You are on page 1of 11

Acta Mech Sinica (2005) 21: 531541 DOI 10.

1007/s10409-005-0072-4

R E S E A R C H PA P E R

Guoyu Luo Mao Sun

The effects of corrugation and wing planform on the aerodynamic force production of sweeping model insect wings*

Received: 31 March 2005 / Revised: 25 July 2005 / Accepted: 26 July 2005 / Published online: 21 November 2005 Springer-Verlag 2005

Abstract The effects of corrugation and wing planform (shape and aspect ratio) on the aerodynamic force production of model insect wings in sweeping (rotating after an initial start) motion at Reynolds number 200 and 3500 at angle of attack 40 are investigated, using the method of computational uid dynamics. A representative wing corrugation is considered. Wing-shape and aspect ratio (AR) of ten representative insect wings are considered; they are the wings of fruit y, craney, droney, hovery, ladybird, bumblebee, honeybee, lacewing (forewing), hawkmoth and dragony (forewing), respectively (AR of these wings varies greatly, from 2.84 to 5.45). The following facts are shown. (1) The corrugated and at-plate wings produce approximately the same aerodynamic forces. This is because for a sweeping wing at large angle of attack, the length scale of the corrugation is much smaller than the size of the separated ow region or the size of the leading edge vortex (LEV). (2) The variation in wing shape can have considerable effects on the aerodynamic force; but it has only minor effects on the force coefcients when the velocity at r2 (the radius of the second moment of wing area) is used as the reference velocity; i.e. the force coefcients are almost unaffected by the variation in wing shape. (3) The effects of AR are remarkably small: when AR increases from 2.8 to 5.5, the force coefcients vary only slightly; oweld results show that when AR is relatively large, the part of the LEV on the outer part of the wings sheds during the sweeping motion. As AR is increased, on one hand, the force coefcients will be increased due to the reduction of 3-dimensional ow effects; on the other hand, they will be decreased due to the shedding of part of the LEV; these
* The project supported by the National Natural Science Foundation of China (10232010 and 10472008) and Ph. D. Student Foundation of Chinese Ministry of Education (20030006022) The English text was polished by Keren Wang. G. Y. Luo M. Sun (B) Ministry-of-Education Key Laboratory of Fluid Mechanics, Institute of Fluid Mechanics, Beihang University, Beijing 100083, China Tel.: 86-10-82313027 E-mail: m.sun@263.net

two effects approximately cancel each other, resulting in only minor change of the force coefcients. Keywords Insect ight Sweeping wing Unsteady aerodynamics Wing corrugation Planform

1 Introduction High aerodynamic force coefcients are required for hovering and low speed ying insects [1]. Experiments and computations simulating hawkmoth [2,3] and fruit y [4,5] hovering demonstrated that the required high force coefcients could be produced. Further computations on apping model fruit y wing [6] showed that high force coefcients could be produced when Reynolds number (Re) is above approximately 100. In the above studies [26], the model wings were atplates and only the wing planforms of hawkmoth and fruit y were considered. The chordal prole of the wings of many orders of insects is corrugated and their wing planform (especially the aspect ratio AR) varies greatly (in the present study, AR is dened as R/c, where R and c are the wing length and mean chord length, respectively). It is of interest to know whether or not the corrugation and the variation in planform have signicant inuence on the aerodynamic force production. Some experiments on the effects of wing corrugation under steady ow conditions have been reported. Rees [7] tested a corrugated wing (a model hovery wing) and a smooth wing (the section of the smooth wing is formed by drawing a smooth envelope through the corner points of the corrugated section). Lift and drag were measured at Reynolds number of 450, 800 and 900. He found little difference between the smooth and corrugated models. His ow visualization test showed that the ow over the corrugated wing seemed to make some uid trapped in the folds, either being stagnant or recirculating, and the main ow seemed to behave as if the folds were solidly inlled. Recently, Kesel [8] tested various corrugated sectional proles with Re = 7 880 and 104 . The

