotopic set-membership estimation to detect faults in the reaction wheel. First, considering the unknown but bound noise, a zonotope-based estimation is designed based on the nonlinear model of the reaction wheel using the mean value theorem. Then, the residuals and the adaptive threshold are based on the a priori zonotopic estimation, used to detect the fault in parameters, including the Coulomb friction torque coefficient, the bus voltage, and the motor torque coefficient. Finally, the effectiveness of the proposed method is verified by numerical simulations. The simulation results illustrate that the proposed methods can accurately detect different fault scenarios.