the effect of corrugated skins on aerodynamic …michael.friswell.com/pdf_files/c331.pdfthe effect...

ICAST2012: 23rd

International Conference on Adaptive Structures and Technologies

October 11-13, 2012, Nanjing, China

1

ICAST2012 #27

The Effect of Corrugated Skins on Aerodynamic Performance

Y. Xia1*, O. Bilgen2, M. I. Friswell1

1 College of Engineering, Swansea University, Swansea, UK

2 Department of Mechanical and Aerospace Engineering, Old Dominion University, Norfolk, USA

Abstract

Corrugated skins provide a good solution to morphing wings due to their highly anisotropic behavior. If

the low stiffness corrugation plane is aligned in the chordwise direction the airfoil shape change is

possible. In contrast to the traditional smooth skin of an airfoil, a corrugated skin influences both the local

and global aerodynamics of a wing. The aim of this study is to investigate the effect of a corrugated skin

on the global aerodynamics of an airfoil, particularly lift and drag characteristics. First, the aerodynamic

analysis of a NACA 0012 airfoil with a smooth profile is conducted, both in the wind tunnel and by CFD,

at different Reynolds numbers, and compared to the data in the literature. The lift and drag coefficients at

different angles of attack, between -10° and 10°, are considered. Next, two NACA 0012 airfoils with

different sized corrugated profiles are investigated both experimentally and by numerical simulation. The

effect of corrugation size and Reynolds number are analyzed, quantified relatively to the standard

NACA0012 airfoil with a smooth skin. Preliminary numerical results are validated using the experimental

data.

1. INTRODUCTION

The concept of morphing aircraft is attracting a lot of attention as it could yield higher aerodynamic

performance than the conventional fixed wing aircraft. Although there is a significant progress in adaptive

structures for morphing wings [1], research on morphing skins [2] is relative behind. Corrugated laminates

offer a solution due to their extremely anisotropic behavior. Compliance in the chordwise direction, which

assumed to be the corrugation direction as well, allows shape change and increase in surface area. In

contrast, stiffness in the spanwise direction (transverse to the corrugation) enables the aerodynamic and

inertial loads to be carried. There are many papers on the estimation of the equivalent stiffness of

corrugated panels [3-12], however the aerodynamic investigation for corrugated skins is rarely found in

the literature, although the aerodynamic performance plays a key role in the application of corrugated

skins.

It is generally believed that the non-smooth surface is not suitable for an aerodynamic profile operating

in high Reynolds number as it has relatively poor aerodynamic performance, generating low lift and high

drag. However in nature some insects, dragonflies, damselflies, and others have corrugated wing profiles

[13] which operate in relatively low Reynolds numbers. Rees [14] indicated that at very low Reynolds

* Corresponding author, [email protected]

ICAST2012: 23rd

International Conference on Adaptive Structures and Technologies October 11-13, 2012, Nanjing, China

2

number around 103 to 10

4, a well-designed corrugated wing profile could reach a similar level of

aerodynamic performance compared to traditional streamlined airfoils.

Lissaman [15] reviewed different types of low Reynolds number airfoils and mentioned a critical

Reynolds (Re) number of about 70,000. Below this number, because of relatively large viscous effects, the

maximum lift-to-drag ratio of rough airfoils is higher than the conventional smooth profiles; however

above this value the roughness of the profiles plays an important role and smooth airfoils have higher

values of lift-to-drag ratio. Although there have been many experimental investigations [16-18] and

computational fluid dynamics (CFD) [19] simulations for the corrugated profiles, the majority of previous

studies were conducted at very low Reynolds numbers. Hu and Tamai [20] investigated the aerodynamic

performance of bio-inspired corrugated airfoils by using the (particle image velocimetry) PIV technique at

Re=34,000. Murphy and Hu [21] conducted further studies at Re number range of 58,000-125,000.

Whitehead et al. [22] and Thill et al. [23] performed wind-tunnel tests and CFD analysis of airfoils with

corrugated skins in the aft 1/3 of the chordwise section at Reynolds numbers between 250,000 and

1,000,000.

