vigneshkaushik

I am trying to simulate a simple wind tunnel test with a continuous wind flow inside a constant volume with one static building object as collision.

I have run it for 240 frames, I find that the length and angle of velocity vector constantly keeps changing through out the volume.

How do we determine when to stop the simulation? Is there a way we can let the simulation settle and converge to a point beyond which there is minimal change in velocity vector?

On the Smoke object in the dynamics network, inisde the Initial data tab, the default for Velocity sampling is “Faces”. I understand that despite being slower, this gives more accurate collisions of smoke with another dynamic object.

I would like to know whether it is the same for static objects like buildings also. If not can we set velocity sampling to “Centers” for a faster simulation?

These are links to some papers I found online about Houdini's smoke and fluid solvers. They seem to suggest that the incompressible Navier-Stokes equation is used as the base for most of these solvers with tweaks here and there to improve realism, speed and computational power. Also there seems to be three main methods used in Houdini; Smooth Particle Hydrodynamics (SPH), Volume Grid approaches based on voxels and FLIP which is a type of hybrid between a particle based and volume based fluid simulations.

http://www.cescg.org/CESCG-2007/papers/TUBudapest-Horvath-Peter/TUBudapest-Horvath-Peter.pdf [cescg.org]

http://www.peterclaes.be/tutorials/pct_0007.pdf [peterclaes.be]

http://www.fxguide.com/featured/the-science-of-fluid-sims/ [fxguide.com]

Has anyone compared the results of these simulations with a more academic CFD software package like Fluent, ScStream etc?