[ThomasN]

Hello,

in my Oil/Air multiphase flow calculation with a rotating domain (only 1° is modeled) i want to add gravity in the radial direction z. The rotating axes is x. If I set the buoyancy model and use

x=0 [m/s^2]
y=0 [m/s^2]
z=9.81 [m/s^2]

the solver automatically counter-rotates the relative Frame gravity vector and so the vector components y and z Change with the Rotation angle. How can i avoid this? Could i use a subdomain and define a momentum source in order to model the gravity?
What do i have to add there? The unit is [Nm]

best regards

Thomas

[ghorrocks]

If the domain in rotating in the X axis and gravity is along z then I cannot see how a 1° segment with periodic or symmetry conditions can model that. You are going to have to model the full 360°.

But to keep gravity pointing the same direction, you could use a momentum source, or you could just make the gravity vector a function of time. They both sound like they would work.

[ThomasN]

Hello Glenn,

i think gravity is the wrong word. In my rotational domain (which could be imagined as a rotational couette flow) I get a state where oil droplets are surrounded by air, so they hover in space with no velocity. With the additional body force I want to get them out of the Domain in the radial z direction.

I still tried to make the gravity (buoyancy) vector as a function of time but the calculation is not exact because my rotational speed also changes with time.

So what I have to add in the momentum source term to apply a Body acceleration in z?

regards

[ghorrocks]

I do not understand how gravity can be used to help in the strange situation you explain. But I do not need to understand, I will just answer your question

If your rotation changes with time then simply link the rotation of the gravity vector to the rotation speed. Isn't that the easiest approach?

I think some tutorials use source terms - not sure which ones, have a look.

[ThomasN]

Thanks for your help! I will figure it out.