Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Vibhav Durgesh
Elifalet Garcia and Hamid Johari
The performance of airfoils at low Reynolds numbers is intriguing due to the complex
fluid dynamics phenomena observed including laminar separated flow, increased transition
susceptibility, and the rapid transition of the separated shear layer to a turbulent flow.
Therefore, the objective of this investigation was to experimentally study the aerodynamic
performance of a symmetric airfoil (NACA-0009) and a cambered (NACA-6409) airfoil at
low Reynolds numbers (specifically below 50,000), as well as to identify the flow structures
responsible for altering aerodynamic performance. Lift and drag force measurements were
performed, along with flow visualization measurements for Reynolds numbers of 20,000,
30,000, 40,000, and 50,000. All measurements for this study were performed in a low-
speed water tunnel facility. A force/torque transducer was used for performing force and
moment measurements, and hydrogen bubble technique was used for flow visualization. The
lift coefficient results and drag polar results indicated that the camber airfoil performed
better than the thin symmetric airfoil. The results also indicated that the aerodynamic
performance of the studied airfoils is significantly different from that at Reynolds number
values of the order of 106 . Furthermore, the flow visualization results showed a strong
correlation between the observed flow structures and aerodynamic performance.
Nomenclature
Angle of attack
c Chord length
Cl Lift coefficient
Cd Drag coefficient
Rec Reynolds number based on chord length
x, y, and z Cartesian coordinate system
I. Introduction
The aerodynamics of airfoils at low Reynolds numbers are complex, since the boundary layer separates
due to the presence of an adverse pressure gradient, and the separated shear layer transitions and reattaches
further downstream as a turbulent boundary layer. The region between laminar separation and turbulent
reattachment is commonly referred to as a laminar separation bubble (LSB). The presence and structure
of a LSB and other associated flow phenomena have a significant influence on the overall aerodynamic
performance of airfoil. Moreover, it has been observed that airfoils exhibit significantly different flow behavior
at low Reynolds numbers as compared to that at high Reynolds numbers.16
Several experimental studies have focused on quantifying the aerodynamic performance of different airfoils
at low Reynolds numbers.1, 713 While studying low-Reynolds number flows, Gaster5 observed the presence of
LSB on airfoils. Since then, several researchers have experimentally and numerically studied LSB on different
Assistant Professor, Department of Mechanical Engineering, and AIAA member.
Undergraduate, Department of Mechanical Engineering, and AIAA Student Member.
Professor, Department of Mechanical Engineering, and AIAA Associate Fellow.
1 of 9
Hydrogen bubble flow visualization was used to identify the flow structures on the suction side of the airfoils.
The results obtained from this study provided critical information about the aerodynamic loads on airfoils
at low Reynolds numbers, which may potentially aid in developing improved flow control applications for
UAVs and MAVs.
B. Models
The two airfoils used for this study were the NACA-0009 and NACA-6409, as shown in Figure 1. The
cambered airfoil (NACA-6409) allowed for studying the impact of camber on aerodynamic performance for
a fixed thickness ratio of 9%. Both airfoils used in this study had a chord length (c) of 6.0in, and were
CNC-machined from a single-piece aluminum block with an industrial-grade smooth surface (hydraulically
smooth). The maximum blockage ratio in the test section was 8% at the highest angle of attack.
C. Instrumentation
The aerodynamic load on the airfoils was measured using a Gamma ATI 6-Axis force/torque transducer,
which was capable of measuring 65N of force along the x and y axes and 130N along the z axis, with a
resolution of 0.025N. This transducer can measure 5Nm of torque with a resolution of 0.00075 Nm. The
force and moment data from the transducer were acquired using a 16 channel NI data acquisition system at
a sampling rate of 200 Hz for 120 seconds. Multiple measurements were performed for each case in order to
test for consistency in the aerodynamic load results. A schematic of the load cell and mounting mechanism
2 of 9
to 5 per second, depending on the Reynolds number. The pulse generator circuit was designed based on
the guidelines provided by Smits and Lin.18 A schematic of the field of view of the hydrogen bubble flow
visualization system, along with coordinate axes, is shown in Figure 2(b).
