Multidisciplinary Multimissions Design And Optimization For Truss Geostationary Orbit Satellite Bus

With the development in the application s of communication satellites, trussgeostationary satellite platform has been a researched focus in order to satisfy the designrequirements of mass and stability of satellite platform. Satellite platform is complexsystem engineering, including many disciplines or subsystems such as geometry, structure,propulsion, control and thermal, which coupled with each other. In order to improve thedesign efficiency, shorten the design cycle and reduce the design cost under considerationof the interdisciplinary couplings, multidisciplinary design and optimization (MDO) fortruss geostationary orbit satellite platform has drawn a lot of attention, and it is of greatresearch value.This thesis studies on design missions of truss geostationary orbit satellite platform. Itresearches modeling and analysis method of design mission, MDO strategy and modularMDO software for truss geostationary orbit satellite bus. The main contents of this thesisare followed by:The research significant and state of art of geostationary orbit satellite platform andMDO are summarized. It provides the basis and assumptions for MDO of trussgeostationary orbit satellite platform, and describes the development of metamodel-baseddesign and optimization, which support the rest research of this thesis.For three design missions, including parameters optimization of geometry-structure,thrusters layout and instruments layout, this thesis researches integrated design system ofgeometry-structure, propellant coupling modeling and analysis and instruments layoutoptimization strategy. The analysis and optimization mathematical models are built, whichprovides model basis for MDO modeling of truss geostationary orbit satellite platform.Through statistics parameters and disciplines of three design missions and analysisrelationship of data, the design structure matrix of truss geostationary orbit satelliteplatform is proposed. Three optimization elements of truss geostationary orbit satelliteplatform are chosen, and then build MDO model.In order to deal with computational-intensive models in truss geostationary orbit satellite platform analysis, this thesis proposed surrogate-based efficiency optimizationstrategy based on sequential radial basis function (SEO-SRBF). As other adaptivemetamodeling methods, this strategy generated new samples around optimal point toimprove the approximation accuracy using significant sampling space method. Thecomparison results of benchmark tests and engineering applications indicate SEO-SRBFshows much better performance on optimization efficiency compared with well-knownMBDO strategies.According to mathematical characteristics of truss geostationary orbit satelliteplatform MDO, this thesis proposed SEO-SRBF-based optimization strategy for problemswith mixed-discrete variables and multiobjectives (SRBF-MDC). SRBF-MDC takescontinuous relaxation instead of mixed-discrete variables, and uses the weighted l2normmethod and modified pareto fittness method to deal with multiobjectives. In addition, thisstrategy builds RBFs for computational-intensive objective and constraints, and updates themetamodels by significant sampling space. Through application of benchmark tests andengineering problems, the comparison results between SRBF-MDC and other well-knownMBDO strategies indicate SRBF-MDC can obtain much better optimal result under thesame computation cost.In order to deal with MDO for truss geostationary orbit satellite platform, based on theresearch of mission modeling and SRBF-MDC, modular multidisciplinary optimizationsoftware for truss structural satellite platform (MMOS-TSSB) is developed. Based onModelCenter, this software packaged the models of disciplianries and integratedSRBF-MDC optimizer, which developed through C#programming. Finally, the MDO fortruss geostationary orbit satellite platform is completed by this software, and theoptimization results satisfy all design requirements.