4.

ANKARA INTERNATIONAL AEROSPACE CONFERENCE

27-29 September, 2007 - METU, Ankara

AIAC-2007-109

AEROELASTIC ANALYSIS OF A TAPERED AIRCRAFT WING

S. Durmaz1, O. Ozdemir Ozgumus 2, M. O. Kaya 3
Istanbul Technical University, Faculty of Aeronautics and Astronautics, 34469, Maslak, Istanbul, Turkey

ABSTRACT
In this study, aeroelastic flutter analysis of a tapered aircraft wing that is modeled as an Euler-Bernoulli
beam is carried out. Applying the Hamiltons principle to the kinetic and the potential energy expressions, the
governing differential equations of motion and the boundary conditions are obtained. The desired results are
obtained by applying both one-term and two-term Galerkin Method. However, only the two-term Galerkin
Method is introduced here since it already covers the one-term method. Additionally, an example that studies
the Goland wing is taken from open literature in order to validate the calculated results and a very good
agreement between the results is observed.
WING MODEL
In this study, a tapered aircraft wing that is given in Fig.1 is considered. Here, the x -axis coincides with
the elastic axis, which is assumed to be straight, and the distance between the elastic axis and inertia axis at
any point x is denoted by y x . The speed of the air flow relative to the wing denoted by U is assumed
to be constant.

Elastic Axis

Inertia Axis

Figure 1. Elastic axis and inertia axis for a tapered cantilever aircraft wing in steady air flow
The wing deformation can be measured by a bending deflection w in the z -direction and a rotation
about the elastic axis. The angle is referred to the local angle of attack. The bending deflection, w , is
positive downward while the rotation, , is positive if the leading edge is up. The chordwise distortion will be
neglected. The combined bending and rotational vibration of the cantilever wing in steady air flow is shown in
Figs.2 (a) and 2 (b). Here the chord length is given by c x 2b x .

Graduated Student in Aeronautics Department, E-mail: durmazseher@yahoo.com

Research Assistant in Aeronautics Department, E-mail: ozdemirozg@itu.edu.tr
3
Associate Professor in Aeronautics Department, E-mail: kayam@itu.edu.tr
2

AIAC-2007-000

Durmaz, Ozdemir Ozgumus, Kaya

x w

w
y

(a)

yc

(b)

Figure 2. (a) Bending deflection of the elastic axis (b) Rotation of the wing cross section
FORMULATION
Equations of Motion and the Boundary Conditions
The kinetic and the potential energy expressions of an Euler-Bernoulli Beam that experience bendingtorsion coupling are taken from Ref.[4]. Applying the Hamiltons principle to these energy expressions, the
governing differential equations of motion and the boundary conditions are obtained as follows
Equations of Motion:

2
x 2

EI

2w
2w
2

my
L 0

x 2
t 2
t 2

(1a)

2w
2
GJ

my

I
M 0

x
x
t 2
t 2

where the lift force,

(1b)

and the aerodynamic moment,

1
L' 2 UbC k w U b a b 2 w U ba
2

are defined as follows

(2a)

1
1
M ' b a 2 UbC k w U b a b 2 w U ba
2
2

1
1 a
b 3 w U b
8 2
2

2
Ankara International Aerospace Conference

(2b)

AIAC-2007-000

Here,

Durmaz, Ozdemir Ozgumus, Kaya

EI is the bending stiffness, GJ is the torsional stiffness, m mass per unit length, I is the mass

moment of inertia per unit length,

is the air density, C L

is the local lift coefficient and

is the time.

Boundary Conditions:

w
0
x

At x 0

At x L

w
w
2w
2w

GJ
EI
0
EI

2
2
x
x
x
x
x

(3a)

(3b)

Exponential Solution Function:

A sinusoidal variation of w( x, t ) and ( x, t ) with a circular natural frequency
the functions are approximated as exponential solutions.

is assumed and

w x, t w x e it

(4a)

x , t x e i t

(4b)

where

w x A1 f1 x eit A2 f 2 x e it

(5a)

x B11 x eit B2 2 x eit

(5b)

