|
[Sponsors] |
July 6, 2013, 09:49 |
|
#201 |
Super Moderator
Tobias Holzmann
Join Date: Oct 2010
Location: Bad Wörishofen
Posts: 2,711
Blog Entries: 6
Rep Power: 52 |
Hi bobi and all,
I compiled OF-2.2.x on my laptop today and tried the flamelet model 2.2.x: ------------------------------------------------------------------------------------------------------------- 1. git clone ... is working fine (no errors) 2. flamelet libs compiled well 3. thermodynamics do not compile That problem is solved by removing a deb file 4. flameletSimpleFoam is not compiling That problem is going to solved. The files are "crashed" there are some nonsence lines in the files. I remove them now and clean the files 5. flameletPisoFoam is compiling 6. tutorial SIMPLE is not working due to changes in the PDF library That problem is going to be solved 7. tutorial PISO is not working due to changes in the PDF library That problem is going to be solved Tobi |
|
July 6, 2013, 12:09 |
|
#203 |
Senior Member
Bobby
Join Date: Oct 2012
Location: Michigan
Posts: 454
Rep Power: 16 |
Many Thanks to my friend Tobi
|
|
July 24, 2013, 02:31 |
|
#204 |
Senior Member
Bobby
Join Date: Oct 2012
Location: Michigan
Posts: 454
Rep Power: 16 |
Hi Tobi
What is the convergence criterion in your simple and piso codes and how can we increase or decrease the convergence accuracy? I mean where should we implement the modifications? Regards] Bobi |
|
July 29, 2013, 07:31 |
convergence
|
#205 |
Super Moderator
Tobias Holzmann
Join Date: Oct 2010
Location: Bad Wörishofen
Posts: 2,711
Blog Entries: 6
Rep Power: 52 |
Hi Bobi,
as you can see in the sandia-flame the residuals never go below 1e-3 - 1e-4. In that case I wait till i reach a "steady-state" residual state; like a horizontal line. But that depend on your case. In my masterthesis it was possible to get a residual to 1e-5 but for that I decreased the relax factor strong. Further more I always save the last time steps (like every 100 iterations) and compare the differences between the slices/Clips/values to be sure that I reached a steady state solution. Problem in most cases I had is the fact, that the combustion Chambers (in my case) never be a steady. I always realize transient movements in my values on certain Points in the geomtry. For the transient (piso) algo I use the time, residuals and visualisation. In the sandia-flame I get a better convergence in my Simulation but that was all I simulated with my transient solver. |
|
August 13, 2013, 05:33 |
About simpleFoamflamelet solver
|
#206 |
Member
sonu
Join Date: Jul 2013
Location: delhi
Posts: 92
Rep Power: 13 |
helo ,
After successfully running the tutorial of Sandia COH2N2 flame in flamelet model under the simplefoamflamelet solver now i want to try it for my geometry and case . i had converted my mesh file into OF , so now can anybody tell me which file i have to copy from the tutorial and kept in my working folder to solve my case ? |
|
August 13, 2013, 12:34 |
|
#207 |
Senior Member
Bobby
Join Date: Oct 2012
Location: Michigan
Posts: 454
Rep Power: 16 |
Hi
You should make your own PDF file, 0 (initial conditions) and constant folder for required coefficients. Just the same as Tobi tutorial. Bobi |
|
August 14, 2013, 04:48 |
|
#208 |
Member
sonu
Join Date: Jul 2013
Location: delhi
Posts: 92
Rep Power: 13 |
thanks alot bobi .....i will try as u said ....
|
|
August 21, 2013, 10:04 |
Mixing of two gases with turbulentFlameletRhoSimplecFoam
|
#209 |
Member
Join Date: May 2013
Location: Netherlands
Posts: 30
Rep Power: 13 |
The turbulentFlameletRhoSimplecFoam creates a lot of opportunities for me for steady state combustion modelling in OpenFOAM. In the end, I am very interested in simulations with chemical reactions and combustion involved and that is what this solver is doing. Also thank you for the good documentation about the chemistry involved in this solver, that is what is missing for some other chemistry solvers on CFD online.
