I'm studying the Benedek 10355 airfoil (Plot), which should be a profile used in air-controlled models. While I was carrying out a boundary layer analysis, I realized the particularity that I have placed in the title. By making an incompressible analysis at Re = 500,000 and alpha = 2 deg, the profile has a laminar bubble on the upper surface (H index which diverges from 2.5 to 4 and then falls to 1.4, negative friction coefficient and plateau of the pressure coefficient) and then it also looks like a bubble on the lower surface. In this last case, the H index starts from 2.5 reaches about 3.5 then drops to values of 2.5, a sign that the boundary layer separates but there is no turbulent transition. The friction coefficient assumes slightly negative values. For Re = 1e6, alpha = 2 deg, the extension of the bubble on the upper surface decreases, while the bubble on the lower surface increases; H starts at 2.5 goes to 4 then drops back to 2.5 (so once again there is no transition on the belly). A Re = 8e6, no bubbles, either on the upper or on the lower surface. What is the cause of this abnormal behavior on the lower surface? It happens at every angle of attack, at every Reynolds. Thanks to those who will be able to help me!
1 Answer
The laminar "bubble" on the bottom is caused by the wavy contour. There is a pressure rise at about 25% on the lower side which causes the boundary layer to thicken, but the subsequent flat pressure profile keeps the boundary layer laminar until it separates at the trailing edge. Here is the plot from XFOIL and I placed a green circle over the spot:
With a higher Reynolds number the flow is less stable, the boundary layer is thinner and the waviness makes itself felt more than at lower Reynolds numbers. That explains the larger "bubble" at Re = 1,000,000. With Re = 8,000,000 the boundary layer transits to turbulent flow ahead of the disturbance and no "bubble" can be observed.
I guess you play around with XFOIL. Go to MDES and PLOT the Qspec, then MODI the contour so it becomes more smooth. The result below (first pressure, then H) should prove my point:
-
$\begingroup$ Wow! Thank you so much, your answer was very helpful! I also noticed that if I play around with the Ncrit, the "bubble" disappears for every Reynolds and every alpha $\endgroup$– leo95nfNov 13, 2020 at 20:20
-
1$\begingroup$ Both your images show the same Reynolds (0.5*10^6) did you mean to post a different image? $\endgroup$– FedericoNov 14, 2020 at 7:43
-
1$\begingroup$ @Federico Yes, that is on purpose. The second one uses a smoothed contour. Compare the incompressible pressure coefficient (dashed line) for the lower side. $\endgroup$ Nov 14, 2020 at 12:02
-
$\begingroup$ ah, I understood that the Re was supposed to be different too, sorry. $\endgroup$– FedericoNov 14, 2020 at 12:17
-
1$\begingroup$ @leo95nf There is a pressure (better: suction) peak for both, but in the viscous case it is smoothed a bit by viscous effects. That the dotted line shows a sharp peak is not relevant, a suction peak is a minimum along the c$_p$ curve. $\endgroup$ Nov 14, 2020 at 20:26