Scaling

Submitted by mytxii on Mon, 07/20/2015 - 15:36

Hello guys,

I want to simulate a small channel (11 micrometers) with small particles (2 micrometers).
Therefore I chose the "CGS" unit system. I tried to build a model with a 11 cm channel and 2 cm particles and it works fine.

But when I scale my region, neighbor, scaling coefficients, velocities, radiuses, blockMeshDict; it doesn't work anymore (Courant number rises very quick).

Do I have to change my timesteps as well ? I don't think so, because of the definition of the Courant number (Co = dt * u/dx).
If I scale my velocity (1cm/s --> 1e-4µm/s) and my dx (1cm --> 1e-4µm), it seems to me not necessary to change my timesteps.

Please let me know, what I'm doing wrong and how I may could solve my problem.

Thank you so much

PS: If I reduce my timesteps, the Courant number stays small (<1). But my particles run off when they arrive the boundary of the processor.

mytxii | Thu, 07/23/2015 - 07:41

Hello again,

if you need any files of my simulation please let me know !

Regards

Nils

j-kerbl's picture

j-kerbl | Thu, 08/06/2015 - 16:58

Hi Nils,

have you tried to compare ErgunTestMPI and ErgunTestMPI_cgs in the cfdem tutorials?
This is where I would look first.

Cheers
Josef

mytxii | Sat, 08/15/2015 - 10:38

Hello Josef,

I compared ErgunTestMPI and ErgunTestMPI_cgs to scale my simulation model to cgs, but I will do it again - maybe there's a mistake in my code.
I will inform you, when I do so.

Regards

Nils