Optimal sensor placement for momentum wheel fault diagnosis using Bayesian network

Abstract

This paper investigates the problem of optimal sensor placement for momentum wheel fault diagnosis. An optimal sensor placement algorithm for momentum wheel is proposed by utilizing Bayesian network and effective independence method. First, a fault diagnosis model that can effectively diagnose faults of the momentum wheel in the spacecraft attitude control system is established based on the Bayesian network. Then, an optimal sensor placement algorithm is proposed by using the mutual information theory and the effective independence method. Finally, to validate the effectiveness of the proposed algorithm, an application to the momentum wheel is carried out to perform the sensor optimization placement.