Research on Knowledge Mining Algorithm of Spacecraft Fault Diagnosis System

Abstract

The change of telemetry data of spacecraft is usually caused by tele-command or fault, which conforms to the causality model of remote-control input and telemetry output under different conditions of spacecraft. Traditional expert system relies on static knowledge of experts to diagnose telemetry parameters. In order to solve the problem of rule-based expert system knowledge acquisition and less manual intervention, considering the characteristics of spacecraft telemetry, this paper proposes an expert knowledge acquisition algorithm based on successful data envelope line and conditional probability from two dimensions of analog and digital quantities respectively. Through data mining of historical telemetry, this algorithm achieves the threshold of analogue quantities and automatic extraction of causal rules at different stages of product life cycle. The experimental results show that the algorithm is effective and the simulation value is more accurate than the product design index and the redundancy of causal rules is less. After knowledge mapping, the algorithm can be applied in the spacecraft fault diagnosis expert system.