Warning: A vortex crosses a pressure opening 1

[Swini]

Hi there.
I'm trying to model natural ventilation using internal analysis of a building. Solver keeps giving the warning "a vortex crosses a pressure opening", which significantly affects my results.

The advice in the documentation is to "extend the duct length" at the opening. However, in doing so i introduce friction losses, which will affect my results. Also, for bouyant flows an extended vertical duct will increase the stack effect and hence drive a bigger flow.

Is there a way to avoid this problem? It seems strange that one should have to go to such bizarre lengths as adding sections of ducting to your model to get Cosmos to deal with the problem. Has anyone else experienced this and what have they done?

Cheers,

Chris Swinburn

 

[cbrn]

Hi,
there will mathematicaly no way to get rid of the warning if you don't modify something in your boundary conditions: the flow solution manifestly finds a vortex-shape pattern at your pressure opening as the only possible way to respect continuity, energy conservation etc. The vortex-shape is not compatible with the constant, uniformly-distributed pressure imposition of the BC.
I think the best thing is to reverse the BCs: place a flow condition where you currently have the pressure condition, and place the pressure condition where you currently have the flow condition. Pressures are relative, so don't worry if you have to work with "unrealistic" pressure values: you will "shift" them afterwards. Moreover, if you decide that the pressure condition has to stay where there is a vortex, you must be aware that it is your BC which is unrealistic (i.e. you "don't have the right" to let it be there).
In cases where vortices exist at both inlet and outlet, then you have no way but to extend your analysis volume on the pressure condition's side, as far as it is needed to encounter a zone of uniformly distributed pressure.
If you put wall roughness to zero in the "extension" part, I believe you would not add friction losses.

Regards

 

[Swini]

Thanks for that cbrn,

The problem is that I don't have a flow boundary condition- just two pressure conditions. All flow is generated by heating of the air inside the solid object. The problem i have with increasing the duct length is that I have a vertical duct in which air is heated. So increasing the duct length also increases stack and hence the flow! If i put in a 90deg bend to solve this i then have bend losses...seems a long way round the problem to me.

Regards
Chris

 

[kellnerp]

So what you have is internal convection and heat transfer.

http://www.engtran.com

http://www.niswug.org