
[Sponsors] 
April 21, 2012, 17:01 
transient

#22 
Super Moderator
Tobias Holzmann
Join Date: Oct 2010
Location: Leoben (Austria)
Posts: 1,947
Blog Entries: 6
Rep Power: 33 
Hi all,
the solver seems to make good results with the steadyState solver and the Sandia flames. I wanna have a try in building my own transient solver for the flamelet model. Its the first time I am doing such an integration. Building and adding new scalars to a solver is not a problem but I think that 's a challenge. Okay I did the follow steps: Step 1: creating a new solver and compile it Code:
mkdir p $HOME/OpenFOAM/shorty2.1.x/applications/solvers/combustion/ cd $HOME/OpenFOAM/shorty2.1.x/applications/solvers/combustion/ cp r $FOAM_SOLVERS/compressible/rhoPimpleFoam . mv rhoPimpleFoam flamletRhoPimpleFoam cp r $HOME/OpenFOAM/libOpenSMOKE/src/applications/rhoSimpleFoam_flamlet1007 . (as reference) Step 2: adding variables and pdfThermo I did this by changing the createField.H file (added the lines need > refer to the original) changing the *.C file Code:
wclean wmake After that I changed the hEqn.H and added the missing files for csiEqn.H writeMass... etc. Code:
wclean wmake > floatingpoint Well the problem is, that the csiEqn.H is not with ddt()... so I added it with ddt(rho, csi) Now the solver workes for more then 100 iterations with error > floatingpoint. I have no experiance with the turbulence modelling like the following line: Code:
fvScalarMatrix csiEqn ( ( fvm::ddt(rho, csi) + fvm::div(phi, csi)  fvm::Sp(fvc::div(phi), csi)  fvm::laplacian(turbulence>alphaEff(), csi) ) );  fvScalarMatrix csiEqn ( ( fvm::ddt(rho, csi) + fvm::div(phi, csi)  fvm::Sp(fvc::div(phi), csi)  fvm::laplacian(turbulence>mut()/sigmat, csi) ) ); I think I have to get the correct equation or? And in which line should I add the Code:
#include csiEqn.H Its a very good exercise for me and I understand more and more  but till too less for that implementation. Any suggestions? Is that the right way to do something like that? Nice weekend tobi 

April 21, 2012, 17:44 

#23 
Super Moderator
Bruno Santos
Join Date: Mar 2009
Location: Lisbon, Portugal
Posts: 10,031
Blog Entries: 39
Rep Power: 110 
Hi Tobi,
Unfortunately I'm unable to comment on changes on the equations. But I can comment on this: did you properly modify "Make/files" on your modified solver? More specifically, the line that starts with "EXE = ", you should use the variable "FOAM_USER_APPBIN", as shown here: https://github.com/wyldckat/libOpenS...007/Make/files Secondly, (like I've told you before) you better have started to learn how to use git! And make good use of it, otherwise you might end up regretting it for accidentally erasing a file or a line, being then unable to go back due to not having some kind of a backup... Best regards and good luck! Bruno
__________________


