conditional solving of transport equation
hi people,
I would like to set up a solver for a solidification process with solving a transport equation for a species only in the liquid phase of the sytem. There is a volScalarField alpha which defines the state of phase (0<alpha<1). alpha is the liquid fraction, alpha = 0 > complete solid. How can I define a solver which works only in the non solid part of the domain like: // definition of eq. {liqEqn = ....} // conditional solving of liqEqn if (alpha != 0) liqEqn.solve(); Is there somewhere a similar case/tutorial/documentation reference? thank you for advice, dzi (I use OF 2.01) 
It is very difficult to only use portions of the mesh for the matrix solution in OpenFOAM. You'd probably have to create a new temporary mesh, and create new variables on it  then you'd have to create new boundaries where it is cutoff... Rather than that, you probably want to work with the full mesh, and modify the matrix so that the portion from the solid cells reduce to a trivial equation.
I'm thinking you could create a custom preconditioner for your matrix. I don't know exactly what you want to do to the matrix to achieve this, though. 
I think the easiest solution is to multiply all terms in the equation with alpha.

Quote:
volScalarField solveEq(pos(alphaalphaCutOff)); volScalarField doNotEq(1solveEq); which is 1 when alpha > alphaCutOff, and 0 elsewhere. Then there are two possible solutions, depending on what you are trying to do (momentum equation or scalar equation?)  Momentum equation: Code:
fvMatrix UEqn Code:
myEqn.setValues(...) Best, 
thank you for the replies,
for me the easiest solution is the suggestion from akidess to multiply all, or parts of the equation with alpha. Looks like it can be solved and I get something out which makes sense. The other suggestions also sound interesting. I will try if I come to a limit with the first solution, but on the first glance they seem to be more sophisticated. Thank you again for helping on this topic! dzi 
Quote:
If you are solving the momentum equation, you will also have to deal with the problem of the central coefficient going to zero, which will lead to a segmentation fault when you calculate H/A. You can check how this problem is addressed in compressibleTwoPhaseEulerFoam. Best, 
Solidification in OpenFOAM
Hi
I am solving a binary alloy solidification problem with OpenFOAM. For the species equation, which seems to be the most critical equation especially at the region where channels form, I have used zero grad boundary conditions; which causes a flow into the wall. Have you guys also used grad(CL) = 0, and grad(C) = 0 at the boundaries? 
conditional solving of transport equation
Hello dear foamers,
I want to revive this old thread with a new question. I wonder if it is possible in OpenFOAM to solve a transport equation only for cells whose variable values are larger than a certain value. Like: solve Eqn. only for values >0… Let me explain: I simulate a fluid consisting of two gas fractions gas1 and gas2. Gas2 does not cover the whole simulation domain but is present in places. So, basically, there are regions where the concentration of gas2 is zero. Consequently, the partial density rho of gas2 in those regions is zero too. In order to simulate a thermal nonequilibrium I need to solve the energy equation for both gases separately. I want to solve the energy equation for gas2 only for regions, where the density of gas2 is larger than a certain value (e.g. 10e9 or something…). My energy equation is taken from rhoCentralFoam, where the internal energy is solved in two steps: Step1: the predictor step without diffusion term: solve ( fvm::ddt(rhoE) + fvc::div(phiEp)  fvc::div(sigmaDotU) ); Than calculating internal energy and correcting boundary conditions: e = rhoE/rho  0.5*magSqr(U); e.correctBoundaryConditions(); thermo.correct(); rhoE.boundaryField() = rho.boundaryField()* ( e.boundaryField() + 0.5*magSqr(U.boundaryField()) ); Step2: the diffusion correction equation. if (!inviscid) { solve ( fvm::ddt(rho, e)  fvc::ddt(rho, e)  fvm::laplacian(turbulence>alphaEff(), e) ); thermo.correct(); rhoE = rho*(e + 0.5*magSqr(U)); } The predictor step is working just fine, but since I have to divide rhoE by rho of gas2 I understandably get the error, because there are cells in the simulation domain with rho = 0. What I did is: forAll(e, celli) { if (rho [celli] > 0.000001) { e [celli] = rhoE [celli]/rho [celli]  0.5*magSqr(U[celli]); } else { e [celli] = e [celli]*0.0; } } But I have my suspicion, that this technique is highly inefficient concerning the computational speed. The next problems occur by executing: e.correctBoundaryConditions(); and thermo.correct(); When executing the command “correctBoundaryConditions” the solver crashes with the following report: #0 Foam::error::printStack(Foam::Ostream&) at ??:? #1 Foam::sigFpe::sigHandler(int) at ??:? ... … ... Floating point exception (core dumped) I took a look at “src/OpenFOAM/</SPAN>fields/</SPAN>GeometricFields/</SPAN>SlicedGeometricField/</SPAN>SlicedGeometricField.C</SPAN>”, where I suppose the function “correctBoundaryConditions” is defined, but my OpenFOAM knowledge is too little to understand the file. My questions are:  Is there a better way than using the “forAll” loop, like I did?  How could I handle the printStack error coming up by calling the correctBoundaryConditionsfunction? Many thanks in advance! </SPAN> 
Hello,
I guess you should also loop over boundaryMesh, not only (internal) cells. regards, olivier 
Hello Olivier, hello all.
Olivier, thank you for the quick reply. I just did it like this: Code:
It compiles just fine. The same problem appears with the command thermo.correct(). What I did there is: Code:
It compiles also fine. I guess, that the function e.correctBoundaryConditions() is only executed for the particular cell “I” at each forAllloop and not over all boundary cells. Is it right? Because otherwise I would get the same error as without using forAll. The next challenge is the corrector step: Code:
Here I have the same error as with the correctBoundaryConditions()function, namely: #0 Foam::error::printStack(Foam::Ostream&) at ??:? #1 Foam::sigFpe::sigHandler(int) at ??:? ... … ... Floating point exception (core dumped) Is there some technique to solve this equation conditionally (for e > 0.00...)? Thanks in advance. Regards Alex 
All times are GMT 4. The time now is 13:05. 