[shashankshankar]

Hello,

1. I have been trying to estimate CL and CD of NACA0012 airfoil. I am constantly getting an underprediction of CL by around 19%. The Reynolds number I am working on is 4.0E5. Details of the setup are as follows:-
(a) 3D Steady in water
(b) Polyhderal and Prismlayer mesher
(c) SST-menter K-w turbulence model. Wall Y+ after simulation is attached.

2. I am unable to understand what I am missing. Is it the mesh? The number of cells is around 1.4 million.

3. Convergence happens without any issues. Further, in SST Menter-K-w model, do we need to activate the low Re damping factor for this case? The starccm manual says only for transitional flows that action needs to be undertakenPrism Layer.jpg

Prism Layer.jpg

Wing_AOA4_Mesh Scene 1.jpg

Wing_AOA4_Mesh Scene 2.jpg

Wing_AOA4_XY Plot 1.jpg

[fluid23]

your mesh is definitely not appropriate. look into wake refinement and conduct a mesh dependency study before determining that your converged result is the best you can do.

[shashankshankar]

Hello Matt,

1. Thank you very much for your suggestion.
2. Since the polyhedral meshing model doesn't have a wake refinement option, I incorporated a volumetric control. However, the drag result looks nice but CL is still being underpredicted.
3. I have set the extents of volumetric control based on the scalar view of velocity. Is there anything else I should incorporate in such cases?

[fluid23]

The polyhedral mesher absolutely does have wake refinement options. A volumetric control is probably not the best way to approach that. Refer to the Star-CCM help files:

Pre-Processing > Meshing > Custom Mesh Controls > Mesh Refinement in a Wake

After that, make sure to do a mesh dependency study. Once you are confident that your result is independent of the spatial discritezation (i.e. your mesh), then you can assess whether your values are 'correct'.

[cwl]

You also can a Table-based Mesh Refinement.

[Chris2337]

The polyhedral only has wake refinement in parts-based meshing, so if you're using region-based meshing, like you would for the trimmer 2D, then switch to parts-based.

[Paulh]

Adaptive mesh refinement should be WAY easier than it is in CCM+, but . . . .
Consider taking a look at the STEVE article I'm hyperlinking below. It might give you some ideas on how to tweak your mesh to get the best predicted lift you can: https://thesteveportal.plm.automatio...pturing-shocks

[cwl]

Originally Posted by Paulh

Adaptive mesh refinement should be WAY easier than it is in CCM+, but . . . .
Consider taking a look at the STEVE article I'm hyperlinking below. It might give you some ideas on how to tweak your mesh to get the best predicted lift you can: https://thesteveportal.plm.automatio...pturing-shocks

How can it be "WAY easier"? - It consists of as few steps as possible:
1) Define where, why (∇P, ∇T, Turbulent Viscosity Ratio etc) and how the mesh should be refined via Field Function.
2) Extract Field Function to a table.
3) Mesh according to the table.

Since not everyone has constant access to the Steve Portal - here is a video on youtube:
Table Based Refinement in STAR-CCM+