Gurney flaps in sailing, do they help?

 2
the flow from the upper and lower surfaces must smoothly merge 
in the wake, then S
2
 tends to the rear end producing an anti-
clockwise vortex, see Fig.1c. Then, by The Helmholtz theorem 
about vorticity, the whole system responds with a clockwise 
vortex around the airfoil producing circulation, see Fig.1d. This 
assumption is inseparable from the well know equation Γ××=
0
V
b
L
ρ  
which allows the calculation of lift if the magnitude of  circulation 
is known. 
 
 
Fig. 1 
 
Then, Liebeck proposed a short region of separated flow directly 
upstream of the Gurney flap and two counter-rotating vortices 
downstream, see Fig 2, which he described as having a turning 
effect on the local flowfield. The result is increasing in circulation 
Γ, hence the downforce on the spoiler. But Liebeck  observed a 
slight reduction in drag too; he proposed that the separation of 
airflow from the  suction surface of the “upside down wing” is 
postponed thanks to the action of sucking by the separation bubble 
born by the two shedding vortices. Additionally, the postponed 
separation on the suction surface allows a higher angle of attack to 
 3
be used before stall, which further enhanced the wing's 
effectiveness.  
 
 
Fig. 2 
 
Recently, Jeffrey et al. [2] have conducted experiments to 
understand the origin of the phenomenon with more modern and 
sophisticated instruments among others the Laser Doppler 
Anemometry (LDA). In particular, he used an untwisted constant-
chord wing of finite span that has an Eppler c423 section, 
optimized for high lift and shares some features with typical race 
car wing elements, for example a significant degree of camber on 
the pressure surfaces. The freestream turbulence level in the tunnel 
wind was of 0.2-0.3%; the freestream velocity of V
0
=40ms
-1 
which 
gave Reynolds number in the range Re
c
=0.75-0.89×10
6
. Then, by 
observing the mean-velocity vectors and streamlines, it was found 
that the time-averaged flow downstream of  the vertical tab 
consists of two counter-rotating vortices, matching Liebeck’ s 
hypothesis. But the spectral analysis of LDA showed clearly that  
the instantaneous flow is a wake of alternately shed vortices in 
particular a von Kármán vortex street of alternately shed vortices, 
see Fig 3. A smoke visualization of the flow downstream of the 
Gurney flap confirmed this flow structure.  
 4
 
 
Fig. 3 
 
To explain how this particular wake contributes to increase the 
general loading on a wing it is first necessary to understand the 
process of vortex shedding for a typical bluff body. According to 
the generally accepted mechanism first postulated by Gerrard [3] 
for a two-dimensional vortex-shedding flow, the boundary layers 
on a bluff body separate at some point to form two shear layers of 
opposing vorticity. This shedding cycle begins as the separating 
shear layer on one side of the body rolls up to form a vortex 
breaking away from the body. Together it draws, very powerfully, 
the separating shear layer over from the other side of the body, 
that contains an opposite vorticity. Then, as the sucked layer 
crosses the wake centreline, its provision of vorticity is interrupted 
by the first eddy; now the second vortex is shed and moves 
downstream while the shear layer on the opposite side has just 
started to roll up repeating the process.  
Jeffrey adds that the Gurney flap works as an obstacle to the shear 
layer of the pressure surface, providing a fixed separation point 
 5
that interacts with the separating point from the suction surface to 
form a vortex street in a manner similar to other bluff bodies. This 
one is not fixed but depends on the angle of attack and Reynolds 
number on the airfoil that works as well as a circular cylinder, as 
already known. Thus, two considerations follow: at low incidences 
the shear layer separation point on the suction surface is located 
on the trailing edge and the upstream possible movement will not 
eliminate any vortex shedding but it will only modify the shedding 
process. 
Obviously a shedding principal frequency f
p
 is associated at the 
shedding cycle: the height of the Gurney flap and the incidence, 
directly connected to the boundary layer thickness, are two 
parameters that modify f
p
. In fact the shedding frequency reduces 
as the distance between two separating shear layers is increased; 
because the time, that the opposite shear layer takes to cross the 
wake centreline and cut off the supply of vorticity from the rolling 
up vortices, increases. Thus, the period of one cycle increases and 
the related frequency, obviously, reduces. This explains why f
p
 