(a) (b)
Figure 2: Schematic of experimental setup: (a) airfoil with force torque transducer and rotating mechanism,
and (b) airfoil with coordinate axes and flow visualization field of view.
3 of 9
A. Lift coefficient
0.6
Cl
0.4
0.2
0
0 3 6 9 12 0 3 6 9 12 0 3 6 9 12 0 3 6 9 12
, , , ,
Figure 3: Lift coefficient as a function of angle of attack, for NACA-0009 airfoil at: (a) Rec =50k, (b)Rec =40k,
(c) Rec =30k, and (d) Rec =20k.The blue dash-dot line represents the 2 curve.
Figure 4(a)-(d) shows Cl behavior of the NACA-6409 airfoil for Reynolds numbers of 50k, 40k, 30k and
20k, respectively. The observed Cl trend is different from that observed for the thin NACA symmetric airfoil
(NACA-0009). As observed in these figures, zero lift occurred at negative values, which is as expected for
the cambered airfoil. However, at Rec values of 50k and 40k, zero lift occurred at zL = 4 , while at Rec
values of of 30k and 20k, zero lift coefficient occurred at zL = 3 . For Rec values of 50k and 40k, the
4 of 9
0.5
0
Downloaded by UNIVERSITY OF COLORADO on January 15, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-0771
-0.5
-5 0 5 10 15 20 -5 0 5 10 15 20 -5 0 5 10 15 20 -5 0 5 10 15 20
, , , ,
Figure 4: Lift coefficient as a function of angle of attack, for NACA-6409 airfoil at: (a) Rec = 50k, (b)Rec
= 40k, (c) Rec = 30k, and (d) Rec = 20k. The blue dash-dot line represents 2 curve.
B. Drag polar
Lift and drag measurements on the airfoils allow for drag polar analysis at the studied Reynolds numbers.
Figures 5 and 6 show the drag polar for the NACA-0009 and NACA-6409 airfoils, respectively. As observed
in Figures 5(a)-(d), the drag polar curves exhibited a similar trend for Rec values of 50k, 40k, 30k, and 20k,
suggesting Cd Cl2 . It should be noted that for Rec of 20k and of 0 , the NACA-0009 airfoil exhibited
higher Cd values as compared to those at greater Rec values (i.e., 50k, 40k, and 30k). Similar to the
symmetric airfoil, the NACA-6409 airfoil exhibited a quadratic relationship between Cl and Cd , as observed
in Figure 6. At higher drag coefficient values and Rec of 50k and 40k, the airfoil displayed no significant
increase in lift coefficient values. For lower Rec values (i.e., 20k) at a higher drag coefficient values (i.e.,
Cd 0.3), there is a decrease in lift coefficient values with further increase in drag coefficients (see Figure
6(d)). Furthermore, a discontinuous variation is also observed in the drag polar at low Rec values as a result
of jumps in the lift curves.
0.6
Cl
0.4
0.2
0
0 0.1 0.2 0 0.1 0.2 0 0.1 0.2 0 0.1 0.2
Cd Cd Cd Cd
Figure 5: Drag polar for the NACA-0009 airfoil at: (a) Rec = 50k, (b)Rec = 40k, (c) Rec = 30k, and (d)
Rec = 20k.
C. Flow visualization
In this subsection, a comparison of flow visualization results for select cases is presented. These flow visual-
ization results underline the differences in the flow field under various scenarios and the observed differences
5 of 9
Figure 6: Drag polar for the NACA-6409 airfoil at: (a) Rec = 50k, (b)Rec = 40k, (c) Rec = 30k, and (d)
Rec = 20k.
Downloaded by UNIVERSITY OF COLORADO on January 15, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-0771
in Cl values.
Figures 7(a) and (b) highlight the difference in the flow field for the NACA-0009 airfoil at = 0 for
Reynolds numbers of 20k and 50k. As observed in these figures, at Rec of 50k, the NACA-0009 airfoil
exhibited the presence of a thin laminar boundary layer, while at Rec of 20k the presence of a thick laminar
boundary layer was observed. Furthermore, the wake is observed to be laminar at both these Rec values
(i.e., 50k and 20k). Periodic oscillation in the near wake region can be observed at Rec value of 50k, but not
at Rec value of 20k, at least in the near wake.