The aim of this study is to investigate the effect of a corrugated skin on the global aerodynamic

response of an airfoil, particularly on lift and drag coefficients. First, the aerodynamic analysis of a NACA

0012 airfoil with a smooth profile is conducted, both in the wind tunnel and by CFD, at different Reynolds

numbers, and compared to NACA0012 data in the literature. The lift and drag coefficients at different

angles of attack, between -10° and 10°, are considered. Next, two NACA 0012 airfoils with different sized

corrugated profiles are investigated both experimentally and by numerical simulation, at different

Reynolds numbers. The effect of corrugation size and Reynolds number are analyzed and compared to the

standard NACA0012 airfoil with a smooth surface.

2. RESEARCH PROCEDURE

2.1 Model

The NACA0012 airfoil is chosen as the base airfoil for this study. Two NACA 0012 airfoils with

different sized corrugated profiles are chosen for further analysis of the effect of the corrugated skin,

shown in Figure 1. The labels COR01 and COR02 are used to indicate the two corrugated profiles. These

models have a chord length c of 178 mm and round corrugation shapes with radius R=0.5%*c for COR01

and 1%*c for COR02.

COR01, R=0.5%*c COR02, 1%*c

Figure 1. Analysis models of NACA0012 airfoil with round corrugations

(Chord normalized airfoil dimentions)

ICAST2012: 23rd



3

2.2 Wind-tunnel Setup

The aerodynamic experiments were conducted in a low speed, open circuit, and closed test section

wind tunnel with octagonal cross test section[24, 25]. At the inlet, an aluminum honeycomb flow-

straightener and a fiberglass mesh is used to condition the flow. Flow velocity during the tests is observed

using four static ports at the inlet of the test section. The test section was converted to a 610 mm by 152

mm rectangular cross section for two-dimensional experiments with the use of two removable splitter

plates shown in Figure 2.

Figure 2. The wind tunnel experimental setup. Note that the upper and lower walls of the test-section

and fairing around the sting are omitted.

All parameters are controlled and measured automatically with a National Instruments (NI) cDAQ data

acquisition system and a personal computer. A total of 16 channels are monitored using four NI 9239 four

channel, isolated, 24-bit voltage input cards. The output signals are generated using two NI 9263 16-bit,

four channel voltage output cards. For each test point, a 20 second data is sampled at 100 Hz and then

averaged to get the mean value.

Barlow et al.[26] suggests several corrections due to the existence of the walls around the airfoil. The

solid blockage term and the wake blockage terms described by Rae et al.[26]. The wind tunnel

wall effects and buoyancy corrections were applied as necessary using the standard techniques found in

Barlow et al.[26]. The reported corrected lift and drag coefficients are calculated by:

( ) ⑴

( ) ⑵

The uncorrected lift and drag coefficients, and , are calculated by:

( ) ⑶

( ) ⑷

The streamwise turbulence of the flow in the empty test section is measured by a standard Hot Wire

Anemometry technique. The probe is placed at the centre of the test section (aligned approximately at the

quarter chord location along the streamwise direction) for all turbulence tests. After proper conversion of

the measured voltages to velocity (V), the turbulence intensity (TI) is calculated by:

√

∑ ( )

⑸

Rotary

Table

Load

Cell

Airfoil

Splitter

Plates

Moment

Arm

‘Sting’

Balance

Assembly

Base

ICAST2012: 23rd


4

where the index represents each sample. Turbulence intensity is measured at several velocities for

different filter settings. Based on the test results, the turbulence intensity of the wind tunnel, for the 2D

test section, is TI= 0.075%±0.01%, which is derived from 2.5 Hz-10 kHz band-pass filtered signal for the

velocity range of 5-30 m/s.

The experimental analysis was conducted at chord Reynolds numbers of around 120,000, 240,000 and

360,000. Each of the three profiles was measured over the angle of attack range from -10° to 10° with the

angle step of 0.5°.

