How to build the axisymmetric model in Abaqus CFD ? 1

[vkopey]

Need help to build the CFD model of the ball valve in Abaqus 6.11-3 (see picture).

The problem is that calculation 3D model takes a long time.

I tried to build 2D-planar model (see right picture) where:
1 - wall, 2 - inlet velocity, 3 - outlet pressure, 4 - symmetry BC (normal velocity = 0)
But such simplification is not correct because in that case we got the cylinder but not the ball.

Is it possible to build the axisymmetric model in Abaqus CFD ? How to set the axisymmetric Boundary Conditions ? I attached the model file.

 

[btrueblood]

The problem is that you have two axes - the pipe axis and the ball axis - at right angles to each other (well, at least for real ball valves which have a hole through the middle of the ball). Pick one to be the axis of symmetry, and you lose the resolution of the other axis, as you found. You won't have a decent model with a rotationally symmetric simulation, you need to use a full 3d sim.

IF, on the other hand, you are trying to model something like a ball check valve, then pick the pipe axis as the symmetric axis.

 

[vkopey]

btrueblood, I try to model the ball check valve. I can not pick the pipe axis as the symmetric axis because Abaqus does not allow it. Abaqus support only 3D and 2D-plane CFD models. The question is how to build the axisymmetric model by setting axisymmetric Boundary Conditions on 3D or 2D-plane models? Is it possible?

For example this simplification of the model is correct?

where 1 - faces with symmetry Boundary Conditions (normal velocity = 0)
These symmetry Boundary Conditions are correct?

 

[btrueblood]

Yes, those would be correct. But why not model the remaining 3/4? Can't be that much more computationally costly...

 

[vkopey]

Even this 1/4 model is calculated more than 8 hours. I need to calculate 10 such models. I search the way how to reduce the calculation time.

 

[btrueblood]

Just realize that you artificially constrain the flow to be symmetric in 90 degree increments, or whatever the slice angle you choose. This would, for example, preclude any 3d swirling flow, or non-steady eddies across the downstream face of the ball.

 

[vkopey]

btrueblood, I understand, thank you.