Composite wing elastic axis for aeroelasticity

optimization design
S.H. Huo, F.S. Wang, Z. Yuan and Z.F. Yue
Advanced Materials Test Center, School of Mechanics Civil Engineering and Architecture,
Northwestern Polytechnical University, Xi’an, People’s Republic of China

Abstract
Purpose – Computational efficiency is always the major concern in aircraft design. The purpose of this paper is to investigate an efficient aeroelasticity
optimization design method. Analysis of composite wing elastic axis is presented in the current study and its application on aeroelasticity optimization
design is discussed.
Design/methodology/approach – Elastic axis consists of stiffness centers. The stiffness centers of eight cross sections are analyzed and the wing
elastic axis is obtained through least-squares procedure. In the analysis of the cross section stiffness center, the wing model is approximated by
assuming the wing cross section as a thin walled structure with a single cell closed section and assuming the composite material to be a 3D anisotropic
material. In aeroelasticity optimization design, objective functions are taken to be the wing weight and elastic axis position. Design variables are the
thickness and area of wing components.
Findings – After aeroelasticity optimization design, the wing weight decreases while the divergent velocity increases. Meanwhile, it can achieve an
expected result but costs much less computational time than the conventional method.
Practical implications – The results can be used for aircraft design or as an initial value for the next detailed optimization design.
Originality/value – The computational time can be dramatically reduced through the aeroelasticity optimization design based on the elastic axis. It is
suitable for engineering applications.

Keywords Air transport engineering, Design, Composite wing, Elastic axis, Divergent velocity, Aeroelasticity optimization design,
Multi-Island genetic algorithm

Paper type Research paper

Introduction developed a particularly true MDO framework for the

aerodynamic shape optimization. Unfortunately, the
Aeroelasticity is always the major concern in aircraft design. computational time of MDO analysis increased drastically.
Aircraft structure is elastic. Elastic deformation, which is the For reducing the time, Zhang et al. (2008) proposed surrogate
result of initial aerodynamic function, will lead to additional models. At the same time, many algorithmic methods which
aerodynamic load. Additional aerodynamic load will lead to a reproduce some natural phenomena or physical processes were
new elastic deformation. Aircraft structure will obtain a balanced proposed, such as Genetic and Evolutionary Algorithms
state or deform gradually until structural damage happens. (Alonso et al., 2009), Swarm Intelligence (Venter and
There is a critical speed for each specified aircraft, which is Sobieszczanski-Sobieski, 2004) and Simulated Annealing
defined as divergent velocity. Both aircraft layout and mass (Jeon et al., 2008).
distribution have a direct impact on the divergent velocity. In aeroelasticity optimization design, the objective function
Higher divergent velocity and lighter weight of the aircraft are the correlates closely with the divergence speed and structure mass.
targets of aircraft design. The optimization design, especially Guo et al. (2003) demonstrated that a significant increase of
multi-disciplinary design optimization (MDO), is widely used in flutter speed can be achieved by optimizing the wing skin
aircraft design. thickness or the laminate layups. Kameyama and Fukunaga
Aeroelasticity optimization design which considers both (2007) investigated the effect of laminate configuration on the
weight and aerodynamic characteristic is a branch of MDO. aeroelasticity characteristics and optimization design of
Many scholars dedicated to research in this field and a lot of composite plate wings using lamination parameters. Barcelos
methods were put forward. Wakayama (2000) showed an initial and Maute (2008) studied a general optimization methodology
structural optimization which has not a proper MDO framework. for fluid-structure interaction problems based on turbulent flow
As the development of computational fluid dynamic (CFD) models. Librescu and Maalawi (2009) presented the
models, Guruswamy and Obayashi (2004) and Yang et al. (2005) aeroelasticity optimization of a wing-type structure through a
novel mathematical approach. Mastroddi et al. (2011) studied
The current issue and full text archive of this journal is available at the initial design of wing structures based on the integrated
www.emeraldinsight.com/1748-8842.htm modeling of structures, aerodynamic, flight dynamics and
aeroelasticity. Chintapalli et al. (2010) investigated a layout

Aircraft Engineering and Aerospace Technology: An International Journal

85/1 (2013) 10– 15 This study is supported by the National Natural Science Foundation of
q Emerald Group Publishing Limited [ISSN 1748-8842] China (No. 50905142) and Shaanxi Provincial Natural Science
[DOI 10.1108/00022661311294030] Foundation (No. 2009JQ1006).

