[sraheja]

Hello everyone,
This is for a short term project for which I am willing to pay.
I am a newbie to CFD and Openfoam. I am currently working on a multiphase simulation in OpenFOAM using the `interFoam` solver to model the segregation of oil and water in a still atmospheric tank without any inlet/outlet flows. The tank dimensions are 4m x 2m with a height of 1m, and it is initially filled with an oil-water mixture where oil comprises 50% of the total volume. Despite setting up the simulation to observe the phases separating over time, I have not seen any segregation between oil and water, as the oil should float to the top due to its lower density.

Here are some details of my setup:
- Solver: interFoam
- Initial Condition: Uniform distribution of alpha.oil = 0.5 across the domain.
- Boundary Conditions: Walls are set to noSlip, and the top is a free-slip boundary at atmospheric pressure.
- Mesh: Structured mesh with 20 cells/m resolution.

I have tried to check the physical properties, boundary conditions, and numerical schemes, but the phases remain mixed without clear separation. Gravity is also applied in the constant/g file in the negative Z direction.

I also tried another simple case of a still-settling tank, but this tank had an altered initial condition: lighter oil was at the bottom and heavier water was at the top. The simulation ran for some time, showing oil moving to the top and water to the bottom, but it became slow after 14 seconds of physical time while working on an adjustable time step as it struggled to maintain a courant number.

I seek insights or suggestions that might help diagnose and resolve this issue. My next goal is to add inlet/outlet nozzles and a weir plate for a realistic storage tank, which I will model after fixing such a simple scheme. The paid project would involve estimating effective settling periods and attainable oil water separation efficiencies in some typical configurations.

Could there be an issue with my solver settings and BCs, or how the initial conditions are applied?

I am attaching the complete case setup files along with the blockmesh dictionary. Any advice on adjustments to the setup or further diagnostic steps would be greatly appreciated! I can discuss further payment terms and conditions with those willing to mentor on such problems using opensource tools like OpenFoam/Gmsh and Salome.

regards

[PrasantKaredla]

Hello,

I am looking into your case.

It seems like the initial condition for oil may be the issue. You must set the initial fields for oil using the setFields utility. you said lighter oil was at the bottom. Do you know to what extent oil was settled at the bottom initially? out of 1m height?

You can reach out to me at prasant.karedla@gmail.com. I have good exposure to multiphase simulations using OpenFOAM.

Best Regards,
Prasant K

[sraheja]

There were two cases which I have tried to simulate:

case-1 : The whole tank is filled with oil water mixture with alpha=0.5. I have now used setFields as suggested but there is no change in the simulation results. there is no segregation happening at all. The setFields dict is as below:

defaultFieldValues
(
    volScalarFieldValue alpha.oil 0.5
);

regions
(
    boxToCell
    {
        box (0 0 0) (4 2 1); // fully filled with oil with alpha 0.5
        fieldValues
        (
            volScalarFieldValue alpha.oil 0.5
        );
    }
);

// ************************************************************************* //

case-2:
In this case the bottom half of the tank was filled with oil (upto 0.5 m height). setFields was used here also. The simulation runs. I also increased oil viscosity this time,this surprisingly helped convergence allowed larger timesteps. Final result after 200s is still weird. Oil is more at the end. (result attached)