Learning Forum

[benjamin.triffaux]

Hello everyone,

I'm currently simulating a water wheel on fluent and I encountered a problem. The water in the blades that exit the water seems stuck and gravity isn't pulling it down. It is weird because the water has no problem getting inside the rotating mesh, but it seems that it is stuck in the interior.
I think it might be due to the interface between the rotating mesh and the stationary mesh, but I used a symmetric interface (I created 2 circular contacts, one with the inner domain (rotating domain) circular boundary as the source and the outer domain circular boundary as the target, and another contact when I switched the source and the target).

I thought that Fluent interpreted the circular boundary of the rotating mesh as a wall, but in that case, the water should still "fall" until it hits the boundary.

but here it seems that there is a force pushing the water up and the latter forms a perfectly horizontal air-water-interface inside the blades that exit the water.

Does someone knows where the problem could come from ?

Regards,

[Rob]

The interface should be seen as an internal boundary assuming there are cells on both sides of it. How well resolved are the buckets? How well matched are the two sides of the interface mesh? How long (in rotations) has the model been running? Looking at the upstream & downstream boundaries you may want to review "open channel" options too. 

[benjamin.triffaux]

Thanks for your answer, I replied in the discussion down below. I can't delete my answer below to answer here.

[benjamin.triffaux]

What do you mean by internal boundary ? I'm new to Fluent, and some concepts are difficult to understand. Regarding the mesh resolution, I increased the number of elements in the interior and at the boundary between the meshes. I don't think the two sides are matched; should I enable the option "matching" in the mesh interface parameters ?
The model has turned over half a turn. The water in the buckets on the top of the wheel was just a test to see what happened if I initialized some cells, if the gravity would work, and on those ones, it worked physically. 
For the open channel part, I only created and pressure inlet and initialized some cells with water (on the same water level as the inlet). For the downstream flow, I left it open. 

[Rob]

Have a look at the various Intro lectures for Fluent on Learning. An interface (what you have) has a wall associated with it for when the two sides don't line up, and otherwise works as an interior/internal boundary (cells on both sides & flow goes through it). 

Cell size needs to be roughly similar on both sides to help the solver maths. And with VOF aim for an aspect ratio as near 1.0 as you can get. You also need a fairly high mesh resolution to capture the free surface. 

Open channel is a VOF boundary option and is designed for channels with a preset height of water: they're covered in the User's and Theory Guides in Fluent's Help system.