Hi, I am doing a simulation in a model with a emission surface in a wind tunnel. I defined an additional variable. The inlet concentration is 1.83E-06kg/m3, but there is small area where the concentration is higher than the inlet concentration before the emission surface and there is also a small area where the concentration is minus (please see the figures). I checked the simulations results in CFX11.0.It converged well to Max RES in 10E-05 and the monitor point value is quite stable.

http://img132.imageshack.us/my.php?image=modelnm4.png http://img132.imageshack.us/my.php?i...trationee6.png

Thank you very much!

Hi,

What discretisation scheme are you using for volume/mass fractions? Some discretisation schemes can generate over/undershoots.

Glenn Horrocks

Hi Glenn Horrocks, thank you for your response. I used second order discretisation scheme. Thanks.

Hi,

Do you get the over/undershoots when using upwinding for the AV?

Glenn Horrocks

Hi Glenn,

I tried the upwind scheme and then the unphysical area disappeared. I can understand why the second scheme can cause the over/undershoots somehow, but do you think it is ok to keep the results with second scheme even though there are over/undershoots because I still think the second scheme can give better results for the transport equation. Thank you very much!

Best Regards, Li

Hi,

Whether the under/over shoot is a problem depends on what you want. Is this a significant region? What are the key parameters for the simulation results and does this region affect them?

You will want to use second order differencing if at all possible as it is much less diffusive.

Glenn Horrocks

Hi Glenn,

Thank you for your responding very much. I will compare the results with the experiments and have a look at the simulation results in detail.

Best Regards, Li

Hi Li,

Use the High Resolution scheme for your scalar. It will give you bounded second order results.

-CycLone