532

G. Y. Luo, M. Sun

proles were constructed using measurements taken from the cross sections at different span location of a dragony forewing. His results showed that with no camber in the corrugated proles, the lift and drag coefcients were close to those of the at-plate wing. The above results are for xed wings under steady ow conditions and are instructive for understanding the gliding ight. But so far, the investigation on the effects of corrugations for wings in apping motion has not been reported. As for the effects of wing planform, Usherwood and Ellington [9,10] conducted force measurement experiments on revolving model wings with wing planforms of bumblebee, hawkmoth and mayy and with planforms based on hawkmoth wing adjusted to AR varying from 2.3 to 7.9. They showed that the aerodynamic force coefcients for the above three wings are not very different and that the effects of considerable variation in AR were relatively small. With no oweld information available in their experiments, some facts concerning the aerodynamic forces (e.g. the force coefcients changed only a little when AR was increased by about three times) were not explained. More investigations are needed to account for the relatively small effects of AR. It is also desirable to show what effects will be produced when the wing shape is varied with AR and R (wing length) xed. In the present study, we address the above questions by studying the effects of corrugation, wing shape and aspect ratio for wings that rotate azimuthally by 160 at a constantspeed and with a constant angle of attack after an initial acceleration, using the method of computational uid dynamics (CFD). The reason for considering this motion is that it can, to a great extent, represent a downstroke (or upstroke) of the apping motion of insect wings. Moreover, this motion is simple and it is easier to identify the ow mechanism and its major causes. The CFD method can provide forces and ows simultaneously. The effects of wing corrugation are investigated by comparing the aerodynamic forces and ows between corrugated and at-plate wings at various Re (care has been taken to ensure that the corrugated wing has no camber). In order to study the effects of wing shape, ten model wings of various wing shape with the same AR and R are made based on the wing planforms of ten species of insects (from ve orders), i.e., the planforms of the ten insect wings are adjusted to the same AR and R. In order to study the effects of AR, a series of model wings are made based on one insect wing (the hawkmoth wing) adjusted to various AR values. To see the combined effects of wing shape and AR, model wings with the planforms of the above ten insects are also used. 2 The methods 2.1 The model wings 2.1.1 The at-plate and corrugated model wings For studying the effects of wing corrugation, we used two model wings. One is a at-plate wing. Its planform (Fig. 1(a))

Fig. 1 (a) The planform of model fruit y wing; (b) The corrugated section and the corresponding sectional grids

is based on the fruit y wing [11] with a thickness of 0.03c1 , where c1 is the chord length. The other wing has the same planform and thickness, but its chordal prole is corrugated. Rees study on the sectional shapes of wings of some insects (two dipterans and a hymenopteran) [12] shows that it is reasonable to model the corrugation by triangular waves. Here, the corrugation is modeled by triangular waves between the leading edge and a point 0.6c1 from the leading edge (Fig. 1(b)). In studying the structural properties of insect wings, Rees [12] also modeled the corrugations as triangular waves. He measured the wavelengths and amplitudes of the corrugations of seven species of insect (see table 1 of Ref. [12]), and the average values of the wavelength and the amplitude are around 0.2c1 and 0.03c1 , respectively. In the present study, we set the wavelength and the amplitude of the triangular wave as 0.2c1 and 0.05c1 , respectively. 2.1.2 Model wings with various planforms Three groups of model wings are used for studying the effects of wing shape; they are at-plate wings with various
Table 1 The geometry parameters of the model insect wings Insect hawkmoth fruit y bumble bee honey bee ladybird droney lacewing hovery craney dragony ID HM FF BB HB LB DF LW HF CF DRF R/mm 51.9 2.02 13.2 9.8 11.2 11.4 14 9.3 12.7 47.4 c/mm 18.26 0.67 4.16 3.07 3.22 3.24 3.85 2.24 2.37 8.7 R/c 2.84 3.0 3.17 3.19 3.47 3.52 3.63 4.15 5.35 5.45 r2 /R 0.53 0.56 0.56 0.57 0.54 0.54 0.59 0.58 0.61 0.58

The effects of corrugation and wing planform on the aerodynamic force production of sweeping model insect wings

533

planforms (with section thickness of 0.03c1 ). One group includes ten model wings, with planforms based on the wings of fruit y (FF), craney (CF), droney (DF), hovery (HF), ladybird (LB), bumblebee (BB), honeybee (HB), lacewing (LW; forewing), hawkmoth (HM), and dragony (DRF; forewing). These planforms are shown in Fig. 2. The planforms of fruit y and dragony are taken from Zanker [11] and Norberg [13], respectively; the planforms of the other eight insects are taken from Ellington [14]. The geometry parameters for these model wings are given in Table 1; they are the wing length R, the mean chord length of the wing c, and the radius of the second moments of wing area r2 , dened as:
2 r2 =

