I know about the theory differences between regular airfoils and thin airfoils, but is there any condition for saying a given airfoil can be analyzed as if it is thin? The extreme case of infinitely thin makes sense, but I am curious if there is some cutoff dimension that determines thin or not thin.
There is no hard boundary. Normally it is around 8% relative thickness, details depending on camber and nose shape.
Typical for a thin airfoil is a stall originating from the nose, with a sudden separation of upper side flow, while thicker airfoils start to stall with a separation starting from the trailing edge and moving gradually forward. This gives thin airfoils a nasty stall behavior while thick airfoils stall in more benign ways. The stall behavior depends not only on thickness but also on camber and the geometry details of the airfoil nose, but can be used to separate thin from thick airfoils.
Thin airfoils make sense in two applications:
- when the local angle of attack is well controlled, such as in flaps and turbo machinery, and
- for trans- and supersonic flight where thickness causes wave drag.
In all other applications thicker airfoils with a blunter nose should be preferred because they allow to store more fuel and to make the load-carrying structure more efficient. The upper limit of practical airfoils is at 20% to 22%, the root thickness of the Davis wing as used in the B-24 and B-29.
Airfoils are usually divided in "thin" and "thick" according to their stall behaviour: trailing edge stall, leading edge stall, and thin airfoil stall. One of the main defining parameters in how the wing stalls is the thickness of the wing profile.
From Torenbeek, both the pictures and the citations:
Trailing edge stall
This type of stall is characteristic of most airfoil sections with thickness/chord ratios of approximately 15% and above. The flow at large angles of attack is characterized by a progressive thickening of the turbulent boundary layer on the upper surface. As the angle of attack is increased to about 10 degrees (B), flow separation starts at the trailing edge and moves gradually forward.
Leading edge stall
Airfoils with thickness/chord ratios of about 9 to 12 percent experience an abrupt separation of the flow near the leading edge. On these sections separation of the laminar boundary layer occur well before the attainment of maximum lift and prior to transition to a turbulent boundary layer. Transition occurs in the shear layer thus formed, and the expansion of the turbulent motion spreads at such an angle that re-attachment of the flow quickly occurs, enclosing a "short bubble" and subsequently forming a turbulent boundary layer (B).
The lift and pitching moment curves exhibit abrupt changes when the angle of attack for maximum lift is exceeded. There is little or no rounding of the lift curve and a sudden negative shift of the pitching moment resulting from the rearward shift of the centre of pressure is observed.
Thin Airfoil stall
On very thin sections of thickness/chord ratios of less than about 6 percent and on round noses a small separation bubble occurs at very small angles of attack (S). At a certain critical angle of attack the short bubble breaks down, but the flow subsequently re-attaches downstream, forming a "long bubble" which causes a slight reduction in the lift-curve slope (B). With increasing angle of attack the point of flow reattachment progressively moves backward until it coincides with the trailing edge and maximum lift is reached at this condition (C)
So according to the categorisation of airfoils in three types of stall, "thin airfoil" means a chord thickness of 6% or less.
2$\begingroup$ As much as I respect Torenbeek, his distinguishing of three types of stall is a bit arbitrary. After all, his second case is just like the first, only that in thinner airfoils the laminar separation bubble rushes forward more quickly with increasing angle of attack. And his "thin airfoil stall" can be observed with airfoils of 10% thickness as well if they have little camber and a small leading edge radius. $\endgroup$ Feb 19, 2018 at 13:50
$\begingroup$ Yes the mechanics are pretty much identical, it's just the stall behaviour that differs, the leading edge stall being very nasty with the sudden drop in $C_L$, the thin airfoil $C_L$ drop being much more benign. Torenbeek does state in the book that there are many more variables, indeed such as the ones you mention. Thanks for the edit by the way. $\endgroup$– KoyovisFeb 19, 2018 at 14:14