CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   itterative running of two sub models (http://www.cfd-online.com/Forums/cfx/23106-itterative-running-two-sub-models.html)

Babu September 27, 2006 04:29

itterative running of two sub models
 
Hi all:

I have a huge 3D (360 deg.) fluid domain. I got strong necessity to simplify the domain by chunking it to two sub domains and also with two different sector angles and run both iteratively till I derive converged solution. To avoid huge mesh size. I got only the inlet and outlet pressures (BCs) of the overall fluid domain. The bi-sectioning of the model is planned to have it at a location where gradients are very less.

The process I would like to implement to this problem is as follows

1. Run Domain-1 separately with pressure BCs (assuming exit pressure BC). 2. Mass outflow from this run is captured and supplied as inlet boundary condition to Domain-2 3. Run Domain-2 with this mass inlet and given exit pressure BC. 4. Extract inlet static pressure from second run and supply back that pressure to the first case. 5. Keep iterating two domains till results across the interface converge.

Is it a right way to do? Is there any other existing process behind?

If this is a correct, how can I map one 3D sector results to another, when they are different?

Please let me know.

Thanks in advance Babu


TB September 27, 2006 22:21

Re: itterative running of two sub models
 
I would suggest you to run a quick solution with a coarse mesh (depends on your machine capability) before you run on a finer mesh. If you have multi-processors in your machine, simply go parallel.

If you know roughly the condition at the intermediate interface, I don't see any reason that you can't divide it into two flow domains. However, it's hard to suggest anything more creative, without a detailed look at your problem.

Babu September 27, 2006 22:55

Re: itterative running of two sub models
 
Thanks for the response

Running a full 360 model is almost an impossible task in my case. To give more insight to this problem, the domain is a combination of a set of rotor-stator blades & a static cavity connected at the middle above the main free path. If I have to run a ful 360 deg model, than I have to work on 72+88 blade passgaes along with a full big static cavity connected above.

Please suggest. I'm strongly looking at a possible solution of running this case by selecting two different sector sizes, stitching and running iteratively

Thanks Babu

TB September 29, 2006 01:54

Re: itterative running of two sub models
 
Is the flow in the blade passage symmetrical? If yes, you could try to model just one blade passage. If not, you should either increase your machine capability or get more information about condition in the intermediate boundary. It wouldn't hurt to simplify your analysis (e.g. in 1D model) and then use the resulting approximated value as the boundary condition for the connecting surface.


All times are GMT -4. The time now is 09:27.