FF wing, with AR of 3; the other nine are based on the wings of CF, DF, HF, LB, BB, HB, LW, HM and DRF adjusted to AR of 3, which are denoted by CF , DF , HF , LB , BB , HB , LW , HM , DRF , respectively. The third group includes four model wings with the same shape but different AR (Fig. 4); they have the shape of hawkmoth wing with their AR adjusted to vary from 2.8 to 5.5 (with r2 /R xed). 2.2 The wing motion The motion is sketched in Fig. 5. OXYZ is an inertial system (xed on the body of an insect in hovering or constantspeed ight) and oxyz is a non-inertial coordinate system xed on the sweeping wing. Since the aerodynamic forces produced by the wings vary periodically, one may expect that the insect body would oscillate periodically and OXYZ would not be an inertial frame. However observations [15] showed that for many insects, the body oscillation is negligibly small. This may be because the wing aps very fast compared to the response time of the body; a recent study [16] on the ight stability of a bumblebee showed that this is true for the bumblebee. Only for insects having very low frequency, such as butteries, the body oscillation is large. The wings sweep (rotate) about Z-axis at constant angle of attack with constant speed after an initial acceleration from rest. This motion resembles the downstroke (or upstroke) of a apping insect-wing. Let denote the azimuthal angle,

1 S

r 2 dS =

1 S

R 0

r 2 c1 dr,

(1)

where r is the radial distance, S is the wing area, and c1 is the local chord length of the wing. AR of these planforms ranges from 2.84 (hawkmoth wing) to 5.45 (dragony forewing). Another group includes ten model wings with the same AR (AR = 3) and R but different shape (Fig. 3). One is the

Fig. 2 The planforms of ten model insect wings

Fig. 4 Model hawkmoth planforms with a range of AR

Fig. 3 The planforms of model wings with aspect ratio (AR) adjusted to 3

Fig. 5 Sketches of the reference frames and wing motion

534

G. Y. Luo, M. Sun

the rotational velocity and 0 the constant-rotational velocity after the initial acceleration. In the initial acceleration, is given as + (2) + = 0.50 [1 cos(/a )], (0 a ) + where + = c/U , 0 = 0 c/U , = tU/c, and a is the non-dimensional time taken for accelerating the wing from + rest to 0 . The reference velocity is taken as U = r2 0 . Note that because of the above choice of reference velocity U , + the non-dimensional rotational velocity of the wing (0 ) is equal to c/r2 , which is specied when the wing geometry is given. As a result, in prescribing the sweeping motion of a wing, two parameters, the geometric angle of attack () and the non-dimensional initial acceleration time (a ) need to be specied. 2.3 The ow equations and the solution method

2.4 Non-dimensional parameters that affect the aerodynamic force coefcients For a wing of given geometry, when its sweeping motion is prescribed, the solution of the non-dimensional NavierStokes equations (equations (3)-(6)) gives the aerodynamic force coefcients CL and CD ; Re is the only non-dimensional parameter in the Navier-Stokes equations that needs to be specied. To prescribe the sweeping motion, as mentioned above, and a need to be specied. Therefore, the aerodynamic force coefcients on a sweeping wing depend on the wing geometry and on three non-dimensional parameters, Re, and a . In the present study, and a are xed at some typical values [15]: a is set such that the wing rotates 20 during the initial acceleration phase and in most of the computed cases, is set as 40 ; the wing geometry (sectional shape, wing shape, aspect ratio) and Re are varied (the effect of various wing length and different frequency is represented in Re, since U is proportional to wing length and frequency).

