Date of Award

Winter 1999

Document Type


Degree Name

Doctor of Philosophy (PhD)


Mechanical & Aerospace Engineering


Aerospace Engineering

Committee Director

Oktay Baysal

Committee Member

Thomas E. Alberts

Committee Member

Osama Kandil

Committee Member

David E. Reubush


Relative-moving boundary problems have a wide variety of applications. They appear in staging during a launch process, store separation from a military aircraft, rotor-stator interaction in turbomachinery, and dynamic aeroelasticity.

The dynamic unstructured technology (DUT) is potentially a strong approach to simulate unsteady flows around relative-moving bodies, by solving time-dependent governing equations. The dual-time stepping scheme is implemented to improve its efficiency while not compromising the accuracy of solutions. The validation of the implicit scheme is performed on a pitching NACA0012 airfoil and a rectangular wing with low reduced frequencies in transonic flows. All the matured accelerating techniques, including the implicit residual smoothing, the local time stepping, and the Full-Approximate-Scheme (FAS) multigrid method, are resorted once a dynamic problem is transformed into a series of “static” problems. Even with rather coarse Euler-type meshes, one order of CPU time savings is achieved without losing the accuracy of solutions in comparison to the popular Runge-Kutta scheme. More orders of CPU time savings are expected in real engineering applications where highly stretched viscous-type meshes are needed.

The applicability of DUT is also extended from transonic/supersonic flows to hypersonic flows through special measures in spatial discretization to simulate the staging of a hypersonic vehicle.

First, the simulations in Mach 5 and Mach 10 flights are performed on the longitudinal symmetry plane. A network of strong shocks and expansion waves are captured. A prescribed two-degrees-of-freedom motion is imposed on the booster and the adapter to mimic the staging.

Then, a 3-D static Euler solver with an efficient edge-based data structure is modified for time-accurate flows. The overall history of aerodynamic interference during the staging in Mach 5 flight is obtained by an animation method, consisting of six static solutions along the assumed stage path. From the animation method, the following conclusions are made. After the booster and the adapter move away from the research vehicle by 60% vehicle length, their effects on the research vehicle are confined to the wake flow of the research vehicle. The aerodynamic forces on the research vehicle converge to the values in free flight when the booster is away from the research vehicle by 1.77 times vehicle length. The aerodynamic interference is a highly nonlinear function in terms of the distance between the vehicle, the booster, and the adapter.

Finally, two dynamic computations are performed when the booster and the adapter are extremely close to the research vehicle. It is observed from these 3-D dynamic computations that as the stage separation advances, the aerodynamic interference becomes less sensitive to further relative motions.