10
 Composite wing elastic axis for aeroelasticity optimization design Aircraft Engineering and Aerospace Technology: An International Journal
S.H. Huo, F.S. Wang, Z. Yuan and Z.F. Yue Volume 85 · Number 1 · 2013 · 10 –15

optimization method of a skin-stringer panel for an aircraft Simplified model

wing-box. They all investigated the accuracy and efficiency of
aircraft aeroelasticity optimization design and some considerable Elastic axis is a position where only bending happens when the
load is applied on it. Elastic axis is composed of many wing cross
methods were proposed. However, computational efficiency still
section stiffness centers. The wing cross section is simplified to a
has an obvious impact on aircraft design, but it is not often
thin walled structure with a single cell closed section. In the two
addressed.
dimensional cross section model, skin and spar web are simplified
Stiffness characteristic has an impact on aeroelasticity.
as some lines while stringer and spar flange are simplified as some
Serious aeroelasticity problems may occur under an irrational
points. Two equivalent processes, area equivalent and material
stiffness characteristic design. Elastic axis plays an important
equivalent, are carried out for the model simplifying.
role in stiffness characteristic. The present work aims at
investigating the composite wing elastic axis and its
(1) Area equivalent
application on the aeroelasticity optimization design.
Cross section area of each component has a direct effect on
stiffness center. For skin, all of the cross section areas are effective.
Analysis of composite wing elastic axis For part of skin which locates within the interval of [a, b],
its effective area can be written as:
Stiffness characteristic of the wing structure has a high Z b
correlation with the location of elastic axis. Hence, an Askin ¼ tds ð1Þ
appropriate location of the elastic axis is important in a
engineering design. Here, the location of elastic axis analysis
is carried out on a composite wing structure. where Askin is skin effective area and t is skin thickness.
For stringer, it has been simplified to an intensive area
above which is shown as a dot. Usually, the dot needs to be
Composite wing configuration lain in the line obtained from the simplified skin. So the
stringer area needs to be reduced, it can be written as:
The composite wing shown in Figure 1 is composed of skin,
spar, rib and stringer. It is a primal finite element model. h21
A*stringer ¼ Astringer ð2Þ
NACA0012 airfoil is adopted in the model. The chord length h20
of the wing root is 2,500 mm while the wing tip is 1,500 mm.
The span which has a back swept angle of 208 is 16,000 mm. where A*stringer and Astringer are the reduced and initial area,
The material of all components is T300/QY8911. Its material respectively, h1 is the distance from the skin to the neutral surface
properties are shown in Table I. Table II shows the specific of airfoil cross section, and h0 is the distance from the stringer
laminate layup and thickness of every component. The length centroid to the neutral surface of the airfoil cross section.
and thickness of spar flange is 100 mm and 6 mm, respectively. For the spar, top and down flanges are reduced to an
The thickness of spar web is 5 mm. For T shaped stringer, intensive area just as the method in stringer treatment while
the length and thickness of flange are 60 mm and 4 mm, the web is treated in the same way as used for the skin.
respectively, and those of web are 80 mm and 4.8 mm.
(2) Material equivalent
Skin, stringer and spar have their own different layups, so it is
Figure 1 Finite element model of composite wing
necessary to select one of them as material sample and others
equivalent to it. A reduction factor w is introduced and
written as:
Ei
w¼ ð3Þ
E0
where Ei is Young’s modulus of practical material and E0 is
that of material sample.
In the present study, composite material is converted into a 3D
anisotropic material based on the stiffness matrix (Dong et al.,
Table I Material property of T300/QY8911 2010). For an orthotropic laminate, the stress-strain relationship
E11/MPa E22/MPa E33/MPa G12/MPa G23/MPa G13/MPa n12 n23 n13 of each layup under 3D stress condition can be written as:
8 9 2 38 9
135,000 8,800 8,800 4,470 3,200 4,470 0.33 0.48 0.15 >
> sx >> Q11 Q12 Q13 >
> 1x >
>
> > > >
>
>
> > 6
>
6 Q21 Q22 Q23
7>
7>
>
1y >
> s y >
> >
> >
>
>
> > 6
> 7>> >
>
>
> > 6 7 > >
< sz > = 6 Q31 Q32 Q33 7>< 1z > =
Table II Laminate layup of every component 6 7
¼6 7 ð4Þ
>
> t > 6 Q44 7> g >
Component Layup Thickness (mm) > yz >
> > 6
> 7>> yz >
> >
>
>
> tzx >
> 6 7> >
>
> >
> 6 Q55 7>> gzx >>
Skin [0/90/0/45/-45/0/90/0/45/-45]S 4 >
> > 4
> 5>>
>
>
>
>
>
: txy ;> >
Q66 : gxy > ;
Spar [0/45/-45/90/0/45/-45/0/45/-45]S –
Rib [0/90/0/45/-45/0/90/0/45/-45]S 4
Stringer [0/45/-45/0/90/0/90/0/-45/45]S – Because the principle axis of each layup may not coincident with
the global coordinate axis, the material stiffness matrix with

