88

Investigation on a multispar composite wing

D G Stamatelos and G N Labeas*
Laboratory of Technology and Strength of Materials, University of Patras, Patras, Greece

The manuscript was received on 22 March 2010 and was accepted after revision for publication on 18 March 2011.

DOI: 10.1177/0954410011406476

Abstract: The potential advantages of an airliner composite wing are investigated by consid-
ering a multispar configuration. An efficient initial sizing methodology, suitable for large-scale
composite components, such as the studied aircraft main wing, is applied for the preliminary
design of a multispar wing. The methodology comprises three basic modules, i.e. the compu-
tational stress analysis of the structure, the comparison of the stress–strain results to design
allowable, and a suitable resizing procedure in order to satisfy all design requirements.
A detailed comparison between optimized designs of conventional (2-spar) and three alterna-
tive wing configurations, which comprise 4-, 5-, and 6-spars for the wing construction is per-
formed. In order to understand the effect of different material properties, as well as the
variation of maximum strain level allowed the total wing mass, parametric analyses are per-
formed for all configurations considered. Comparing the conventional and the multispar
arrangements, it arises that under certain conditions, the multispar configuration demonstrates
significant advantages over the conventional design. In addition, for all the configurations
investigated, i.e. the 2-, 4-, 5- and 6-spar designs, the mass breakdown in the wing subcompo-
nents is presented, such that the maximum mass saving potentials are identified for each
multispar configuration.

Keywords: composite wing, multispar, stress analysis, optimization, finite element analysis

1 INTRODUCTION studies on the possible advantages of composite mul-

tispar wing configurations have not performed so far.
The application of composite materials in primary Most researchers are focusing their studies on other
aeronautical structures is progressively increasing, specific problems of composite wing design, such
e.g. about 25 per cent of the new Airbus 380 aircraft as aeroelasticity, flutter, and local optimization of
is produced of composites (mainly the rear fuselage, panels or areas with cut-outs, etc. [4–6]. Regarding
central wing box, horizontal, and vertical tail plane), the detailed design of a composite wing, the majority
while the next aircraft generation is expected to have of research works are focused on local structural
at least 50 per cent of composite material. To fully problems, e.g. buckling issues or lay-up optimization
exploit the advantages of composite material for obtaining lighter designs [7–12]. The first work
design, apart from other developments, investigation referring to the analysis of a multispar concept was
on the possible advantages of non-conventional published in 1932 by Sunger [13] and presents the
design configurations is essential. static analysis of a metallic multispar wing structure.
A literature review has revealed that investigation In reference [14], Herbeck and Wilmes have focused
on multispar wing structures is rare [1–3], while in 2002 to the selection of design criteria for a com-
posite wing. The most recent publication relevant to
*Corresponding author: Laboratory of Technology and Strength of this investigation is that of Chintapalli and Elsayed
Materials, University of Patras, Panepistimioupolis Rion, Patras [15], in which a methodology for optimal design of
26500, Greece. stiffened skin panels of an aircraft metallic wing is
email: labeas@mech.upatras.gr presented.

Proc. IMechE Vol. 226 Part G: J. Aerospace Engineering

Downloaded from pig.sagepub.com at NANYANG TECH UNIV LIBRARY on May 23, 2015
 Multispar composite wing 89

