Time step independence?
 

Basic stuff here.

I cannot get consistent results from pisoFoam. I have a simple cavity with a cross flow inlet / outlet. I vary the time step from 0.001 to 0.0000001 s and the resulting pressure fields vary by 20-60%.

Does anyone have any ideas?

So far I've observed:

• Changing the cavity size has no effect (between 0.1m and 10m)
• Changing the mesh density has no effect (10x10 vs 100x100)
• Turbulence model has no effect (kEplison vs disabled)

Yeah, I'm an OpenFOAM developer... not a user.

-Dave

Hi David,
are you comparing averaged pressure or just pressure at some time ?

Regards,
Awais

Dave,

Are you sure that there is convergence between all of your fields at each timestep? If not, then this will have a big affect on your pressure field. Adding outer-correction will ensure convergence at each timestep. Have you looked at using pimpleFoam instead of pisoFoam? This is suggested on http://www.cfd-online.com/Forums/ope...ependency.html

@Awais

I am calculating an RMS error across the entire pressure field. Average pressure doesn't vary nearly as much.

@Dan

Thanks for the suggestions, I will try implementing pimpleFoam into the solver.

I've attached some pictures of the pressure and flow field. I'm getting a pressure fluctuation at the inlet, and discontinuities at the corners. I'm sure this is a rookie problem, but it mystifies me. These images are for velocity driven flow. When I do pressure driven flow, though, it smooths out, but I still can't get timestep independence.

Thank you all for your suggestions. I will continue troubleshooting this between my full-time job and parenthood!

-Dave