Fault Diagnosis Of Satellite Ground Station Receiving System Based On CBR And XGBOOST

The satellite ground station receiving system as an important part of the satellite communication system is responsible for receiving the radio frequency signals from the satellites.It is extremely essential that the system runs safely and stably.If any devices of ground station breaks down,and can't be eliminated effectively it will not only break the normal safe and stable work of the ground station system,but also cause the failure of the entire ground station system.Therefore,it is of great significance for satellite station equipment to perform fault location and fault diagnosis.However,due to the complex structure of the receiving equipment at the ground station,it is difficult to establish an accurate and complete model.Once the system fails,the positioning and diagnosis of fault are difficult.The traditional fault diagnosis method is inefficient and it is difficult to meet the requirements of rapid fault diagnosis and positioning of satellite ground station equipment.This thesis has carried out relevant research on the fault diagnosis and positioning of the ground station receiving system.The main work is as follows:1.For the effective feature extraction of fault feature signals,the use of feature extraction technology based on wavelet packet transform to solve the traditional wavelet transform and Fourier transform cannot be mutated The problem of effective feature extraction is that the aperiodic and non-stationary signals are extracted,and the simulation signal is used to verify the fault feature extraction method based on wavelet packet transform.2.Aiming at the problem of the low accuracy of the traditional single system-level fault diagnosis method,this thesis proposes a method of fault diagnosis for ground station receiving system based on joint case-based reasoning(CBR)and fault tree analysis(FTA).For the inefficiency of case reasoning stage and,a secondary search index strategy was proposed and a multi-level case library organization structure for satellite ground station failure was established.In case of incomplete case base and complex analysis of fault tree in case reasoning,it is proposed to use case reasoning to retrieve faults and then use fault tree diagnosis.Finally,taking the phenomenon of carrier lost-lock failure in a satellite communication ground station as an example,the method steps in this thesis are presented in detail,and the faults that do not exist in the case library are accurately diagnosed.After this search,the number of cases in the case base was increased,and the efficiency of subsequent search was improved,which validated the efficiency of the method.3.Aiming at the low efficiency of component-level fault diagnosis,a fault diagnosis method based on wavelet packet transform and XGBOOST algorithm is proposed.To solve the problem that the low-noise amplifier soft fault feature of the ground station is difficult to extract,the ‘db4' wavelet is used to extract the 8-bit fault feature vector from the voltage signal of the circuit output.This paper proposes the application of XGBOOST algorithm to the fault diagnosis of analog circuits.It automatically uses the CPU multi-threads for distributed computing,and integrates multiple weak classifiers to form a strong classifier.Simulate software simulates 11 different fault conditions of the low-noise amplifier circuit to verify the experiment.The experimental results show that using the wavelet packet transform and XGBOOST algorithm in the test reach 98 % in accuracy,recall rate,F1 value,significantly better than the traditional fault feature extraction methods and fault diagnosis methods.4.Design software system based on CBR and FTA to improve the operability of this method.Based on case-based reasoning and fault tree analysis,the system application layer,business logic layer,and knowledge data layer were constructed.The intelligent fault diagnosis of the satellite ground station receiving system was realized.