[stefanke]

Niklas,

I try to run dieselEngineFoam in parallel with a simple wedge geomerty. It crashes everytime when fuel is injecting. Without fuel injection all is ok! This bug is very easy to reproduce.

When I run this case nonparallel all works fine including fuel injection. So it has something do with the parallel computation!

[dhebert]

Hello All,

I am having the same problem as Wladimyr Mattos da Costa Dourado above in trying to restart dieselFoam. I have tried both latestTime and startTime (with appropriate startTime) in controlDict, and simulation crashes when evolving spray. Is there a work-around or fix to this problem that I am missing? Note I am using downloaded binary of OF version 1.3.

Thanks for your help,

David

[dhebert]

Hello All,

I am having the same problem as Wladimyr Mattos da Costa Dourado above in trying to restart dieselFoam. I have tried both latestTime and startTime (with appropriate startTime) in controlDict, and simulation crashes when reading temperature field:
.
.
.
Evolving Spray

--> FOAM FATAL ERROR : attempt to use janafThermo<equationofstate> out of temperature range 200 -> 6000; T = 3.88739e+252


Is there a work-around or fix to this problem that I am missing? Note I am using downloaded binary of OF version 1.3. Also note all fields are written in binary.

Thanks for your help,

David

[hjasak]

Wouldn't you say the temperature of T = 3.88739e+252 Kelvin is a bit high? How did you get such a high temperature?

Hrv

[dhebert]

Hrvoje,

I agree with you, the high temperature value made me think that the problem was with reading the stored temperature field. I also set the maxCo to 0.01 and obtained the same error. The error seems like it is a binary precision error (single instead of double, or vice versa). I checked the .dep files and they all seem to be double.

Note that I get the same error when I try to restart the tutorial dieselFoam/aachenBomb case. I ran until time 0.0006, then stopped. I attempted to restart and got the same temperature error when evolving spray.

Thanks again for you assistance,

David

[lucchini]

David,
there is a bug in the spray class which does not read correctly the lagrangian field when a new simulation is restarted. Send me an e-mail and I will send you an updated version of the lagrangian class.
Bye
Tommaso

[chris1980]

Tommaso, it is possible to post your updated version of the lagrange class. I think a lot of people are interested in.

Thx

[dhebert]

Tommaso (and anyone else with new lagrangian class)

Are you able to compile decomposePar with the corrected spray/lagrangian class you sent? When I try to compile decomposePar (after wclean of decomposePar and decompositionMethods), I get the following error:



In function `Foam::Cloud<foam::indexedparticle>::Cloud(Foam::p olyMesh const&)':
decomposePar.C: undefined reference to `Foam::Cloud<foam::indexedparticle>::readFields()'
collect2: ld returned 1 exit status


....

This was done on a single processor AMD64 machine.

Thanks for the help,

David

[dhebert]

Hello everyone,

I tried to run the dieselFoam aachenBomb tutorial with a single processor. After a little time running, I received the following error message:

--> FOAM FATAL ERROR : NO_READ specified for read-constructor of object meshPhi of class IOobject

Any pointers as to why this would suddenly happen, and how to fix?

Thanks for all your help,

David

[lucchini]

Hi David,
this usually happens if you have switched on the runTimeModifiable option and clock skew.
Try to switch runTimeModifiable off in the controlDict file and see what happens.
Bye
Tommaso

[dhebert]

Thanks Tommaso for your suggestion. I will give that a try shortly.

I would like to post a question to the board regarding compressible flow in an expansion. I am trying to use dieselFoam to model a spay generated from an injector used in a rocket application. For now I am trying to model a cold-flow case, where air and water are injected at 293K.

The air inlet velocity is Mach 0.7. Water is injected at 5m/s. It seems that when the air reaches the expansion the temperature drops (presumable due to isentropic expansion), and I get a JANAF table temperture error. Does anyone have suggestions for IC/BC? So far I have tried

1) Initial T = 300 and 1400K
2) Pressure 1atm
3) inlet fixed velocity and timeVaryingUniformFixedValue (to try to reduce any pressure waves)
4) Initial interrior U 0, 25, and 200m/s
5) Outlet pressure fixed at 1atm.
6) laminar and realizableKE turb model


Any insight into a flow of this type is greatly appreciated.

Thank you,

David

[lucchini]

David,
what is the temperature value you are getting when you have the fatal error?
are you running with adjustTimeStep switched on? If so, what is the maximum Courant number?
try to run the case with maxCo among 0.05 and 0.2 and see what happens.

[dhebert]

The temperature value at failure seems to vary. Sometimes it is in the range of 168K, other times it is -24K (which, of course, is not physically possible).

I do run with the adjustTimeStep on. We have been trying to run with a high maxCo. I have been between 0.7 and 0.9. This might explain the wide range in temperature failure. I was able to run with such a high maxCo for another simulation with Ma=0.3, so I was thinking that it would be ok here. I will set the maxCo lower as you suggest and see if that solves the problem. I am concerned that with isentropic expansion with such a high Mach number that the temperature will always drop out of range.

Also, I tried running the case using sonicTurbFoam (same gas phase setup, no spray), and got the same problem. Again, I was using a high maxCo of 0.9.

Thanks again for your assistance,

David

Hej,

For my model I need temperature range from 300 K up to 20 000 K as a B.C. . I am getting error when I start the case buoyantSimpleFoam. Could someone to help me to uderstand : am I doing something wrong or OpenFOAM could not working with such big value of temperature?

Thank u,
Rita

[gschaider]

Hi Rita

May I ask what you are simulating? My sources say (havn't been there myself) that the temperature at the surface of the sun is 5800 Kelvin, so your temperature range is quite impressive.

The problem is not that that OpenFOAM can't handle such temperature range, I just doubt whether the approximations done by buoyantSimpleFoam are right for such a case (is it steady? etc)

So if you tell us more about the problem you want to simulate, maybe someone can point you to a fitting solver

[dhebert]

Greetings everyone,

When I try to run the aachenBomb case in parallel, I am only getting a lagrangian/positions file when I reconstructPar, as reported by Stefan above.

In looking through the processor directories, I notice that some have a full lagrangian directory, while others only have positions. Could this be because some processors do not have lagrangian particles in their domain? And if so, should reconstructPar take care of this? Is there a fix I missed somewhere?

Thanks for any help,

David

Hej,

For my model I need temperature range from 300 K up to 20 000 K as a B.C. . I am getting error when I start the case buoyantSimpleFoam. Could someone to help me to uderstand : am I doing something wrong or OpenFOAM could not working with such big value of temperature?

Thank u,
Rita

[hjasak]

Hello Margarita,

I would guess you are using JANAF tables for material properties, which have an upper limit of 5000 Kelvin on the fitting functions. Therefore, if you wish to use this kind of temperature, you need another source of material properties or you need to fix the coefficients yourself (awfully high, though).

Enjoy,

Hrv

Hej,

I am trying to model a weld process. At the begining I took just momentum and energy equations without electromagnetic part (temperatures range 300K-20000K,steady,argon gass is injected and I'm using buoyantSimpleFoam). I have got very strange results, which are so far from reality.
Maybe someone could help understand what I did wrong?

Thank you for your answers or explanations.
Rita

[joakim]

Hi Magarita

I can understand why things looks strange for T=20000K, but do the result also look strange for 300K?

Regards

/Joakim