Aerodynamic and Aeroelastic Applications of a Parallel Multigrid Unstructured Flow Solver

The objective of this study is to develop a flow solver that can be used by aerodynamic design environments to analyze a variety of candidate designs in a relatively short period of time. To meet this goal, a combination of flow solvers developed by the second author, AIRPLANE and FLO77, which are well known for their robustness and accuracy, was parallelized. These flow solvers use unstructured grids to implement the numerical discretization of the governing equations and hence offer the flexibility of handling a variety of geometries through an automated grid generation process. The parallel flow solver was used to analyze aircraft configurations in transonic flight, and also extended to incompressible flows to perform aerodynamic simulations of sail geometries used in the Americas Cup. The turnaround time of the flow solver is typically under 5 minutes and the lift and drag converge in about 50-75 multigrid cycles. To obtain more realistic simulations for the sail configurations, aeroleastic simulations were performed by coupling the flow solver to the commerical finite element package, NASTRAN. Introduction Numerical experiments to analyze fluid flow problems are finding increasing use in engineering environments. The development of robust and accurate numerical methods to obtain steady and unsteady flow solutions 12–14 has enabled design environments to incorporate computational methodologies into the design cycle. Growth in computing power and parallel computing environments have added fuel to the development of new and more accurate numerical algorithms and coupled with their ability to provide detailed descriptions of the quantities of interest, computational methods have become an important component in the armor of modern day designers. RANS based simulations are gaining increasing acceptance within the aeronautical community as an alternative to experimental methods, but the lack of completely automated grid generation tools and an incomplete understanding of the phenomenon of turbulence inhibits their widespread use. Inviscid assumptions on the governing flow equations have helped aerodynamic design environments to take advantage of the developments in grid generation and flow solution methods. The accuracy and convergence properties of Finite Volume 1 , Finite Element 2, 3 and Finite Difference 4 approximations to the Euler equations have received much attention over the last two decades and hence robust and efficient numerical algorithms are available for inviscid calculations. Generation of multi-block structured grids around complete aircraft configurations , while seemingly tedious, is still a tractable problem 5, 6. Unstructured grid generation methods have become more robust …