[cons013]

I am trying to validate an s1223 airfoil vs experimental data found here:

https://m-selig.ae.illinois.edu/pd/p...vol2/S1223.LFT

I am trying to do this in 2d and 3d to compare results and speed, for the purposes of comparing turbulence models. My 3d model consists of a wing with two symmetry walls at either side, with two no shear walls above and below, using Re=200,000. The 2d model is an unstructured tet mesh and all simulations are steady state with second order schemes.

In 3D, am noticing that the k-w sst model is underpredicting lift compared to k-e, and for some reason k-w sst does not handle post stall properly at all. The lift figures nosedive after 13 degrees aoa and I don't see why this would happen. 2D is even worse, even pre-stall the omega sst model is underpredicting lift by a long way, whereas k-e is showing good convergence. I have ensured y+ of less than 1 with sufficient prism growth, but nothing works.

I noticed that, in 3D, if my mesh is too fine, convergence isn't achieveable. I coarsened the mesh to see if this would artifically dampen any vortex shedding at 13 degrees and it worked well, giving a Cl of around 2. To test if this would help further, I tried again using these same mesh settings at 18 degrees, expecting a Cl of around 2.1, but ended up with around 1.55. Again k-e gave a much better result at around 1.8.

I am very confused here, what could be going wrong? Experimental data shows stall at around 19 degrees, yet the CFD is giving me much earlier stall. Am I just wasting my time trying to get steady-state convergence here, considering that around stall we will have shedding and other transient effects happening? What confuses me is the sheer difference with k-e and k-w sst. I had believed that the omega sst model was better and predicting and handling separation, but my results show this to not be the case. What could I be doing wrong??