Some notes about

Comparing GU & Savier airfoil equipped half canard

In S4 wind tunnel (France)
Matthieu Scherrer
Adapted from Charlie Pujo & Nicolas Gorius work
Contents
1 Test conditions 3
1.1 S-4 Wind-tunnel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 The team . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 The tested canards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 Limitation : a comparison work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2 Qualitative work : ow visualization 5
2.1 Wool string visualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Laser ow visualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.3 Oil ow visualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.4 Visualization report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3 Quantitative work (1/2) : polar measurements 7
3.1 Data for comparing VariEze Canard equipped with GU and Savier airfoil . . . . . . . . . . . . . 7
3.1.1 Canard equipped with GU airfoil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.1.2 Canard equipped with Savier airfoil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.1.3 Comparison of canards equipped with GU and Savier airfoil . . . . . . . . . . . . . . . . . 16
3.2 3D visualisation of lift coecient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.2.1 Canard equipped with GU airfoil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.2.2 Canard equipped with Savier airfoil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.3 Eect of wetted surface (rain) and transition on canard equipped with Savier Airfoil . . . . . . 25
4 Quantitative work (2/2) : extrapolation from polar measurements 29
4.1 Extrapolated data for airfoils alone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
4.2 Data adapted to ight condition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
4.2.1 Result of in ight measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.2.2 Comparison of GU and Savier airfoil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
5 Conclusion 41
2
Chapter 1
Test conditions
1.1 S-4 Wind-tunnel
The S-4 wind tunnel used to be a commercial low speed wind tunnel until three years ago. Now property
of ENSICA school since two years, it is used in a pedagogic way, for studies realized by student. But the
measurement equipment is still the original high quality equipment.
Figure 1.1: S4 wind tunnel elliptic test section.
The main data are :
Elliptic free test section 2x3m.
Speed between 10 and 40m/s.
Canard mounted on three mast, with 6 axis measurement (only 3 used).
Laser and UV light equipment for ow visualisation or measurement.
Camera and light for lming each test.
3
1.2 The team
This study was conducted as a pedagogic project (PIR) for two French SUPAERO young students : Charlie
Pujo and Nicolas Gorius. they produced a (French) report, highlighting some of the main points observe
throughout this study.
Figure 1.2: The full team within test section.
This project was initiated by Fabrice Claudel, helped by Patrick Lefebvre. The new canard was build
according to Klaus Saviers work and drawings.
The study was directed by Paul Claude Dufour and Matthieu Scherrer.
1.3 The tested canards
Full size half-canards were tested into the wind tunnel. A small at plate was put to create a symmetry axis.
Both GU and Savier canards were similarly intalled in the wind tunnel.
The wind speed was set to the maximum available, that is 40m/s. this will gives correct values of Reynolds
number for the low speed enveloppe of the VariEze, but too low for high speed.
1.4 Limitation : a comparison work
Here are some limitation about the work that was done.
Only half a canard, without fuselage inuence, was tested. The symmetry plane was not a perfect one.
It was possible to measure and then subtract only a part of drag created by the ttings and attachment
device.
As a consequence, the absolute value concerning C
d0
and C
lmax
might not be accurate. But the ttings and
attachment system were similar for both canard.
This work is then a Comparison of the two canards.
Chapter 2
Qualitative work : ow visualization
There were many way available to visualize the airow over the canards. We tried characterize the canards with
each of those.
NB : this was done partly in a pedagogical objective.
2.1 Wool string visualization
It was possible to move a wool string all aver the canard. Then functionning of the ap gap or wing tip vortex
could be observed.
2.2 Laser ow visualization
Wing tip ow was through laser tomography. A laser beam was split into a plan, and smoke was send. Then
the wing tip vortex could be studied.
There as no qualitative dierence in aspect between the vortex of the two dierent canard.
2.3 Oil ow visualization
The oil ow visualization were the most interesting qualitative testing. It was possible to visualize and analyze
the laminar to turbulent transition on the extrados, and the quality of this transition.
Shortly summed up, we may say :
GU Canard Savier canard
C
l
0.5, no ap de.
Position of transition 60-65% 40%
Laminar bubble wide (10% of chord) & thick (oil) short (3-4% of chord) & thin (no oil)
C
l
1.2, no ap de.
Position of transition - 25%
Laminar bubble - disappearing
Intrados was not much observed.
It was tried to trip the laminar boundary layer to turbulent on the GU airfoil, to kill the laminar bubble.
This does work with quite thick scotch tape.
5
Other visualization were made to observed the wing tip ow. The path of the vortex wing tip can be clearly
seen.
2.4 Visualization report
Each visualization run was fully lmed.
A 3 minutes sequences was assembled by Charlie Pujo and Nicolas Gorius, and compressed as an *.AVI
lm. It represents the best visualization report. Indeed, photograph are too static to represents properly those
visualisations.
This *.AVI lm ( 50 Mo) might be available on the internet or on an ftp site (to be dened).
I will also issue a compilation of some of the photographs taken during those visualization trial.
Chapter 3
Quantitative work (1/2) : polar
measurements
3.1 Data for comparing VariEze Canard equipped with GU and
Savier airfoil
Here are displayed the results of measurement on the two canards. Three axis were measured :
Lift force values.
Drag force values.
Pitching moment values, at 25% of chord (that aerodynamical center).
First are presented values for GU canard, and then for Savier canard. Finally comparison are made and some
conclusion drawn.
7
3.1.1 Canard equipped with GU airfoil
Lift coecient
10 5 0 5 10 15 20 25
1
0.5
0
0.5
1
1.5
2
2.5
GU Airfoil
Alpha (deg)
C
l

