[ELwardi]

Hi all,

I'm trying to solve the basic incompressible, immiscible two-phase flow that can be modeled with the following equations in a coupled manner:


SwEqn: (water equation)



pEqn: (oil equation)



For now, what works for me is to solve:
1- SwEqn explicitly for Sw: All coeffs and are calculated from previous time.
2- The equation (SwEqn+pEqn) implicitly for p (ddt(Sw) eliminated from pressure equation only if the flow is incompressible ).


is a function of Sw and is calculated explicitly.


Of course, the explicit calculation of Sw sets some constraints on deltaT and introduces instabilities when attempting to model source terms in a semi-implicit manner.



My first attempts to couple these two equations (using foam-extend-4.0) through the terms:
1- pInSw:
2- SwInp:
did not prove to be successful .


The general layout of the coupled solver looks like this:


1- Create the block matrix for the working variable pSw.
2- Assemble pEqn and SwEqn without the coupling terms
3- Calculate coupling terms separately.
4- Add equations and coupling terms to the block matrix

5- Solve it and retrieve solution.


Till now, the testing case works fine with the segregated solver but
gives weird results with the coupled version.


I'm currently investigating the boundary conditions but i don't think there is a problem there.



Any help/orientation to reference material is greatly appreciated

[Cyp]

Hi ELwardi,

You can have a look at the porousMultiphaseFoam package https://github.com/phorgue/porousMultiphaseFoam that solves the multiphase Darcy equations with the IMPES algorithm.

The corresponding paper is there : https://www.sciencedirect.com/scienc...10465514003403

Cheers,
Cyprien

[ELwardi]

Originally Posted by Cyp

Hi ELwardi,

You can have a look at the porousMultiphaseFoam package https://github.com/phorgue/porousMultiphaseFoam that solves the multiphase Darcy equations with the IMPES algorithm.

The corresponding paper is there : https://www.sciencedirect.com/scienc...10465514003403

Cheers,
Cyprien

Thanks for your reply,



I have already seen the awesome toolkit , what I'm looking for is to create a fully coupled solver though.


A very simple case where IMPES methods fall short is when applying a semi-implicit well rate (BHP-controled well, you may improve deltaT calculation for this if you want).