Research On Key Technologies Of Satellite System Fault Diagnosis

In recent years,satellite network technology has been continuously improved,and the internal structure of satellites has become more complicated.The high-speed flight and on-board load limitation of different orbiting satellites have brought many new problems for satellite fault diagnosis and recovery.In such a satellite system,when a single satellite fails due to an external attack or internal software abnormality,it needs to collect information through the ground station,and spends a lot of time and resources to find faults within the satellite and repair them,and Network routing and service execution throughout the satellite system are affected.Aiming at the above problems,this thesis proposes an expert system fault diagnosis method based on improved confidence rule base.This method optimizes the rule structure of the confidence rule base and uses the improved K-nearest neighbor algorithm to correct the misdiagnosis rate of fault diagnosis.Reduce computational complexity and improve retrieval efficiency.At the same time,the fuzzy dynamic fault tree method is used to analyze the reliability of the system,and a more detailed and accurate system reliability model is established,so that the internal components respond quickly after fault diagnosis and implement the recovery strategy.The work of this thesis is as follows:(1)For the fault rule base system,the fault analysis reasoning is required.For each input,it is necessary to traverse the entire rule base to calculate the rule weight,which leads to the system's reasoning efficiency is not ideal.A clustering analysis method based on weighted principal component distance is proposed.The fault diagnosis of the rule base structure can avoid redundant search and improve the efficiency of reasoning.Finally,the improved K-nearest neighbor algorithm is used to feedback the correct rate of diagnosis,and the regression data and classification data are experimentally verified to reduce the misdiagnosis rate.(2)Perform fault information statistics on the component information inside the system,and determine the component failure frequency and fault propagation based on the fault conclusion confidence distribution of the rule set in the fault diagnosis expert system,so as to determine the reliability of each component of the system.Due to the limited hardware resources of the system,the components with low reliability and faults are selected,and the fault diagnosis algorithm is adopted first.When a fault occurs inside the node,fault diagnosis is performed according to the component information in time to achieve the purpose of rapid response and recovery,reduce the time of fault recovery,and improve recovery efficiency.By combining the fault diagnosis mechanism of the satellite node system,it is possible to quickly diagnose the fault and provide a corresponding recovery strategy when the satellite system is attacked or an abnormality occurs in the internal components,thereby shortening the lengthy of the entire satellite network due to node failure.Restore time and increase the availability of satellite networks.