April 21, 2012, 17:52 

#24  
Super Moderator
Tobias Holzmann
Join Date: Oct 2010
Location: Leoben (Austria)
Posts: 1,947
Blog Entries: 6
Rep Power: 33 
Quote:
Well the solver is running and using the flamelet library Code:
OpenSMOKE_PDF_NonAdiabaticFlamelet_Library File name: PDFLibrary/LookUpTable.out Number of Enthalpy defects: 1 Minimum enthalpy defect: 0 kJ/kg Maximum enthalpy defect: 0 kJ/kg Adiabatic flamelets: 1 Temperature fuel: 292 K Temperature oxidizer: 290 K Density fuel: 0.843587 kg/m3 Density oxidizer: 1.206918 kg/m3 Enthalpy fuel: 2196299.457 J/kg Enthalpy oxidizer: 109289.5676 J/kg Initialize basic fields... Initialize enthalpy field (field)... pdfThermo h(const scalarField& T, const label patchi)... end.. pdfThermo h(const scalarField& T, const label patchi)... end.. pdfThermo h(const scalarField& T, const label patchi)... end.. pdfThermo h(const scalarField& T, const label patchi)... end.. pdfThermo h(const scalarField& T, const label patchi)... end.. pdfThermo h(const scalarField& T, const label patchi)... end.. pdfThermo h(const scalarField& T, const label patchi)... end.. pdfThermo h(const scalarField& T, const label patchi)... end.. Calculate (first time)... Fuel enthalpy: 2196299.457 J/kg Oxid. enthalpy: 109289.5676 J/kg Fuel temperature: 292 K Oxid. temperature: 290 K Preparing additional scalar fields (references) Reading field U Reading/calculating face flux field phi Creating turbulence model Selecting RAS turbulence model kEpsilon kEpsilonCoeffs { Cmu 0.09; C1 1.47; C2 1.92; C3 0.33; alphah 1; alphak 1; alphaEps 0.76923; muLimiter on; Lsgs 0.0002; sigmak 1; sigmaEps 1.3; Prt 1; } Reading flamelet dictionary Preparing field Qrad (radiative heat transfer) Creating field dpdt Creating field kinetic energy K Starting time loop Courant Number mean: 8.44541524e07 max: 2.911575313e05 Time = 1e08 diagonal: Solving for rho, Initial residual = 0, Final residual = 0, No Iterations 0 DILUPBiCG: Solving for Ux, Initial residual = 1, Final residual = 1.708306365e14, No Iterations 2 DILUPBiCG: Solving for Uy, Initial residual = 0.9999999911, Final residual = 5.36071748e09, No Iterations 1 DILUPBiCG: Solving for Uz, Initial residual = 0.9999999879, Final residual = 1.929270108e15, No Iterations 2 DILUPBiCG: Solving for H, Initial residual = 0.9999999378, Final residual = 1.912266547e08, No Iterations 1 Correct thermodynamics Updating lookup table extractions... Updating mass fraction extractions... DILUPBiCG: Solving for csi, Initial residual = 1, Final residual = 1.93099693e08, No Iterations 1 Algebraic equation for csiv2... csiv2 max/min : 1502.115168 0 DICPCG: Solving for p, Initial residual = 0.9999987876, Final residual = 2.299981717e12, No Iterations 1 diagonal: Solving for rho, Initial residual = 0, Final residual = 0, No Iterations 0 time step continuity errors : sum local = 1.351669305e16, global = 7.808351117e17, cumulative = 7.808351117e17 rho max/min : 1.206986433 0.8436347068 DICPCG: Solving for p, Initial residual = 7.611717111e08, Final residual = 1.20032573e19, No Iterations 1 diagonal: Solving for rho, Initial residual = 0, Final residual = 0, No Iterations 0 time step continuity errors : sum local = 1.272683967e16, global = 7.714220478e17, cumulative = 1.55225716e16 rho max/min : 1.206986436 0.8436347095 DILUPBiCG: Solving for epsilon, Initial residual = 7.774178526e06, Final residual = 1.096653757e13, No Iterations 1 DILUPBiCG: Solving for k, Initial residual = 1, Final residual = 7.314882102e09, No Iterations 1 Calculating massFlow ExecutionTime = 0.23 s ClockTime = 0 s Courant Number mean: 8.445415785e07 max: 2.916317493e05 Time = 2e08 diagonal: Solving for rho, Initial residual = 0, Final residual = 0, No Iterations 0 DILUPBiCG: Solving for Ux, Initial residual = 7.9735332e06, Final residual = 8.601544082e14, No Iterations 1 DILUPBiCG: Solving for Uy, Initial residual = 0.2414552065, Final residual = 1.260364316e09, No Iterations 1 DILUPBiCG: Solving for Uz, Initial residual = 0.001292184778, Final residual = 6.911558089e14, No Iterations 1 DILUPBiCG: Solving for H, Initial residual = 0.001240166818, Final residual = 1.676567063e13, No Iterations 1 Correct thermodynamics Updating lookup table extractions... DILUPBiCG: Solving for csi, Initial residual = 0.001293280751, Final residual = 3.408964691e14, No Iterations 1 Algebraic equation for csiv2... csiv2 max/min : 0.03973394674 0 DICPCG: Solving for p, Initial residual = 0.003765993377, Final residual = 7.857355414e15, No Iterations 1 diagonal: Solving for rho, Initial residual = 0, Final residual = 0, No Iterations 0 time step continuity errors : sum local = 1.28577143e16, global = 7.288222584e17, cumulative = 2.281079418e16 rho max/min : 1.207054867 0.843722596 DICPCG: Solving for p, Initial residual = 5.596609331e08, Final residual = 8.830044552e20, No Iterations 1 diagonal: Solving for rho, Initial residual = 0, Final residual = 0, No Iterations 0 time step continuity errors : sum local = 1.225076996e16, global = 7.161084698e17, cumulative = 2.997187888e16 rho max/min : 1.207054878 0.8437226037 DILUPBiCG: Solving for epsilon, Initial residual = 5.762029595e06, Final residual = 5.398702051e15, No Iterations 1 DILUPBiCG: Solving for k, Initial residual = 0.001755283888, Final residual = 3.878694754e13, No Iterations 1 Calculating massFlow ExecutionTime = 0.3 s ClockTime = 0 s Courant Number mean: 8.445415844e07 max: 2.916591386e05 Time = 3e08 But you are right. I should make use of git to get back to older files or so on But for today I am very tired. 

