Re LES mode

harry · January 18, 2013, 00:10

Hi All,

I recently plan to put a LES model into my multiphase code to simulate gas-solid flow. I never have experience on LES model. I know FVM method for incompressible flow and have experiences in the development of unstructured based code but based on KE model. Is there any open source or free academic code which allows me to quickly take into my code (for gas-solid flow based on discrete element method coupled with a unstructured CFD for incompressible flow)? Your helps would be greatly appreciated.

Cheers,

Harry

afeworkimussie · January 18, 2013, 03:32

Yea, you can try OpenFOAM's icoFoam solver.

harry · January 18, 2013, 05:47

Quote:

Originally Posted by afeworkimussie

Yea, you can try OpenFOAM's icoFoam solver.

thx your reply. open foam is too complicated to me. my current position does not give me much time for my topic. furthermore, open foam may need lots of improvements to solve gas solid flow based on discrete element method.

FMDenaro · January 18, 2013, 05:51

Quote:

Originally Posted by harry

thx your reply. open foam is too complicated to me. my current position does not give me much time for my topic. furthermore, open foam may need lots of improvements to solve gas solid flow based on discrete element method.

sorry to say that you can not hope to work in LES without experience with such a constraint ...

sbaffini · January 18, 2013, 12:59

Is your code 3D and unsteady?

Does it allow non dissipative convective schemes of 2nd or higher order without loosing stability?

Is the time integration of higher order too?

In this case, your first basic implementation may be just as simple as a mixing length eddy viscosity model (a.k.a. Smagorinsky model in LES) with the length provided trough some linear measure of the grid cell volumes. This approach is known as implicit filtering and is the one mostly used in practice.

Then you can further improve it by somehow changing the constant in the model (either by near wall dumping or by the dynamic procedure), or change the model to some more advanced.

However, quoting Filippo, time restrictions are not friends of LES practicioners neither... less than ever are compatible with LES beginners

harry · January 20, 2013, 19:16

Thank you all for your comments, in particular for sbaffini's details. Here, I would appreciate Sbaffini further comments for the following points.

In my simulations, time step is alway small, say less than 5e-6. Does it mean that I should focus on other things.

1) accuracy of numerical approximation. Do I have to use second order approximation of convective, diffusion and time terms in NS equations?

2) I are going to simulate some domain with a little bit complex boundary, such as pipe. Are there some special attentions to the simple Smagorinsky model? I noticed some papers discussed this issue.

3) How fine the grid would be? Simulation time is so far a big issue to me due to that I mainly employed sequential simulations which last for around one month. More grids, definitely more time, mainly caused by the coupling between particles and gas flow from my side.

So far, I have to persist my own code, which I have testes by many applications in relation to different gas/liquid solid flows, leading to a few journal papers. The problem is that the code uses one order intergration of time and it does not contain some advanced turbulent models. I have a few cases which cannot be reproduced by my model where I believe turbulence plays a key role. I need to test my thoughts with les models or the like.

Best regards,
Harry

January 18, 2013, 00:10	Re LES mode	#1
harry New Member Harry Join Date: Mar 2009 Posts: 27 Rep Power: 17	Hi All, I recently plan to put a LES model into my multiphase code to simulate gas-solid flow. I never have experience on LES model. I know FVM method for incompressible flow and have experiences in the development of unstructured based code but based on KE model. Is there any open source or free academic code which allows me to quickly take into my code (for gas-solid flow based on discrete element method coupled with a unstructured CFD for incompressible flow)? Your helps would be greatly appreciated. Cheers, Harry

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Switching on LES in 2D mode of fluent	Josef Dubsky	FLUENT	27	March 16, 2016 13:49
LES Filtering: how are the small and large scales equations solved?	atmcfd	Main CFD Forum	38	March 14, 2016 14:52
Turbulence dampening due to magnetic field in LES and RAS	eelcovv	OpenFOAM	0	June 8, 2010 11:35
Inlet Temperature Fluctuation in CFX-5.7 LES Mode	Halim Choi	CFX	0	December 23, 2004 22:05
Some Questions about LES.	Bin Li	Main CFD Forum	2	February 20, 2004 09:58

January 18, 2013, 03:32		#2
afeworkimussie New Member Mussie Join Date: Feb 2011 Posts: 7 Rep Power: 15	Yea, you can try OpenFOAM's icoFoam solver.

January 18, 2013, 12:59		#5
sbaffini Senior Member Paolo Lampitella Join Date: Mar 2009 Location: Italy Posts: 2,152 Blog Entries: 29 Rep Power: 39	Is your code 3D and unsteady? Does it allow non dissipative convective schemes of 2nd or higher order without loosing stability? Is the time integration of higher order too? In this case, your first basic implementation may be just as simple as a mixing length eddy viscosity model (a.k.a. Smagorinsky model in LES) with the length provided trough some linear measure of the grid cell volumes. This approach is known as implicit filtering and is the one mostly used in practice. Then you can further improve it by somehow changing the constant in the model (either by near wall dumping or by the dynamic procedure), or change the model to some more advanced. However, quoting Filippo, time restrictions are not friends of LES practicioners neither... less than ever are compatible with LES beginners

January 20, 2013, 19:16		#6
harry New Member Harry Join Date: Mar 2009 Posts: 27 Rep Power: 17	Thank you all for your comments, in particular for sbaffini's details. Here, I would appreciate Sbaffini further comments for the following points. In my simulations, time step is alway small, say less than 5e-6. Does it mean that I should focus on other things. 1) accuracy of numerical approximation. Do I have to use second order approximation of convective, diffusion and time terms in NS equations? 2) I are going to simulate some domain with a little bit complex boundary, such as pipe. Are there some special attentions to the simple Smagorinsky model? I noticed some papers discussed this issue. 3) How fine the grid would be? Simulation time is so far a big issue to me due to that I mainly employed sequential simulations which last for around one month. More grids, definitely more time, mainly caused by the coupling between particles and gas flow from my side. So far, I have to persist my own code, which I have testes by many applications in relation to different gas/liquid solid flows, leading to a few journal papers. The problem is that the code uses one order intergration of time and it does not contain some advanced turbulent models. I have a few cases which cannot be reproduced by my model where I believe turbulence plays a key role. I need to test my thoughts with les models or the like. Best regards, Harry