Analytical redundancy based fault detection of gyroscopes in spacecraft applications

Abstract

Spacecraft are generally designed with very high reliability to operate for many years through fault avoidance practices. But, in spite of fault avoidance, faults do occur. Hence there is a need for on-board fault detection and isolation without creating significant degradation in the spacecraft services. Analytical redundancy or model reference method of fault detection provides such a solution as most spacecraft are now equipped with an on-board computer, which is an ideal platform for implementing the analytical redundancy based fault detection algorithms. In this paper, analytical redundancy based fault detection technique has been applied for detecting the faults in gyroscopes used in a three-axis stabilized low earth orbiting satellite like the Indian Remote Sensing (IRS) spacecraft. The scheme proposed is an observer based on a dynamic model having attitude and rates as states and the gyro sensed rates and the horizon sensor outputs as measurements. The torque computed by the AOCS processor is used as the control input for this observer. Eigen-structure assignment approach has been applied to the design of this dedicated fault detection observer. Design exercise has been carried out using MATLAB tools. Spacecraft dynamics and kinematic equations, the models for gyroscopes and the other sensors, the on-orbit normal mode controller have been included in the simulations. Extensive simulation studies have been conducted to validate the design and the results are presented. It is believed that this approach will help in achieving the larger goal of autonomy in spacecraft.