Experiences in accurately predicting time-dependent flows

Abstract

As computational fluid dynamics matures, researchers attempt to perform numerical simulations on increasingly complex aerodynamic flows. One type of flow that has become feasible to simulate is massively separated flow fields, which exhibit high levels of flow unsteadiness. While traditional computational fluid dynamic approaches may be able to simulate these flows, it is not obvious what restrictions should be followed in order to insure that the numerical simulations are accurate and trustworthy. Our research group has considerable experience in computing massively separated flow fields about various aircraft configurations, which has led us to examine the factors necessary for making high-quality time-dependent flow computations. The factors we have identified include: grid density and local refinement, the numerical approach, performing a time-step study, the use of sub-iterations for temporal accuracy, the appropriate use of temporal damping, and the use of appropriate turbulence models. We have a variety of cases from which to draw results, including delta wings and the F-18C, F-16C, and F-16XL aircraft. Results show that while it is possible to obtain accurate unsteady aerodynamic computations, there is a high computational cost associated with performing the calculations. Rules of thumb and possible shortcuts for accurate prediction of massively separated flows are also discussed.

Aeroacoustic Simulation of a Nose Landing Gear in an Open Jet Facility using FUN3D

“…[5][6][7][8][9][10][11] For the current work, we make use of an unstructured-grid flow solver FUN3D, 12 where the Navier-Stokes equations supplemented by turbulence models are solved to simulate the unsteady flow field. The FUN3D code has been used for many large-scale applications in the past including simulation of unsteady flow past single-and tandem-cylinder configurations, 13 where time-averaged and perturbation quantities were shown to be in fairly good agreement with experimental data and earlier computations from the well-established, structured-grid flow solver CFL3D.…”

Computational approximation of nonlinear unsteady aerodynamics using an aerodynamic model hierarchy

“…Likewise the Euler predictions, a time step of 5 × 10 −5 was used. For more details of time step selection, the reader is referred to the work of Cummings et al [8]. This results in a computational cost equivalent to that of 40 RANS steady-state calculations and a total cost (both Euler and RANS) of 88 RANS steady-state calculations.…”

Assessment of Sting Effect on X-31 Aircraft Model Using CFD

“…Figure 4 shows Power Density Spectra (PDS) of the lift of the wing. The spectrum is rather noisy, however some important frequencies corresponding to certain features in the flow (see Cummings et al 26 for details) can be distinguished, mainly: Strouhal numbers in range of S h ≈ 0.05 and S h ≈ 0.11 corresponding to vortex breakdown, several distinctive peaks in range of S h ≈ 0.2 and S h ≈ 0.4 corresponding to vortex shedding at high α and S h ≈ 1 and S h ≈ 2 corresponding to Helical mode instability. Several peaks can be located above the value of Strouhal number S h > 3 possibly corresponding to Helical mode instability and shear layer instabilities.…”