11
 Composite wing elastic axis for aeroelasticity optimization design Aircraft Engineering and Aerospace Technology: An International Journal
S.H. Huo, F.S. Wang, Z. Yuan and Z.F. Yue Volume 85 · Number 1 · 2013 · 10 –15

respect to the material coordinate system needs to be converted Position of elastic axis
into that under the global coordinate system. Introducing
the stress transfer matrix [T ] and the router matrix [R ], the One of the wing cross sections is selected for introducing the
stiffness center computation method (Liu et al., 2006). The wing
stress-strain relationship can be rewritten as:
8 9 8 9 8 9 cross section is composed of ten stringers, four spar flanges, two
> sx > > s1 > > 11 >
>
> >
> >
> >
> >
> >
> spar webs and skin among two intensive area positions. Table III
>s >
> > >
> >
> >
> >
>
>
> >
> >
> s >
> >
> 1 >
> shows the area and coordinate of the initial wing intensive area.
>
>
y >
> >
> 2 >
> >
> 2 >
>
>
> >
> >
> >
> >
> >
> In Table III, Bi and Si are the intensive area of spar flange and
>
> > > > > >
< sz >= < s3 >
> = < 13 >
> = stringer, respectively.
21 21
¼ ½T
¼ ½T
½Q
The area of each wing cross section component and the
> > > > > >
> tyz >
>
> >
> > t23 >
>
> >
> > g23 >
>
> >
> area centroid can be obtained
I through:
>
> >
> >
> >
> >
> >
> X
>
> >
> >
> >
> >
> >
>
t
> zx >> t
> 13 >
> > > 13 >
> g >
> A ¼ wi t i ds þ wi Ai
>
> >
> >
> >
> >
> >
>
: txy >
> ; :t >
> ; : g12 >
> ;
i
12 H P
8 9 ð5Þ wi t i xi ds þ wi Ai xi
> 1x > i
>
> >
> j0 ¼ ð10Þ
>
> >
>
>
> 1y >>
>
> H
A
>
> >
> P
>
> > wi t i yi ds þ wi Ai yi
> 1z >
< >
= i
¼ ½T
21 ½Q
½R
½T
½R
21 h0 ¼
> > A
> gyz >
>
> >
> Where wi is the reduction factor, ti is the thickness of skin, A is
>
> >
>
>
> gzx >
> the total area of all component cross section, and j0 and h0
>
> >
>
>
> >
> are x and y coordinate value of centroid, respectively.
>
: gxy >
;
Because the computation of stiffness center will be carried
out in inertia axis, it is necessary to transfer the coordinate,
Therefore, stiffness matrix under partial coordinate system can which can be done by subtracting the value of centroid. Then,
be written as: the area moment of inertia of the reduced cross section with
respect to the inertia axis can be obtained as:
½Q
21 ¼ ½T
21 ½Q
½R
½T
½R
21 ð6Þ I X
J x ¼ y2i wi t i ds þ wi Ai y2i
Based on equation (6), the stiffness matrix of each layup under i
the global coordinate system can be written as: I ð11Þ
X
2 3 Jy ¼ x2i wi t i ds þ wi Ai x2i
Q11 Q12 Q13 0 0 Q16 i
6 7
6 Q12 Q22 Q23 0 0 Q26 7 The moment of the area of reduced cross section with respect
6 7
6 7 to the inertia axis can Ibe obtained as:
6 Q13 Q23 Q33 0 0 Q 36 7 X
6 7
½Qij
k ¼ 6 7 ð7Þ S x ¼ yi wi t i ds þ wi Ai yi
6 0 0 0 Q44 Q45 0 7
6 7 i
6 7 I
6 0 0 0 Q45 Q55 0 7 X ð12Þ
4 5 Sy ¼ xi wi t i ds þ wi Ai xi
Q16 Q26 Q36 0 0 Q66 i
Based on the structural mechanics theory, the stiffness center
The laminate equivalent stiffness matrix can be written as: of a cross section can be obtained as:
PN H !
½Qij
k d k Z
½Qij
¼ k¼1 ð8Þ 1 ðS x =GtÞds
t X¼ S x rds 2 V H
Jx s ðds=GtÞ
where N is the number of layup, dk is the thickness of the kth layer 0 1
Z H ð13Þ
and t is the thickness of laminate. For a skew-symetric laminate, 1 @ ðS y =GtÞds
Q16 ¼ Q26 ¼ Q36 ¼ Q45 ¼ 0 because they are odd functions of Y¼ S y rds 2 V H A
Jy ðds=GtÞ
the layup angle. Then, the stiffen matrix can be written as: s