(

)
Flap setting
27deg
15deg
7.5deg
0 deg
7.5deg
15deg
24deg
Figure 3.1: Lift coecient vs incidence. Note the tricky hysteresis behavior for high ap deec-
tion.
Drag polar
0 0.1 0.2 0.3 0.4 0.5 0.6
1
0.5
0
0.5
1
1.5
2
2.5
GU Airfoil
Alpha (deg)
C
l

(

)
Flap setting
27deg
15deg
7.5deg
0 deg
7.5deg
15deg
24deg
Figure 3.2: Lift coecient vs drag coecient.
Drag polar (detail)
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
0.4
0.2
0
0.2
0.4
0.6
0.8
1
1.2
GU Airfoil
C
d
()
C
l

(

)
Flap setting
27deg
15deg
7.5deg
0 deg
7.5deg
15deg
24deg
Figure 3.3: Lift coecient vs drag coecient, detailed around minimum drag.
Moment coecient (at 25% chord)
10 5 0 5 10 15 20 25
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
0.05
0.1
GU Airfoil
Alpha (deg)
C
m

(

)
Flap setting
27deg
15deg
7.5deg
0 deg
7.5deg
15deg
24deg
Figure 3.4: Moment coecient vs incidence, detailed around minimum drag.
3.1.2 Canard equipped with Savier airfoil
Lift coecient
10 5 0 5 10 15 20 25 30 35
1
0.5
0
0.5
1
1.5
2
2.5
Savier Airfoil
Alpha (deg)
C
l

(

)
Flap setting
15deg
10deg
8deg
3deg
0deg
15deg
30deg
40deg
Figure 3.5: Lift coecient vs incidence. Contrary to GU airfoil, there is no hysteresis behavior
for high ap deection.
Drag polar
0 0.1 0.2 0.3 0.4 0.5 0.6
1
0.5
0
0.5
1
1.5
2
2.5
Savier Airfoil
Cd ()
C
l

(

)
Flap setting
15deg
10deg
8deg
3deg
0deg
15deg
30deg
40deg
Figure 3.6: Lift coecient vs drag coecient.
Drag polar (detail)
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
0.4
0.2
0
0.2
0.4
0.6
0.8
1
1.2
Savier Airfoil
Cd ()
C
l

(

)
Flap setting
15deg
10deg
8deg
3deg
0deg
15deg
30deg
40deg
Figure 3.7: Lift coecient vs drag coecient, detailed around minimum drag.
Moment coecient (at 25% chord)
10 5 0 5 10 15 20 25 30 35
0.3
0.25
0.2
0.15
0.1
0.05
0
0.05
0.1
Savier Airfoil
Alpha (deg)
C
m

(

)
Flap setting
15deg
10deg
8deg
3deg
0deg
15deg
30deg
40deg
Figure 3.8: Moment coecient vs incidence.
3.1.3 Comparison of canards equipped with GU and Savier airfoil
We shall now plot the C
l
, C
d
and C
m
values on the same graphics.
Some results Here are the main results concerning lift :
The Savier canard can reach the same C
l
range as GU canard, including very high values for C
lmax
.
This C
l
range is obtained for dierent ap setting and incidence for GU or Savier canard. Hence the
importance of a correct setting of the rigging angle on the fuselage.
The Savier has apparently no problem with hysteresis for high ap deection, comparing to the GU. It
can be hoped that Savier canard might not present tricky behavior with high loading as GU sometime
does.
Then some results concerning drag :
The Savier canard has a lower minimum drag coecient than GU canard. This corresponds to lower C
d
at low loads, that is high speed conguration.
Minimum drag for the Savier canard is reached for -3
o
ap defelction
The polar for Savier canard is more rounded and closed than for GU, than is C
d
gets higher for the
Savier than for the GU when its load gets higher.
My two cents about this : my impression is that Savier canard is more adapted to high speed, even if less
laminar, and then would have less C
lmax
with no ap deection.
But thanks to a carefully design slat, it can reach as high C
lmax
as GU canard , while being more performing
for high speed.
Lift coecient
10 5 0 5 10 15 20 25 30 35
1
0.5
0
0.5
1
1.5
2
2.5
GU & Savier Airfoil
Alpha (deg)
C
l