The Cl values vary significantly with change in Rec value from 50k to 20k and of 3 for the NACA-
0009 airfoil. Therefore, the flow visualization images at this value are analyzed further. Figures 8(a)
and (b) show typical flow visualization images for the highest and lowest studied Rec values of 50k and 20k,
respectively. The wake exhibited turbulent behavior for Rec value of 50k, which is in contrast to the behavior
for Rec value of 20k. As observed in Figure 8(a), for Rec of 50k, the separated flow follows the airfoil profile
closely on the suction side, which aided in suction pressure recovery at this Rec . However, at Rec of 20k, the
separated flow moved away from the airfoil suction side, as observed in Figure 8(b). It should be noted that
for Rec value of 50k, the separation location had close to the leading edge, while the separation location has
moved towards the trailing edge for Rec value of 20k.
(a) (b)
Figure 7: Flow visualization on NACA-0009 airfoil for = 0 at: (a) Rec = 50k, and (b) Rec = 20k.
(a) (b)
Figure 8: Flow visualization on NACA-0009 airfoil for = 3 at: (a) Rec = 50k, and (b) Rec = 20k.
The flow visualization results at of 6 for the NACA-0009 airfoil display separated flow behavior
different from those presented earlier. Moreover, this particular case was of interest because at this value,
the Cl values are almost equal for all studied Rec values. Figures 9(a) and (b) show flow visualization results
6 of 9
(a) (b)
Figure 9: Flow visualization on NACA-0009 airfoil for = 6 at: (a) Rec = 50k, and (b) Rec = 20k.
Downloaded by UNIVERSITY OF COLORADO on January 15, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-0771
Key aspects of the aerodynamic performance of the NACA-0009 airfoil are evident in Figures 7(a), 8(a),
and 9(a). As observed in these figures, the separated flow exhibited distinctly different behavior, (1) the
separated flow stayed away from the airfoil for of 0 , (2) the separated flow closely followed the airfoil
suction surface for of 3 , and (3) the separated flow reattached to the airfoil surface for of 6 . For
Rec value of 20k, the separated flow exhibits similar behavior, except for the case where the separated flow
closely followed the airfoil surface.
The cambered airfoil studied here also showed significant deviation in Cl and drag polar behavior with
increase in Rec value from 20k to 50k. The following paragraphs discuss in detail the flow visualization
results for the studied cambered airfoil.
The flow visualization images in Figures 10(a) and (b) show the flow behavior over the NACA-6409 airfoil
at Cl values close to zero. As observed in the figure, the flow separated close to the leading edge and stayed
separated over the airfoil with a turbulent wake. Furthermore, for both cases presented here, the flow on the
pressure side of the airfoil separated, transitioned to fully turbulent flow, and vortex rolling over the surface
is observed.
(a) (b)
Figure 10: Flow visualization on NACA-6409 airfoil for = 3 at: (a) Re = 50k, and (b) Re = 20k.
The flow visualization results for the cambered airfoil are analyzed for Rec values of 50k and 20k and
of 2 , as shown in Figure 11. At this , there was significant deviation in the Cl values for the studied Rec
values of 50k and 20k. These flow visualization results illustrated the difference in flow structure leading
to the differences observed in Cl values. The separated flow stayed close to the airfoil suction surface and
reattached near the trailing edge for Rec value of 50k. This may help in recovery of suction pressure on the
airfoil. In contrast, the separated flow for Rec value of 20k did not reattach to the airfoil, thereby leading
to a significant loss in suction pressure.
(a) (b)
Figure 11: Flow visualization on NACA-6409 airfoil for = 2 at: (a) Rec = 50k, and (b) Rec = 20k.
7 of 9
(a) (b)
Downloaded by UNIVERSITY OF COLORADO on January 15, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-0771
Figure 12: Flow visualization on NACA-6409 airfoil for Rec = 20k at: (a) = 10 , and = 12 .
