Did openoam implicitly implement Gamma differencing shceme

harry · January 30, 2007, 17:42

I am not sure I post this massage on the right place. If not, I apologize for it.
I try to implement Gamma in my code in deferred correction mode. I did not see much improvement of my code. Is there something wrong?
Acoording to Hrvoje's theis, Gamma's big virtue is that it dose not require defferred correction due to its compact computational molecule.
So Openfoam should implicitly introduce Gamma? what's openfoam's treatment? I would appreciate your hints.
At current stage, OpenFoam is a very big challege for me as I know nothing about C++. Hence I still work on my own code. But my final taget is OpenFoam

hjasak · January 30, 2007, 18:02

Yes, this is where I did all my work on the Gamma differencing scheme in the first place. What you do is as follows:

- imagine your scheme as a blend between upwind and central differencing
- look at the paper to find out how to calculate the blending factor from the currect solution for each face (using cell and face gradients)
- ... and now you're free to do whatever you wish.

OpenFOAM implements my scheme directly, i.e. it comes up with interpolation factors that use the blending between upwind and central.

Star-CD used to do it in a deferred correction mode, but I was never to happy with it - deferred correction distorts temporal accuracy. For Mickey Mouse cases, there's no much difference in any case. If you are interested in validation of the numerics rather than chasing LES and similar (sensitive) physics, you'll be all right any way.

Enjoy,

Hrv

harry · February 5, 2007, 02:11

Thanks for Dr.Hrv's answer. In your thesis, you talked a lot about discretised methods for diffusion term. there methods based on the decomposition of surface vector are investigated. Something bother me is that what's the difference between these methods with the one given by Ferziger and Peric in computational methods for fluid dynamics. If any difference, I would appreciate your comments. My code was developed via Ferziger and Peric's text book. Now it seems I need a professional advice. Sorry this topic is not about OpenFoam.Thanks in advance.

January 30, 2007, 17:42	I am not sure I post this mass	#1
harry New Member Harry Join Date: Mar 2009 Posts: 27 Rep Power: 17	I am not sure I post this massage on the right place. If not, I apologize for it. I try to implement Gamma in my code in deferred correction mode. I did not see much improvement of my code. Is there something wrong? Acoording to Hrvoje's theis, Gamma's big virtue is that it dose not require defferred correction due to its compact computational molecule. So Openfoam should implicitly introduce Gamma? what's openfoam's treatment? I would appreciate your hints. At current stage, OpenFoam is a very big challege for me as I know nothing about C++. Hence I still work on my own code. But my final taget is OpenFoam

January 30, 2007, 18:02	Yes, this is where I did all m	#2
hjasak Senior Member Hrvoje Jasak Join Date: Mar 2009 Location: London, England Posts: 1,905 Rep Power: 33	Yes, this is where I did all my work on the Gamma differencing scheme in the first place. What you do is as follows: - imagine your scheme as a blend between upwind and central differencing - look at the paper to find out how to calculate the blending factor from the currect solution for each face (using cell and face gradients) - ... and now you're free to do whatever you wish. OpenFOAM implements my scheme directly, i.e. it comes up with interpolation factors that use the blending between upwind and central. Star-CD used to do it in a deferred correction mode, but I was never to happy with it - deferred correction distorts temporal accuracy. For Mickey Mouse cases, there's no much difference in any case. If you are interested in validation of the numerics rather than chasing LES and similar (sensitive) physics, you'll be all right any way. Enjoy, Hrv Tushar@cfd likes this. __________________ Hrvoje Jasak Providing commercial FOAM/OpenFOAM and CFD Consulting: http://wikki.co.uk

February 5, 2007, 02:11	Thanks for Dr.Hrv's answer. In	#3
harry New Member Harry Join Date: Mar 2009 Posts: 27 Rep Power: 17	Thanks for Dr.Hrv's answer. In your thesis, you talked a lot about discretised methods for diffusion term. there methods based on the decomposition of surface vector are investigated. Something bother me is that what's the difference between these methods with the one given by Ferziger and Peric in computational methods for fluid dynamics. If any difference, I would appreciate your comments. My code was developed via Ferziger and Peric's text book. Now it seems I need a professional advice. Sorry this topic is not about OpenFoam.Thanks in advance.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Examples of implicitly coupled domains	nadine	OpenFOAM Running, Solving & CFD	3	August 15, 2008 18:53
?to implement gamma differencing scheme	Gonski	Main CFD Forum	15	January 28, 2007 17:00
The terms that should be treated implicitly in LES	ben	Main CFD Forum	3	January 28, 2005 03:32
hybrid differencing	alice	FLUENT	0	July 14, 2004 07:45
Temporal differencing for LES	Jongtae Kim	Main CFD Forum	0	June 2, 1999 23:09