Although metallic multispar arrangements (e.g. properties. The FE results are introduced in a sizing
delta wing) are usually used in supersonic military routine, which enables the design optimization of all
aircraft, in order to fulfil the supercritical airfoil subcomponents of the structural layouts. The sizing
requirements, all multispar designs studied so far routine includes an integrated set of design con-
refer to metallic structures and military applications. straints for sizing the entire composite wing, which
A composite multispar wing designed for a commer- at a certain level are also considered as flexible, such
cial aircraft, which could be an efficient alternative that their influence to the design is determined.
solution to the conventional 2-spar wing, has neither The preliminary comparison of the 2-spar to the 4-,
been studied, nor been optimised with respect to the 5-, and 6-spar arrangements demonstrates that the
sizing of its structural components. The motivation of multispar configurations can reach up to the signifi-
this study is to investigate if a multispar composite cant mass-saving levels of 10–13 per cent, especially
wing layout can offer advantages, such as stronger when appropriately selected material systems are
and stiffer structural arrangement, reduction of the used in the design. Therefore, multispar design has
number of stringers and cover thickness, and less the potential to be exploited in an airliner composite
structural mass, thus reduction of manufacturing wing construction.
cost.
Based on the above motivation, in this article, 2 DESCRIPTION OF THE COMPUTATIONAL
multispar composite wing layouts comprising 4-, ANALYSIS MODEL
5-, and 6-spar configurations are preliminary sized
and compared to a conventional 2-spar design, in A composite wingbox comprising the exact geometry
order to fully understand the potential advantages of Airbus A330 airliner wing, scaled at 70 per cent of
of multispar configurations. The flow chart of Fig. its original size, is considered for this study. This
1 illustrates the different stages of the optimization scaled wingbox size has been selected to be consid-
process followed during the initial sizing procedure ered for all the investigations in the frame of research
of the conventional, as well as the multispar wing project ALCAS [17], as the wing platform of a future
structure. European full composite airliner. In order to ensure
Numerical finite element (FE) models of a 2-, 4-, 5-, that this assumption does not affect the results, the
and 6-spar wing configuration are developed using comparative study of all the multispar wing configu-
ANSYS FE code [16]. The computational models are rations examined hereafter, refers to the same scaled
fully parametric, so that the determination of stress, wing geometrical layout. Therefore, the calculated
strain, and displacement fields of numerous wing mass of each configuration is directly comparable
structural variations are derived for different geomet- to all the others. The length of the analysed wing
rical sizing parameters (i.e. thickness or areas of the is 21 m (wing span of 44 m) and total wing area
various wing subcomponents), or varying material is 94 m2. In order to reduce the number of sizing
parameters, it is assumed that the position and the
number of ribs are constant for all the configurations;
this assumption does not significantly affect the
results of this study, as justified later by the small
mass percentage of the wing ribs over the total
wing mass. However, it should be mentioned that
in a later design stage (e.g. in the detailed design
phase), a study of the optimum rib number and loca-
tion can be useful.
The covers are modelled as monolithic by suitably
smearing the stiffeners into the cover, while no addi-
tional area is considered in the junction of the spar
flanges to the covers (landing area). The FE models
mainly comprise the 2-, 4-, 5-, or 6-spars (spar flanges
and spar webs are modelled separately), the upper
and lower covers, as well as the ribs, as presented in
Figs 2 and 3(a) to (d).
‘Shell 181’ element type of the ANSYS element
library is used to model the covers, the ribs, and the
Fig. 1 Flow chart of a multispar composite wing spar web panels; it is a four-noded element with six
optimization process degrees of freedom at each node. This element is

Proc. IMechE Vol. 226 Part G: J. Aerospace Engineering

Downloaded from pig.sagepub.com at NANYANG TECH UNIV LIBRARY on May 23, 2015
90 D G Stamatelos and G N Labeas

