[AUCH]

Hello,
I have been trying to simulate the 0021 airfoil at a Reynolds number of 120,000. The Reynolds number is transitional and according to the experiments, there is a clear plateau in the pressure coefficients. This indicates the presence of the separation bubble and hence transition.
I used the transition SST model to capture the transition. My wall y+ was under 1. With an O-grid topology, the farfield is approximately 20 chord lengths from the airfoil. The velocity inlet and pressure outlet boundary conditions were used. The turbulence specification method was set to Intermittency, k and omega. The values were specified such that the turbulence intensity in the vicinity of the airfoil was around 0.8% (as per the experiments). The gauge pressure at the pressure outlet was kept zero.
I also did a mesh independence study at an angle of 4 degs at steady state conditions and found that 250k elements (880x284) were suitable for the 2D simulation. I have conducted steady state and transient simulations.
What I found was that the bubble was captured fairly accurately for both steady state and transient simulation. Furthermore, there were no problems in convergence for any case. For convergence I monitored the lift and drag coefficients as well as the residuals.
The main problem is that the lift is highly underpredicted as compared to the experimental value at 8 degs. As far as I understand, the peak suction pressure at the leading edge is not captured accurately and is consistently underpredicted. This leads to the smaller lift values. I found that if we specify the correct N_crit value in XFOIL, the results were approximately the same as that of the CFD. Both underpredict the leading edge suction pressure and hence the lift. Also the XFOIL results are slightly more accurate as compared to the CFD results.

I have also tried three dimensional simulations using the SST model but the lift is still underpredicted.

Any help in getting improved lift predictions would be highly appreciated.

Thanks.