The governing equations of the ow are the three-dimensional incompressible unsteady Navier-Stokes equations. In the inertial coordinate system OXYZ, after being non-dimensionalized, 2.5 Code validation and grid resolution test they are as follows: The code used in this study is the same as that in Sun and v w u + + = 0, (3) Tang [5]. It was tested by measured unsteady aerodynamic X Y Z forces on a apping model fruit y wing [17]. The calculated drag coefcient agrees well with the measured value (see u u u u Fig. 2A and Fig. 2C of Sun and Wu [17]); for the lift coef+u +v +w X Y Z cient, in the translation phase during the middle, and in the 1 2u 2u 2u p rotation phase at the end, of each half-stroke, the computed + + + , (4) value agrees well with the measured value, whereas in the = X Re X 2 Y 2 Z 2 beginning of the stroke, the computed peak value is much v v v v smaller than the measured value (see Fig. 2B and Fig. 2D +u +v +w of Sun and Wu [17]). The discrepancy might be because the X Y Z 2 2 2 CFD code does not perform satisfactorily for the complex 1 v v v p + + + , (5) ow at the stroke reversal. There is also the possibility that = Y Re X 2 Y 2 Z 2 it is due to variations in the precise kinematic patterns, especially at the stroke reversal. Wu and Sun [6] made a further w w w w +u +v +w test of the code using the recent experimental data of Ush X Y Z erwood and Ellington [9,10] on revolving model wings. In 1 2w 2w 2w p + + + , (6) the whole range (from 20 to 100 ), the computed CL = Z Re X 2 Y 2 Z 2 agrees well with the measured values; both have approxiwhere u, v and w are three components of the non-dimen- mately sinusoidal dependence on . The computed CD also sional uid velocity and p is the non-dimensional uid pres- agrees well with the measured values except when is larger sure. In the non-dimensionalization, U , c, and c/U are taken than approximately 60 . The above comparisons show that as the reference velocity, length and time, respectively. Re in there still exist some discrepancies between the CFD simEqs.(4)-(6) is dened as Re = cU/, where is the kine- ulations and the experiments, but in general, the agreement matic viscosity of the uid. The numerical method used to between the computational and experimental aerodynamic solve equations (3)-(6) is the same as that in Sun and Tang forces is good. We think that the present CFD method can calculate the unsteady aerodynamic forces and ows of the [5]. Once the Navier-Stokes equations are numerically solved, model insect wings with reasonable accuracy. The effects of the grid density, the time step and computhe uid velocity components and pressure at discretized grid points for each time step are obtained. The aerodynamic tational-domain size on the computed solutions are considforces (lift L and drag D) acting on the wing are calculated ered. For the at-plate model fruit-y wing, the sensitivity from the pressure and the viscous stress on the wing surface. of the computed ow to the spatial and time resolution and The lift (CL ) and drag (CD ) coefcients are dened as fol- to the far-eld boundary location is evaluated for the case of lows: CL = L/0.5U 2 S, CD = D/0.5U 2 S, where is the Re = 3 400 (this Re is the highest among the cases considuid density and S is the wing area. ered in this study). Calculations were performed using three

The effects of corrugation and wing planform on the aerodynamic force production of sweeping model insect wings

535

different grid systems. Grid 1 had dimensions of 534841 (around the wing section, in the normal direction and in the spanwise direction, respectively), grids 2 and 3 had dimensions of 777061 and 1099378, respectively. The spacing at the wall was 0.003, 0.002, and 0.001 5 for grids 1, 2, and 3, respectively. The far-eld boundary for these three grids was set at 20c away from the wing surface in the normal direction and 8 chord lengths away from the wing-tips in the spanwise direction. The grid points were arranged densely toward the wing surface and toward the wake. The differences between the computed lift coefcients using the three grids are small (there is almost no difference between the lift coefcients computed using grid 2 and grid 3). Computations using grid 2 and two time step values, = 0.02 and 0.01, were conducted. Discrepancies between the computed aerodynamic forces and vorticity elds using the two time steps are very small. Finally, the sensitivity of the solution to the far-eld boundary location was considered by calculating the ow in a large computational domain. In order to isolate the effect of the far-eld boundary location, the boundary was made farther away from the wing by adding more grid points to the normal direction of grid 2. The calculated results show that it is not necessary to put the far-eld boundary farther than that of grid 2. From the above analysis, it is concluded that grid 2 and a time step value of = 0.02 are appropriate for the at-plate model fruit y wing. For at-plate model wings with larger AR than the fruit y wing, more grid points are used in the spanwise direction such that the grid density in the spanwise direction is the same as that of the model fruit y wing. For the corrugated wing, the grid points around the wing section are increased (from 77 to 116 points) to better represent the geometry of the folds. 3 Results and discussion 3.1 The effects of wing corrugation CL and CD vs. (sweeping angle) for the at-plate and corrugated wings at = 40 and at Re = 200 are given in Fig. 6 ( varies linearly with time except during the initial phase). CL and CD of the corrugated wing (at = 40 ) are almost the same as their counterparts of the at-plate wing. Figure 7 gives the surface pressure and surface viscous stress distributions at half-wing length at = 150 for the wings (the viscous stress is more than one order of magnitude smaller than the pressure and the aerodynamic force on the wing is mainly due to the surface pressure). It is seen that in the region of the folds of the corrugated wing, the surface pressure oscillates around that of the at-plate wing, showing that compared to the at-plate wing, the corrugation only produces some local change in surface pressure but does not affect the total force. Figure 8 shows the contour plots of the spanwise component of vorticity at half-wing length at two time instances (i.e. at two sweeping angler positions) for the two wings. It is seen that the ow is separated, the vorticity elds around the corrugated wing are similar to their counterparts of the

