Future high-speed air vehicles will be lightweight, flexible, and reusable. Ve- hicles fitting this description are subject to severe thermal and fluid dynamic loading from multiple sources such as aerothermal heating, propulsion sys- tem exhaust, and high dynamic pressures. The combination of low-margin design requirements and extreme environmental conditions emphasizes the occurrence of fluid-thermal-structural coupling. Numerous attempts to field such vehicles have been unsuccessful over the past half-century due par- tially to the inability of traditional design and analysis practices to predict the structural response in this flight regime. In this thesis, a high-fidelity computational approach is used to examine the fluid-structural response of aerospace structures in high-speed flows. The method is applied to two cases: one involving a fluid-thermal interaction problem in a hypersonic flow and the other a fluid-structure interaction study involving a turbulent boundary layer and a compliant panel. The coupled fluid-thermal investigation features a nominally rigid alu- minum spherical dome fixed to a ceramic panel holder placed in a Mach 6.59 laminar boundary layer. The problem was originally studied by Glass and Hunt in a 1988 wind tunnel experiment in the NASA Langley 8-Foot High Temperature Tunnel and is motivated by thermally bowed body panels designed for the National Aerospace Plane. In this work, the compressible Navier-Stokes equations for a thermally perfect gas and the transient heat equation in the structure are solved simultaneously using two high-fidelity solvers coupled at the solid-fluid interface. Predicted surface heat fluxes are within 10% of the measured values in the dome interior with greater differ- ences found near the dome edges where uncertainties concerning the exper- imental model's construction likely influence the thermal dynamics. On the flat panel holder, the local surface heat fluxes approach those on the wind- ward dome face due to a dome-induced horseshoe vortex scouring the panel's surface. Comparisons with reduced-order models of heat transfer indicate that they perform with varying levels of accuracy around some portions of the geometry while completely failing to predict significant heat loads in re- gions where the dome-influenced flow impacts the ceramic panel. Cumulative effects of flow-thermal coupling at later simulation times on the reduction of panel drag and surface heat transfer are quantified. The second fluid-structure study investigates the interaction between a thin metallic panel and a Mach 2.25 turbulent boundary layer with an ini- tial momentum thickness Reynolds number of 1200. A transient, non-linear, large deformation, 3D finite element solver is developed to compute the dynamic response of the panel. The solver is coupled at the fluid-structure interface with the compressible Navier-Stokes solver, the latter of which is used for a direct numerical simulation of the turbulent boundary layer. In this approach, no simplifying assumptions regarding the structural solution or turbulence modeling are made in order to get detailed solution data. It is found that the thin panel state evolves into a flutter type response char- acterized by high-amplitude, high-frequency oscillations into the flow. The oscillating panel disturbs the supersonic flow by introducing compression waves, modifying the turbulence, and generating fluctuations in the power exiting the top of the flow domain. The work in this thesis serves as a step forward in structural response prediction in high-speed flows. The results demonstrate the ability of high- fidelity numerical approaches to serve as a guide for reduced-order model improvement and as well as provide accurate and detailed solution data in scenarios where experimental approaches are difficult or impossible.
展开▼
