| In the advanced launch vehicle design community, there exists considerable interest in fully reusable, two-stage-to-orbit vehicle designs that use ‘branching trajectories’ during their missions. For these reusable systems, the booster must fly to a predetermined landing site after staging occurs.; The solution to this problem using an industry-standard trajectory optimization code typically requires at least two separate computer jobs—one for the orbital branch from the ground to orbit (in some cases, this can be broken into two computer jobs) and one for the flyback branch from the staging point to the landing site. These jobs are tightly coupled and their data requirements are interdependent. In addition, the objective functions for each computer job differ and conflict.; This research produces a method to solve these distributed branching trajectory problems with respect to an overall system-level objective while maintaining data consistency within the problem. This method is used to solve the trajectories of two relevant two-stage-to-orbit vehicles: the Kistler K-1 and the Stargazer launch vehicles. Both of these vehicles require a powered flyback. Thus, optimization contingent on the feedback of the flyback fuel is a relevant part of this study.; The solutions of the branching trajectory problems via traditional methods, termed ‘One-and-Done’ and manual iteration, are compared with those involving the multidisciplinary design optimization techniques of fixed-point iteration, optimization-based decomposition, and collaborative optimization. Optimization-based decomposition was used to solve each problem; the K-1 trajectory includes a fixed-point iteration solution. The use of collaborative optimization as an solution technique for branching trajectories is introduced in the solution to each problem.; Results show that proposed method involving collaborative optimization and optimization-based decomposition performed well for both the K-1 and Stargazer branching trajectories. The use of these methods for the Kistler K-1 problem shows that an increase in payload weight of 1.0%, on average, could be obtained. Similarly, a reduction in Stargazer's dry weight of approximately 0.8% was achieved through the MDO methods. Conclusions concerning the method outline, comparisons of the method with differing solution techniques, staging flight path angle trends, and the automation of the optimization process are included. |