However, I am also interested in steady state modelling of the mixing between two different gases (fuel and air). So a cold flow case without the actual chemical reaction which shows the mass or volume fractions of the different species. Is it possible to do this with turbulentFlameletRhoSimplecFoam and what do I need to change for this (e.g. in the library). I am using OpenFOAM 2.1.1. |
|
August 21, 2013, 12:23 |
|
#210 |
Senior Member
Bobby
Join Date: Oct 2012
Location: Michigan
Posts: 454
Rep Power: 16 |
Hi TBO
My suggestion for you for simulating cold flows is using RhoPimpleFoam and RhoSimpleFoam solvers of OpenFoam, cause their flow solvers are exactly the same as libOpenSmoke modified versions. LibOpenSmoke and its modified bversions are steady flamelet solvers which automatically spread combustion according to mixture fraction value in the domain. Good Luck Bobi |
|
August 22, 2013, 03:36 |
|
#211 | |
Member
Join Date: May 2013
Location: Netherlands
Posts: 30
Rep Power: 13 |
Quote:
Hi Bobi, Thank you for your reply, please correct me if I'm wrong, since I'm not very experienced with OpenFOAM so far. I am already using rhoSimplecFoam for cold flow modelling, however what I would like to do is calculate the mixture fractions of different species in the domain. So for example the mixing of methane jets in an air flow. As far as I know rhoSimplecFoam can only cope with one species (if anyone can tell me how to implement several species in rhoSimplecFoam, I am very satisfied). I already looked into other solvers, e.g. reactingFoam and there you have the possibility to switch off chemical reactions (however this solver is transient, resulting in extreme calculation times). Is there any possibility to switch off the chemical reactions (or modify the library) for the libOpenSmoke? |
||
August 22, 2013, 06:29 |
|
#212 |
Senior Member
Bobby
Join Date: Oct 2012
Location: Michigan
Posts: 454
Rep Power: 16 |
Hi TBO
HTML Code:
if anyone can tell me how to implement several species in rhoSimplecFoam, I am very satisfied). HTML Code:
Is there any possibility to switch off the chemical reactions (or modify the library) for the libOpenSmoke? I hope you find this helpful Bobi |
|
August 22, 2013, 06:46 |
|
#213 |
Member
Join Date: May 2013
Location: Netherlands
Posts: 30
Rep Power: 13 |
Hi Bobi,
Thank you for your answer. Indeed I mean various inlet constituents without having combustion. I already looked into fireFoam (as well as reactingFoam) before, however the main problem with these solvers is that they are transient and I would like to model steady state (since this saves a lot of computational time). I already looked into several other options (alternateSteadyReactingFoam, edcSimpleFoam). However results for these solvers are not satisfying jet and documentation/CFD online discussions are very limited. Regards, |
|
August 22, 2013, 12:15 |
|
#214 |
Senior Member
Bobby
Join Date: Oct 2012
Location: Michigan
Posts: 454
Rep Power: 16 |
Hi TBO
HTML Code:
I already looked into fireFoam (as well as reactingFoam) HTML Code:
I would like to model steady state HTML Code:
documentation/CFD online discussions are very limited Bobi |
|
September 16, 2013, 06:41 |
|
#215 |
New Member
parsa
Join Date: Sep 2013
Posts: 3
Rep Power: 13 |
Dear All,
I'm doing also the same job with OpenFoam, but until now, i didn't get the semialer result as I got in Transient mode. I made several steady-State Solvers, but seems that No Combustion happens! I'm using the 'psi Combustion Model' which is used in reactingFoam Solver as well. If any of you could run the Transient Solver in Steady-State mode ( with delta T=1 for example , and not with deltaT= 10^-6) or already created a new steady-state solver , I would be so grateful if help me this way, looking forward to your replies, Sincerely, Parsa |