In this study, both one-term and two term Galerkin Methods are applied. However, only the two-term
Glerkin Method is introduced here since it already covers the one-term method.
Substituting Eqs. (4a)-(5b) into Eqs. (1a) and (1b) gives

d2

d 2 f1
b2

A1 Z 2 EI

mf

C k if 1 b 2 f 1
1

2
k
dx
dx

d2

d 2 f2
b2

A2 Z 2 EI

mf

C k if 2 b 2 f 2
2

2
k
dx
dx

b3
b3
b3
1

B1 my 1 2 2 C k 1 2 C k a i 1 i 1 b 3 a 1
k
k
k
2

b3
b3
b3
1

B2 my 2 2 2 C k 2 2 C k a i 2 i 2 b 3 a 2 0
k
k
k
2

3
Ankara International Aerospace Conference

(6a)

AIAC-2007-000

Durmaz, Ozdemir Ozgumus, Kaya

A1 2 my f 1 2 Ub 2 C k a if 1 2 b 3 a f 1 2 b 3 f 1
2
2

A2 2 my f 2 2 Ub 2 C k a if 2 2 b 3 a f 2 2 b 3 f 2
2
2
2

d 1
d
1

1
1

B1
GJ
2 I 1 2 U 2 b 2 C k a 1 2 Ub 3 C k a a i 1
dx
2

2
2

dx

1 a
a i 1 2 b 4 a a 1 b 3Ui 1 2 b 4 1
2

8 2

b 3U

d 2
d
1

1
1

2
2 2
3
B2
GJ
I 2 2 U b C k a 2 2 Ub C k a a i 2
dx
2

2
2

dx

1 a
a i 2 2 b 4 a a 2 b 3Ui 2 2 b 4 2 0
2

8 2

b 3U
(6b)

where k

1
b
and Z 2 1 ig .
U

Tapered Beam Formulas:

In this study, a beam model that tapers in one plane, xy plane, is considered. Therefore, the general
equations for the half of the veter length, b x , the cross-sectional area, A x , the distance between the
elastic axis and inertia axis, y x , the elastic axis location (positive rearward), a x , the moment of

inertia, I y x , the polar moment J x and the mass moment of inertia per unit length I of a beam, are
given by [2, 3].

x
x

b x b0 1 cb , a x a 0 1 cb ,
L
L

y x y 0 1 cb ,
L

x
x
x

I y x I y 0 1 cb , J x J 0 1 cb , I x I 0 1 cb
L
L
L

A x A0 1 cb
L

(7)

Here, the subscript 0 denotes the values at the root section of the tapered beam and c b represents the
breadth taper ratio which can be expressed as follows

cb 1

b
b0

(8)

These tapered beam formulas are used in the related parts of the expressions given by Eqs. (6a) and
(6b).
4
Ankara International Aerospace Conference

AIAC-2007-000

Durmaz, Ozdemir Ozgumus, Kaya

APPLICATION OF THE GALERKIN METHOD

In this section, application procedure of the two-term Galerkin Method is described. Following the
procedure, Eq. (6a) is multiplied by f 1 and the resultant expression is integrated from 0 to L which gives
the following equation

