Research On Multi-source Information Bayesian Melding Method For Satellite Uncertainty Modelling

There is a large amount of uncertainty in the development and operation of satellites,which directly affects the satellite performance and lifetime.It is necessary to fully consider various types of uncertainties in the design phase and carry out the reliability-based design optimization for comprehensive benefits.Among them,how to accurately model various uncertainties is the primary problem to be solved.Aiming at the characteristics of the satellite multi-level system integration,this paper proposes a multi-level multi-source information integration method,which provides a new idea for the satellite uncertainty modeling.The main research contents are as follows:Firstly,the research progress of the satellite uncertainty modelling and multi-source information integration methods is reviewed at home and abroad.In particular,the Bayesian Melding Method is deeply analyzed from the aspects of distribution modelling,information integration and updating.Besides,the drawbacks of the existing methods are discussed in terms of the method itself and applicable conditions.Secondly,the multi-level multi-source information integration method is proposed.The existing problems of the iterative Bayesian Melding method are improved.Discrete and continuous distributions are considered.The proposed method can fully integrate the multi-source information for the multi-level systems.On this basis,this paper proposes a multi-source information integration strategy for the system with partial distribution information missing.These methods are tested by numerical examples and actual satellite uncertainty modeling problems.The results verify the feasibility and effectiveness of the proposed methods.Finally,based on the proposed method,the application research of satellite reliability-based design optimization is carried out.The satellite reliability model is reasonably established by integrating the expert experience information or on-orbit failure data of satellite components.Based on the reliability modelling method,the satellite reliability-based design optimization is carried out,which effectively improves the system reliability under multiple design constraints,and provides a basis for designers to carry out further analysis.In this paper,the multi-level system information integration is deeply studied,which effectively solves the problem of multi-source information integration in the satellite uncertainty modeling,and provides the uncertain distribution model basis for the further satellite reliability-based and robustness-based design optimization.