[Fadih]

hello guys,

I am simulating a 3D VAWT, i have created my geometry in ansysand used ansys meshing for the mesh. i am having a problem in fluent convergence of the continuity equation while all other criterions converge. i am using the SAS model for better accuracy of results and actually i am required to use this model. The first thing i did to try solving the issue is refine my mesh, i am up to 2.5 million nodes and the mesh propagation through the bodies is quite satisfying since the mesh is matched at interfaces so transition from a body to another should not be the problem, i did get an error in ansys meshing saying that inflation have created stair step mesh but when i launched fluent the mesh check quality didn't report any error also their is no divergence issue and even when i ran my simulation no warning showed up so i think that this is not the source of the problem. i chose my time step so the turbine rotates 0.27 deg/timestep.
i have no clue how to solve this, continuity keeps oscillating up and down.
Does anyone have any idea, would be greatly appreciated.
i have uploaded pics of the geometry and the mesh at the interfaces ( its just tetra mesh)

[Fadih]

here are the remaining screen shots

[hyugakojiro999]

Is your convergence criteria 10^0-6 or 10^-3 ?

Which of them is adequate for your situation?

----------------------------------------------------------

What about mesh quality? Are they adequate including Y+ ?

----------------------------------------------------------

Had u done validation before performing this analysis?

[Fadih]

mesh quality is something around 0.03 however the average quality is 0.85.
criterion is 1e-3 for all equations
yes i checked the quality and no warnings or errors showed up
y+ is around 80, in my case i don't need a low y+ because for the turbine Cp is mainly generated from pressure and not shear and does not require much clustering.
the bad quality mesh is around the blades due to stair syep mesh, i struggeled with that for 2 days but couldn't solve it but fluent didn't seem to have any problem with that, and the warning repair wall distance didnt even show up
i struggeled quite a bit with the mesh but i need to maintain a max of 3 million nodes so can't really insert a very low inflation which js currently at 1e-4

[Fadih]

but what is weird is that Cm that i am monitoring is stable and close to experimental data

[hyugakojiro999]

Are u using any UDF or UDmaterial? Check it twice
-----------------------------------------------

Try only default meshing to see if continuity will be converged.
-----------------------------------------------

Check even BCs.

[Fadih]

no it doesn't that' s why i refined , no UDF just regular settings