[DanielBarreiro]

Good day everyone,

At the moment I'm perfoming some routine simulations of 2D Airfoils in CFX to determine its polars at a given Reynolds Number. The airfoils I'm simulating at the moment belong to a wind turbine blade. One of them is a more aerodynamical profile, while the other is a thick airfoil.

The structured meshing for both airfoils has been done in ICEM CFD, using a O-Grid and with sufficient resolution in the region near the wall for an accurate boundary layer prediction. The y+ value of the mesh is around 1. The meshing procedure has been done exactly the same way for both airfoils, and it can be seen in the first picture attached in this post.

When simulating the airfoils in CFX, the main characteristics of the setup are the following:

Steady Simulation
-Turbulence Model: SST (K-W SST for the Fluent set up)
-Time Scale Control: Autoscale Conservative
-Fluid: Air at 25ºC
-Residuals: 1*10^-6

The boundary coinditions are a velocity inlet with V=150 m/s (for a Reynodls number of 1*10^7 in an airfoil with a 1m chord length), a Pressure Outlet, and the Airfoil Wall as well as the 2 side symmetries.
The domain has the recomended size of 15-20 chord lengths downstream and more than 5 upstream.

Mesh quality is as well above 0.95 (checked in ICEM) and the skewness and aspect ratios are all within the acceptable limits.

When I simulate the aerodynamical airfoil, the polars for both the Lift Coefficient and the Drag Coefficient versus the Angle of Attack are almost in perfect relation with the experimental data, with the error percentaje in any case being 5-10%.

The problem I'm having is with the LIFT COEFFICIENT of the thick airfoil, the one that can be seen in the second image.
For this case, running THE EXACT SAME SET UP to obtain the polars, the drag coefficient shows good values in accordance with the experimental data, whereas the LIFT is vastly underpredicted, at around 40-60% of error when compared to the supposed values it should have.
As well it should be mentioned that for the thick aifoil, the simulation does not always converge, and the residuals become flat in time before the set convergence target value.

I'm out of ideas of why this is happening, since the meshing procedure is the same for both airfoils, as well as the set up in CFX, but apparently, the lift coefficient in the thick airfoil is not predicted well.

This same problem was run on Fluent, and although for this software there are no convergence issues, the results remain the same. Very good polars for the thin airfoil, but for the thick airfoil, the lift polar is once again off by 40-60%, completely underpredicted.

Any ideas why this is happenning? Any suggestions to fix this issue and obtain a good lift polar for the thick airfoil? How can there be such a big change in the lift prediction just by switching to a thicker airfoil?

I insist on the fact that the mesh, and the setup for both airfoils are exactly identical.

Thank you very much in advance.

[JBeilke]

1) run it transient
2) switch to 3d, so that real vorticies can develope


and for the fun part:


http://blog.gridpro.com/the-art-and-...hing-airfoils/

[DanielBarreiro]

Originally Posted by JBeilke

1) run it transient
2) switch to 3d, so that real vorticies can develope


and for the fun part:


http://blog.gridpro.com/the-art-and-...hing-airfoils/

1)Why would the transient simulation provide better results than the steady simulation? Is there any particular reason you are suggesting this approach?

2) I have data as well for the Cl and Cd values for a 3D airfoil and the results are sligthly different, but the 2D simplification is still a valid approach since the vortex do not affect that much the polars.

Thank you for the meshing techniques link!

Do you have any other suggestions about why this problem is happening?

[JBeilke]

Originally Posted by DanielBarreiro

1)Why would the transient simulation provide better results than the steady simulation? Is there any particular reason you are suggesting this approach?

Because the flow is transient and 3d.

Originally Posted by DanielBarreiro

2) ... but the 2D simplification is still a valid approach since the vortex do not affect that much the polars.

If you think so.

[AidealZohary]

Hi, this is an old thread. But if anyone is still interested to run a simulation on the NACA 0012, NACA 4415, FX 61-184 E420 and S1223 airfoil, please take a look at:

Numerical Investigation on the Pressure Drag of Some Low-Speed Airfoils for UAV Application.

https://doi.org/10.37934/cfdl.13.2.2948

Unsteady 3-equation k omega intermittency SST was used. Good comparison with XFOIL and experimental data. Transition features also shown through cf and cp plots.

Learn how I designed the mesh here:

https://www.youtube.com/watch?v=qZRqBu9Ss2U