suitable for analysing thin to moderately thick shell configurations comprises approximately 40 000–
structures and is selected because the cover thick- 50 000 elements, depending on the number of spars.
nesses of composite wings are relatively thicker A detail of the FE models discretization in the area of
compared to the respective metallic structures. The the lower covers access holes is shown in Fig. 4(a). It
flange spars are modelled using ‘Link 8’ element type, should be mentioned that all geometry and material
which is a three-dimensional spar element with three input data are introduced parametrically in the
degrees of freedom at each node, capable of repre- model, in order to enable variations of these param-
senting uniaxial tension–compression behaviour. eters during the optimization procedure. More speci-
The entire FE model of the 2-, 4-, 5-, and 6-spar fically, the material system properties, the cover and
web thickness, the flange areas, and bolt–hole dimen-
sions are all considered independent parameters of
the wing sizing procedure. Between two spars and
two ribs, a cover panel is defined, the thickness of
which is not considered constant. Each panel is
divided into five areas; as it is shown in Fig. 4(b)
each cover panel area can be sized with a different
thickness value, allowing the cover panel tapering
chordwise; therefore, maximum cover mass reduc-
tion can be achieved.
The minimum allowed wing cover thickness is
assumed to be 3.5 mm and the minimum spar
flange area is assumed to be 250 mm2, due to produc-
tion limitations. This lower bounds considered in the
design will be shown to play a very important role in
Fig. 2 External wing structure the total mass of the wing, in the case of high number

Fig. 3 Internal wing geometry for the (a) 2-spar, (b) 4-spar, (c) 5-spar, and (d) 6-spar configurations

Proc. IMechE Vol. 226 Part G: J. Aerospace Engineering

Downloaded from pig.sagepub.com at NANYANG TECH UNIV LIBRARY on May 23, 2015
 Multispar composite wing 91

Fig. 4 (a) Upside-down view, indicating the discretization of lower cover in the area of access holes
and (b) chordwise panel thickness variation

of spars. The boundary conditions of all FE models the wing. Each member of the wing has to satisfy
comprise fixation of all degrees of freedom in the pre-set constraints. More specifically:
wing root.
(a) its strain value should not exceed maximum
The apparent material properties of the Carbon
strain limitations;
Epoxy composite system, which is used as refer-
(b) the compressively loaded members should be
ence, are E11 ¼ E22 ¼ 52 GPa, the shear modulus
loaded at compressive stresses which are lower
G12 ¼ 26 GPa, and the Poisson ratio n12 ¼ 0.4; this
to buckling limits;
basic material system is referred hereafter as MS1.
(c) the riveted/bolted joint areas should be capable
Variations of the basic material system (MS1) are con-
to withstand bearing–bypass loads.
sidered in the multispar investigation, as it is
described in the following sections. The basic set of strain level limits referred hereafter as
As this study aims to a basic comparison of the pre- ‘SL1’ is þ6000 mE for tension, 6000 mE for compres-
liminary design of a conventional 2-spar wing with dif- sion, and 7000 mE for shear loading. It should be men-
ferent multispar wing configurations, only the most tioned here, that these strain limits are far lower than
important and critical wing loading scenario is the strain-to-failure limits of the material. The reason
selected for the analysis, which comprises the aerody- for setting design limits much lower than the material
namic lift and drag pressures combined to the weight failure limits comes from ‘no growth of damage’
of the powerplant. All secondary loading conditions, requirements, e.g. to ensure that a delamination
e.g. landing gear loads, secondary aerodynamic loads, damage will not propagate at the entire lifetime of
etc. are ignored. The parametric FE models have been the structure. As these strain limits are related to the
built using the Ansys FE code parametric design lan- damage tolerance assumptions and maintenance
guage (APDL), which is a scripting language that can procedures, variations of this basic SL1 set, referred
be used to automate processes and build a FE model as SL2 and SL3, which will be discussed in the follow-
in terms of parameters (variables). It includes a wide ing sections, are also considered. In Table 1, the type
range of features, such as repeating commands, of allowable applicable in each structural member of
macros, if-then-else branching, do-loops, as well as the lateral wing is presented.
scalar, vector, and matrix operations. The developed A post-processing routine is developed as an
parametric FE models are solved linearly and the stres- ANSYS macro-routine, in order to perform an auto-
ses, strains, and displacements of the wing structural mated comparison of the analysis of the calculated
components are calculated. results to appropriate allowable. The macro-routine
is used to extract strain/stress results and necessary
3 DESIGN ALLOWABLE CONSIDERED geometric and material data out of the post-proces-
sing model database; consequently, magnitudes of
In order to perform a preliminary structural sizing interest are derived and compared to the pre-set
of the different configurations of the lateral wing limits of the allowable, i.e. maximum permitted
components, the numerically calculated static analy- strain levels, critical buckling loads, bearing–bypass
sis results (stress, strain, and displacement) are com- strength, etc.
pared to respective allowable, in order to check if the The optimization process applied to this investiga-
sizing fulfils the design constraints. Different allow- tion consists of four consequent steps (a) sizing of the
able are considered for the different components of entire wing according to the pre-set strain limit levels,

