Aeroelastic Divergence and Free Vibration of Tapered Composite Wings
Aeroelastic Divergence and Free Vibration of Tapered Composite Wings
Aeroelastic Divergence and Free Vibration of Tapered Composite Wings
and ribs are simulated as the core resisting the boundary conditions can be expressed in matrix
transverse shear and the force normal to the face. form as follows:
With the transverse stiffening components, it is
F ′ − F0 + p = I 0 d (2)
usually assumed that the wing chordwise section is
strong enough to avoid the deformation. So that
some aerodynamic parameters are constants as the F = K 1d + K 2 d ′ , F0 = K 0 d + K 1T d ′ (3)
airfoil shape of the wing is fixed. Besides, because
the wing cross section must have a streamline shape
commonly referred to as an airfoil section, the K 1d + K 2d′ = Fˆ on y = y f and y = y0 (4)
thickness of the sandwich will be a function of the
where F, p and d are, respectively, the section force,
airfoil and the thickness is not too small to neglect
surface force and displacement vectors; K0, K1 and
the transverse shear deformation. Based upon these
K2 are stiffness matrices related to the extensional,
considerations shown above, a comprehensive
coupling and bending stiffness of the composites; I0
model was developed [17]. In this model, the
is a matrix related to the mass, gravity center and
displacement field can be expressed as follows:
moment of inertia. The overdot and the prime denote,
z respectively, the differentiation with respect to time
and space. The overhat stands for the prescribed
value on the boundaries. Please refer to Hwu and
Gai [17] for detailed explanation.
2.2 Deduction of the Elementwise Comprehensive
x Model
In order to refine the comprehensive model, the
model is rederived from Hamilton’s principle [18] as
shown by the following equation:
y t2
(π − κ )dt = 0
Fig. 1. Geometry of the composite wing [16]
δ ∫t1
(5a)
∫ (Tˆ u )dS
⎛1 ⎞
w(x, y, z , t ) = w f ( y, t ) − xθ ( y, t )
π= ∫ ⎜⎝ 2 σ
V
ij ε ij − f i ui ⎟dV −
⎠ Sσ
i i (5b)
(1c)
where u, v, and w are the displacement components
∫
1
κ= ρu i u i dV (5c)
in the direction x (chordwise), y (spanwise), z 2 V
(thicknesswise), respectively (Fig. 1); t is the time
variable, v0 the midplane displacement in the y By introducing Eq. (3) into Eq. (5b) and (5c) and
direction, wf the displacement (positive upward) in subtracting Eq. (5c) from Eq. (5b), we may obtain
the z direction measured along the reference axis, π −κ
∫ [d′ K d′ + 2d K d′ + d K d − d I d − 2p d]dy
and θ the rotation angle with respect to x axis due to =
1 T T T T T
2 1 0 0
twist around the reference axis (positive nose up), 2 y (6)
that is, βx=θ. Furthermore, βf is the rotation angle of −F d[ ]
ˆT yf
y0
x-z plane with respect to the y axis measured at the
reference axis and βr is the variation rate of βf along The governing equations and boundary conditions
the x axis, hence βx=βf+xβr. Through the above derived from Eq. (6) agree with Eq. (2) and Eq. (4)
interpretation, it is obvious that there are five basic derived in [17].
deformation functions for the stiffened composite
wing structures as follows, v0, wf, θ, βf, and βr. The
equations of motion, constitutive equation, and
2
AEROELASTIC DIVERGENCE AND FREE VIBRATION OF TAPERED
COMPOSITE WINGS
ye Le 1
qn ≡ ρV n2 = q cos 2 Λ (10)
2
Node 1 Node 2 Node 3 where ρ , V n , q , and Λ are the density of the
airflow, airflow velocity normal to the leading edge,
Fig. 2. Local coordinate of each element dynamic pressure, and the angle of sweep
respectively. The lift curve slope coefficient a is a
Because y-axis is the primary coordinate aerodynamic parameter defined as
variable in the wing structure. Thus the wing
structures are simulated as a beam divided in n dC L AR
a≡ = a0 (11)
elements connected along the y axis. Each element dα AR + 4 cos Λ
has 3 nodes (Fig. 2) and the shape function of each
where a0 is the corresponding 2D lift curve slope and
node is expressed by 2nd order polynomials. By
AR is the aspect ratio.
following the procedures stated between eqns.(5)-(6),
With the aerodynamic functions listed above,
the energy expression of each element can be written
the external aerodynamic force vector fe can be
in terms of the nodal displacement vector ue as
simplified further as
π −κ = ∫
⎧1 T 2
[ ]T ⎫
⎨ u e g K e − g J e u e − g (t )u e f e ⎬dy f e = N T ( y )p( y )dy = q n (p 0 + K a u e )
∫
2
(7)
ye ⎩2 ⎭ (12)
in which g is a function of time, and Ke, Je and fe are,
where the value of p 0 is constant, but the value of
respectively, element stiffness matrix, element
inertia matrix and element external force vector the aerodynamic stiffness matrix K a varies with the
determined through the following relations: interaction between the aerodynamics (L, T) and the
displacement function (ue).