(

)
GU airfoil
Savier airfoil
27
15
7.5
0
7.5
15
24
15
10
8
3
0
15
30
40
Figure 3.9: Lift coecient vs incidence. Contrary to GU, Savier airfoil has no hysteresis behavior
for high ap deection. C
l
slope is more regular for Savier than for GU.
Drag polar
0 0.1 0.2 0.3 0.4 0.5 0.6
1
0.5
0
0.5
1
1.5
2
2.5
GU & Savier Airfoil
Cd ()
C
l

(

)
GU airfoil
Savier airfoil
Figure 3.10: Lift coecient vs drag coecient. Eect of ap deection may seem higher for
Savier airfoil.
Drag polar (detail)
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
0.4
0.2
0
0.2
0.4
0.6
0.8
1
1.2
GU & Savier Airfoil
Cd ()
C
l

(

)
GU airfoil
Savier airfoil
Figure 3.11: Lift coecient vs drag coecient, detailed around minimum drag. Minimum drag
is smaller for Savier than for GU airfoil.
Moment coecient (at 25% chord)
10 5 0 5 10 15 20 25 30 35
0.3
0.25
0.2
0.15
0.1
0.05
0
0.05
0.1
GU & Savier Airfoil
Alpha (deg)
C
m

(

)
24
27
15
7.5
0
7.5
15
15
10
8
3
0
15
30
40
Figure 3.12: Moment coecient vs incidence. Absolute values of C
m0
are smaller for Savier than
for GU airfoil.
3.2 3D visualisation of lift coecient
3.2.1 Canard equipped with GU airfoil
C
l
according to incidence and ap setting
10
5
0
5
10
15
20
25
30
35
20
0
20
40
1
0.5
0
0.5
1
1.5
2
2.5
Incidence (deg)
GU Airfoil
Flap setting (deg)
Figure 3.13: Lift coecient as a function of both incidence and ap setting. Tri-dimensionnal
view.
C
l
according to incidence and ap setting
10 5 0 5 10 15 20 25 30 35
30
20
10
0
10
20
30
40
GU Airfoil
Incidence (deg)
F
l
a
p

s
e
t
t
i
n
g

(
d
e
g
)

0
.
5

0
.
5
0
0
0
0
0
.
2
5
0
.
2
5
0
.
2
5
0
.
2
5
0
.
5
0
.
5
0
.
5
0
.
5
0
.
5
0
.
5
0
.
7
5
0
.
7
5
0
.
7
5
0
.7
5
0
.
7
5
0
.
7
5
1
1
1
1
1
1
.
2
5
1
.
2
5
1
.2
5
1
.
2
5
1
.
5
1
.
5
1.5
1
.
5
1
.5
1.75
1
.
7
5
1
.
7
5
1.75
1
.8
1
.
8
1
.
8
2
Figure 3.14: Lift coecient as a function of both incidence and ap setting. Levels view.
3.2.2 Canard equipped with Savier airfoil
C
l
according to incidence and ap setting
10
5
0
5
10
15
20
25
30
35
20
0
20
40
1
0.5
0
0.5
1
1.5
2
2.5
Incidence (deg)
Savier Airfoil
Flap setting (deg)
Figure 3.15: Lift coecient as a function of both incidence and ap setting. Tri-dimensionnal
view.
C
l
according to incidence and ap setting
10 5 0 5 10 15 20 25 30 35
30
20
10
0
10
20
30
40
Savier Airfoil
Incidence (deg)
F
l
a
p