Fig. 6 Lift and drag coefcients vs. sweeping angle (Re = 200)

Fig. 7 Non-dimensional pressure (a) and sectional component of nondimensional viscous stress (b) on upper and lower surface at half-wing length for at-plate and corrugated wings at = 150 ( = 40 ; Re = 200)

at-plate wing, except in the folds of the corrugation, showing that the main ow of the corrugated wing behaves as if the folds do not exist. The negative vorticity (dotted lines) shed from the trailing edge of the corrugated wing (the starting vortex) is almost identical to that of the at-plate wing (see plots at = 20 in Fig. 8; the trailing-edge vortex at = 150 is far from the wing and is out of the plots). Since the total vorticity is conserved, this indicates that the corrugated and at-plate wings have approximately the same total positive vorticity around the wing. On the basis of the relation between the aerodynamic force and the vorticity moment [18], the two wings will have approximately the same aerodynamic force. We thus see that because of the ow separation at the large and/or low Re, the size of the vorticity region is much larger than that of the corrugation, resulting in the rather small effects of wing corrugation. Insects do not y with small angle of attack. But for reference, we compute a case of small ( = 5 ). CL and CD

536

G. Y. Luo, M. Sun

Fig. 8 Vorticity plots at half-wing length for at-plate (left) and corrugated (right) wings. = 40 ; Re = 200. Solid and broken lines indicate positive and negative vorticity, respectively. The magnitude of non-dimensional vorticity at the outer contour is 1 and the contour interval is 3

Fig. 9 Vorticity plots at half-wing length for at-plate (left) and corrugated (right) wings. = 5 ; Re = 200. Solid and broken lines indicate positive and negative vorticity, respectively. The magnitude of non-dimensional vorticity at the outer contour is 1 and the contour interval is 3

for = 5 are also given in Fig. 6. At this small , CL of the corrugated wing is also almost the same as that of the at-plate wing, but CD is about 17% larger than that of the at-plate wing. The corresponding contour plots of the spanwise component of vorticity are given in Fig. 9. It is seen that although the ow is not separated, due to the low Re, viscous layers on wing surfaces are very thick and the main ow of the corrugated wing is quite similar to that of the at-plate wing, except that the viscous region of the corrugated wing is a little thicker than that of the at-plate wing. The physical reason for the slightly larger drag of the corrugated wing is that the corrugation increases the thickness of the viscous region and a thicker viscous region corresponds to a larger drag. Since the corrugation is symmetric (no camber is introduced by the corrugation), the thicknesses of viscous layer on the upper and lower wing surface are increased to approximately the same extent. As a result, the lift is not increased by the corrugation.

Fig. 10 Lift and drag coefcients vs. sweeping angle (Re = 3 500)

The effects of corrugation and wing planform on the aerodynamic force production of sweeping model insect wings

537

Fig. 11 Vorticity plots at half-wing length for at-plate (left) and corrugated (right) wings. = 40 ; Re = 3 500. Solid and broken lines indicate positive and negative vorticity, respectively. The magnitude of non-dimensional vorticity at the outer contour is 1 and the contour interval is 3

Fig. 12 Vorticity plots at half-wing length for at-plate (left) and corrugated (right) wings. = 5 ; Re = 3 500. Solid and broken lines indicate positive and negative vorticity, respectively. The magnitude of non-dimensional vorticity at the outer contour is 1 and the contour interval is 3