Proc. IMechE Vol. 226 Part G: J. Aerospace Engineering

Downloaded from pig.sagepub.com at NANYANG TECH UNIV LIBRARY on May 23, 2015
92 D G Stamatelos and G N Labeas

Table 1 Design constraints considered for each wing structural component

Wing Sizing
members constraints
Spars Spar webs Strain limits Shear buckling Bending Combined shear
buckling and bending buckling
Spar flanges Strain limits Bearing–bypass
Covers Upper cover Strain limits Compression Shear Combined compression Bearing–bypass
buckling buckling and shear buckling
Lower cover Strain limits Bearing–bypass
Ribs Strain limits Brazier loading

constrains. The design phase is considered complete,

when a reserve factor higher than unity (but close to
it) is achieved (RF 5 1) for all structural members and
for all design constraints.

4 COMPARISON OF 2-, 4-, 5-, AND 6-SPAR

CONFIGURATIONS

In the initial optimization for all wing configurations,

the basic material system MS1 and the basic strain
level limits set SL1 are considered. The calculated
total wing mass versus the spar number is presented
Fig. 5 Wing mass versus number of spars for MS1 and in Fig. 5. It can be observed from Fig. 5, that as the
Strain Level 1 number of spars increases the total weight of the wing
decreases, which indicates that multispar configura-
tions are potentially beneficial with respect to weight
saving. It should be also observed that the 6-spar
arrangement is slightly heavier than the 5-spar
arrangement, which is due to the structural arrange-
ment at the wing tip. More specifically, in the 6-spar
arrangement, at the wing tip area, the spars are too
closely placed, which is practically unnecessary, con-
sidering the extremely low loading in this area.
However, this situation results to a small mass pen-
alty of the 6-spar design, compared to the 5-spar
design, which is imposed by the lower bounds
imposed to the skin thickness and the flange area
values, due to production limitations, as explained
previously. This small weight penalty is expected to
vanish if the wing tip area is locally designed with one
Fig. 6 Mass of wing covers and spars versus total or two false spars.
number of main spars In Fig. 6, the detailed mass breakdown of the indi-
vidual structural members of the four basic structural
(b) sizing of the upper cover panels according to com- configurations, as sized by the optimization proce-
pression buckling allowable, (c) sizing of the webs dure described in sections 2 and 3, is presented. To
according to shear and bending buckling allowable, better understand the way mass savings are achieved
and (d) sizing of the bolted joint areas according to by increasing the number of spars, the required cover
bearing–bypass analysis constraints. At each optimi- mass and spar mass for each wing structural config-
zation step, after sizing a specific type of structural uration investigated (i.e. 2-, 4-, 5-, and 6-spars) is
members (e.g. spar flanges) according to one con- plotted in Fig. 6. It can be observed from Fig. 6 that
straint, all other structural members (e.g. cover as the number of spars increases the mass of the
panels, spar webs, etc.) are resized with respect to upper cover decreases, reaching up to a 50 per cent
all applicable constraints, in order to avoid oversizing mass reduction in the case of 6-spar arrangement.
and achieve an optimized structure. Several iterations It is evident that the cover panels are relaxed due
are required in order to satisfy all the pre-set to the increase of the stiffness of each wing

Proc. IMechE Vol. 226 Part G: J. Aerospace Engineering

Downloaded from pig.sagepub.com at NANYANG TECH UNIV LIBRARY on May 23, 2015
 Multispar composite wing 93

cross-section, due to the increase of spar number.