s
e
t
t
i
n
g

(
d
e
g
)
0
0
0
0
0
.
2
5
0
.
2
5
0
.
2
5
0
.
2
5
0
.
5
0
.
5
0
.
5
0
.
5
0
.
5
0
.
5
0
.
7
5
0
.
7
5
0
.
7
5
0
.
7
5
0
.
7
5
0
.
7
5
1
1
1
1
1
1
1
1
1
.
2
5
1
.
2
5
1
.
2
5
1
.
2
5
1
.
2
5
1
.
2
5
1
.
2
5
1
.
2
5
1
.
5
1
.
5
1
.
5
1
.
5
1
.
5
1
.
5
1
.
7
5
1
.
7
5
1
.
7
5
1
.
7
5
1
.
8
1
.
8
1
.
8
1
.
8
2
2
Figure 3.16: Lift coecient as a function of both incidence and ap setting. Levels view.
3.3 Eect of wetted surface (rain) and transition on canard equipped
with Savier Airfoil
Articial transition was created by carborandum located at 5% chord.
Wetted surface referred to measurements with visualization oil on the canard.
One conclusion is that if transition occurred early, performance is really aected. But it seems Savier airfoil
is resistant to wetted surface, since wetted surface did not seem to create transition. This point should be
detailed further.
Eect of rain & transition on lift coecient
10 5 0 5 10 15 20 25 30 35
0.8
0.6
0.4
0.2
0
0.2
0.4
0.6
0.8
1
1.2
Savier Airfoil
Alpha (deg)
C
l

(

)
Natural airfoil
Wetted extr.
Transition extr. 5%
Figure 3.17: Lift coecient vs incidence. Savier airfoil seems to be not really hurt by wetted
surface, whereas genuine transition reduced greatly maximum lift.
Eect of rain & transition on drag polar
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
0.8
0.6
0.4
0.2
0
0.2
0.4
0.6
0.8
1
1.2
Savier Airfoil
C
d
()
C
l

(

)
Natural airfoil
Wetted extr.
Transition extr. 5%
Figure 3.18: Lift coecient vs drag coecient.
Eect of rain & transition on moment coecient
10 5 0 5 10 15 20 25 30 35
0.01
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
Savier Airfoil
Alpha (deg)
C
m

(

)
Natural airfoil
Wetted extr.
Transition extr. 5%
Figure 3.19: Moment coecient vs incidence.
Chapter 4
Quantitative work (2/2) :
extrapolation from polar measurements
From raw data, some calculation work can be done for extrapolating some complementary data.
At rst, the part of the airfoil within drag can be extracted.
Then, we can look carefully the wind tunnel data on each functioning point that is encountered in steady
ight.
29
4.1 Extrapolated data for airfoils alone
In short, drag measured for the half canard can be split into :
Mounting drag C
dMount
, that is drag causes by the tting device into the wind tunnel.
Airfoil drag C
dAirf
Induced drag C
di
C
d
= C
dMount
+C
dAirf
+C
di
(4.1)
Assuming the left plate is a correct plan of symmetry, we can say that the eective aspect ratio is about
= 5. This is coherent with the C
l
slope measured. Then the induced drag can be expressed as C
di
=
1

C
l
2
,
and subtracted from the total drag coecient.
Then we can compare airfoil, since Mounting drag C
dMount
is comparable for both half canards.
NB : so keep in mind that the minimum drag values C
d0
plotted contains C
dMount
, that is ttings and
attachment system. So that they are not genuine airfoil C
d0
.
Drag polars
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
1
0.5
0
0.5
1
1.5
2
2.5
GU Airfoil
Alpha (deg)
C
l

(

)
Flap setting
27deg
15deg
7.5deg
0 deg
7.5deg
15deg
24deg
Figure 4.1: Lift coecient vs drag coecient, for airfoil only.
Drag polars
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
1
0.5
0
0.5
1
1.5
2
2.5
Savier Airfoil
Alpha (deg)
C
l

(

)
Flap setting
15deg
10deg
8deg
3deg
0deg
15deg
30deg
40deg
Figure 4.2: Lift coecient vs drag coecient, for airfoil only.
Drag polars comparison
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
1
0.5
0
0.5
1
1.5
2
2.5
GU & Savier Airfoil
C
d
()
C
l