The above analysis is for the case of Re = 200. Results for a higher Re (Re = 3 500) are shown in Figs. 1012. The above analysis for the case of Re = 200 also applies to the case of Re = 3 500. It is well known that the corrugation of an insect wing is related to the advantages of low mass, high stiffness and low membrane stress [12]. The results here show that the corrugation of an insect wing, which usually operates at large angle of attack, has no obvious aerodynamic shortcomings. That is, the corrugation brings the above advantages without producing negative aerodynamic effect. The present results also show that when investigating the aerodynamic forces of apping insect wings by experiment or by computational simulation, a at-plate wing could be employed. 3.2 Wings of various shapes (but with the same AR and R) We rst study the model wings shown in Fig. 3, which have different shapes but with the same AR and R, hence the same
Fig. 13 Non-dimensional lift and drag vs. for model insect wing with various shape but same AR (velocity at wing tip is used as reference velocity, which is same for all wings) ( = 40 ; Re 200)

538

G. Y. Luo, M. Sun

Fig. 14 Lift and drag coefcients vs. sweeping angle for model wings with various shape but same AR (velocity at r2 is used as reference velocity, which is different for different wing) ( = 40 ; Re 200)

Fig. 16 Lift and drag coefcients vs. sweeping angle for model hawkmoth wings with a range of AR ( = 40 ; Re = 3 500)
r2 , its CL and CD are larger. The physical reason for this is clear. For a wing with a given wing-area and wing-length, increasing r2 means that a larger part of the area is moved to the outer part of the wing (see equation (1)). For a rotating wing, the relative velocity at the outer part of the wing is larger than that at the inner part. Therefore, when r2 is larger, a larger part of the wing area would have larger aerodynamic forces. In propeller aerodynamics, the velocity at r2 is used in the denition of force coefcients, which is based on the bladeelement theory. In this theory, it is assumed that the lift and drag on a section at r (denoted as dL and dD, respectively) are proportional to the square of the relative velocity (0 r) and the area (c1 dr) of the section: dL 0.5(0 r)2 c1 dr, (7) 0 r)2 c1 dr. dD 0.5( (8)

Fig. 15 Lift and drag coefcients vs. sweeping angle for model hawkmoth wings with a range of AR ( = 40 ; Re = 200)

Then, one has 2 L 0.5 0 2 D 0.5 0


R 0 R 0

area (the effects of AR are studied below). We let the wings rotate (sweep) at the same angular speed. Since the wings have the same R and AR and rotate at the same speed, any differences in the aerodynamic forces among them must be due to the wing shape variation. Figure 13 gives the non-dimensional lift (CL ) and drag (CD ). In the non-dimensionalization here, the velocity at the wing tip is used as the reference velocity, which is the same for all the wings in this section. It is seen that a signicant difference in CL and in CD is produced by the variation in the wing shape. The HM wing has the smallest CL and CD (about 0.45 for CL and about 0.38 for CD ); the CF wing has the largest CL and CD (about 0.6 for CL and about 0.48 for CD ). The CL of CF wing is 33% larger than that of the HM wing. We thus see that the shape of a wing has a great effect on its aerodynamic forces. From the wing geometrical data in Table 1 and the CL and CD results in Fig. 13, we see that for a wing with a larger

r 2 c1 dr, r 2 c1 dr.

(9) (10)

2 The integrals in the above equations are equal to r2 S (see equation (1)), therefore, L and D of the wing (or blade) are proportional to 0.5(0 r2 )2 S. That is, when the velocity at r2 is used as the reference velocity, CL and CD will not be very much affected by the wing shape (the effect of wing shape is included in the denition of r2 ). In the denition of force coefcients of insect apping or rotating wings, following the practice in propeller aerodynamics, the velocity at r2 is also used as the reference velocity. It is reasonable to expect that when the velocity at r2 is used as the reference velocity in the denition of force coefcients, the force coefcients will be much less affected by the variation of wing shape. Figure 14 shows CL and CD

The effects of corrugation and wing planform on the aerodynamic force production of sweeping model insect wings