At of 12 , the NACA-0009 airfoil exhibited soft stall behavior and the NACA-6409 airfoil had Cl values
almost two times larger than those observed for the NACA-0009 airfoil. Therefore, flow visualization images
for this particular case are analyzed. As observed from Figures 13(a) and (b), the flow separated close to
the leading edge for both airfoils. However, the separated shear layer became fully turbulent and did not
reattach to the airfoil, thereby leading to a loss of suction pressure for the NACA-0009 airfoil. On the other
hand the separated flow stayed close to the suction surface of the airfoil for the NACA-6409 airfoil and
may have aid in recovery of some suction pressure over the airfoil. This behavior of the separated flow is
responsible for better performance of the thin camber airfoil than the thin symmetric airfoil.
(a) (b)
Figure 13: Flow visualization at Rec = 50k and = 12 : (a) NACA-0009, and (b) NACA-6409.
V. Conclusions
For this investigation, lift and drag coefficient measurements, along with flow visualization measurements
were performed for two airfoils (i.e., NACA-0009 and NACA-6409) at Reynolds number of 20k, 30k, 40k, and
50k. The lift coefficient and drag polar plots for the studied airfoils showed that aerodynamic performance
of the airfoils at low Reynolds numbers is different compared to the behavior observed at Reynolds number
of the order of 106 . The cambered airfoil (i.e., NACA-6409) had higher lift coefficient values than the
symmetric airfoil (NACA-6409) at these low Reynolds numbers. At Reynolds numbers of 20k, both airfoils
demonstrated non-linear Cl behavior, and the non-linear trend decreased with increase in Reynolds number.
Furthermore, the drag polar showed that there was a quadratic relationship between drag coefficient and lift
coefficient values with a Reynolds number dependence.The flow visualization images indicate a correlation
between the separated flow condition and observed aerodynamic performance of the to studied airfoils.
References
1 Carmichael,B., Low Reynolds number airfoil survey, Vol. 1, NASA Technical Report, 1981.
2 Yarusevych,S., Sullivan, P. E., and Kawall, J. G., On vortex shedding from an airfoil in low-Reynolds-number flows,
Journal of Fluid Mechanics, Vol. 632, 2009, pp. 245271.
8 of 9
measurements on a low Reynolds number airfoil in three facilities, AIAA paper , Vol. 5149, No. 1, 2005, pp. 2005.
5 Gaster, M., The structure and behaviour of laminar separation bubbles, Citeseer, 1969.
6 Ravi, S., Watkins, S., Watmuff, J., Massey, K., Peterson, P., and Marino, M., Influence of large-scale freestream
turbulence on the performance of a thin airfoil, AIAA Journal, Vol. 50, No. 11, 2012, pp. 24482459.
7 Boutilier, M. S. H., Experimental investigation of transition over a NACA 0018 Airfoil at a low Reynolds number ,
separation bubble on an SD7003 airfoil, Experiments in Fluids, Vol. 44, No. 4, 2008, pp. 609622.
9 Fitzgerald, E. J. and Mueller, T. J., Measurements in a separation bubble on an airfoil using laser velocimetry, AIAA
Document, 2000.
12 OMEARA, M. and Mueller, T., Laminar separation bubble characteristics on an airfoil at low Reynolds numbers,
Downloaded by UNIVERSITY OF COLORADO on January 15, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-0771
separation bubbles, AIAA Journal, Vol. 45, No. 6, 2007, pp. 13461356.
16 Burgmann, S. and Schroder, W., Investigation of the vortex induced unsteadiness of a separation bubble via time-resolved
and scanning PIV measurements, Experiments in Fluids, Vol. 45, No. 4, 2008, pp. 675691.
17 Olson, D. A., Katz, A. W., Naguib, A. M., Koochesfahani, M. M., Rizzetta, D. P., and Visbal, M. R., On the challenges
in experimental characterization of flow separation over airfoils at low Reynolds number, Experiments in Fluids, Vol. 54, No. 2,
2013, pp. 111.
18 Smits, A. J. and Lim, T., Flow Visualization, Techniques and Examples, Imperial College Press, 2000.
9 of 9