Introduction to transient turbulence modeling (RAS,LES) - Part 2

[@giovannifilomeno]

Hi guys. For everyone using a more recent version of OpenFOAM (I'm using v2012) there are some little changes in Smagorinsky model. First of all, you must rename nutSgs into nut. Looks like the LESProperties file is no more necessary, instead you need to modify the turbulenceProperties file as follows: simulationType LES; LES { turbulence on; LESModel Smagorinsky; [...] } and substitute the [...] with the entries present in the LESProperties file (delta, printCoeffs, cubeRootVolCoeffs ecc...). Hope it helped since I've lost almost 1 hour to figure out what the problem was :) Thanks József for the great effort!

[@giovannifilomeno]

Similar adjustments need to be done for the One Equation Model. Rename nutSGS into nut and modify the turbulenceProperties file as follows: simulationType LES; LES { turbulence on; LESModel kEqn; //oneEqEddy in older versions [...] }

[@OpenFOAMJozsefNagy]

Thank you!

[@samy20042004]

Perfect , I learnt more than whatever I waste on reading papers. Thank you so much Jozsef.

[@MrMangkokoo]

how is your CFD skill now :)?

[@lgargalis]

Great tutorial!!! Keep up the good work. Thank you Jozsef!!!!

[@andreabray07]

Great work! Your lessons are always clear and accurate. Compliments from Rome!

[@pablofdez.1467]

Thank you for your amazing work, it is really helpful for beginners like me. Keep it up!

[@OpenFOAMJozsefNagy]

I am glad to help.

[@domenicolahaye303]

Wonderful. Thanks.

[@smrutimayapattanaik4077]

Awesome ....

[@MrMangkokoo]

hi how's your skill now?

[@emanbagheri4375]

Great video as always. Could you please also do a tutorial on using mapped boundary condition on the velocity inlet to account for the oscillations in the velocity profile?

[@maqiitg]

Hi, It was very informative tutorial on turbulence modelling using LES. I am hoping if you can guide us more through tutorial on combustion models specially non premixed diffusion flame.

[@OpenFOAMJozsefNagy]

Well, that's a very specific topic and it's most definitely nothing for starters. Maybe in the future.

[@samuelperrin8666]

Hi Josef, Thank you for your work. It very helpfull for understand CDF. If I well understood, you saied for LES simulation you need to visualise the viscous sublayer, so the yPlus need to be really small in the wall and is it not the case on your mesh. Could we considert that your simulation is valide? Regards and thank you again for your amazing work :) Samuel

[@OpenFOAMJozsefNagy]

No, this mesh is just a tutorial case. For real LES simulations, your mesh has to be finer.

[@Roselen1993]

Could you please make a tutorial of DES simulation and implications of grid refinement? I'd appreciate it so much!

[@ruizeng3076]

for open foam v6 it seems we have to keep nut in 0 folder

[@ruizeng3076]

Also oneEqn is called kEqn model now

[@OpenFOAMJozsefNagy]

@@ruizeng3076 Exactly! Thank you for your input.

[@davisalesbarreira2884]

Hello József, Congrats on your videos, they are very good! I was wondering about the boundary conditions related to the Velocity Field (U). Dont you have to generate some velocity variation in the inlet or use that mappedPatch, like in the pitzDailyMapped? I've been reading about the subject, but it's still not very clear to me how to set a proper boundary condition for U.

[@OpenFOAMJozsefNagy]

+Davi Sales Barreira Of course you have to think about your BC-s. Don't forget, this is a tutorial, I cannot go into too much detail. The videos are already long enough with this much content. But with the BC for k you already assume some amount of turbulent kinetic energy of the velocity oscillations.

[@siyeli2173]

Thank you Jozsef! I meet some problem when i use the incompressible/pisoFoam/LES/pitzDaily in OpenFoam-7. 1 There are the s, nut, k, P, U and nuTilda dictionaries in 0/,but not the nutSgs dictionary. 2 In fvSchemes, ddtSchemes is backward and div(phi, k) uses Gauss limitedLinear 1 scheme. 3 Why in fvSolution we add the pfinal into solvers Many thanks for your work, Keep it up!

[@OpenFOAMJozsefNagy]

The syntax changed since then. nuSgs was merged to nut, so you use for RAS and LES nut. pFinal is the entry for your last iteration in each time step. You can use the same as for all other p iteration, but you can change it here.

[@siyeli2173]

​@@OpenFOAMJozsefNagy Thank you very much, i looked into the xxxFoam.c and createfield.H dictionaries and found the pFinal term. By the way, could you talk about something about the overset grid and adaptive grid with OpenFoan ? Sincerely hope! Best wishes to you !!!