539

Fig. 17 Vorticity plots at two spanwise locations at = 60 for model hawkmoth wings of various AR. ra and rb denote locations 0.4R and 0.6R from the wing root, respectively. = 40 ; Re = 3 500. Solid and broken lines indicate positive and negative vorticity, respectively. The magnitude of non-dimensional vorticity at the outer contour is 1 and the contour interval is 3

of the ten model wings when the velocity at r2 of each wing is used as the reference velocity. It is seen that the differences in CL and CD among these wings are very small; the CL and CD of a wing is different from that of another wing by less than 5%. (The above computations are for Re = 200; computations for Re = 3 500 have also been conducted and similar results are obtained.) From the above results, it can be stated that the variation in wing shape has considerable effects on the aerodynamic

force of a wing, but it has only minor effects on the aerodynamic force coefcients when the velocity at r2 of the wing is used as the reference velocity. 3.3 Wings of various AR (but with the same shape) The planform of a wing is, to a great extent, characterized by the aspect ratio (AR). For the ten insect wings shown in

540

G. Y. Luo, M. Sun

Fig. 20 Non-dimensional lift distribution along wing length at = 60 ( = 40 ; (a) Re = 200 and (b) Re = 3 500) Fig. 18 Instantaneous streamline plots for model hawkmoth wings of various AR at = 60 ( = 40 ; Re = 3 500)

Fig. 21 Lift and drag coefcients vs. sweeping angle for ten model insect wings. HM, hawkmoth; FF, fruit y; BB, bumblebee; HB, honeybee; LB, ladybird; DF, droney; LW, lacewing (forewing); HF, hovery; CF, craney; DRF, dragony (forewing) (Re = 200) Fig. 19 Pressure contours on the upper surface of the model hawkmoth wings of various AR at = 60 ( = 40 ; Re = 3 500)

Fig. 2, AR varies considerably, from 2.84 (hawkmoth wing) to 5.45 (dragony forewing), a 92% of difference. In order to isolate the effects of AR, we use planforms based on hawkmoth wings adjusted to AR values of 2.8, 3.5, 4.5 and 5.5 (with r2 /R unchanged, see Fig. 4). Figure 15 shows CL and CD vs. for these model wings at Re = 200 and Fig. 16 shows the results at Re = 3 500.

It is seen that when AR changes from 2.8 to 5.5 (nearly doubled), the change in force coefcients, on the average, is less than 10%. When AR of a wing is increased, one would expect that the three-dimensional (3D) ow effects would be reduced and the force coefcients would increase. But the results here show that the force coefcients make very small change for considerable variations of AR. This is remarkable. Here we explain why.

The effects of corrugation and wing planform on the aerodynamic force production of sweeping model insect wings

541

Figure 17 shows the contour plots of the spanwise component of vorticity at two spanwise locations at the middle of the stroke ( = 60 ) for wings of various AR at Re = 3 500. The two spanwise locations, denoted by ra and rb , are 0.4R and 0.6R from the wing root, respectively. When AR is relatively large (i.e. AR=4.5 and 5.5, Figs. 17 (c) and (d)), the vorticity contours on the upper surface at rb tends to move downstream from the wing (Figs. 17 (c), (d)), which indicates that a part of the LEV, which is at the outer part of the wing, is shed. This can also be seen in the plots of instantaneous streamlines shown in Fig. 18. As a result, the aerodynamic force produced by the outer part of the wing of large AR will be reduced, because the suction pressure due to the LEV there disappears. Figure 19 shows the pressure contours on the upper surface of the wings of various AR at the middle of the stroke. When AR increases, the region of minimum pressure on the upper surface moves inward, reducing the suction force of the outer part of the wing. The above is also true for the cases of Re = 200. These ow eld and surface pressure data suggest that the partial shedding of the LEV may explain why the force coefcient does not increase with AR. That is, when AR is increased, on one hand, the aerodynamic force coefcients are increased due to the reduction of 3D ow effects, and on the other hand, they are decreased due to the shedding of the part of the LEV at the outer part of the wing; these two effects may approximately cancel each other, resulting in only minor change of the force coefcients. This explanation is supported by the data in Fig. 20, in which the non-dimensional lift distributions along the wing length are shown. In the inner part of the wing (r/R 00.55), the lift coefcient for AR = 5.5 is larger than that for AR = 2.8, because the 3D ow effect for AR = 5.5 is smaller. But in the outer part of the wing (r/R 0.551), the lift coefcient for AR = 5.5 is smaller than that for AR = 2.8, because of the shedding of the outer part of the LEV. 3.4 The model insect wings The ten model insect wings shown in Fig. 2 are considerably different both in shape and in AR. The effects of shape and AR have been considered in sections 3.3 and 3.4, respectively. On the basis of the results in those two sections, it is not unreasonable to expect that the aerodynamic force coefcients (using the velocity at r2 as the reference velocity) for these wings would not be very different. Figure 21 shows CL and CD vs. for the ten model insect wings at Re = 200. CL and CD vary only slightly among these wings, as is expected. This is also true for the case of Re = 3 500. 4 Conclusions (1) The corrugated and at-plate wings produce approximately the same aerodynamic forces. This is because for a sweeping wing at a large angle of attack, the length scale of the corrugation is much smaller than the size of

