A Traceability Technique of Spacecraft Interference Fault Based on Bayesian Networks

Because of the small sample size and complexity of the tracing data about spacecraft interference fault, it is not possible to perform identification evaluation relative to the large sample. However, it gains a lot of expert experience, on-orbit analysis data as prior information. So a traceability technique of spacecraft interference fault based o Bayesian networks is proposed in this paper. Firstly, analyzing and classifying the phenomena of spacecraft interference fault to determine the source location and interference mode of interference fault. Secondly, the Bayesian network learning algorithm and historical data are used to establish a traceability model for spacecraft interference fault. Then Fuse the available prior information to update the current model. Finally, based on the updated model, the Bayesian network inference method is used to identify the source and mode of disturbance from the measured sample data. It shows that the technology can quickly and effectively track the source of spacecraft interference and provide strong support for the healthy operation and fault diagnosis of spacecraft.