Research On Prediction Of Residual Use Life For Attitude Control System Of Satellite Based On Dynamic Fault Tree

The research on the life prediction of satellites becomes popular with the development of the long-life and high-reliable spacecrafts. In this paper, the research on the residual useful life(RUL)prediction of Attitude Control System(ACS) and critical units has been done. The failure-mechanism model of ACS is established based on the Dynamic Fault Tree(DFT); a method is proposed to predict the RUL of units which has different work-conditions to switch; an improved method which aim at eliminating the problem of the solution for spare gate is proposed to descript the failure probability of spare-unit; besides, a method which takes the configuration and work-mode into consideration is proposed to predict the RUL of ACS, and the numerical simulation is done to verify the effectiveness.Firstly, the failure model which is based on the DFT is established by analyzing the failure-mechanism of ACS. In order to reduce the complexity, top-down method is employed to analyze and model the subsystem of ACS, and then model the failure of ACS.Secondly, because the unit may have different work-conditions to switch, the method which takes work-conditions and switch-time into consideration is proposed to predict the RUL of the unit, and the numerical simulation of momentum wheel has demonstrated the effectiveness.Thirdly, in order to get the RUL of ACS, the modular approach is employed to solve the DFT model in order to reduce the complexity of solution. The BDD and DTBN are employed to solve the static module and dynamic module respectively. In this paper, a modified method which aims to solve the problem that the cumulative density function is unequal to 1 when time is infinite for spare unit is proposed to descript the failure probability density of spare unit, Eventually, an approach based on the established DFT model and the above methods to solve is proposed to predict the RUL of the ACS with different configurations, and numerical simulation of wheels subsystem with different configurations has demonstrated the effectiveness.Finally, taking the effect of system's work-mode on its RUL, The corresponding work-conditions and switch-time of units are analyzed in the given system's work-mode, and then the approach is proposed to predict the RUL of the systems with different work-mode based on the method proposed above to predict the RUL of units which has different work-conditions to switch. Eventually, the numerical simulation of four-axis skew wheels has demonstrated the effectiveness.