the separated ow region or the size of the leading edge vortex (LEV). (2) The variation in wing shape can have considerable effects on the aerodynamic force; but it has only minor effects on the force coefcients when the velocity at r2 (the radius of the second moment of wing area) is used as the reference velocity; i.e. the force coefcients are almost unaffected by the variation in wing shape. (3) The effects of AR are remarkably small: when AR increases from 2.8 to 5.5, the force coefcients vary only slightly; the oweld results show that when AR is relatively large, the part of the LEV on the outer part of the wings sheds during the sweeping motion. As AR is increased, on one hand, the force coefcients will be increased due to the reduction of 3-dimensional ow effects; on the other hand, they will be decreased due to the shedding of part of the LEV; these two effects approximately cancel each other, resulting in only minor change of the force coefcients. References
1. Ellington, C.P.: The aerodynamics of hovering insect ight. I. Quasi-steady analysis. Phil. Trans. R. Soc. Lond. B, 305, 115 (1984) 2. Ellington, C.P., Van Den Berg, C., Willmott, A.P.: Leading edge vortices in insect ight. Nature 384, 626630 (1996) 3. Liu, H., Ellington, C.P., Kawachi, K., Van Den Berg, C., Willmott, A.P.: A computational uid dynamic study of hawkmoth hovering. J. Exp. Biol. 201, 461477 (1998) 4. Dickinson, M.H., Lehman, F.O., Sane, S.P.: Wing rotation and the aerodynamic basis of insect ight. Science 284, 19541960 (1999) 5. Sun, M., Tang, J.: Unsteady aerodynamic force generation by a model fruit y wing in apping motion. J. Exp. Biol. 205 5570 (2002) 6. Wu, J.H., Sun, M.: Unsteady aerodynamic forces of a apping wing. J. Exp. Biol. 207, 11371150 (2004) 7. Rees, C.J.C.: Aerodynamic properties of an insect wing section and a smooth aerofoil compared. Nature 258, 141142 (1975) 8. Kesel, A.B.: Aerodynamic characteristics of dragony wing sections compared with technical aerofoils. J. Exp. Biol. 203, 3125 3135 (2000) 9. Usherwood, J.R., Ellington, C.P.: The aerodynamics of revolving wings. I. Model hawkmoth wings. J. Exp. Biol. 205, 15471564 (2002) 10. Usherwood, J.R., Ellington, C.P.: The aerodynamics of revolving wings. II. Propeller force coefcients from mayy to quail. J. Exp. Biol. 205, 15651576 (2002) 11. Zanker, J.M.: The wing beat of Drosophila melanogaster. I. Kinematics. Phil. Trans. R. Soc. Lond. B 327, 118 (1990) 12. Rees, C.J.C.: Form and function in corrugated insect wings. Nature 256, 200203 (1975) 13. Norberg, RA: The pterostigma of insect wings an inertial regulator of wing pitch. J. Comp. Physiol. 81, 922 (1972) 14. Ellington, C.P.: The aerodynamics of hovering insect ight. II. Morphological parameters. Phil. Trans. R. Soc. Lond. B 305, 1740 (1984) 15. Ellington, C.P.: The aerodynamics of hovering insect ight. III. Kinematics. Phil. Trans. R. Soc. Lond. B 305, 4178 (1984) 16. Sun, M., Xiong, Y.: Dynamic ight stability of a hovering bumblebee. J. Exp. Biol. 208, 447459 (2005) 17. Sun, M., Wu, J.H.: Aerodynamic force generation and power requirements in forward ight in a fruit y with modeled wing motion. J. Exp. Biol. 206, 30653083 (2003) 18. Wu, J.C.: Theory for aerodynamic force and moment in viscous ow. AIAA Journal 19, 432441 (1981)

You might also like