[ELwardi]

Hi all,


I've been fiddling with this for quite some time now without a clear success.I really do hope someone could help with this problem of mine.


I was solving a coupled problem in a traditional segregated manner:

fvScalarMatrix SwEqn
(
     porosity*fvm::ddt(Sw) + fvc::div(phiw)
);

// phiw = phiP + phiG



// Explicit Sw

SwEqn.solve();


fvScalarMatrix pEqn
(   
     fvm::laplacian(-Mf, p) + fvc::div(phiG)
);


pEqn.solve();

phiP = pEqn.flux();


// Then Update fields with phiP and proceed in time

The previous approach sets some constrains on deltaT, so a fully coupled solver is desired.


I'm trying to choose between:
1- Solving these equations (where phiw comports an implicit term) with Newton-Raphson classic method, and
2- Solving them using the block coupled library supplied with foam-extend 4.0


I decided to go step-by-step about this and try to solve the "segregated" problem first ( This may not be possible at all, but it may give valuable insight into matrix construction/solving ).

// pSw is a 

fvBlockMatrix<vector2> blockM(pSw);



// Define pEqn and SwEqn as defined previously 

// Then



blockM.insertEquation(0,  SwEqn);
blockM.insertEquation(1,  pEqn);


blockM.solve();
// retrieve solution and update BCs

This is still not working for me so I thought there is someone who can help here! Any help is much appreciated.