∫ {N′ K N′ + 2N K }
T T T
Ke = 2 1 N′ + NT K 0 N dy (8a) Because divergence is a static phenomenon in
ye
aeroelasticity, we can neglect the time effect by
eliminating the inertial terms Je. Equation (7) can
∫
1 therefore be rewritten as follows
Je = NT I 0 Ndy (8b)
2 ye
π −κ =
1 T
2
(
u e K e u e − u Te q n p 0 + K a u e ) (13)
fe =
1
2 ye∫
NT pdy + NT Fˆ [ ] yf
y0 (8c) After substituting Eq. (13) into Eq. (5a), the
equilibrium equation of the divergence analysis is
where N(y) is the shape function matrix. expressed as
2.2.1 Divergence
In this article, we will use the aerodynamic
⎡1
( T
) ( T
⎢ 2 K e + K e − qn K a + K a )⎤⎥u e = qnp 0 (14)
⎣ ⎦
strip theory and the known results for two-
dimensional flow to approximate the lift and the The determinant of the coefficient matrix of ue
pitching moment [19]. The relation between the should be zero to avoid the trivial solutions, which
structure deformation and aerodynamic forces (lift will then lead to the result of divergence,
force L and pitching moment T) can be expressed as
[ (
L( y ) = q n acθ 0 + ac θ ( y ) − w′f tan Λ )] (9a)
1
2
( ) (
K e + K Te − qn K a + K Ta = 0 ) (15)
[( ) ( )]
Obviously, it is an eigen problem and the eigen-
T ( y ) = qn aceθ0 + c 2Cm, ac + ac eθ ( y ) − ew′f tan Λ (9b) value and eigen-vector indicate the normal
divergence dynamic pressure and the nodal
where c is the chord length, θ 0 the initial angle of displacement vector respectively. After the
attack, Cm,ac the pitching moment coefficient about superposition of all the elements accomplishing, the
the aerodynamic center, and e the distance between eigen problem can be solved.
the lines of aerodynamic and flexural centers. The
normal dynamic pressure qn is
3
Mengchun Yu, Chyanbin Hwu
3.1 Divergence
In this section, the calculating efficiency of the Fig. 4. Comparison with the existing data
elementwise comprehensive model will be shown
first. And the accuracy of the present model will be
examined by comparing with the existing numerical
solutions provided by other papers.
Before investigating the tapered effect in wing Table 1 shows the natural frequencies got from
structures, the taper ratio rt is denoted as the present model in this article well agree with the
solution of Hwu & Gai and ANSYS. It deserves to
be mentioned that the calculating efficiency is also
croot − ctip improved as shown in Table 2.
rt = (18)
croot
Table 1. Natural frequencies comparison for the
where croot and ctip are the chord lengths of the wing wing structure with uniform span (Unit:Hz)
root and wing tip, respectively. Mode No. Present
Hwu & Gai
ANSYS
The examples in a series of tapered ratios are [17]
solved for AR = 4 and 5. The numerical results are I 16.89 16.15 15.75
II 99.12 96.45 93.84
shown in Fig. 5. It shows that once the value of III 108.69 110.34 114.94
tapered ratio increases, the value of divergence IV 256.83 252.45 245.18
pressure will increase, too. V 325.23 333.85 343.82
3.2 Free Vibration
Table 2. Time consumed comparison
In this section, the present model is first Hwu & Gai
Present ANSYS
examined by comparison with the existing numerical [17]
data of the wing structure with uniform span Time consumed
2 25 30
(minute)
evaluated by Hwu and Gai [17]. The geometry
properties of the wing are chordwise length c=0.1m
and spanwise length L=0.4m shown in Fig. 6 (with Table 3. Effects of the taper ratio on the basic
croot = ctip ) where croot and ctip means the chord natural frequencies (Mode I) (Unit:Hz)
Taper ratio Mode No. Present ANSYS
length at the wing root and wing tip respectively. I 16.89 15.25
The detailed material properties of wing are listed in II 99.12 90.35
[17]. 1.0 III 108.69 113.16
IV 256.83 234.64
ctip V 325.23 275.61
I 17.38 14.37
II 99.88 85.11
0.95 III 113.85 117.66
IV 256.83 224.07
V 330.82 280.47
Wing Tip I 19.08 16.47
II 101.88 88.43
0.80 III 132.73 125.47
IV 256.89 224.35
V 350.35 277.64
y I 20.49 17.13
L=400mm II 103.39 87.02
0.70 III 148.41 133.02
IV 257.09 217.85
V 366.40 271.96
I 22.21 17.91
Wing Root II 105.23 85.39
0.60 III 167.08 155.67
IV 257.58 210.67
V 385.92 268.10
I 35.50 22.90
II 122.80 80.28
0.20 III 267.76 120.55
x IV 282.12 180.53
croot V 465.98 213.95
5
Mengchun Yu, Chyanbin Hwu
6
AEROELASTIC DIVERGENCE AND FREE VIBRATION OF TAPERED
COMPOSITE WINGS