|
September 16, 2013, 13:03 |
|
#216 |
Super Moderator
Tobias Holzmann
Join Date: Oct 2010
Location: Bad Wörishofen
Posts: 2,711
Blog Entries: 6
Rep Power: 52 |
Dear all,
this thread is for libOpenSMOKE and not for other combustion solver. To have a clear thread, please make new threads with your problems Thanks Tobi |
|
September 16, 2013, 15:28 |
|
#217 |
Senior Member
Bobby
Join Date: Oct 2012
Location: Michigan
Posts: 454
Rep Power: 16 |
Hi Tobi
I have an issue with k parameter in simulating bluff-body flames. what should I take for this value? For instance; In your Sandia_CO/H/N2 you have mentioned as: Code:
outlet { type zeroGradient; } axis { type empty; } sidewall { type compressible::kqRWallFunction; value uniform 1; } burnerwall { type compressible::kqRWallFunction; value uniform 1; } Code:
bluffbody { type compressible::kqRWallFunction; value uniform 17.34; } back { type wedge; } outlet { type inletOutlet; inletValue uniform 1; value uniform 1; } My specific issue is its value for the bluff-body boundary as a kind of wall. It is the initial value, So is it important choosing the exact right value. I suggest maybe just estimating a value and starting my simulation. Am I right Buddy? Regards Bobi Last edited by babakflame; September 17, 2013 at 05:22. |
|
September 17, 2013, 07:10 |
|
#218 | |
Super Moderator
Tobias Holzmann
Join Date: Oct 2010
Location: Bad Wörishofen
Posts: 2,711
Blog Entries: 6
Rep Power: 52 |
Quote:
Hi Parsa, just one question. Are you using the libOpenSMOKE tool and the flamelet solver or a other solver? It s confusing cause you mentioned you are using the ''psi Combustion Model" ? Tobi |
||
September 17, 2013, 07:16 |
|
#219 | |
Super Moderator
Tobias Holzmann
Join Date: Oct 2010
Location: Bad Wörishofen
Posts: 2,711
Blog Entries: 6
Rep Power: 52 |
Quote:
you are right. The values you choose are just initial values. After the first iteration the value change. But if you choose your value very good your simulation get to a steady state solution faster. Additionally the first timesteps could be hard for openfoam to calculate due to bounding things and stuff like that. So its good to have a accurate value. In my tutorial case its just a standard value if I do not know a correct one I always begin with 0.01 to 1 ... due to the velocity and geometry. The inlet value can be calculated (not exact but good for initialising). For the wall I could use 10 or 100 too So just start with 0.1 or 1 and check the first 10 or 50 timesteps / iterations ... if you have no problems with the wall value then everything is okay. I read in an other thread that People set the k and Epsilon values to 1e5 or sth. like that Well in my opinion you should set a correct dimension. So if you know that the inlet has k=0.0032 then I choose the same Dimension for the walls like 0.001 ... but not 10 Hope this answer will help you for your Problem. Tobi |
||
September 18, 2013, 15:11 |
|
#220 |
Member
sonu
Join Date: Jul 2013
Location: delhi
Posts: 92
Rep Power: 13 |
greetings oll ,
i have a small doubt that in 0 folder the value of each species (like CH4) in boundary fields is in mole fraction . rite ?? not the mass fraction correct me if i am wrong Thank you , Regards sonu |
|
Thread Tools | Search this Thread |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Numerical treatment of the source term in combustion equations | Tobi | Main CFD Forum | 37 | September 15, 2020 14:42 |
[openSmoke] flameletSmoke + new ODESolver (by Alberto Cuoci) | Tobi | OpenFOAM Community Contributions | 1 | November 21, 2017 19:24 |
Unsteady solver with Flamelet Model (libOpenSMOKE) | francesco_capuano | OpenFOAM Running, Solving & CFD | 11 | November 26, 2013 05:50 |
LibOpenSmoke, getting the species in ParaFoam | Christoph_84 | OpenFOAM | 1 | May 31, 2012 15:42 |