A1

d2
dx 2

Z
0

A2

EI

d2
dx 2

d 2 f1
b2
2

f
dx

mf
dx

C k if12 dx b 2 f12 dx
1
1

dx
k
0
0
0

EI

d 2 f2
dx 2

f1dx mf1 f 2 dx 2

b2
C k if1 f 2 dx b 2 f1 f 2 dx
k
0

b3
b3
1

B1 my 1 f1dx 2 2 C k 1 f1dx 2
C k a i1 f1dx
k
k
2

0
0
0
L

b3
b3
i1 f1dx b 3 a1 f1dx B2 my 2 f1dx 2 2 C k 2 f1dx
k
k
0
0

2
0

b3
b3
1

C k a i 2 f1dx
i 2 f1dx b 3 a 2 f1dx 0
k
k
2

0
0

(9)
Writing the terms of Eq.(9) as matrix coefficients gives

i
i
1

A1 Za111 a112 C k a113 a114 A2 Za121 a122 C k a123 a124 B1 b111 C k 2 b112
k
k
k

i
1
i

C k b113 b114 b115 B2 b121 C k 2 b122 C k b123 b124 b125 0

k
k
k

(10)
where
L

a111
0

d2
dx 2

EI

d 2 f1
f1dx
dx 2

a114 b 2 f12 dx
0

a123 2 b 2 f1 f 2 dx
0

a112 mf12 dx
0

d2
a121 2
dx
0

d 2 f2
EI
f1dx
dx 2

a124 b 2 f1 f 2 dx
0

5
Ankara International Aerospace Conference

a113 2 b 2 f12 dx
0

a122 mf1 f 2 dx
0

b111 my 1 f1dx
0

AIAC-2007-000

Durmaz, Ozdemir Ozgumus, Kaya

b112 2 b 1 f1dx
3

b113

b115 b a1 f1dx
3

b121 my 2 f1dx
0

b123

2 b
a 1 f1dx
2

0
3

2 b
a 2 f1dx b124 b 3 2 f1dx
2

0
0
3

b114 b 31 f1dx
0

b122 2 b 3 2 f1dx
0

b125 b 3 a 2 f1dx
0

As described above, Eq. (6a) is multiplied by f 2 and the resultant expression is integrated from 0 to
which gives the following equation

i
i
1

A1 Za 211 a 212 C k a 213 a114 A2 Za 221 a 222 C k a 223 a 224 B1 b211 C k 2 b212
k
k
k

i
1
i

C k b213 b214 b215 B2 b221 C k 2 b222 C k b223 b224 b225 0

k
k
k

(11)
Eq. (6b) is multiplied by 1 and the resultant expression is integrated from 0 to
following equation

L which gives the

i
i

A1 a 311 C k a 312 a 313 a 314 A1 a 322 C k a 322 a 323 a 324 B1 b311 b312
k
k

1
i
i
i
1

C k 2 b313 C k b314 b315 b316 b317 b318 B 2 b321 b322 C k 2 b323

k
k
k
k
k

C k

i
i
i

b324 b325 b326 b327 b328 0

k
k
k

Eq. (6b) is multiplied by 2 and the resultant expression is integrated from 0 to

following equation

L which gives the

i
i

A1 a 411 C k a 412 a 413 a 414 A2 a 422 C k a 422 a 423 a 424 B1 b411 b412
k
k

1
i
i
i
1

C k 2 b413 C k b414 b415 b416 b417 b418 B 2 b421 b422 C k 2 b423

k
k
k
k
k

C k

(12)

i
i
i

b424 b425 b426 b427 b428 0

k
k
k

Eqs. (10)-(13) can be written in a

4 4 matrix form as follows :

6

Ankara International Aerospace Conference

(13)

AIAC-2007-000

Durmaz, Ozdemir Ozgumus, Kaya

i
Za111 a112 C k k a113 a114

Za a C k i a a
211
212
213
114

a311 C k i a312 a313 a314

i
a411 C k a412 a413 a414
k

i
a123 a124
k
i
Za221 a222 C k a223 a224
k
i
a322 C k a322 a323 a324
k
i
a422 C k a422 a423 a424
k
Za121 a122 C k

1
i
b C k b113 b114 b115
2 112
k
k
1
i
b211 C k 2 b212 C k b213 b214 b215
k
k
1
i
i
i
b311 b312 C k 2 b313 C k b314 b315 b316 b317 b318
k
k
k
k
1
i
i
i
b411 b412 C k 2 b413 C k b414 b415 b416 b417 b418
k
k
k
k
b111 C k

7
Ankara International Aerospace Conference

AIAC-2007-000

Durmaz, Ozdemir Ozgumus, Kaya

1
i
b121 C k 2 b122 C k b123 b124 b125
k
k
1
i
b221 C k 2 b222 C k b223 b224 b225
k
k
1
i i
i
b321 b322 C k 2 b323 C k b324 b325 b326 b327 b328
k
k k
k
1
i i
i
b421 b422 C k 2 b423 C k b424 b425 b426 b427 b428
k
k k
k

A 0
1
A2 0
B 0
1
B2 0

(14)

SOLUTION PROCEDURE
Determinant of Eq.(14) is taken and made equal to zero. Solving this characteristic equation for a
certain k value, four roots which are Z 1 , Z 2 , Z 3 , Z 4 , are found. If the imaginary part of any root is
positive, the wing experience flutter.
Assuming that the imaginary part of the second root, Z 2 , is positive and knowing that Z
it can be written that

Z 2 i

1
1 ig
2

1
1 ig ,
2

(15)

Considering Eq.(15), the frequency value at which flutter occurs is calculated as follows

1
1

flutter
2

rad. / sec.

