Design of a robust fault-tolerant control algorithm based on failure mode effects criticality analysis for a three-axis satellite. (January 2019)

• Record Type: Journal Article
• Authors: Abedini, Mohammad Reza; Abedi, Mostafa
• Is Part Of: Proceedings of the Institution of Mechanical Engineers. Volume 233:Number 1(2019)
• Page Start: 91
• Page End: 110
• Publication Date: 2019-01
• Subjects: Three-axis satellite -- attitude control system -- adaptive -- sliding mode -- fault correction -- unwinding -- singularity -- estimation; Aeronautics -- Periodicals; Astronautics -- Periodicals; Airplanes -- Design and construction -- Periodicals; Aerospace industries -- Periodicals; 629.1
• DOI: 10.1177/0954410017727300
• Languages: English
• ISSNs: 0954-4100

Abstract

This paper proposes a robust fault-tolerant control algorithm for a three-axis satellite. In this regard, an adaptive sliding attitude control algorithm is suggested, which has the capability of fault estimation in the satellite actuators and correction of their effects. For this, the disturbances due to environmental effects and actuator failures and also the satellite unknown parameters are estimated by the adaptive updating law; the sliding mode algorithm compensates the errors due to estimation process. In the suggested design process, the sliding surface is selected so that the unwinding and singularity problems are solved, and also a compensator part is included to remove unstable equilibrium points. In this paper, the failure mode effects criticality analysis have been done to classify different failure modes of reaction wheel according to their severity and probability of occurrence. Accordingly, the critical failure modes and their effects at the control system level are derived. It is shown that the derived critical failures lead to small or severe variations in the generated torques of reaction wheels for which a supervision level will be proposed to correct their effects. Finally, different simulations are conducted to validate expected performance of the suggested algorithms.