However, it is worth to mention that as the number
of spars increases, the increase of spar mass is lower
than the respective decrease in cover mass, leading to
a total mass reduction. This can be explained, if the
sizing criteria of the upper covers are fully under-
stood. The sizing driver of the upper cover is basically
compression buckling as described by equation (1)

2 
N0 ¼ D11 m4 þ 2ðD12 þ 2D66 Þm2 n 2 ða=bÞ2
m2 a 2

þða=bÞ4 D22 ð1Þ

Fig. 7 Detailed mass breakdown of the four wing

In equation (1), Dij are the bending stiffness matrix configurations (MS1 and SL1 considered)
terms, m and n the buckling mode integers, and (a/b)
the panel aspect ratio. As the panel aspect ratio (a/b) Table 2 Comparison of maximum wing tip
increases due to the increase of spar number in the displacement (MS1 and SL1)
cross-section, the critical buckling load gets higher. Maximum
Hence, the panels can be designed with less thick- Examined displacement
ness, resulting to mass savings for the covers up to configurations (mm)
50 per cent. Taking into account that the weight 2-spar 3562
penalty required to fulfill upper cover buckling con- 4-spar 3557
5-spar 3532
straints of the 4-spar arrangement is only 10 per cent 6-spar 3565
of the upper skin mass, it can be understood that the
margin for a potential weight saving relevant to panel
compression buckling is quite essential. achieved, if the sizing is based exclusively on strain
The total mass of the spars increases as the accom- allowable.
modated spar number in the wing box increases, as
presented in Fig. 7. However, there is also a mass 5 EFFECT OF STRAIN LIMIT ALLOWABLE ON
redistribution between spars. It can be observed THE MULTISPAR CONFIGURATION
from Fig. 7, that increase of the spar number results
in a decrease of the front and the rear spar mass up to It is evident that the total mass of the wing depends
60 per cent; this can be justified considering that the strongly on the strain allowable limit values. The
intermediate spars are more loaded, as they are strain limit values considered in this study are far
located closer to the loaded areas. In general, the lower than the material static strain-to-failure
reduction of the skin mass in the multispar arrange- values; the low values selected are exclusively related
ments is relatively higher to the mass penalty caused to ‘damage no growth’ requirements during the entire
by the addition of intermediate spars in the multispar lifetime of the structure and can be relaxed signifi-
configuration, which results to an overall advantage cantly, if modern damage tolerance principles are
of the multispar arrangements. applied and/or if the maintenance intervals are mod-
Finally, in Table 2, the wing tip (maximum) dis- ified accordingly. The first set of strain levels used is
placement of all investigated multispar arrangements ‘SL1’ and corresponds to damage ‘no growth’ design;
is presented. It may be observed from Table 2, that the according to this, the strain is limited to 6000 mE for
maximum lateral displacement of each configuration tension, 6000 mE for compression, and 7000 mE for
is approximately the same, indicating that the total shear loading. In order to investigate if a relaxation
bending stiffness of the wing remains almost con- of strain limits has a strong benefit to multispar
stant in all multispar configurations. This important configurations, two additional strain limit sets have
remark is better understood if a wing cross-section is been considered. The second set of strain levels used
considered; it is evident that when the number of considers limits of þ6500 mE for tension, 6500 mE for
spars increases, the design optimization requires a compression, and 7500 mE for shear loading, and is
mass movement from the skins towards the spar referred hereafter as ‘SL2’. The third set of strain
flanges, with a tendency to keep the overall second levels considers limits of þ7000 mE for tension,
moment of inertia almost constant. This observation 7000 mE for compression, and 8000 mE for shear load-
indicates that no significant mass savings can be ing and is referred hereafter as ‘SL3’. To isolate the

