Hi

What is the importance of setting y+ of 1 with the sst model? The reason i ask is that when setting the inflation layer on my airfoil with a y+ of 1 ( 8 layers ) the airfoil will not stall, however if it is set by thickness of 10mm (5 layers) the stall shows and seperation occurs with recirculation. The tet mesh density conected to the y+ version is also higher. Why would the stall not show?

Hope that makes sense

Steve

I should also add that in a previous 2D study with the same airfoil seperation did occur however in the 3d study it does not. Just to clear that up

Steve

Hi,

Whether y+<=1 is required or not is problem dependant. Do a sensitivity study on your case to find out whether it is important or not.

Glenn Horrocks

Thanks Glen

I think i should pose the question differently, I have carried out two simulations on an airfoil, the first was 2d with y+ of 1 and seperated as expected, however when a 3d simulation was carried out with y+ also one the flow does not seperate. Are there common reasons for this? Is sst suitable for 3d? Does it need to be transient? Or should is use a different model?

Thanks again

Steve

did you used prism layers near the walls, Also Please check the mesh jump an refinement near to the area of interest,SST is sensitive in such cases

Im using 8 layers of prism elements and the face spacing is set at min 0.9mm and max 5mm the area above the airfoil also has line spacings set at 5mm so there is not much expansion at all.

Steve

Hi,

Are there common reasons? No, it is proplem dependant.

Is SST suitable for 3D? Yes.

Does it need to be transient? No. In fact there are limitations on the applicability of RANS turbulence models to transient flows. RANS models are based on steady state flows.

Should you use a different model? SST is probably about as good as you'll get without getting a lot more complex with things like LES or DES.

It looks like you have just shown in your case you need a mesh with y+ around 1 and that the normal logarithmic wall functions are not suitable in your case.

Glenn Horrocks

Hi again!

This problem has me pulling my hair out! This is the current situation. I have gone back to an aifoil which is extruded by a small amount between two walls to simulate '2D' effects. It stalls and shows seperation, now i change the geometry so the encloseure is bigger in the z direction (hence making it 3D) with exactly the same parameters all over save chaging one wall to be a symetry plane. It no longer stalls!!! How can this be!? A finite wing (3D) should stall sooner than the 2D calcuation due to induced drag and downwash.

Any ideas as im at a dead end here

Steve

Im an idiot. I dont know where i got the idea that finite wings stall before infinite wings. What a plonker!!!

Steve