THIS IS AN AUTOMATED MESSAGE, DO NOT REPLY.
The following task has a new comment added:
FS#134 - reintroduce the physical-functions
User who did this - Philipp Nuske (PhilippNuske)
----------
I introduced some sort of physicalness for the mpnc model.
So, just as an idea, this is what I do currently: call the physicalness after
the fluidstate was updated. Check whatever I like. Print location and some more
variables to the screen, if a criterion was violated. Throw numerical problem,
i.e. proceed with smaller timestep.
There is a precompiler macro around the call to the physicalness function. This
way it is not called if I am explicitly not debugging.
In the end the results with and without physicalness are comparable. Without
physicalness function the matrix becomes singular and the timestep is reduced.
With physicalness the timestep gets reduced and we proceed with a very small
step.
However, I am using this mainly for debugging. This way I can see, WHERE things
go wrong. This was very helpful for me.
Benefits:
- more meanignful output: I see what goes wrong.
- more meanignful output: I see where sth goes wrong.
- simulation runs a little longer, because unphysical results are rejected from
the start.
in volume variables:
#ifndef NDEBUG
const GlobalPosition & globalPos =
fvGeometry.subContVol[scvIdx].global ;
physicalness(fluidState_, globalPos);
#endif
const void physicalness(const FluidState & fs,
const GlobalPosition & globalPos) const
{
const Scalar eps = 1e-6 ;
for(int phaseIdx=0; phaseIdx<numPhases; phaseIdx++){
for (int compIdx=0; compIdx< numComponents; ++ compIdx){
const Scalar xTest = fs.moleFraction(phaseIdx, compIdx);
if (not std::isfinite(xTest)
or xTest < 0.-eps
or xTest > 1.+eps ){
std::stringstream positionString;
positionString << \"Here:\";
for(int i=0; i<dim; i++)
positionString << \" x\"<< (i+1) << \"=\"
<< globalPos[i] << \" \" ;
positionString << \"\\n\";
std::stringstream message ;
message << \"Unphysical value found: x[\"<<phaseIdx<<\"][\"
<<compIdx<<\"]=\"<< xTest <<\"\\n\" ;
message << positionString.str() ;
message << \"with \" ;
for(int phaseIdx=0; phaseIdx<numPhases; phaseIdx++)
for (int compIdx=0; compIdx< numComponents; ++ compIdx)
message << \" x\"
<<\"_\"<<FluidSystem::phaseName(phaseIdx)
<<\"^\"<<FluidSystem::componentName(compIdx)<<\"=\"
<< fs.moleFraction(phaseIdx, compIdx) ;
for(int energyEqIdx=0; energyEqIdx<numEnergyEqs;
energyEqIdx++)
message << \"T\"
<<\"_\"<<FluidSystem::phaseName(energyEqIdx)<<\"=\"<<
fs.temperature(energyEqIdx) ;
// Ts is only known in the energyVolVars
DUNE_THROW(NumericalProblem, message.str());
}
}
}
}
----------
More information can be found at the following URL:
http://www.dumux.org/flyspray/index.php?do=details&task_id=134#comment322
You are receiving this message because you have requested it from the Flyspray
bugtracking system. If you did not expect this message or don't want to
receive mails in future, you can change your notification settings at the URL
shown above.
_______________________________________________
Dumux mailing list
[email protected]
https://listserv.uni-stuttgart.de/mailman/listinfo/dumux