reduces as the height of the Gurney flap is increased: an higher 
flap means longer distance to travel between the two layers. 
Secondly, under the same condition, the boundary layer thickness 
increases with the incidences. But a thicker boundary layer means 
a weaker velocity gradient across it, hence, a lower vorticity; then 
it will occur more time to have sufficient opposite vorticity to cut 
off the supply to the rolling up vortex; the consequence is, clearly, 
a reduction in principal frequency. Lastly, further experimental 
calculations show that the boundary layer thickness has a weaker 
effect on the shedding frequency than the distance between the 
shared layers. 
About  the trailing edge suction, it appears that the increased 
suction acting on the downstream face of the Gurney flap and 
hence at the trailing edge of the airfoil, is encouraged by the 
vortex shedding. Also, an increase in maximum suction does not 
generate premature boundary layer separation at low incidence 
 6
because in this manner the pressure recovery demand reduces. 
Loose relationships seem to reveal that increasing the height of the 
Gurney flap increases the trailing edge suction while the opposite 
happens if wing incidence decreases.  
About the trailing edge pressure, the chordwise pressures clearly 
show that the Gurney flap increases the pressure at the trailing 
edge of the airfoil because the upstream face of the flap 
decelerates the flow. For a flat plate immersed in a turbulent 
boundary layer, Good and Joubert [4] say that the maximum 
pressure measured upstream of the plate increases as the height of 
the disturbance is increased but these effects do not scale directly 
with the height of the device. Because the bigger Gurney flaps 
cause a relatively large increase in pressure, they work just as a 
bluff body in a ground plane that decelerates the flow which 
separates at some point upstream of the trailing edge,  then 
reattaches at an other point on the upstream face of the body. The 
Gurney flap introduces a pressure difference acting at the trailing 
edge of the airfoil and it is this that generates an increase in 
loadings over the entire airfoil. This one can be modelled in a 
simple two dimensional model by modifying the Kutta-Joukowski 
condition. In fact the increase in suction surface velocity and the 
reduction in pressure surface velocity can be treated theoretically, 
as a vortex point placed at the trailing edge which will increase the 
total circulation hence the lift acting on the wing. 
One the nature and the physic of the phenomenon have been 
explored, the next step is inspecting the latest studies about the 
performances and the effectiveness of the Gurney flaps. 
Bao et al. [5] have presented a research concerning the power and 
efficiency augmentation of an horizontal axis wind turbine system 
by adding small flaps to its blades. NACA62
2
-215 was selected as 
the airfoil to be tested and the Reynolds number was fixed at 
2.4×10
5
 based on airfoil chord. The angle of attack varied from 0° 
to 40° and the heights of the flaps was 1.0%, 1.5%, 2.0% and 
2.5% of the airfoil chord. The effects of different deflection angles 
 7
of the trailing edge flap were compared: 0°, 45°, 90° (i.e. Gurney 
flap), 135°. From the results of the experiments the following 
conclusion can be deducted: all tested types of flap significantly 
increase the lift coefficient with little or no change in drag 
coefficient and also the best increment in performance was by the 
addition of Gurney flaps in particular with the 2.0%c high flap. In 
this case, the stall angle was 2° earlier than that of the airfoil 
without flap. Bloy and Harrison [6] have investigated the effect of 
freestream turbulence intensity as a factor in the performance of 
the Gurney flap. They placed a turbulence generating screen in the 
test section of a wind tunnel to increase freestream turbulence 
intensity (from 0.2% to 2.5%) and repeated tests were made at the 
same Reynolds number on chord of 0.48×10
6
 on the airfoil section 
NACA5414; a geometry from which wind-turbine blades are 
obtained. Increasing in turbulence intensity, significant increases 
in maximum lift and profile drag were remarked in the plane wing 
but only with little effect on the performance of the devices. The 
explanation is that an increased turbulence produces just little 
changes in the boundary layer thickness which is the dominant 
parameter in determining the flap drag characteristics according 
with the studies of Giguere et al. [7]. In fact he has affirmed that 
for acceptable drag penalty, the flap should be submerged in the 
wing boundary layer. Giguere has added that a relatively long flap 
with low deflection angle is comparable with a shorter one but 
fitted perpendicular. In summary, among small flaps the Gurney  
is the best to increase the lift, it does not feel the influence of the 
freestream turbulence and to reduce the drag it has to be immersed 
in the airfoil boundary layer that is just till around 2.0%c. 
One of the wider range research was conducted probably by 
Storms and Young [8]; with the admitted target to provide an 
experimental database on the performance of the Gurney flaps on 
a single element airfoil. Measurements for a NACA4412 fitted 
with 0.5%, 1.0%, 1.5% and 2.0%c Gurney flaps were obtained; 
tests were without turbulence reducing screen in the circuit of 
 8
wind tunnel and with a Re
c
=2×10
6
. Here, this data base is 
reported. 
The comparison of pressure distribution for various device heights 
shows that Gurney flap considerably increases the aft loading of 
the airfoil but it also noted that much of the lift increment is 
caused by a general increase in loading and by a higher suction 
peak See Fig 4a and 4b to compare pressure distribution on the 
clean and modified airfoil, also as the Gurney flap height is 
increased higher loading is noted along the entire airfoil. 
The lift and drag coefficient, respectively C
l
 and C
d
, are presented 
in Fig. 5a and 5b, with an angle of attack from 0° to 15°. The 
addition of the Gurney flap produces a considerable lift increment 
compared to the baseline configuration. Larger flap produce larger 
increment in maximum C
l
, but not proportionally: 13% for the 
smallest and 32% for the biggest.  
Fig. 6 shows the drag polar for the same configuration. At low to 
moderate C
l
 , there is a drag penalty which increases with flap 
height; however, at higher C
l
 the drag is significantly reduced. The 
effect on the maximum L/D is small, but the C
l
 for a given L/D is 
significantly increased.   
Finally, the nose-down pitching moment coefficient C
m
 about 
0.25c is shown in Fig. 7, it increases with Gurney flap height 
although not proportionally. 
With the idea to stow the flap during cruise to eliminate drag 
increments at lower C
l
, Storm and Jang have considered a 
miniature split flap with a 90° deflection, see Fig. 8. Two 
configurations were tested for a 1.25%c split flap with a hinge line 
located forward of the trailing edge by 1.0 and 1.5 flap height, 
respectively. Fig. 9 illustrates that these configurations yielded 
essentially the same results as the 1.25%c Gurney flap without 
degradation in flap effectiveness. 
 9
          