[@dramaswamy]

These are wonderful tutorials. But please, can you clear some of my doubts? What are the 'time step continuity errors' which are displayed during the run time of simulations? Do they affect our solution? And how can they be corrected/removed? Thanks in advance :)

[@OpenFOAMJozsefNagy]

That's a bit more involved. I will post a video on more tutorials in the next couple of weeks, please stay tuned!

[@gaming4life25]

Hi Jozef! I'm learning a lot from you. But I have trouble understanding the codes (I know only the basics of C++ and also took one class of CFD and one combustion in my last semester). I want to know how to get an in depth understanding of OpenFoam, so that I can experiment with it and use it for my masters thesis? Any reference would help, thanks again. Cheers :)

[@ilc15341]

Mr Nagy, basically the difference between these RANS and LES simulations is the ability to see or not the eddies forming correct? Would you say that the kind of turbulence model choosen would differ too much if we were interested in obtaining general overall pressure loss in a system? Thank you for your video.

[@OpenFOAMJozsefNagy]

The overall pressure loss should be equal, but it would be interesting to find out.

[@vigneshgopinathan2612]

hi, Your video is very informative. I have a doubt. I am using openfoam v7 and in the pitzdaily tutorial of LES there's a file "s" in the "0" folder. Can you please brief on it!?

[@bowenxu8129]

Thank you for all of this gorgeous work! I am simulating your examples in the version 4.1. There are some changes in the setup, and results are a little different. Unlike your models, my results of kEpsilon is very close to the other two models, is this normal? Thanks again!

[@OpenFOAMJozsefNagy]

I will post a video on more tutorials (also on updates of my tutorials) in the next couple of weeks, please stay tuned!

[@dodgewang700]

When dealing with Smagorinsky model, looks like streamlines file in system folder needs to be deleted, otherwise there will be an error said 'cannot find field k', I'm using v2206.

[@nsd6297]

Hello sir, want to know how to model surface roughness in turbulence. plz make a tutorial on surface roughness.

[@OpenFOAMJozsefNagy]

It is the nut wallfunction called nutkRoughWallFunctionFvPatchScalarField. Check out the file ~/OpenFOAM/OpenFOAM-v2012/src/TurbulenceModels/turbulenceModels/derivedFvPatchFie lds/wallFunctions/nutWallFunctions/nutkRoughWallFunction/nutkRoughWallFunctionFvPatchScalarField.H for more info.

[@foambuilder]

Hello Jozsef, many thanks for your work. I have a question: I always asked me what the meaning of writing output like "bounding k, min: ... max: ... average: ..." or "bounding epsilon ..." and so on... Not every time OF write this kind of comments, could they be issiuses for convergence?

[@attilafelfoldi7062]

tubulence kinetic energy or dissipation is too high due to low mesh quality somewhere in your domain

[@deadmau5e]

I love you

[@fulfill3d]

Hello Mr. Jozsef, Thank you so much for all tutorials you created. I learned a lot thanks to them and want to advance my ability a bit more. Thus, i want to create my own 2D CAD and own mesh. I use gmsh to do that but struggle to import my mesh to openFOAM. It gives me some errors but i have not found any explanations about them on the forums. Could you please create a video about that (for example creating a simple 2D geometry and mesh on gmsh and import to openFOAM) or give me some clue? Thanks.

[@OpenFOAMJozsefNagy]

I am sorry, but I have never used gmsh. I think on wiki.openfoam.com there is a tutorial with gmsh. Maybe that helps.

[@Vyacheslav7794]

Hi Thank you for your video. I have any questions. I'd like to simulate mixing of two gases without reactions (two supersonic flow). What kind of solver should be used?

[@OpenFOAMJozsefNagy]

one of the compressible ones with species transport

[@ashutoshsingh-et7vm]

Sir how to insert energy equation in simple or pisofoam

[@OpenFOAMJozsefNagy]

There is already rhoSimpleFoam and rhoPimpleFoam for that.

[@justinparker5513]

Hi! I seem to have a trouble in the smagorinsky case.. I deleted nut and retained nuSgs as you guided, but on running pisoFoam I got an error saying it cannot find nut. So I renamed nuSgs as nut, but the end result for the sub grid scale viscosity doesn't match with yours, it's completely uniform with very mild traces of eddies. The same happened with oneEq case as well. Am I doing something wrong? Thanks in advance!

[@OpenFOAMJozsefNagy]

This is true. In 2.3.0 the subgrid scale viscosity was nuSgs, but for newer versions this was rename into nut in order to have a more consistent syntax. They also changed the models themselves, so this is why you have slightly different results. Unfortunately I cannot update videos on KZbin, but I do update the tutorials with comments on wiki.openfoam.com/Transient_turbulence_modeling_by_Jozsef_Nagy so please check out the text there. I hope this helps.

[@justinparker5513]

Ahh! I thought so! Thank you so much!

[@OpenFOAMJozsefNagy]

I am glad I can help.

[@tyagiaryan82]

bhai bada krde yr