|
In this simulation the incompressible flow about a submarine is computed in parallel. The domain decomposition used for the parallel calculation is shown here. The pressure contours on the surface of the submarine can be seen here. |
 |
These series of simulations are for the DARPA SUBOFF model. Runs were conducted for the model without a fairwater for 0, 8, and 16 degrees angle of attack. A definition of the SUBOFF model can be seen here and here. The surface mesh may be seen here, with a detailed surface mesh and mesh in cut plane x=0.2 of SUBOFF model here. Surface pressure contours for 0, 8 and 16 degrees angle of attack can be seen here, here and here. The corresponding surface shear vectors and velocity contours in 8 cut planes can be seen here, here and here. Vorticity contours in 3 cut planes can be seen here, here and here. |
 |
The same series of runs was conducted for the DARPA SUBOFF model with fairwater. A definition of the SUBOFF model can be seen here and here. The surface mesh can be seen here, and a detailed surface mesh and mesh in cut plane x=0.2 here. Surface pressure contours for 0, 8 and 16 degrees angle of attack can be seen here, here and here. The corresponding surface shear vectors and velocity contours in 8 cut planes can be seen here, here and here. Vorticity contours in 3 cut planes can be seen here, here and here. |
| This simulation considers the vortical field in the vicinity of the sail for a typical submarine. The mesh used may be discerned from this image. Surface pressure contours and pressure contours in 4 cut planes can be seen here. The development of the vortices is clearly visible. |
 |
This is a simulation of a colloid suspension. The fluid surface mesh and a cut through the grid are shown here. The fluid flow velocity is shown on different cutting planes here. The fluid velocity on the surface of the fluid domain and the pressure on the spheres here. Finally the force on the spheres together with the fluid velocity are shown here. |
 |
This is a simulation of a the flowfield past a NACA0012 wing at an angle of attack of 10 degress (incompressible Euler). The surface mesh and a cut through the grid are shown here and here. The surface pressure, as well as the velocity in 4 cut planes and the vortex core are shown here. This case was run with vorticity confinement. |
 |
This example considers a realistic car configuration, as is commonly encountered for external car aerodynamics. The Reynolds-number based on the length of the car is approximately Re=3.2e+6. The Smagorinsky turbulence model was used. The resulting flow is quasi-steady and shows the development of a vortex train behind the body. The surface mesh employed, as well as the pressure and velocity field obtained can be seen here, here and here. Note the boundary layer mesh (the complete mesh consisted of 3.7Mtets and 658Kpts). This problem was run in two stages. First, local timestepping with a local Courant-nr. of C=4.1 and LU-SGS-GMRES relaxation was used to achieve quickly the overall flowfield. Then, a fully implicit scheme with a timestep of dt=0.01 was employed to run the time-accurate simulation. |
 |
This example considers a wind turbine configuration. An overlapping grid approach with two grids was used to treat the relative movement of parts. The resulting flow if quasi-periodic and may be seen here, here, here, here using xanim under Linux, and here, here, here, here using MS Windows. |