2 3 where r is the distance from the area centroid to the tangent of

Q11 Q12 Q13 0 0 0
6 7 each component, and V is two times the enclosed cross section.
6 Q12 Q22 Q23 0 0 0 7 Through the method above, it can be obtained that the
6 7
6 7 distance from the stiffness center to the leading edge is
6 Q13 Q23 Q33 0 0 0 7
6 7 44.41 percent of the chord length.
½Qij
¼ 6 7 ð9Þ
6 0 0 0 Q44 0 0 7 Eight wing cross sections are selected and their stiffness
6 7
6 7 centers are obtained. Then the wing elastic axis position can
6 0 0 0 0 Q55 0 7
4 5 be obtained through least-squares procedure.
0 0 0 0 0 Q66

Hence, the composite material has been converted into 3D

Effects of elastic axis on aeroelasticity
anisotropic material and its Young’s modulus can be obtained Take a two dimensional wing for example, bending and torsion
from the material stiffen matrix. of the wing were simulated as sink-float and pitch of the airfoil

12
 Composite wing elastic axis for aeroelasticity optimization design Aircraft Engineering and Aerospace Technology: An International Journal
S.H. Huo, F.S. Wang, Z. Yuan and Z.F. Yue Volume 85 · Number 1 · 2013 · 10 –15

Table III Area and coordinate of the initial wing intensive area
Top skin Down skin
Serial number x/mm y/mm Area/mm2 Serial number x/mm y/mm Area/mm2
B1 2,394.48 139.08 600 B3 2,392.18 2 69.54 600
B2 3,505.33 70.53 600 B4 3,502.67 2 49.53 600
S1 2,556.49 144.34 624 S6 2,551.51 2 81.38 624
S2 2,737.94 140.64 624 S7 2,732.94 2 85.69 624
S3 2,932.04 129.41 624 S8 2,927.35 2 83.05 624
S4 3,130.37 112.56 624 S9 3,126.22 2 74.96 624
S5 3,324.28 92.23 624 S10 3,320.84 2 63.21 624