                                            Fig. 4a                                                                                       Fig. 4b  
 
 
               
                                          Fig. 5a                                                                                             Fig. 5b 
 
 
      
                                                Fig. 6                                                                                      Fig. 7                                                             
 10
 
 
 
Fig. 8 
 
 
 
 Fig. 9 
 
Briefly, to show that Gurney flaps also work on airfoils made with 
the latest design philosophy to have already the maximum C
l
 at 
low Re regime, the research of Selig and Guglielmo is reported 
[9]. The key element about the airfoils is to make use of a concave 
pressure recovery with aft loading; so three codes for airfoil 
design and analysis were used to design the new S1223 high-lift 
airfoil for a Re
c
 of 2×10
5
. In wind tunnel tests the new airfoil 
yielded a maximum lift coefficient C
l max
 of 2.2 with an increase of 
25% relatively to older airfoils; with a 1% chord Gurney flap it 
 11
goes to 2.3, adding an ulterior 5% without modifying stall 
characteristics, which is very sensitive factor in these applications.  
A specific study has been performed on a single-element high-lift 
wing fitted with Gurney flaps in ground effect by Zerihan and 
Zhang [10]. A modification of a NASA GA(W) profile, used as 
front wing on Tyrrell 026 Formula1 car, with two sizes of Gurney 
flap i.e. 1.45% and 2.9%c, was tested in a wind tunnel in which 
Re
c
 was approximately 0.46×10
6
. The conclusions of this research 
say that the aerodynamic performance of the Gurney flap on a 
wing in ground effect has similar effects if compared in freestream 
circumstance. In fact the Gurney flap increases the downforce, 
even if disproportionately with its height. As the ride height is 
reduced in the force enhancement region for a fully attached flow, 
the gain in downforce with the Gurney flap increases. This one 
compares with results when the incidence of the wing is increased 
in a freestream; however, the gain in downforce with the Gurney 
flap in ground effect can be twice the gain in the freestream. 
 