2.3 Computational Procedure

The ANSYS/Fluent 14.0 software was used for the CFD analysis, and ANSYS/Gambit 2.4.6 software

was used for the mesh generation. A total of 553,990, 1,731,200 and 2,086,400 quadrilateral cells were

generated for the NACA0012, COR01 and COR02 models respectively. Sufficient numbers of cells were

used so that the mesh was fine enough to ensure convergence. The mesh in the near-wall area, which

needs to meet the y+ constraint based on the turbulence model, is generated within the boundary layer of

each model. Figure 3 shows the mesh for COR01.

As the Mach numbers in this study are low, incompressible flow is considered for the fluid model.

Furthermore, steady flow is considered. Velocity inlet and pressure outlet boundary conditions are used in

the far-field. The k-ε turbulence model with enhanced wall treatment was used to estimate the viscous

effect. The SIMPLE scheme was used to solve the pressure-velocity coupling. For the spatial

discretization, PRESTO! was used for pressure, and a second order upwind scheme is chosen for the

momentum equation. Only the 2-D solver was used for the analyses.

The calculations were performed at Reynolds numbers of around 120,000, 240,000 and 360,000, for

comparison with the experiment study. Each of the three profiles were simulated over the angle of attack

range from -10° to 10° with a step 1° for each Reynolds number.

Figure 3. Mesh of COR01 (1,731,200 cells)

ICAST2012: 23rd



5

3. RESULTS AND DISCUSSION

3.1 Validation

The experimental and computational results of the NACA 0012 are compared to the experimental

results published in the literature. As there is a limited amount of literature with aerodynamic data at the

Reynolds numbers, we use the drag and lift factor data in Ref. [27] at the Reynolds number of 1.0×106 for

comparison of the trend, which are higher than the test environment for the current study.

Figure 4. 2D lift and drag coefficient comparison of NACA0012 at different Re

Figure 5. Drag characteristics of NACA0012 at different Re

Figure 4 show the lift and drag characteristics of the NACA0012 from both wind tunnel tests and CFD

analyses. The experimental lift coefficient results are well predicted by the CFD results. The slope of the

lift coefficient curves (∂Cl/∂α) are predicted correctly in the test range and the lift-to-drag ratio increases

with increasing Re number. These trends matched with the literature [27].

3.2 The Effect of Corrugation Size

Figures 6(a)-8(a) present the lift coefficient response of the NACA0012, COR01, COR02 airfoils at

different Reynolds numbers. For each positive angle of attack, from smooth airfoil (NACA0012) to the

ICAST2012: 23rd


6

small-corrugation airfoil (COR01), then to the large-corrugation airfoil (COR02), the lift coefficient is

reduced. From these figures we conclude that the slope of the lift coefficients curve (∂Cl/∂α) decreases

with increasing the size of the corrugation for the range of corrugations that are examined. Figures 6(b)-

8(b) present the lift-drag ratio response of the NACA0012, COR01, COR02 airfoils at different Reynolds

numbers. For each positive angle of attack, the drag coefficient increases with the increased size of the

corrugation.

Loftin and Smith [27] mentioned that the minimum drag coefficient of the airfoil with roughed leading

edge is higher than the airfoil with smooth leading-edge. It is also reported in Ref. [27] that the lift-curve

slope of an airfoil with smooth surface is larger than the airfoil with roughness. In our study, as expected,

the corrugations reduced the lift-curve slope (∂Cl/∂α) and lift-drag ratio (∂Cl/∂Cd).

(a) lift coefficient (b) lift-drag ratio

Figure 6. 2D lift coefficient and lift-drag ratio comparison at Re=120,000


Figure 7. 2D lift coefficient and lift-drag ratio comparison at Re=240,000

ICAST2012: 23rd



7


Figure 8. 2D lift coefficient and lift-drag ratio comparison at 360,000

3.3 The Effect of Reynolds Number

Figures 9-10 present the comparison of the lift coefficient and lift-drag ratio for the COR01 and

COR02 airfoils at different Reynolds numbers. The results suggest that a slightly larger lift-curve slope

(∂Cl/∂α) is obtained by increasing the Reynolds number from 120,000 to 360,000. Unlike smooth airfoils,

the aerodynamic performance of corrugated airfoils at low angles of attack, quantified in terms of lift-to-

drag ratio, has tiny variation as the Reynolds number is increased.