(Palacios and Cesnik, 2005). Because aerodynamic load caused Optimization parameters
by sink-float is feeble, only pitch is considered in airfoil
aeroelasticity problems. Practical airfoil attack angle a is Wing weight and the distance from the stiffness center to the
composed of initial attack angle a0 and torsion angle u under the leading edge point are two objective functions in the present
study. The optimum objective function can be written as:
function of aerodynamic. Moment of aerodynamic on the
stiffness center, Me, can be written as: W 1G W 2e

 F¼ þ ð18Þ
Gm l
›CL
M e ¼ qd ScC MAC þ qd S ða0 þ uÞ e ð14Þ
›a where G is the wing weight after optimization, Gm is the
maximum wing weight in design variable span, e is the distance
where qd is the dynamic pressure, S is the airfoil area, c is the from the stiffness center to the leading edge, l is the cross section
cross section chord length, CMAC is the moment coefficient, chord, and W1 and W2 are weight coefficients.
(›CL)/(›a) is the slope of lift curve and e is the distance from the Constraint conditions are selected based on structural
stiffness center to the aerodynamic center. characteristics. The maximum torsion angle, maximum elastic
Two springs are adopted to simulate constraints of the axis flexivity and stability are constraint conditions. Table IV
airfoil. Equilibrium equation of aerodynamic moment and shows their specific values.
spring elastic moment is written as: The stress is not uniformly distributed over the wing. So,

 different region thickness values advantage wing design. Figure 2
›C L shows the regions divided in the composite wing. As shown in
qd S ða0 þ uÞ e þ qd cSC MAC ¼ ka u ð15Þ
›a Figure 2, the finite element model of the wing structure is divided
into 20 regions. In each region, skins, spars, ribs and stringers
where ka is the spring elastic coefficient. So the torsion angle u have their own independent design variables. Thickness is the
can be obtained and written as: most important design variable in the present work.
ðqd SÞ=ðka Þðeð›C L =›aÞa0 þ cC MAC Þ
u¼ ð16Þ
1 2 ððqd eSÞ=ka Þð›C L =›aÞ Optimization process
The torsion angle becomes infinite when the denominator of Whole optimization process is carried out using the software
equation (16) equals to zero, indicating that the wing is in a iSIGHT. Multi-island genetic algorithm is used for wing
optimization design. Static strength and stability are analyzed
divergent state. Hence, the divergent velocity, VD, can be
after setting initial model design parameters. It can be judged
written as:
from the result that whether the model meets constraint
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi conditions. If it does, the wing weight and the position of elastic
2ka
VD ¼ ð17Þ axis are output. If not, design parameters need to be modified
ra eS ›C L =›a and computed again. Static strength and stability are analyzed
through FEM software NASTRAN while the position of elastic
From equation (17), it can be seen that divergent velocity axis in the FORTRAN code is based on the introduction above.
decreases with the increase of e. Hence, the location of the Figure 3 shows the optimization process. As shown in Figure 3,
stiffness center has a direct impact on the wing divergent the stop condition can be expressed as:
velocity. A stiffness center closer to the aerodynamic center by
changing the wing layout is preferred in wing design. Table IV Values of constraint conditions
Constraint condition Value
Optimization design based on elastic axis
Maximum tensile strain j1t max j , 4,500
For general multi-disciplinary optimization problems, it can be Maximum compression strain j1c max j , 4,500
described by an optimum objective function Maximum shear strain j1xy max j , 4,000
Fð f ðx1 Þ; f ðx2 Þ; . . . ; f ðxn ÞÞ, constraint condition ðg j ðXÞ , 0 Maximum torsion angle umax , 2.08
ð j ¼ 1; 2; . . . ; mÞÞ or hj ðXÞ ¼ 0 ð j ¼ m þ 1; . . . ; nÞ and Maximum elastic axis flexivity Umax , 1.3 m
design variables X L # X # X U , X ¼ ðX 1 ; X 2 ; . . . ; X n ÞT Stability lmin . 1.0
(Riche and Haftka, 1993).

