CFD simulation of vortex shedding

CFD simulation of square obstruction with streaklines at 96.0 seconds 2400 steps. Re=1000. 102 x 45 mesh. This was done with SLFCFD.

• Category

  Science & Technology

[flabber404725]

hello.. how to make this video actually?

[slffea]

SLFCFD uses OpenGL to create the individual frames, then ffmpeg to compile the frames into animation.

[dtiydr]

How accurate is this simulation compared to an real enviroment?

[slffea]

I based my software on the paper, "A numerical study of vortex shedding from rectangles", by Davis, R. W., and E. F. Moore. There are comparisons with actual experimental data in terms of the Strouhal number in the original paper.  My results were not quite as good as theirs, but were of the same order as the experimental results.

[BeenuZz]

hey can you give me the dimensions, fluid characteristics and velocity so i check if LS-DYNA gives the same answer?

Thanks

[slffea]

The dimensions of the domain are 27 x 12. It is a uniform finite difference grid. dx = dy = .266. dt = .04. Re = 1000. The square is located at about (5,5) and about 1 x 1 in size. All units were non dimensionalized. You can read my paper "Re-examining the QUICKEST Algorithm for Two-Dimensional Incompressible Fluids" in Int. J. Numer. Meth. Fluids 2011; 67:1470-1499 to have the full details.

[Lavabug]

Interesting, do you use this for aeronautical engineering? I'll probably be taking a CFD course in grad school (physics).

[slffea]

It's actually based on laminar flow. The code will remain

stable for Re=50,000 and beyond, but you have probably crossed over into turbulence by then and will need a more sophisticated code.

[Marco-Felipe King]

Interesting. Which turbulence model is this?

[slffea]

There were probably no comments because most people watched this video embedded from my website. You can get the code there as well. Unfortunately I can't link from here. It would also be easier if you have Linux/Unix.