Gambit boundary condition for Rankine vortex in 2d
 

Hi,

I'm trying to model a Rankine vortex in Fluent/Gambit in 2 dimensions.

What kind of Boundary Condition should I use for the Rankine vortex? I already tried to use a cirlce as a velocity inlet but this doesn't work and while exporting the mesh there's the error message: "Continuum or boundary condition specified is invalifd for fluent 6".

I'd appreciate any help!

Markus

Re: Gambit boundary condition for Rankine vortex i
 

Hi Markus

I take it from what youve said your trying to simulate a cyclone separator or combustor using the seggregated (pressure based) solver under axisymmetric conditions. I assume the inlet is tangential, not that it really matters for a 2D axi-swirl case. Using a circle as a velocity inlet means that the inlet is a face is 2D but when rotated for axi-swirl it becomes a torus volume and a volume can not be using as a BC condition (except for fluid and solid also ignoring source terms). What you can do is define a line of length the same as the hydraulic diameter of the inlet and when rotated for the axi-swirl condition you will have a face as the inlet. Despite it being a ring and does not accurately model multiple tangential injectors its the best your going to get in 2D. If you carry on modelling in 2D expect some difficulties in obtaining good residual convergence of the swirl term in Fluent, it may not occur but I had this problem and I could not resolve it. Switching to 3D may be alot more work but is alot easier to get converged solutions in my experience. If running 3D under the pressure based solver with a decent mesh it doesn't take that long to converge.

Hope this helps

Neil

Re: Gambit boundary condition for Rankine vortex i
 

Hi,

I'm trying to simulate vortex interaction in 2d using Rankine vortices, I followed your instrductions and used a line as boundary for the vortex and meshed it. But when I tried to export the mesh, the error "boundary entity does not contain any valid entity" occurred. How can I solve this problem?

Thanks. Markus