(16)

8
Ankara International Aerospace Conference

AIAC-2007-000

Durmaz, Ozdemir Ozgumus, Kaya

Using the calculated flutter frequency and knowing that k

flutter b
b
k U flutter
U
k

b
, the flutter speed is found as follows
U

ft / sec.

(17)

RESULTS
The computer package Mathematica is used to write a computer program for the expressions obtained in
the previous sections. Calculation of the flutter speed is made both for a uniform and for a tapered wing
model. The effects of several parameters are examined and in order to validate the calculated results, datas
of the Golands wing whose parametric values are given below are used.

GJ 0 2.39 10 6 lb
2

EI 0 23.6 10 6 lb

ft

ft

b0 3 ft

I 0 1.943 slug
ft

y 0 0.19973 ft

L 20 ft

a 0 1 / 3

m 0.746 slug/ft

Firstly, the flutter frequency and the flutter speed of a uniform Goland wing are found as k 0.47 and
U flutter 446.462 ft/sec. which are very close to the ones found by Ref.[5]. In order to validate the
calculated results, a comparison is made in Table 1.
Table 1. Flutter Speed, Flutter Frequency and Reduced Frequency of a Uniform Goland Wing
Flutter Speed

Flutter Frequency

f rad / sec

Reduced Frequency (
k)

447

69

0.47

One-Term
Galerkin

443.610

69.883

0.4727

Two-Term
Galerkin

446.462

70.035

0.4706

U f ft / sec

Analytical Study
(Ref. [4])

Present
Study

In Fig. 3, variation of both the bending and the torsion frequencies of a uniform wing with respect to the
free stream velocity, U is introduced. Here it is noticed that as the free stream velocity increases, the
frequencies get closer to each other which increases the possibility of flutter.
In Fig. 4, variation of the flutter speed, U f with respect to the location of the elastic axis, a , is
introduced. Here it is noticed that as the elastic axis gets closer to the leading edge (forward), it gets more
possible for the wing to get into flutter because the flutter speed gets lower.

9
Ankara International Aerospace Conference

AIAC-2007-000

Durmaz, Ozdemir Ozgumus, Kaya

In Fig. 5, variation of the flutter speed, U f with respect to the taper ratio, c b is introduced. Here it is
noticed that as the taper ratio increases, the flutter speed increases which makes it less possible for the wing
to experience fluttr.

100

Bending Frequency

80

60

Torsional Frequency
40

20

U
200

400

Figure 3.

U plot

10
Ankara International Aerospace Conference

600

800

AIAC-2007-000

Uf

Durmaz, Ozdemir Ozgumus, Kaya

800

700

600

500

400

300

-0.4

-0.2

Figure 4. Elastic Axis Location, a , and Flutter Speed,

11
Ankara International Aerospace Conference

0.2

U f , Relation

AIAC-2007-000

1200

Durmaz, Ozdemir Ozgumus, Kaya

Uf

1000

800

600

cb

400
0 .0

0 .2

0 .4

0 .6

0 .8

REFERENCES
[1] Ozdemir Ozgumus, O. and Kaya, M.O., Energy expressions and free vibration analysis of a
rotating double tapered Timoshenko beam featuring bendingtorsion coupling, International Journal of
Engineering Science, 45, 562-586, 2007.
[2] zdemir . ve Kaya M.O., Flapwise bending vibration analysis of a rotating tapered cantilevered
Bernoulli-Euler beam by differential transform method, Journal of Sound and Vibration, 289, 413420, 2006.
[3] Hodges, D.H. ve Pierce, G.A., Introduction to Structural Dynamics and Aeroelasticity, Cambridge
Aerospace Press, 2002.
[4] Lin, J. And Lliff, K.W., Aerodynamic Lift and Moment Calculations Using A closed Form Solution of
The Possio Equation, NASA, 2000

12
Ankara International Aerospace Conference