(a)COR01 (b)COR02

Figure 9. Computational and experimental 2D lift coefficient comparison

ICAST2012: 23rd


8

(a)COR01 (b)COR02

Figure 10. 2D lift-drag ratio comparison of COR02

3.4 Flow Behavior of Corrugated Airfoils

Figure 11 shows the static pressure coefficient distribution around the corrugated airfoil (COR01) at an

angle of attack of 5° and at Re=240,000. Figure 12 illustrates the local streamlines around the

corrugations. These figures show that the local flow sustains an attached flow (outside the corrugation

troughs). The eddies fill the troughs of the corrugations and 'smooth' the shape of the corrugated structure

so that the flow outside the corrugation is similar to that around streamlined airfoils.

Figure 11. Static pressure coefficient distribution (COR01) at the angle of attack 5°, Re=240,000

ICAST2012: 23rd



9

Figure 12. Local streamlines for COR01 at the angle of attack 5°, Re=240,000

3. CONCLUSION

This article investigated the two-dimensional aerodynamic effect of chordwise corrugations in a two-

dimensional NACA 0012 airfoil in terms of lift and drag coefficients by using both experimental and

computational methods. It was found that the lift-curve slope (∂Cl/∂α) decreased and the minimum drag

coefficient (Cd0) increased with the increasing size of corrugation (i.e. the roughness of the skin). The

results for the corrugated airfoils suggest that a slightly larger lift-curve slope (∂Cl/∂α) is obtained by

increasing the Reynolds number from 120,000 to 360,000. Unlike smooth airfoils, the aerodynamic

performance of corrugated airfoils at low angles of attack, quantified in terms of lift-to-drag ratio, has tiny

variation as the Reynolds number is increased. The CFD flow field plots show that the local flow sustains

an attached flow (outside the corrugation troughs). The eddies fill the troughs of the corrugations and

'smooth' the shape of the corrugated structure so that the flow outside the corrugation is similar to that

around streamlined airfoils.

Based on this study, in-depth simulations and wind-tunnel experiments will be conducted to further

understand the effect of a) corrugation size and geometry; b) Reynolds number; c) turbulence intensity and

the choice of turbulence model; d) aerodynamic performance at large angles of attack.

Acknowledgements

The authors acknowledge funding from the European Research Council through Grant Number 247045

entitled Optimisation of Multiscale Structures with Applications to Morphing Aircraft.

References

1. Barbarino, S., Bilgen, O., Ajaj, R.M., et al., A Review of Morphing Aircraft. Journal of

Intelligent Material Systems and Structures, 2011. 22(9): p. 823-877.

2. Thill, C., Etches, J.A., Bond, I.P., et al., Morphing Skins - A Review. The Aeronautical

Journal, 2008. 112: p. 117-139.

3. Briassoulis, D., Equivalent Orthotropic Properties of Corrugated Sheets. Computers &

Structures, 1986. 23(2): p. 129-138.

ICAST2012: 23rd


10

4. Kress, G. and Winkler, M., Corrugated Laminate Homogenization Model. Composite

Structures, 2010. 92: p. 795-810.

5. Kress, G. and Winkler, M., Corrugated laminate analysis: A generalized plane-strain

problem. Composite Structures, 2011. 93(5): p. 1493-1504.

6. McFarland, D.E., An Investigation of the Static Stability of Corrugated Rectangular

Plates Loaded in Pure Shear. 1967, University of Kansas: Lawrence, KS.

7. Thill, C., Etches, J.A., Bond, I.P., et al., Composite Corrugated Structures for Morphing

Wing Skin Applications. Smart Materials and Structures, 2010. 19(124009).

8. Thill, C., Etches, J.A., Bond, I.P., et al., Investigation of trapezoidal corrugated

aramid/epoxy laminates under large tensile displacements transverse to the corrugation

direction. Composites Part a-Applied Science and Manufacturing, 2010. 41(1): p. 168-

176.

