205 lines
6.3 KiB
C
205 lines
6.3 KiB
C
/*---------------------------------------------------------------------------*\
|
|
========= |
|
|
\\ / F ield | OpenFOAM: The Open Source CFD Toolbox
|
|
\\ / O peration |
|
|
\\ / A nd | Copyright (C) 2004-2007 Hrvoje Jasak
|
|
\\/ M anipulation |
|
|
-------------------------------------------------------------------------------
|
|
License
|
|
This file is part of OpenFOAM.
|
|
|
|
OpenFOAM is free software; you can redistribute it and/or modify it
|
|
under the terms of the GNU General Public License as published by the
|
|
Free Software Foundation; either version 2 of the License, or (at your
|
|
option) any later version.
|
|
|
|
OpenFOAM is distributed in the hope that it will be useful, but WITHOUT
|
|
ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
|
|
FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
|
|
for more details.
|
|
|
|
You should have received a copy of the GNU General Public License
|
|
along with OpenFOAM; if not, write to the Free Software Foundation,
|
|
Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
|
|
|
|
Application
|
|
elasticThermalSolidFoam
|
|
|
|
Description
|
|
Transient/steady-state segregated finite-volume solver for small strain
|
|
elastic thermal solid bodies. Temperature is solved and then coupled
|
|
displacement is solved.
|
|
|
|
Displacement field U is solved for using a total Lagrangian approach,
|
|
also generating the strain tensor field epsilon and stress tensor
|
|
field sigma and temperature field T.
|
|
|
|
Author
|
|
Philip Cardiff UCD
|
|
|
|
\*---------------------------------------------------------------------------*/
|
|
|
|
#include "fvCFD.H"
|
|
#include "constitutiveModel.H"
|
|
#include "thermalModel.H"
|
|
|
|
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //
|
|
|
|
int main(int argc, char *argv[])
|
|
{
|
|
# include "setRootCase.H"
|
|
# include "createTime.H"
|
|
# include "createMesh.H"
|
|
# include "createFields.H"
|
|
# include "readDivSigmaExpMethod.H"
|
|
|
|
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //
|
|
|
|
Info<< "\nStarting time loop\n" << endl;
|
|
|
|
while(runTime.loop())
|
|
{
|
|
Info<< "Time: " << runTime.timeName() << nl << endl;
|
|
|
|
# include "readSolidMechanicsControls.H"
|
|
|
|
int iCorr = 0;
|
|
scalar initialResidual = 1.0;
|
|
scalar relResT = 1.0;
|
|
scalar relResU = 1.0;
|
|
lduMatrix::solverPerformance solverPerfU;
|
|
lduMatrix::solverPerformance solverPerfT;
|
|
lduMatrix::debug = 0;
|
|
|
|
// solve energy equation for temperature
|
|
// the loop is for non-orthogonal corrections
|
|
Info<< "Solving for " << T.name() << nl;
|
|
do
|
|
{
|
|
T.storePrevIter();
|
|
|
|
fvScalarMatrix TEqn
|
|
(
|
|
rhoC*fvm::ddt(T) == fvm::laplacian(k, T, "laplacian(k,T)")
|
|
);
|
|
|
|
solverPerfT = TEqn.solve();
|
|
|
|
T.relax();
|
|
|
|
# include "calculateRelResT.H"
|
|
|
|
if (iCorr % infoFrequency == 0)
|
|
{
|
|
Info<< "\tCorrector " << iCorr
|
|
<< ", residual = " << solverPerfT.initialResidual()
|
|
<< ", relative res = " << relResT
|
|
<< ", inner iters = " << solverPerfT.nIterations() << endl;
|
|
}
|
|
}
|
|
while
|
|
(
|
|
relResT > convergenceToleranceT
|
|
&&
|
|
++iCorr < nCorr
|
|
);
|
|
|
|
Info<< "Solved for " << T.name()
|
|
<< " using " << solverPerfT.solverName()
|
|
<< " in " << iCorr << " iterations"
|
|
<< ", residual = " << solverPerfT.initialResidual()
|
|
<< ", relative res = " << relResT << nl
|
|
<< "ExecutionTime = " << runTime.elapsedCpuTime() << " s"
|
|
<< ", ClockTime = " << runTime.elapsedClockTime() << " s"
|
|
<< endl;
|
|
|
|
// solve momentum equation for displacement
|
|
iCorr = 0;
|
|
volVectorField gradThreeKalphaDeltaT =
|
|
fvc::grad(threeKalpha*(T-T0), "grad(threeKalphaDeltaT)");
|
|
surfaceVectorField threeKalphaDeltaTf =
|
|
mesh.Sf()*threeKalphaf*fvc::interpolate(T-T0, "deltaT");
|
|
|
|
Info<< "Solving for " << U.name() << nl;
|
|
do
|
|
{
|
|
U.storePrevIter();
|
|
|
|
# include "calculateDivSigmaExp.H"
|
|
|
|
// Linear momentum equaiton
|
|
fvVectorMatrix UEqn
|
|
(
|
|
rho*fvm::d2dt2(U)
|
|
==
|
|
fvm::laplacian(2*muf + lambdaf, U, "laplacian(DU,U)")
|
|
+ divSigmaExp
|
|
);
|
|
|
|
solverPerfU = UEqn.solve();
|
|
|
|
if (aitkenRelax)
|
|
{
|
|
# include "aitkenRelaxation.H"
|
|
}
|
|
else
|
|
{
|
|
U.relax();
|
|
}
|
|
|
|
gradU = fvc::grad(U);
|
|
|
|
# include "calculateRelResU.H"
|
|
|
|
if (iCorr == 0)
|
|
{
|
|
initialResidual = solverPerfU.initialResidual();
|
|
}
|
|
|
|
if (iCorr % infoFrequency == 0)
|
|
{
|
|
Info<< "\tCorrector " << iCorr
|
|
<< ", residual = " << solverPerfU.initialResidual()
|
|
<< ", relative res = " << relResU;
|
|
if (aitkenRelax)
|
|
{
|
|
Info << ", aitken = " << aitkenTheta;
|
|
}
|
|
Info<< ", inner iters = " << solverPerfU.nIterations() << endl;
|
|
}
|
|
}
|
|
while
|
|
(
|
|
iCorr++ == 0
|
|
||
|
|
(//solverPerfU.initialResidual() > convergenceTolerance
|
|
relResU > convergenceToleranceU
|
|
&&
|
|
iCorr < nCorr)
|
|
);
|
|
|
|
Info<< "Solved for " << U.name()
|
|
<< " using " << solverPerfU.solverName()
|
|
<< " in " << iCorr << " iterations"
|
|
<< ", initial res = " << initialResidual
|
|
<< ", final res = " << solverPerfU.initialResidual()
|
|
<< ", final rel res = " << relResU << nl
|
|
<< "ExecutionTime = " << runTime.elapsedCpuTime() << " s"
|
|
<< ", ClockTime = " << runTime.elapsedClockTime() << " s"
|
|
<< endl;
|
|
|
|
# include "calculateEpsilonSigma.H"
|
|
# include "writeFields.H"
|
|
|
|
Info<< "ExecutionTime = "
|
|
<< runTime.elapsedCpuTime()
|
|
<< " s\n\n" << endl;
|
|
}
|
|
|
|
Info<< "End\n" << endl;
|
|
|
|
return(0);
|
|
}
|
|
|
|
|
|
// ************************************************************************* //
|