Prediction Of The Remaining Life Of Satellite Key Components

Space satellites provide strong support to the fields of national defense,national security,national economy and people's livelihood,but long-term work in complex external environments such as thermal vacuum,strong electromagnetic radiation,and space debris will gradually degrade the performance of satellite components,which will lead to various failures.In order to ensure the safe and reliable operation of the satellite during the orbit mission,it is of great significance to study the remaining life prediction of the key satellite components.By analyzing the existing satellite component life prediction methods,this thesis finds that a single information source,a single degradation amount,and a single degradation model cannot accurately describe the performance degradation process of complex products such as satellites,which seriously affects the accuracy of the remaining life prediction.In response to these problems,this thesis uses a variety of methods based on information fusion to predict the remaining life of the key satellite components,which improves the accuracy of the prediction results.The specific research content is as follows:(1)Because a single information source only contains part of the information about the performance degradation of the key satellite components,the prior distribution of the degradation parameters is not unique.In order to reduce this uncertainty,this thesis proposes a weighted fusion model based on the three algorithms of correlation function,gray correlation analysis and average mutual information entropy.Through fusion processing of multi-source reliable information,the prior distribution of degradation parameters is obtained,and estimate the reliability of satellite components based on the distribution.Through the simulation experiments of 5 sets of momentum wheel axle temperature telemetry data,it is proved that the degradation parameter distribution obtained by the fusion method proposed in this thesis is*The research work in this thesis was supported by the National Natural Science Foundation of China(No.:61773016)and the Open Fund of the Key Laboratory of On-orbit Fault Diagnosis and Maintenance of Spacecraft(No.:SDML-OF2015004),the author expresses her heartfelt thanks.close to the actual distribution,and the estimation accuracy of satellite component reliability based on this parameter distribution has also been significantly improved,verifying the validity of the model.(2)For a single degradation model that cannot accurately describe the multi-stage characteristics of the equipment during the degradation process,this thesis proposes a multi-model fusion algorithm based on Bayes method which uses different degradation models to model the different degradation stages of the equipment during the degradation process.After weighted fusion of the state values and posterior probabilities of different models,the current operating state value of the device is obtained,and the model parameters can be predicted by using the Expectation Maximum(EM)algorithm to update the remaining life of the device.The gyroscope drift data and the vibration coupler vibration amplitude data are used for verification.The results show that the multi-model fusion algorithm proposed in this thesis can describe the degradation process of these two devices more accurately.(3)Aiming at the complex correlation in some degraded features of equipment,and the single degraded features cannot fully describe the problem of equipment degradation.This thesis proposes a multi-degradation fusion model to predict equipment life.Based on the main degradation information contained in the multiple degradation amounts,an evaluation index of the overall degradation trend of the equipment is constructed,and then the remaining life of the equipment is predicted based on this index.Experiment on the degradation data of bearing temperature,current,and lubricant of the satellite momentum wheel shows that the model can extract the main characteristics of the momentum wheel degradation from a variety of degradation amounts,and predict the remaining life of the momentum wheel based on this feature,which proves that the feasibility of the model.(4)In order to meet the actual needs of the project,this thesis designs and implements a life prediction software system for satellite key components based on mixed programming with C#and MATLAB.