May 3, 2012, 09:17 

#25 
Super Moderator
Tobias Holzmann
Join Date: Oct 2010
Location: Leoben (Austria)
Posts: 1,947
Blog Entries: 6
Rep Power: 33 
Hi Tom,
are you still testing the flamelet solver and are you using the standard 2006 transport and thermodynamic files? I used the newer one but there is an error while executing the chemkinInterpreter. Well I contact the authors but they are not writing back And further more, do you know what the Code:
SOLVE CYCLE REFINELEANSIDE Tobi 

May 3, 2012, 10:08 

#26 
Senior Member

Hi Tobi.
In fact I am still looking at it from time to time. I have started looking at the GRI3.0 mechanism, and I managed to use that in order to create a flamelet library to redo the project I mentioned earlier, checking if I can get similar results, or maybe improve the result. Maybe you can check the validity of the Chemkin files with the standard OpenFOAM tools, like chemkinToFoam? About the refineleanside, solve and cycle I am also not entirely sure, but it seems to me that it represents some numerical integration parameters in order to solve the flamelet equations and adaptively refine the grid based on the result that was obtained. For my library I just used the default setup from the tutorial and I got a result for the GRI3.0 mechanism, transport and thermodynamic data, which at least seems to give appropriate intermediate results, but I am not yet far enough to comment on the accuracy. If it is possible to comment on it I will let you know. I Hope you will receive some more information from the original authors. Regards, Tom 

May 3, 2012, 16:10 

#27  
Super Moderator
Bruno Santos
Join Date: Mar 2009
Location: Lisbon, Portugal
Posts: 10,031
Blog Entries: 39
Rep Power: 110 
Hi Tobi and Tom,
It looks to me like the authors might be very busy with the project(s) related to their presentation(s) this year at OFW7  http://extendproject.de/ofw7technicalprogram: Quote:
Best regards, Bruno
__________________
Last edited by wyldckat; May 29, 2012 at 15:20. Reason: fix broken link 

May 3, 2012, 16:17 

#28  
Super Moderator
Tobias Holzmann
Join Date: Oct 2010
Location: Leoben (Austria)
Posts: 1,947
Blog Entries: 6
Rep Power: 33 
Quote:
Yes that would be very nice, but I think my english is too bad to discuss with the guys and there is no time for me  shit! 

