CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   OpenFOAM Running, Solving & CFD (https://www.cfd-online.com/Forums/openfoam-solving/)
-   -   Unsteady MRFSimpleFoam? (https://www.cfd-online.com/Forums/openfoam-solving/124843-unsteady-mrfsimplefoam.html)

Ruli October 14, 2013 13:09

Unsteady MRFSimpleFoam?
 
Hi Foamers,
I want to simulate a rotating wind turbine for my MA thesis, but before I start I would like to ask some question to set the direction of my efforts ;)

Because of avialable computing ressources I would like to use MRFSimpleFoam, but as I am interested in unsteady effects, I am afraid I have got a problem here.

In another post renyun0511 writes about his problems tying to get MRFSimpleFoam unsteady:

http://www.cfd-online.com/Forums/ope...implefoam.html

Does anyone know if this is possible? I know, that otherwise I would have to use GGI/AMI.

Another question: As I read opposed opinions, I am not sure if MRFSimpleFoam works in parallel (in one of the newer releases). Does anyone have a clue?

Thanks for any answers
Julian

TReviol October 15, 2013 03:31

Hi Julian,

first your question about parallel computing: I often use MRFSimpleFoam (since OF 2.2.1. simpleFoam with fvOptions) for parallel jobs. I don't know any problems.

Then your question about unsteady MRFSimpleFoam: MRFSimpleFoam creates multi reference frames. The rotational behaviour of the flow will result from a rotating reference frame, the mesh will be frozen - so this method is also called frozen rotor. With these assumes, I m not sure about the sense of an unsteady frozen rotor. Why don't you use pimpleDyMFoam?

Regards

Thomas

Ruli October 15, 2013 04:33

Hi Thomas,
Thanks for the fast reply :)

I know about the basic concept of MRV, but as I understand the solution of the unsteady NS equation should not depend on wether the mesh is "solid" or not. Is this right? Therefore I would like to catch the unsteady flow effects with the MRF approach, but in SIMPLE this is not possible. Is there a solution to this problem?

Quote:

Originally Posted by TReviol (Post 456971)
Why don't you use pimpleDyMFoam?

For the beginning of my project I would like to test MRFSimpleFoam, as I can confine my mesh to ONE blade (in combination with periodic boundary conditions), so I don`t have to generate a mesh of the whole rotor. This should save a lot of computing time.

Bye
Julian

TReviol October 15, 2013 08:06

Hi Julian,

of course the NS equation can be solved for unsteady problems with steady mesh and for problems with a moving mesh, and of course you can imagine to do this for your problem. But I don't believe, that the solution of a frozen rotor with unsteady flow will lead to a physical solution, that make sense. Many effects will be neglected.
But finally I never tried to solve such a problem and I don't know the relevance of the mistakes they come with your method.

And yes, it is right. With MRFSimpleFoam you will reduce your effort enormously. So I understand your motivation.

I'm sorry. I can't help you. Maybe someone else has an idea?

Thomas

Ruli December 11, 2013 06:27

Dear Foamers,
any new thoughts on that? Would MRFInterFoam work? I guess, if I use only one fluid, it should give me a transient solver, with the cost off one additional transport equation for alpha, wouldn`t it?

Best regards
Julian

Kaskade December 13, 2013 03:46

If you are interested in unsteady effects use an unsteady solver. Everything else will be a waste of your time.

If the flow is truly unsteady and you run it using (MRF)simpleFoam, you will end up with high residuals, if the case runs stable at all.

If you use the current OF-version you can enable MRF in the fvOptions und run pimpleFoam. But the only computational effort you really save this is way is the mesh motion.

Have a look at the propeller tutorial. If you have further questions regarding pimpleDyMFoam, I may be of help. I am dealing with highly unsteady flows myself. Though it took me a long time to realize that simpleFoam was not appropriate for my type of problems.

Ruli December 13, 2013 04:32

Quote:

Originally Posted by Kaskade (Post 466223)
If you use the current OF-version you can enable MRF in the fvOptions und run pimpleFoam.

Thanks, I wasn`t aware of that! Should be written in the OF manual ;) There it only mentions simple and interFoam for MRF solving.

Best regards
Julian

Kaskade December 13, 2013 04:46

I think it works. I am fairly certain I tried it before finally giving up and switching to DyM.
The manual still mentions MRFSimpleFoam which doesn't exist anymore. With OF it's always best to carefully read the release notes and to have a look at the tutorials that might be important for your line of work.

I simulated a very small HAWT for my diploma thesis. What are the effects you are interested in?

Ruli December 13, 2013 04:49

Quote:

Originally Posted by Kaskade (Post 466236)
I simulated a very small HAWT for my diploma thesis. What are the effects you are interested in?

I am working on cross flow and passive flow control for reduction of cross flow.

Best regards
Julian

Kaskade December 13, 2013 05:15

Winglets? Or am I mistaken? Do simulate you the whole turbine or just the blade tips.

Ruli March 30, 2014 09:55

Dear Foamers,

just to let you know:
I got MRF running in combination with pimpleFoam in piso mode, as mentioned above by simple using fvOptions with the pimple solver.

Though, one problem occurs, which i have not solved yet:
The simulation seems to converge per time step (initial residuals down to 10^-6 to 10^-8). But between two following time step the calculated force seems to switch between to states, which differ +- by the same magnitude from the steady solution. My max. Courant number is 0.8. Could this be the reason?

Did anyone else encounter this problem?

Best regards
Julian

Kaskade March 30, 2014 14:46

How much time has passed in your simulation?

If possible post your fvOptions, fvSchemes and fvSolution.


All times are GMT -4. The time now is 04:04.