[holger_marschall]

Hi FOAMers,

this is just to let you know that Xuan Cai (Karlsruhe Institute of Technology) and me (Center of Smart Interfaces, TU Darmstadt) are working on the implementation of phase field methods / diffuse interface models in FOAM.

In particular, we are aiming at conservative and bounded FV-based methods for the Allen-Cahn and Cahn-Hilliard equations for Direct Numerical Simulations of two-phase flows. The project's topic is on dynamic wetting, where diffuse interface modeling alleviates the known moving contact-line paradox (stress singularity at no-slip walls). However, with phase-field methods, also one could think of other (more exotic) interfacial physics (foaming, crystal growth, etc.), which certainly makes this model class very appealing.

The implementation employs the well-known strategy design pattern. We have chosen the difference of phase volume fractions as an order parameter, such that there are physical bounds within [-1, 1]. The Cahn-Hilliard equation is of fourth order so solution strategy comes into play (segregated/split vs. coupled), which directly bias the boundedness of the order parameter. While the Cahn-Hilliard equation exhibits inherent volume-conservation property, the Allen-Cahn equation needs to be appropriately constrained to this property. However, currently, we are able to obtain a relative volume-error of 1.3E-6 for a spreading droplet after 1.2E-2 seconds.

As an appetizer, allow me to attach some screenshots. At this stage, any comments/hints/ideas on implementation techniques you have heard of and on other test case (currently we examine the 'spreading droplet' and 'rising capillary' cases) are really appreciated...


best regards,
Holger

[cfd_explorer]

Hi FOAMers,

I'd like to join Holger for appreciating your inputs/advices in advance

Best regards,
Xuan

[hrushi.397]

Hi,

How did you manage to keep the parameter within bounds? How fine is your mesh?

Regards,

Hrushikesh

[holger_marschall]

Dear all,

just to give you a brief update. We are currently writing a paper describing the conservative Finite Volume discretisation of the Navier-Stokes Allen-Cahn and Cahn-Hilliard equations for two-phase flows. As for some test cases, we opted for dynamic wetting cases.

Please drop me a mail, if you are interested in a preprint. I will keep you posted on the progress.

best regards,
Holger.

PS.:

Originally Posted by hrushi.397

Hi,
How did you manage to keep the parameter within bounds? How fine is your mesh?

Dear Hrushikesh,

this is a natural outcome if you consider coupling in the momentum equation in order to take into account the relative density flux due to diffusion of components. ONLY this allows you to use conservative fluxes! With this, the boundedness property is then found to be consistently enforced (as an inherent property of the phase field transport equations).

The current implementation of this look very clean

    fvVectorMatrix UEqn
    (
        fvm::ddt(rho, U)
      + fvm::div(rhoPhi, U)
      - fvm::laplacian(mu, U)
      - (fvc::grad(U) & fvc::grad(mu))

      //-HM Coupling term 
      - phaseField.diffRhoPhi(U)

      //- buoyancy and surface tension term
      //- tansferred to pressure equation
    );

The nuts and bolts behind this will be described in the paper. I am also planing a release to Foam-Extend.

Hope this helps,
Holger

[RaghavendraRohith]

Hi Holger,

I am trying to model a surface tension and wetting model in a closed capsule geometry using CLSVOF. I have tried different contact angles (Dynamic and static). The complexity of my problem lies in the incompressible nature of phase1 which is solid and wetting occurs post melting of the solid. The velocity is generally very low @ 1e-5 m/s and the also the pressure difference between two phases is also at the range of 30 Pa. Will your model help me in enhancing the nature of wetting in my defined problem?

Please find the attached Jpg.

VG,
Rohith

[nero235]

Hello,

any news regarding the current state of the new phaseFieldFoam solver? I am very interested in working with this solver.

Kind regards, Sebastian

[eberberovic]

Hi Sebastian,

I have found the solver phaseFieldFoam, which is I believe adopted by you:
https://github.com/11101011/phaseFieldFoam

Have you tested it, how does it perform? Exactly which paper is it referred to?

Regards,
Edin

[eberberovic]

Hi Holger,

Have you already released the diffuse interface solver, or when do you think it will be released?

Regards,
Edin

[nero235]

Originally Posted by eberberovic

Hi Sebastian,

I have found the solver phaseFieldFoam, which is I believe adopted by you:
https://github.com/11101011/phaseFieldFoam

Have you tested it, how does it perform? Exactly which paper is it referred to?

Regards,
Edin

Hello Edin,

yeah well this is the solver from Adam Donaldson, which I ported to a OF 2 version and uploaded it then to github However the version from Holger Marschall is very different, so I heard. I don't know if he has released the solver yet, that is why I am asking.

Kind regards,

Sebastian

[holger_marschall]

Hi Edin, hi Sebastian,

yes definitely, we will release the source. It is planned for this year. I just want to write up a paper describing the model derivation and numerics.

P.S.: If somebody is familiar with phase-field modeling and works with FOAM please contact me. I am eager to see the framework working for other physics than for a capillary interface (i.e. two-phase flow of two immiscible Newtonian fluids taking into account interfacial energy). The code structure should be prepared and flexible enough for that.

best regards,
Holger

[Cyp]

Hi Holger,

I am just curious, is your phase-field model subject to spurious currents at the gas/liquid interface?

It is still a bottleneck for VOF simulation in porous structures at very low capillary number (Ca = 10^-6 and below). The state of art techniques use filtering method, but as of today, nothing is really convincing.

Cheers,
Cyprien

[JULIEN MAES]

Hi Holger,

Did you end up releasing the code?

Cheers,

Julien

[winsway]

i have some confusion with phase field and hope you solve my confusion.I don't know whether phase field is suitable for the calculation and simulation of multi-phase turbulence. Most of the literature is VOF and level set. I hope you can provide some help~

[hdixit]

Hi Holger,

I was curious to know if you did release the phase-field code? We are very keen to use the code for some wetting calculations.

best regards,
Harish