Hi all, a very simple question: I§m running simulation of flow about symetric shape, but resultant flow is very assymetric behind the object. What can I improve in my model? Also the resultant force is quite far away from zero. I'm not using the symmetry BC a The trub. model is SST.

P.S.: I know that I can't get "fully symetric" result, but this is too much.

You know the rules: Command file Mesh pic Geometry pic

I§m sorry, relly don't understand your answer

Hi,

Joe is saying that we cannot really help you unless you give us some more information. Sounds like a picture of the geometry and the Reynolds number it is operating at would be a good start. Also your simulation may not be converged properly. Lots of possibilities.

The effect might be realy - symmetric geometries do not necessarily lead to symmetric flows. Have a look at flow around a cylinder. At very low Re you get symmetric flow, mid Re you get asymmetric laminar flow, at moderate Re you get asymmetric turbulent flow and at high Re you get symmetric turbulent flow (in a time averaged sense).

Regards, Glenn

Hi I'm very sorry, there you are more informations. Any help would be great for me.

Shape - 3D Circular Cone with some grooves. Very simple geometry, about 10 m middle radius

Re - 3e7 - 7.5e7; Turb model - SST (or BSL will be better?); Turb on Inlet - Zero GradientMaterial - Ideal gas, air; Steday State, Timestep - I tried variety from 0.1 sec up to 20 min, no effect; Mesh - unstructured, 25 layers in the BL, 1.375 exp. ratio; BC - Inlet with normal spped, outlet static pressure

You will see my problem from picture bellowwhere the forces are monitored. The Drag force (Fx) never reach any final value. The other ones are "wild".

In the second picture you can see my problem clearly. I would expect the turb.kinetic energy symmetric about the x axis.

"3D Circular Cone with some grooves." Is this completely symmetric? Is the inlet BC flow symmetric relative to the geometry?

Set explicit turbulence conditions on the inlet BC as a test. No gradient can be problematic if there is very little communication of internal turbulence levels to the inlet BC.

Your BL expansion seems high. Your mesh seems coarse judging from your pic. Mesh assymetry could encourage flow assymetry if the flow is finely balanced.

Try a 2D section of the model through the middle of the cone and see if that produces asymetry too.

As the cone thing is seemingly symetric the way in which you place your inlet and outlet BCs could introduce assymetry in the flow.

This problem may be transient due to vortex shedding. The fact that a transient run didnt "fix" the assymtry issue doesnt mean the flow isnt actually transient.

The flow could simply be poorly converged.

The geometry and BC are fully symmetric, the mesh as far as possible by unstructured mesh. 2D case seems to be ok. I'll try to make mesh finer

I saw a paper at a CWE conference about 5 years ago where people were modelling a cube on the ground in an atmoshperic boundary layer. Flow around the cube was not symetric even though it was steady state, and a hex mesh. This was an interesting find, with most people. Why not just apply a symetry plane ? You are already idealising the flow, so why not just introduce this ? When we modelled the problem, usung unstructured mesh, we experienced an asymetrc flow. By Packing in more mesh, the effects were limited, but not removed.

Good luck

Stuart