CFD - Nabla Logo
Simulation Software and Consultancy
CFD - home CFD - solutions CFD - software CFD - support CFD - other CFD - contact
CFD - Nabla Logo Explained
  Advanced Search
  Send us a comment/query
  Back to Solutions
  Multi-phase/free surface flows
  DNS, LES and aeroacoustics
  - Aeroacoustics with LES
  - Flow around wing mirror
  - DNS of flame growth
  Reacting flows
  Complex rheology
  General purpose CFD
  Solid mechanics
  Complex meshes
  Mesh motion
  ©2000-2007 Nabla Ltd.
Flow around a wing mirror
A simplified car wing-mirror, is mounted on a flat plate in an open wind tunnel. The geometry is a half-cylinder of diameter 20 cm with hemi-spherical free end. The far-field flow speed is 40 m/s. All calculations were done using DES methods, using Spalart-Allmaras and turbulent energy transport equation (Yoshizawa). The mesh size was approx. 2.4 million cells. Animation shows pressure on the mirror baseplate, with large scale shedding, a ‘horseshoe’ vortex and wake details.
Wing mirror vortices
Click image to view the animation
Below are comparisons of velocities from LES simulation (black arrows) with LDA measurements (coloured arrows, after A.M.K.P Taylor, Imperial College London). The direction and magnitude of the velocities is predicted well both upstream and downstream of the mirror. The prediction of reattachment in the wake is within 10% of the measured results.

Flow velocities

The graph below compares static pressure measurements at points on the mirror surface with calculated values using different LES models and meshes. A countour plot of the static pressure on the front of the mirror is also shown below. The final image shows time averaged vortices.

Mean static pressure
Pressure Time averaged vortices

The adjacent image shows the RMS of the pressure fluctuations on the baseplate around the mirror. The vortex shows strong vibrational motion, with the strongest noise sources seen to originate from the mirror trailing edges and reaching a maximum some distance downstream in the wake. High surface fluctuations are due to a combination of strong shedding vortices from the trailing edge and turbulence production in the shear layer between the seperation bubble and the free stream.
Below shows Fourier transforms of pressure trace for a point on the back of the wing mirror. The transformed trace is analogous to the noise produced at that point on the surface according the the Lighthill hypothesis. The comparison is between experimental data (black), Spalart Allmaras (S-A, green) and SGS turbulence energy transport DES (1eq, blue). Excellent agreement can be seen, especially for the SA model up to nearly 4kHz.