|
[Sponsors] |
![]() |
![]() |
#1 |
Senior Member
|
Dear Foamers,
Two questions about LES model: 1- What is U or P in this model. is it filtered one or real one. i mean U and P which are results and be written in folders and their initials which is at 0 folder. 2- It is not a filteration process in oneEqEddy or smagorinsky models but there is available for DynSmagorinsky. Why? http://foam.sourceforge.net/doc/Doxy...agorinsky.html http://foam.sourceforge.net/doc/Doxy...agorinsky.html Thanks in advance, |
|
![]() |
![]() |
![]() |
![]() |
#2 |
Member
Charlie
Join Date: Dec 2010
Location: USA
Posts: 85
Rep Power: 16 ![]() |
1. I think the U and P in the result are real one rather than filtered, filtering process is internal process, and basically we are combing resolved and sub-filtering scale to present the real result.
2. since the smagorinsky and oneEqnEddy model are two-level filtering model, dynSmagorinsky model are three filtering model. To avoid the a possible loss of efficiency of classical Germano-Lilly dynamic procedure, the scale-dependent dynamic procedure is obtained by considering a second test filtering level. please see "Large Eddy Simulation For Incompressible Flow" Third Edition, Page142, by Pierre Sagaut. Above is my shallow understanding of your question, if anything wrong, please let me know, thanks! |
|
![]() |
![]() |
![]() |
![]() |
#3 |
Member
Farshad
Join Date: Oct 2010
Posts: 76
Rep Power: 16 ![]() |
Hi Foamers,
Thanks for this nice discussions. I appreciate you. please show me where the real U and P are constructed by the filtered one. My poor understanding is because only the LES pointer is selected by runtime mechanism ,only and only the subgrid scale model is written in terms of filtered variables and therefore the overall equation only contains filtered variables. Please correct me if i am wrong. And also the boundary variables should be set after manual filteration! Thanks alot Farshad |
|
![]() |
![]() |
![]() |
![]() |
#4 |
Member
Charlie
Join Date: Dec 2010
Location: USA
Posts: 85
Rep Power: 16 ![]() |
Hi,
Nice discussion really! As far as I know, in LES model, the effect of the resolved subfilter-scale (RSFS) and subgrid-scale (SGS) motions on the resolved velocity field must be modeled. The RSFS contribution can be theoretically reconstructed, and the SGS stress must be modeled, in DynSmagorinsky we use Germano-Lilly eddy-viscosity SGS model in which the total parameter C*delta is calculated dynamically. So the solution has filtered and subfiltered components in itself. Again, it's just my understanding on that, please let me know if I'm wrong. ![]() |
|
![]() |
![]() |
![]() |
![]() |
#6 |
Member
Charlie
Join Date: Dec 2010
Location: USA
Posts: 85
Rep Power: 16 ![]() |
Hi,
since there are many different type of filters, take simpleFilter as an example. you can refer to the following link: http://foam.sourceforge.net/doc/Doxy...8C_source.html ![]() |
|
![]() |
![]() |
![]() |
Thread Tools | Search this Thread |
Display Modes | |
|
|
![]() |
||||
Thread | Thread Starter | Forum | Replies | Last Post |
possible interview questions | atturh | Main CFD Forum | 1 | February 21, 2012 09:53 |
NACA0012 Validation Case Questions | ozzythewise | Main CFD Forum | 3 | August 3, 2010 15:39 |
[Other] Some perhaps stupid questions about calculix | lynx | OpenFOAM Meshing & Mesh Conversion | 11 | May 17, 2010 07:48 |
So many questions in DPM & UDF - Help! URGENT! | Prashanth | FLUENT | 0 | March 3, 2009 23:26 |
turbulence modeling questions | llowen | Main CFD Forum | 3 | September 11, 1998 05:24 |