[tsh850227]

Hi All,

I am modelling a gas-liquid system in a stirred tank with CFX.

Several spargers were located at the bottom of the tank, and impeller was in separated domain with a specified rotating speed. the buoyancy model was activated as density difference. turbulent models were specified as k-e for fluid at inlet with intensity of medium. The top of the tank is specified as degassing condition. My case is nearly exactly the same as the tutorial provided by ANSYS. I followed the setup in the tutorial.

However, this simulation suffered very strange errors: with steady state mode, calculation diverges and failed with a notice of overflow. I monitored the the gas exchange at the sparger and top of the tank. it seems that after several initial iterations, there would be some gas trying to flow into the domain from the top. I think this is the main reason why the calculation failed.

so anyone can advise?

thanks in advance

[ghorrocks]

Have you investigated why you appear to be getting reverse flow? That seems to be the key to the problem.

[tsh850227]

Originally Posted by ghorrocks

Have you investigated why you appear to be getting reverse flow? That seems to be the key to the problem.

hi Ghorrocks,

thank you very much for your reply, that is why I cannot understand on this error. i set an inlet boundary at the spargers, and degassing condition at the top of the tank. domain is initialized with a volume fraction of 0 for gas and automatic for liquid. all the other boundaries were specified as wall or asymmetry or periodic. I checked the pressure in the tank before solving goes to divergence, in my view, it is fine. so, any suggestion?

thanks again for your reply

shuai

[ghorrocks]

I am not an expert on the degassing boundary but after reading the documentation it appears that the boundary can handle if the free surface is above the boundary, but it is not clear what will happen if the free surface attempts to go below the degassing boundary. So if your fluid motion is enough that the free surface level could go below the degassing boundary location at some points then you might have some convergence problems.

[tsh850227]

Originally Posted by ghorrocks

I am not an expert on the degassing boundary but after reading the documentation it appears that the boundary can handle if the free surface is above the boundary, but it is not clear what will happen if the free surface attempts to go below the degassing boundary. So if your fluid motion is enough that the free surface level could go below the degassing boundary location at some points then you might have some convergence problems.

hi Ghorrocks,

thanks a lot for your reply. I followed your suggestion, and set the free liquid surface a bit below (5 cm). At the beginning, simulation goes well and no reverse flow of air happened. however, after several iterations, reverse flow occurred again and simulation failed very soon.

thanks again for your guidance

regards,

shuai

[ghorrocks]

Rather than just randomly moving the surface around, how about you look at the flow and work out what you expect the free surface to do. Then place the degassing condition below that.