May 3, 2012, 17:12 

#29 
Senior Member

Hi Bruno,
Thanks a lot for the information. I will be there, so I think I will go and have a chat with him/them. I compiled the library on our cluster today, just have some problem reading the flamelets while running the solver there, which is a shame since it means a large reduction in computational resources, but have to investigate on that. Tom 

May 3, 2012, 17:13 

#30 
Super Moderator
Bruno Santos
Join Date: Mar 2009
Location: Lisbon, Portugal
Posts: 10,031
Blog Entries: 39
Rep Power: 110 
Quick post  I'm searching for projects on sourceforge.net that are somehow related to OpenFOAM and I tripped on this one: http://sourceforge.net/p/ofca/code/  although it looks like it's for the variant 1.6ext only...
__________________


May 3, 2012, 17:20 

#31  
Super Moderator
Tobias Holzmann
Join Date: Oct 2010
Location: Leoben (Austria)
Posts: 1,947
Blog Entries: 6
Rep Power: 33 
Quote:
i am very interessted in your chat! It would be very nice if you report me some details about your chat. Thx tobi 

May 3, 2012, 17:41 

#32 
Senior Member

Well I wonder if I can get something out of the chat. I'll see.
The edcSimpleFoam also sounds interesting. No problem if it is 1.6ext, we have at least 1.52.1 and 1.5dev/1.6ext still lying around in some dusty old harddrive, just in case. More stuff on the todolist. 

May 29, 2012, 09:18 

#33 
Super Moderator
Tobias Holzmann
Join Date: Oct 2010
Location: Leoben (Austria)
Posts: 1,947
Blog Entries: 6
Rep Power: 33 
Hi all,
just one question. Is it possible to get the source code for LibOpenSMOKE using doxygen? I am reading a few days in the code but it would be easier to use doxygen and get graphics for the classes and functions. Is that possible? tobi I SOLVED IT! Last edited by Tobi; May 29, 2012 at 09:39. 

May 29, 2012, 11:41 

#34  
Super Moderator
Tobias Holzmann
Join Date: Oct 2010
Location: Leoben (Austria)
Posts: 1,947
Blog Entries: 6
Rep Power: 33 
Quote:
Hi bruno, i can not open the link! Is that event in Darmstadt? Tobi PS: Tom did you tried the binarys for other kinetic schemes ? 

May 29, 2012, 11:55 

#35 
Senior Member

Hi Tobi,
You mean the tools (shell scripts) to generate different flamelet libraries from alternative kinetic/thermophysical databases? Than yes, I used those, following the recipe from the user guide for libOpenSMOKE. However, I am not entirely sure about the validity of the resulting library from the scripts. Qualitatively it is satisfying however. Regards, Tom 

May 29, 2012, 12:16 

#36  
Super Moderator
Tobias Holzmann
Join Date: Oct 2010
Location: Leoben (Austria)
Posts: 1,947
Blog Entries: 6
Rep Power: 33 
Quote:
Hi tom, i mean the other files you can download from CRECK. Like a kinetic scheme for biomass combustion or sth. like that. I created a bashscript for generating flamelets very simply but I am not able to use other schemes, thermodynamics and transportProperties 

May 29, 2012, 15:24 

#37  
Super Moderator
Bruno Santos
Join Date: Mar 2009
Location: Lisbon, Portugal
Posts: 10,031
Blog Entries: 39
Rep Power: 110 
Hi Tobi and Tom,
I've fixed the link. Here it is once again: http://extendproject.de/ofw7technicalprogram As for the location: Quote:
Best regards, Bruno
__________________


May 30, 2012, 02:53 

#39 
Senior Member

Hi Tobi,
I have not checked those files, however I did manage to get a flamelet library based on the GRI 3.0 mechanism, so in principle it should work I guess. I do not know about the details however. 

May 30, 2012, 05:00 

#40 
Super Moderator
Tobias Holzmann
Join Date: Oct 2010
Location: Leoben (Austria)
Posts: 1,947
Blog Entries: 6
Rep Power: 33 
Hi Tom,
yea i thought that too, couse the files are the same,... just having more species in it. Well in two weeks when I have my new flat I ll be able to try around with the files at home. Its a pitty that I am too late for the 7th Workshop. It would be very nice to meet you too, tom. Tobi 

Thread Tools  
Display Modes  

