Date of Award

Spring 1994

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Mechanical & Aerospace Engineering

Program/Concentration

Aerospace Engineering

Committee Director

Oktay Baysal

Committee Member

Colin P. Britcher

Committee Member

David S. Miller

Committee Member

Osama A. Kandil

Abstract

A computational method is developed to solve the coupled governing equations of an unsteady flowfield and those of rigid-body dynamics in six degrees-of-freedom (6-DOF). This method is capable of simulating the unsteady flowfields around multiple component configurations with at least one of the components in relative motion with respect to the others. Two of the important phenomena that such analyses can help us to understand are the unsteady aerodynamic interference and the boundary-induced component of such a flowfield. By hybridizing two dynamic domain decomposition techniques, the grid generation task is simplified, the computer memory requirement is reduced, and the governing equations of the rigid-body dynamics are simplified with certain assumptions. Three dimensional, unsteady Navier-Stokes equations are solved on each of the subdomains by a fully-vectorized, finite-volume, upwind-biased, and approximately-factored method. These equations are solved on the composite meshes of hybrid subdomain grids that can move with respect to each other. Hence, the present method combines the advantages of an efficient, geometrically conservative, minimally and automatically dissipative algorithm with the advantages and flexibility of the domain decomposition techniques. Several measures that reduce the numerical error are studied and compared with the exact solution of a moving normal shock in a tube. This solution compares very well with the analytic solution of the isentropic equations. It is concluded, that as a minimum measure, the connectivity of nonconservative overlapped scheme needs to be second-order accurate for spatial and temporal discretizations, as well as for the moving subdomain interpolations. Furthermore, the CFL numbers should be restricted to below unity, if affordable, for flows with high flow gradients. The method is further scrutinized by simulating the flow past a sinusoidally pitching airfoil, and the flow past a sinusoidally pitching and plunging airfoil. The results of the former case are successfully compared with the experimental data. The final two-dimensional case is the separation of a store from an airfoil along a prescribed path. As the first three dimensional case, the flowfield past an oscillating cylinder near a vertical wall is simulated. Prior to coupling it with the flowfield equations, the 6-DOF trajectory method is validated by successfully comparing the path it predicts with the one used in a captive trajectory testing. Finally, a rigid-body dynamics method is used to predict the aerodynamically determined trajectory of a store dropped from its initial position under a wing. The results of the present investigation contribute to the understanding of the unsteady aerodynamic interference and the boundary-induced component of such a flowfield. However, its main contribution is the newly proposed computational method for flows involving relative boundary motions.

DOI

10.25777/zfx4-8p98

Share

COinS