Multidisciplinary Design Optimization Of Two-Stage-To-Orbit Reusable Launch Vehicle

Multidisciplinary Design Optimization (MDO) is a design architecture developed since the last two decades, which is mainly focused on solving the design problems of large-scale complicated engineering systems. Because of its excellent computation characteristics, MDO has played an important role in the field of aeronautic & astronautic engineering design. Collaborative optimization (CO), as a typical MDO method, is well suited to multidisciplinary design environment, especially large-scale distributed analysis problem. When used in multidisciplinary environment, CO has several advantages over standard optimization method: reduction of the information transfer; elimination of large iteration loop; allowance of the use of corresponding subspace optimizers in different disciplinary analysis; a parallel optimization architecture which is readily operable on a suite of heterogeneous equipments; more natural fit to the current organization structure found in most institutes of aerospace and aeronautic design; participation of the disciplinary experts to best deal with specific disciplinary models. This paper is mainly focused on CO method. And its application to the design optimization of reusable launch vehicle (RLV) is realized satisfactorily.The major work in this paper includes:a) The development and mathematical framework of CO are introduced and its feasibility is argued by two problems. Two improving schemes are presented to solve the computational difficulty of CO.b) The feasibility of CO in the field of aeronautics & astronautics is validated through the application to the design optimization of ZN-1 meteorologic rocket, and its potential obtains further proof.c) Analysis models of aerodynamics, trajectory, mass and propulsion system are established. Based on these models, the initial values for the design variables are chosen.d) The collaborative optimization framework for RLV is set up. Based on this framework together with those disciplinary analysis models, the collaborative optimization of RLV is finally realized.