|
[Sponsors] |
March 16, 2011, 07:49 |
div(phi,U) GammaV ?
|
#1 |
Member
fisch
Join Date: Feb 2010
Posts: 97
Rep Power: 16 |
Hello,
i read some parts about the Gamma Scheme in Jasak's Thesis (pages 107-110) and wanted to try it now for my openFOAM simulations. My question is now, what the Greek_Gamma value should be if i define for example: div (phi,U) GammaV [Greek_Gamma] In the Documentation there is a recommendation for Greek_Gamma = 1 ( i think for maximal stability). But where is the Greek_beta_m value of the chapter in the thesis? Or is this Greek_Gamma this Greek_beta_m? Or is it not possible to define this Greek_beta_m? What are the best values for this factor in your simulations to get max accuracy (in the best case 2nd order) and still a good convergence rate??? Thanks a lot, Rupert |
|
March 16, 2011, 08:19 |
|
#2 |
Senior Member
Anton Kidess
Join Date: May 2009
Location: Germany
Posts: 1,377
Rep Power: 29 |
What you call Greek_Gamma is beta_m / 2.0. There is no general rule what value to pick, you'll have to play around and find what works best for you. You should not specify a value smaller 0.2 though.
|
|
March 16, 2011, 09:31 |
|
#3 |
Member
fisch
Join Date: Feb 2010
Posts: 97
Rep Power: 16 |
Thanks for your reply.
Jasak is writing in his thesis that beta should be between 0.1 and 0.5. So do you mean that the Greek_Psi (sorry for the Gamma mistake) is two times beta? this would fit then with the bounds to Jasaks description ( beta [0.1;0.5] would lead to Psi [0.2;1]). Or is it really 1/2* beta??? Thanks |
|
March 16, 2011, 10:23 |
|
#4 |
Senior Member
Anton Kidess
Join Date: May 2009
Location: Germany
Posts: 1,377
Rep Power: 29 |
Oh yes, sorry about that! I did mean to say that the value you specify in the fvSchemes file will be divided by 2, which gives you the proper beta_m for the limiter (see also Gamma.H in the C++ documentation).
|
|
March 17, 2011, 01:47 |
|
#5 |
Member
fisch
Join Date: Feb 2010
Posts: 97
Rep Power: 16 |
thank you very much
|
|
August 16, 2013, 15:13 |
|
#6 | |
Senior Member
saeideh mohamadi
Join Date: Aug 2012
Posts: 229
Rep Power: 14 |
Quote:
we set the the grad, div, and laplacian schemes in this way: "gradSchemes : default cellMDLimited Gauss linear 0.5; " "divSchemes : div(phi,U) bounded Gauss GammaV 1.0;" "laplacianSchemes : default Gauss linear limited 0.5; " my Question is : when we want play around to found that what value is a good limiter for GammaV , e.g we decrease the limiter of GammaV tp 0.5, we also should decrease the limiters for grad, and laplacian scheme? or they have no relate to each other? Thank you |
||
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Difference SFCD and GammaV | bboonacker | OpenFOAM Running, Solving & CFD | 7 | March 20, 2014 11:51 |