(

)
GU airfoil
Savier airfoil
Figure 4.3: Comparison of drag polars for airfoils only. Savier Airfoil has a less laminar behavior
(more rounded polar), but lower C
dmin
.
4.2 Data adapted to ight condition
An attempt to get the eective in ight parameters over the canards was made. Every parameter were not
known, but this is a good start for the comparison of both canard.
Indeed, wind tunnel data have to be treated to be used with the dierent parameters
Measurement of ap deection and fuselage attitude was installed into three aircraft (F-PYHR (Savier
canard), F-PYOP (GU Canard), F-PYSM (GU Canard)).
Similar measurements were made on the three aircraft to reduce scatter eect.
Measurements were made for aircraft wing-loadings close to each other.
Data were collected at rst on a steady, constant attitude path, for basic values of incidence and ap
deection. Then some measurements were also made with constant bank and altitude, for higher load on
the canard.
From measurements, calculations were made to get the eective aerodynamic incidence on the canard,
inuenced by the wing.
Some interpolations of the wind tunnel results were made to get the exact functioning parameters of the
canards within the collected data.
We should keep in mind the following points :
A weak point of this study is that we could not get precise CG location for each aircraft. CG location
inuences canard load, hence its performance. We supposed that empty CG location were comparable on
the dierent aircraft, according to the building plan.
It is a pity we could not make the measurements of both canards on the same aircraft. This will be soon
possible on F-PYIB (Savier & GU canard) which is currently grounded, and this study will be released.
Reynolds number for wind tunnel data corresponds to low speed.
Some results
If you look only on the eective drag polar for the canard of trimmed aircraft, Savier Canard has a lower
drag coecient only for low C
l
values.
If you plot C
d
as a function of speed of the trimmed aircraft, Savier Canard has lower drag coecient for
a larger speed domain.
If you plot value of the drag force, as a function of speed of the trimmed aircraft, Savier Canard saves
really quite a lot drag at high speed. This is caused partly to the magnifying eect of V
2
value.
4.2.1 Result of in ight measurements
Measured ap setting and incidence according to speed for the canard of trimmed aircraft
(Based on in ight measurement)
100 150 200 250 300 350
10
5
0
5
10
15
Speed (km/h)
(
d
e
g
)
GU, Flap setting
GU, Incidence
Savier, Flap setting
Savier, Incidence
Figure 4.4: Measured ap setting and incidence on aircraft equipped with both canard. Flap
setting was directly measured, whereas incidence was calculated from fuselage attitude, including
deection and rigging angle of the canard.
Calculated eective C
l
for the canard of trimmed aircraft according to speed
(Based on in ight measurement)
100 150 200 250 300 350
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
V (km/h)
C
l

(

)
Resulting GU
Resulting Savier
Figure 4.5: C
l
for trimmed aircraft vs speed, on aircraft equipped with both canard.
4.2.2 Comparison of GU and Savier airfoil
Eective C
l
for the canard of trimmed aircraft
(Calculation based on in ight measurement plus wind tunnel interpolation)
10 5 0 5 10 15 20 25 30 35
0.5
0
0.5
1
1.5
2
Incidence (deg)
C
l

(

)
Resulting GU
Resulting Savier
Figure 4.6: C
l
for trimmed aircraft vs incidence and ap setting, as put on the wind tunnel
measurement.
Eective drag polar for the canard of trimmed aircraft
(Calculation based on in ight measurement plus wind tunnel interpolation)
0.05 0.1 0.15 0.2 0.25
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
C
d
()
C
l

(

)
Resulting GU
Resulting Savier
Figure 4.7: C
l
vs C
d
for the canard of trimmed aircraft, interpolated from wind tunnel measure-
ment.
Drag coecient for the canard of trimmed aircraft according to speed
(Calculation based on in ight measurement plus wind tunnel interpolation)
100 150 200 250 300 350
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
V (km/h)
C
d

(

)
Resulting GU
Resulting Savier
Figure 4.8: C
d
for the canard of trimmed aircraft vs speed, interpolated from wind tunnel mea-
surement. Savier airfoil equipped canard has higher drag coecient for intermediate speed,
whereas at low and high speed C
d
is lower.
Drag values in Newton (N) for the canard of trimmed aircraft according to speed
(Calculation based on in ight measurement plus wind tunnel interpolation)
100 150 200 250 300 350
60
70
80
90
100
110
120
130
V (km/h)
D
r
a
g

(
N
)
Resulting GU
Resulting Savier
Figure 4.9: Drag values in Newton (N) for the canard of trimmed aircraft vs speed, interpolated
from wind tunnel measurement. Drag value is 1/4 lower for the Savier Canard at high speed.
Chapter 5
Conclusion
This paper has presented the measurements made in a pedagogical project, performed by french students.
In this summary, experimental parameters were described. Then presentation of raw data were performed.
Some extrapolation were also made to use the raw data in in ight conditions.
From those data, others work could be done.
The measurement chain was top quality, and we are condent about the measurements that were made. We
are really happy to have benet from S-4 wind tunnel.
The main regret we have is that the wind tunnel could not produce more than 40m/s. It would have been nice
to measure hinge moment on the ap, since reduction allowed by Savier Canard is a key point observed in ight.
We hope this work will help being condent about the Savier Canard.
41