13
 Composite wing elastic axis for aeroelasticity optimization design Aircraft Engineering and Aerospace Technology: An International Journal
S.H. Huo, F.S. Wang, Z. Yuan and Z.F. Yue Volume 85 · Number 1 · 2013 · 10 –15

Figure 2 Regions divided in the composite wing

Figure 3 Optimization process Table V Optimization results under different weight coefficients
Divergent
Begin
Wing weight e/l velocity
Value Value Value
Optimization Optimization results Trend (%) Trend (%) Trend (%)
Load Condition
W1 5 0.4, W2 5 0.6 Down 20.38 Down 10.86 Up 5.92
W1 5 0.5, W2 5 0.5 Down 21.57 Down 9.74 Up 5.26
Initial Model W1 5 0.6, W2 5 0.4 Down 23.21 Down 9.23 Up 4.96
Design Variables
ten minutes, which is much faster than the traditional wing
Static Strength and Update aeroelasticity optimization design method.
Stability Analysis Design Variables Table VI shows the optimization results of one cross section.
Strength It can be seen from Table VI that all of design parameters
Meet the No under different weight coefficient decrease, but they have their
Stiffness own scale under each weight coefficient. Comparing the results
Constraint Condition?
of the top skin with those of the down skin, it can be seen that
Stability Yes No the thickness of top skin is generally larger than that of the
down skin. It is because that compression load is mainly applied
Wing Weight Obtain on top skin, and a larger thickness value can avoid buckling.
Stop Condition
Fn+1 – Fn Moreover, it can be seen that the thickness of leading spar web is
Position of
F ≤e? larger than that of the trailing spar web. It is because that the
Elastic Axis n weight increase of the leading edge is conducive to the forward
Yes moving of the elastic axis.
We conclude that wing weight decreases and divergent
End
velocity increases through aeroelasticity optimization design
method based on elastic axis. Meanwhile, it can achieve an
  expected result but costs much less computational time than the
F nþ1 2 F n 
 #1 ð19Þ conventional method. Computational efficiency is always the
 Fn  major concern in aircraft design. The aeroelasticity optimization
design method based on elastic axis can be well used to solve the
Where Fn is the optimum objective value at the n time iteration, problem, especially for some large-scale structures.
1 is used to judge the convergence of the iteration and
1 ¼ 1 £ 102 6 in the current optimization.
Conclusions
Optimization result The composite wing elastic axis is analyzed and an aeroelasticity
optimization design method based on the elastic axis is put
Table V shows optimization results under different weight forward in the present work. Some useful conclusions can be
coefficients. From Table V, it can be seen that both wing drawn as follows:
weight and e/l after optimization decrease under different weight .
The wing cross-section was simplified to a thin walled
coefficients, whereas divergent velocity increases. Comparing structure with a single cell closed section and the composite
results under different weight coefficient values, it can be seen material was converted into a 3D anisotropic material based
that weight coefficient has a direct impact on optimization results. on the material stiffness matrix with respect to the elastic axis.
It needs to be selected based on practical engineering problem. .
The location of the wing elastic axis has a direct impact on
One-step of the optimization just includes static strength and the divergent velocity. A wing elastic axis closer to the
stability analyzes using the FEM software NASTRAN and the leading edge by changing the wing layout is preferred in
elastic axis computation using a FORTRAN code, it costs about wing design.

14
 Composite wing elastic axis for aeroelasticity optimization design Aircraft Engineering and Aerospace Technology: An International Journal
S.H. Huo, F.S. Wang, Z. Yuan and Z.F. Yue Volume 85 · Number 1 · 2013 · 10 –15

Table VI Optimization results of one cross section

Design parameter Initial model W1 5 0.4, W2 5 0.6 W1 5 0.5, W2 5 0.5 W1 5 0.6, W2 5 0.4
Top skin thickness/mm 4 3.8 3.5 3.1
Down skin thickness/mm 4 2.4 2.6 2.9
Leading spar web thickness/mm 5 4.0 4.0 3.9
Trailing spar web thickness/mm 5 2.7 3.0 3.1
Top stringer area/mm2 624 602 595 592
Down stringer area/mm2 624 493 497 463

