[huyidao]

Hello,Foamers.
I am trying to use the chtmultiregionsimpleFoam right now,I go through the cases in the tut,but in the fvSolution files ,there are no convergence criterion settings like other solvers in the residualcontrol subdict.When I run the tut cases,I found that it ended at the given end time.
Can I set the convergence criterion in the chtmultiregionSimpelFoam?How can I do it?

[zfaraday]

Hi huyidao,

The answer is no, sadly no convergence criteria can be used to stop the solver once the criteria has been met... It always stops when the endTime is reached.

Regards,

Alex

[huyidao]

Thank you so much ,Alex.
As there is no such convergence ctriterio,then I have to add it myself.
I will post my code here when I finish,feel free to point out any of my errors.
Regards,
huyidao

[zfaraday]

It will be a good enhancement if you finally can develop such a thing! If you do it, post it to the bug tracker also so that they can implement it to the code and all the users will be able to enjoy your work!

Regrads,

Alex

[huyidao]

Hello,Alex!
It takes me some time to do this,because I am not quite familiar with
C++.
Here is what I have done.
1:
I defined a new class named chtSimpleControl,it is almost the same as
simpleControl,but some sentences in the function loop() have been changed.

bool Foam::chtSimpleControl::loop()
{
read();

Time& time = const_cast<Time&>(mesh_.time());

if (initialised_)
{
if (criteriaSatisfied())
{
//Info<< nl << algorithmName_ << " solution converged in "
// << time.timeName() << " iterations" << nl << endl;

// Set to finalise calculation
//time.writeAndEnd();
storePrevIterFields();
return false;
}
else
{
storePrevIterFields();
return time.loop();
}
}
else
{
initialised_ = true;
storePrevIterFields();
return time.loop();
}


//return time.loop();
}

2:
The file createChtSimpleControl.H is created.

List<autoPtr<chtSimpleControl> > fluidConvergence(fluidRegions.size());
List<autoPtr<chtSimpleControl> > solidConvergence(solidRegions.size());

forAll(fluidConvergence,i)
{
fluidConvergence[i].set
(
new chtSimpleControl(fluidRegions[i])
);
}


forAll(solidConvergence,i)
{
solidConvergence[i].set
(
new chtSimpleControl(solidRegions[i])
);
}

3:
The createChtSimpleControl.H is included in the chtMutiRegionSimpleFoam.C,and the iteration
loop is redefined as fllowed.

bool run=true;
while (run)
{
Info<< "Time = " << runTime.timeName() << nl << endl;

forAll(fluidRegions, i)
{
Info<< "\nSolving for fluid region "
<< fluidRegions[i].name() << endl;
#include "setRegionFluidFields.H"
#include "readFluidMultiRegionSIMPLEControls.H"
#include "solveFluid.H"
}

forAll(solidRegions, i)
{
Info<< "\nSolving for solid region "
<< solidRegions[i].name() << endl;
#include "setRegionSolidFields.H"
#include "readSolidMultiRegionSIMPLEControls.H"
#include "solveSolid.H"
}

runTime.write();

Info<< "ExecutionTime = " << runTime.elapsedCpuTime() << " s"
<< " ClockTime = " << runTime.elapsedClockTime() << " s"
<< nl << endl;

run=false;
forAll(fluidConvergence,i)
{
if(fluidConvergence[i]().loop())
{
run=true;
break;
}
}

forAll(solidConvergence,i)
{
if(solidConvergence[i]().loop())
{
run=true;
break;
}
}
}
runTime.writeAndEnd();

Info<< "End\n" << endl;

return 0;

I tested it in the tut heatExchanger,It seems that it works fine,but this may not be the best way to do this and there may be still some bugs feel free to point out any of my errors.

And I still have a question,where is the time increasment defined like runTime++ or something else?The runTime++ method is not called explicitly in this iteration loop.

[zfaraday]

Thanks for your effort huyidao, this a very good job! I don't know if it may be done in a better way as you point out or not, my knowledge in C++ is much weaker than yours. However, I recomend you to report your work to the bug tracker so that they can put it in the repository and more people will be able to enjoy it and, in case it is necessary, they will be able to report bugs in your code in orther to improve it.

Great job! It's a pitty that nobody did it before so that I could use it during the developement of my final thesis...

Many thanks again. Best regards,

Alex

[jmdf]

Originally Posted by huyidao

Hello,Alex!
It takes me some time to do this,because I am not quite familiar with
C++.
Here is what I have done.
1:
I defined a new class named chtSimpleControl,it is almost the same as
simpleControl,but some sentences in the function loop() have been changed.

Hi huyidao,

can you give some more information on the chtSimpleControl class? How did you create it or the files you changed?
I'm trying to implement this procedure but with no luck. I'm asking about the files because i'm not working on my machine and I can't change the source files if that is a must.

Also, do you have the files needed to compile the solver?

Regards,
João