Reliability Evaluation And Life Prediction Of Inertial Actuator Based On Degradation Model

The inertial actuator is the core of the attitude orbit control system of satellites and other inertial models,and it is also the basis for accurately controlling the attitude of the satellite,ensuring the stable operation of the satellite in orbit,and completing various design functions.Momentum wheel,a commonly used inertial actuator for small and medium-sized satellites,has the characteristics of small sample size,high cost,complex failure mechanism,and missing life data.The research work on the reliability evaluation and life prediction methods of the momentum wheel is of great significance to the full utilization of the resources loaded on the satellite,the planning of spacecraft missions and the reduction of research funding.This thesis takes the typical inertial actuator—momentum wheel as the research object,deeply analyzes its potential failure modes,failure causes and failure mechanisms,and determines the performance degradation indicators of its key failure modes.On this basis,complete the research on the reliability evaluation and life prediction methods of inertial actuators based on the degradation model.The main research contents are as follows:(1)Based on the structure and working principle of the momentum wheel,conduct failure mode,mechanism,impact and hazard analysis to determine the "weak link" and key performance degradation indicators of the momentum wheel.(2)Carry out research on reliability modeling and evaluation of inverse Gaussian process considering individual differences.Establish a degradation model based on the standard inverse Gaussian process for the remaining amount of lubricant in the momentum wheel,and introduce random variables into the degradation model to characterize the individual differences of different samples: use the normal distribution to describe the second parameter of the standard inverse Gaussian process.The Bayes method is used to estimate the parameters of the standard inverse Gaussian model and the inverse Gaussian model considering the individual differences of products,and finally complete the reliability evaluation of the momentum wheel.(3)Research on reliability modeling and evaluation under multiple degradation conditions based on Copula function.Aiming at the correlation of multiple degradation indicators,the momentum wheel degradation models based on single,two and three performance indicators are established respectively.The maximum likelihood estimation method is selected to estimate the parameters in the Copula function,and the optimal Copula function is selected to complete the reliability modeling and evaluation research based on the Copula function under the condition of multiple degraded variables.(4)Life prediction method based on EMD-SVR and degradation data.Aiming at large sample size degradation data,a life prediction method based on empirical mode decomposition(EMD)and support vector regression(SVR)is proposed.The original sequence is decomposed by EMD method and the SVR model is established,and the support vector regression function corresponding to each eigenmode component is obtained.The function value sequence corresponding to the support vector regression function is cumulatively reconstructed and the new sequence is established to establish an SVR model to further obtain the support vector regression function.Using the support vector regression function and the current degradation data of the momentum wheel,the life prediction of the momentum wheel based on EMD-SVR is realized.