Simulating the Behavior of Rotating Fluids: CFD vs ALE

[exophysix]

If I were to rotate a basin of fluid at a given angular speed, I would have two acceleration components.. one gravity and one centrifugal. What would happen if I added gravity to both components of acceleration (from tilting the basin). The fluid should move to one side correct? But, what if I just speed the basin up? Would the fluid even itself out at an equilibrium point?

Thanks!

 

[mikeph]

Two acceleration components on what? The fluid?

The fluid in this drum has a body force due to gravity and also shear forces from friction with the wall of the basin. The relative strength of these two forces will dictate how the fluid settles in equilibrium. When the basin is level and spinning, the fluid has some steady-state angular momentum. This angular momentum will be conserved in the fluid until some torque acts on it. When you tilt the basin you are applying a torque, but this is applied at the boundary of the fluid and not the whole fluid. The result is a big mess where the outer parts of the fluid continue to rotate with the tilted basin, but the inner parts rotate as they did before.

Basically it's a very complicated question, and even the qualitative answer depends on the details of the experiment (i.e. turbulent v laminar flow).

 

[exophysix]

Two acceleration components on what? The fluid?

The fluid in this drum has a body force due to gravity and also shear forces from friction with the wall of the basin. The relative strength of these two forces will dictate how the fluid settles in equilibrium. When the basin is level and spinning, the fluid has some steady-state angular momentum. This angular momentum will be conserved in the fluid until some torque acts on it. When you tilt the basin you are applying a torque, but this is applied at the boundary of the fluid and not the whole fluid. The result is a big mess where the outer parts of the fluid continue to rotate with the tilted basin, but the inner parts rotate as they did before.

Basically it's a very complicated question, and even the qualitative answer depends on the details of the experiment (i.e. turbulent v laminar flow).

Thank you mike.

Do you know of a good way to simulate this behavior (CFD versus ALE) [by chance]?