.
After aeroelasticity optimization design, the wing weight Kameyama, M. and Fukunaga, H. (2007), “Optimum design
is decreased while the divergent velocity is increased. of composite plate wings for aeroelastic characteristics using
Meanwhile, it costs much less computational time than the lamination parameters”, Computers and Structures, Vol. 85,
conventional methods and can achieve an expected result. pp. 213-24.
.
All of the design parameters under different weight coefficients Librescu, L. and Maalawi, K.Y. (2009), “Aeroelastic design
decrease through aeroelasticity optimization design. In the optimization of thin-walled subsonic wings against
optimization results, the thickness of the top skin is larger than divergence”, Thin-Walled Structures, Vol. 47, pp. 89-97.
that of the down skin while the thickness of the leading Liu, C.Y., Huang, G.N. and Wang, L.J. (2006),
spar web is larger than that of the trailing spar web. “Computation of stiffness characteristic of a high aspect
ratio wing”, Airplane Engineer, Vol. 4, pp. 21-4.
References Mastroddi, F., Tozzi, M. and Capanolo, V. (2011), “On the
Alonso, J.J., LeGresley, P. and Pereyra, V. (2009), “Aircraft use of geometry design variables in the MDO analysis of
design optimization”, Mathematics and Computers in wing structures with aeroelastic constraints on stability and
Simulation, Vol. 79 No. 6, pp. 1948-58. response”, Aerospace Science and Technology, Vol. 15,
Barcelos, M. and Maute, K. (2008), “Aeroelastic design pp. 196-206.
optimization for laminar and turbulent flows”, Computer Palacios, R. and Cesnik, C.E.S. (2005), “Static nonlinear
Methods in Applied Mechanics and Engineering, Vol. 197, aeroelasticity of flexible slender wing in compressible flow”,
pp. 1813-32. AIAA-2005-1945.
Chintapalli, S., Elsayed, M.S.A., Sedaghati, R. and Abdo, M. Riche, R.G.L. and Haftka, R.T. (1993), “Optimization of
(2010), “The development of a preliminary structural design laminate stacking sequence for buckling load maximization
optimization method of an aircraft wing-box skin-stringer by genetic algorithm”, AIAA, Vol. 31 No. 5, pp. 951-6.
panels”, Aerospace Science and Technology, Vol. 14, pp. 188-98. Venter, G. and Sobieszczanski-Sobieski, J. (2004),
Dong, Y.P. and Wang, P.Y. et al. (2010), “Optimization “Multidisciplinary optimization of a transport
analysis of the composite stiffness based on beam aircraft wing using particle swarm optimization”,
elements”, Journal of Materials Science & Engineering, Structural and Multidisciplinary Optimization, Vol. 26
Vol. 28 No. 6, pp. 857-61. No. 11, pp. 121-31.
Guo, S., Banerjee, J.R. and Cheung, C.W. (2003), “The Wakayama, S. (2000), “Blended-wing-body optimization
effect of laminate lay-up on the flutter speed of composite problem setup”, AIAA Paper 2000-4740.
wings”, Proceedings of the Institution of Mechanical Engineers, Yang, Q., Liang, Q. and Yang, Y.N. (2005), “Static aeroelastic
Part G (Journal of Aerospace Engineering), Vol. 217 No. 3, optimization with CFD method”, Acta Aerodynamica
pp. 115-22. Sinica, Vol. 23 No. 1, pp. 16-20.
Guruswamy, G.P. and Obayashi, S. (2004), “Study on the use Zhang, K.S., Han, Z.H., Li, W.J. and Song, W.P. (2008),
of high-fidelity methods in aeroelastic optimization”, “Coupled aerodynamic/structural optimization of a
Journal of Aircraft, Vol. 41 No. 3, pp. 616-19. subsonic transport wing using a surrogate model”, Journal
Jeon, K.S., Lee, J.W. and Byun, Y.H. (2008), of Aircraft, Vol. 45 No. 6, pp. 2167-70.
“Multidisciplinary design approach using repetitive
response surface enhancement and global optimization”,
Proceedings of the 2008 Ninth ACIS International Conference
Corresponding author
on Software Engineering, Artificial Intelligence, Networking and
Parallel/Distributed, pp. 25-8. S.H. Huo can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected]

Or visit our web site for further details: www.emeraldinsight.com/reprints

15