Design and simulation of hypersonic vehicles require simultaneous consideration of a variety of disciplines due to the highly coupled nature of the flight regime. In order to capture all of the potential effects on vehicle dynamics, one must consider the aerodynamics, aerodynamic heating, heat transfer, and structural dynamics as well as the interactions between these disciplines. While high-fidelity modeling techniques exist for each of these disciplines, the use of such techniques is computationally infeasible in a vehicle design and control system simulation setting for such a highly coupled problem. Many iterations of analyses may need to be carried out as the vehicle design matures, thus requiring quick analysis turn-around time. Additionally, the number of states used in the analyses must be small enough to allow for efficient control simulation and design. As a result, alternative approaches must be considered for vehicle simulations. This dissertation presents a fully coupled, reduced-order aerothermoelastic framework for the modeling and analysis of hypersonic vehicle structures. The reduced-order transient thermal solution used to obtain the instantaneous temperature distribution is based on the projection of the governing equations onto a modal subspace which is obtained via the proper orthogonal decomposition (POD). The proper orthogonal decomposition is used for the thermal problem due to its optimality properties which are described in the dissertation. The reduced-order structural dynamic solution is also based on projection of the governing equations onto a modal subspace. However, for the structural dynamics, the modal subspace is composed of a set of Ritz modes which include both free vibration modes and load- dependent Ritz vectors. In order to avoid the need to reassemble the temperature dependent stiffness matrix and thermal load vector at each time step, a technique is developed for directly updating these quantities as a function of the POD modal coordinates.
展开▼