9. Thill, C., Etches, J.A., Bond, I.P., et al., Experimental and Parametric Analysis of

Corrugated Composite Structures for Morphing Skin Applications, in 19th International

Conference on Adaptive Structures Technology. 2008: Ascona, Switzerland.

10. Winkler, M. and Kress, G., Deformation limits for corrugated cross-ply laminates.

Composite Structures, 2010. 92(6): p. 1458-1468.

11. Xia, Y., Friswell, M.I., and Saavedra Flores, E.I., Equivalent models of corrugated

panels. International Journal of Solids and Structures, 2012. 49(13): p. 1453-1462.

12. Yokozeki, T., Takeda, S.-I., Ogasawara, T., et al., Mechanical Properties of Corrugated

Composites for Candidate Materials of Flexible Wing Strucures. Composites: Part A,

2006. 37: p. 1578-1586.

13. Rees, C.J.C., Form and function in corrugated insect wings. Nature, 1975. 256(5514): p.

200-203.

14. Rees, C.J.C., Aerodynamic properties of an insect wing section and a smooth aerofoil.

Nature, 1975. 258(5531): p. 141-142.

15. Lissaman, P.B.S., Low-Reynolds-Number airfoils. Annual Review of Fluid Mechanics,

1983. 15: p. 223-239.

16. Kesel, A.B., Aerodynamic characteristics of dragonfly wing sections compared with

technical aerofoils. Journal of Experimental Biology, 2000. 203(20): p. 3125-3135.

17. Sunada, S., Yasuda, T., Yasuda, K., et al., Comparison of wing characteristics at an

ultralow Reynolds number. Journal of Aircraft, 2002. 39(2): p. 331-338.

18. Thomas, A.L.R., Taylor, G.K., Srygley, R.B., et al., Dragonfly flight: free-flight and

tethered flow visualizations reveal a diverse array of unsteady lift-generating

mechanisms, controlled primarily via angle of attack. Journal of Experimental Biology,

2004. 207(24): p. 4299-4323.

19. Bourdin, P. and Friswell, M.I., Virtual Shape Definition of a Very Low Reynolds Number

Airfoil Using Vortex Shedding, in RAeS Conference on the Aerodynamics of Novel

Configurations: Capabilities and Future Requirements. 2008.

20. Hu, H. and Tamai, M., Bioinspired Corrugated Airfoil at Low Reynolds Numbers. Journal

of Aircraft, 2008. 45(6): p. 2068-2077.

21. Murphy, J.T. and Hu, H., An experimental study of a bio-inspired corrugated airfoil for

micro air vehicle applications. Experiments in Fluids, 2010. 49(2): p. 531-546.

22. Whitehead, D.S., Kodz, M., and Hield, P.M., Reduction of Drag by Corrugating Trailing

Edges. 1982, Cambridge University, Engineering Department.

23. Thill, C., Downsborough, J.D., Lai, S.J., et al., Aerodynamic study of corrugated skins for

morphing wing applications. Aeronautical Journal, 2010. 114(1154): p. 237-244.

ICAST2012: 23rd



11

24. Bilgen, O. and Friswell, M.I., Implementation of a Continuous-Inextensible-Surface

Piezocomposite Airfoil. Journal of Aircraft, 2012.

25. Bilgen, O., Saavedra-Flores, E.I., and Friswell, M.I., Implementation of a Continuous-

Inextensible-Surface Piezocomposite Airfoil, in 20th AIAA/ASME/AHS Adaptive

Structures Conference. 2012: Honolulu, HI.

26. Barlow, J.B., Pope, A., and Rae, W.H., Low speed wind tunnel testing. 3rd ed. ed. 1999,

New York ; Chichester: Wiley.

27. Loftin, L.K., Jr; and Smith, H., Aerodynamic characteristics of 15 NACA airfoil sections

at seven Reynolds numbers from 0.7 x 106 to 9.0 x 10

6. 1949.

the effect of corrugated skins on aerodynamic …michael.friswell.com/pdf_files/c331.pdfthe effect...

Documents