[hansihans]

Hello,

For my bachelor thesis I want to simulate fuel cells.
I installed openfuelcell according to the official website on ubuntu 14.04 with
openfoam 6 and it works great on single run (make srun).
But if I want to start the case quickTest in parallel (4 cores) I have two problems:

1: "make parprep": the script parprep.csh has some commands (remove commands that do not work in script but manually, and those batch files which run correctly but will not close/exit by themselves)

-> I modified the parprep.csh file a bit so it will work (but batch files do not run in parallel now)

2: "make run": there comes a error message "--> FOAM FATAL ERROR: cannot find porous cellZone cathode"
"From function Foam:orousZone:orousZone(const Foam::word&, const Foam::word&, const Foam::fvMesh&, const Foam::dictionary&, const Foam::word&)
in file porosityModel/porosityModel/porousZone.C at line 142."

-> I have no idea what the problem is and how to solve it

The thing is I want to make e.g. parameter variation using pyFoam and therefore I need the ability to run in parallel. otherwise it would take ages for simulating.

is there any patch or another version that works ?
does parallel run work in older versions ?
do I need to install something (like cshell for the parprep.csh file) ?
can I make it using pyFoam ?

If there is a possibility to get the old V3 version that I think worked with parallel run then please let me know.


Thanks a lot



***Edit:
I installed openfoam 5 to give it a try, but I could not compile openfuelcell.
So I make the mesh and the parprep part with OF5, copied the file to my OF6 directory and made the parallel run with OF6.
The only difference with OF5 in the parprep part was, that I used the modified parprep.csh file from OF6 (but let the "setset" with batch files run in parallel).
=>In OF5 these commands exit/close by themselves if they are finished but in OF6 they do not exit/close by themselves.

So in this status it is not possible to plan some parameterVariations in parallel run because it would take too long to always prepare the case in OF5 then copy the case and run in in OF6.

I also made a graph comparison of the srun and prun of the stack voltage over the timesteps: while it looks normal in srun, the voltage of the parallel run always swings up and down over the time. (I don´t know how to send pictures here).

I try to contact the developers to get more information

[Swagga5aur]

Hello Hans Xaver,

I have been working with the openFuelCell code for quite a bit now (mainly openFoam 6), I haven't

It seems that the cellzone for the cathode is not in your parallel case as you have it in the single processor case I guess its the parprep which is an issue.

I had a similar issue with the setSet commands never closing, I would suggest changing them to topoSet commands instead to generate the different zones.

Such as the following

mpirun -np $NPROCS foamExec topoSet -region fuel -parallel
mpirun -np $NPROCS foamExec topoSet -region air -parallel

This of course requires you to add a topoSet dict for each region in the system directory.

Let me know if you have any questions regarding.

Best regards,

Lasse