 12
1.2 Effects of Gurney Flaps on a NACA0012 Airfoil [11] 
It is worth conducting a separate discussion about this report as 
not only is it the latest published research but it also deals with a 
symmetric airfoil which we will encounter along the route again. 
We can fit this simple strip perpendicular to the pressure surface 
along the trailing edge of an airfoil. But with the idea to stow the 
flap we can consider something like a miniature split flap with a 
90° deflection; a previous research shows this configuration yields 
essentially the normal one. The instantaneous flow downstream 
the airfoil is a well known wake, a von Karman vortex street of 
alternately shed vortices. The f
p
 associated at the shedding cycle 
can be increased by a reduction in angle of attack or, more 
significantly, with a reduction in height of the Gurney flap. About 
the trailing edge suction we can say the increased suction acting 
on the downstream face of the rear end of the airfoil is encouraged 
by the vortex shedding; about the trailing edge pressure the 
Gurney flap increases the pressure at the rear end of the airfoil 
because the upstream surface of the little device decelerates the 
flow. Then the Gurney flap introduces a pressure difference acting 
at the trailing edge of the airfoil and it is this that generates an 
increase in loading over the entire airfoil. 
About the performances of the Gurney flap on the airfoil section a 
0012NACA with a chord length of 1m was used; the Re
c
 was 
2.1×10
6
 thus a boundary layer thickness at α=0° of about 15 mm.  
Also, if we imagine the airfoil immersed into a water flow, a 
velocity of  4.5 knots has to be considered. Gurney flaps with a 
height of 0.5, 1, 1.5, 2 and 3% on chord were examined. 
 
 
Fig. 10 
 13
On the C
l
 versus α  
With higher flaps increments in C
l
 max but not  proportionally.
  
Height on chord in  % 0.5 1 1.5 2 3 
Increase in % 10 11 18 21 27 
 
The stall angle of attack decreased, in countertendency with      
other researches. The results suggest that the effect of the  devise 
is to increase the effective camber of the airfoil. 
 
 
Fig. 11 
 
On the C
d
 versus α 
The significant increase in C
l
 with a 3% on chord height flap 
comes at costs of substantially increased drag; in agreement with 
previous researches we can fix the limit of the flap height at 2% of 
the chord. 
 
 
Fig. 12 
 14
On the polar L\D versus C
l 
At low to moderate Cl there is a drag penalty associated with the 
Gurney flap height, but at higher Cl the L\D is substantially 
increased. Then the effect on L\D is small but the Cl for a given 
L\D is considerably increased. When the Cl is above 1.2 all the 
flaps provided increased L\D although the L\D max decreased 
increasing in flap height. In aircraft applications it means it could 
be desirable to close the flap during the cruise so it can be at its 
best in conditions such as take off and landing. In countertendency 
with some previous studies an overall reduction in drag was not 
found. 
 
 
Fig. 13 
 
On the nose-down pitching moment versus α 
It is increased with the Gurney flap height, this suggests the 
effective camber is increased with the use of the device. 
 
                
Fig. 14 
 15
 
On the pressure distribution (with α=0) 
The increased pressure exactly on the trailing edge of both 
surfaces confirms the theory about a von Karman vortex street and 
also an increased suction is evident everywhere on the relative 
surfaces as well as the lower surface experiencing increased 
pressure. 
 
Fig. 15 
 
On the boundary layer velocity profile taken at 90% of chord 
The boundary layer thickness is about 1.5% of chord in height 
near the trailing edge at α=0°, as the figure below shows. Thus a 
Gurney flap with a height of about 2% of chord or less would not 
significantly increase the drag since most of device works within 
the airfoil boundary layer. Moreover at α=6° the thickness is 
2.5%, at α=10 it is 4%. 
 
Fig. 16