Proc. IMechE Vol. 226 Part G: J. Aerospace Engineering

Downloaded from pig.sagepub.com at NANYANG TECH UNIV LIBRARY on May 23, 2015
94 D G Stamatelos and G N Labeas

effect of the applied strain allowable variation, mate- of 12 per cent, 13 per cent, 15.3 per cent, and
rial system MS1 is kept to all wing configurations in 9.25 per cent, respectively, can be observed, which
this sensitivity study. indicates that the 5-spar configuration is once again
The computational models described in section 2 optimal.
along with the sizing procedure of section 3 are again In Fig. 9, the calculated wing mass, as accumulated
utilized in the preliminary sizing of the 2-spar and the by considering the three different strain level limits
innovative multispar configurations. In Fig. 8, a detail for increasing spar number, is presented. It may be
mass breakdown of the main wing components, con- observed from Fig. 9, that the maximum mass saving
sidering SL2 and SL3 stain limit allowable, is when SL1 is considered is 2.9 per cent (93 kg), which is
presented. relatively small. By relaxing the strain limits, the total
A comparison between the results, as summarized mass saving in the 5-spar arrangement is significantly
in Fig. 8, for the three different allowable strain levels increased, i.e. it becomes 4.2 per cent for SL2 and
yield that the respective wing mass reduces, as 6.7 per cent for SL3.
the strain allowable level relaxes, as expected. From It became obvious from the parametric study, that
Fig. 8, it is also observed that higher reduction can be the value of the allowable strain has a significant role
achieved for the lower cover and spar flanges of the to the final mass of the wing, indicating the signifi-
wing, as they are exclusively sized according to strain cance of the strain constrains over other constrains,
limitations, while less or no influence is achieved for such as buckling or bearing–bypass. More specifi-
the upper covers and the spar webs which are sized cally, as the allowable strain level is relaxed, the
mostly by buckling constraints. strains achieved in the wing structural members
Comparing the sizing of the 2-, 4-, 5-, and 6-spar increase, while the total wing mass decreases.
arrangements for SL1 and SL3, a total mass reduction With respect to production cost issues, a multispar
solution apparently includes more components, e.g.
spars; therefore, its production method should be
carefully selected. One-shot automated manufactur-
ing approaches should be used in order to keep the
cost at the same level of the conventional design.
However, if modern modular manufacturing tech-
niques are considered, a multispar wing design can
have advantages also from a manufacturing cost
point of view.

6 EFFECT OF MATERIAL PROPERTIES ON THE

MULTISPAR CONFIGURATION

It has been understood from the initial multispar

Fig. 8 Mass breakdown for all wing configurations investigations, that at every wing cross-section,
(SL2, SL3, and MS1) when the number of spars increases, a mass move-
ment from the covers towards the spars occurs, with a
final positive balance in the total mass arising when
all design criteria are fulfilled. It can be observed from
Fig. 7, that the spar flanges comprise 27–32.5 per cent
of the total mass of the multispar arrangements,
which means that design should focus on these ele-
ments if mass saving is foreseen. The fact that a high
percentage of the total load is carried by the spars in
the multispar configurations, together with the fact
that the spar flange loading is mostly axial (tensile or
compressive), highlighted that the pseudo-isotropic
material system (MS1), which has been considered
so far in the entire wing volume, cannot lead to an
optimum solution. In this direction, a lay-up which
provides more stiffness in the axial flange direction
Fig. 9 Total wing mass for different strain limit is selected. The apparent material properties for the
allowable spar flange orthotropic material system, referred

Proc. IMechE Vol. 226 Part G: J. Aerospace Engineering

Downloaded from pig.sagepub.com at NANYANG TECH UNIV LIBRARY on May 23, 2015
 Multispar composite wing 95

hereafter as MS2, are E11 ¼ 85 GPa, E22 ¼ 30 GPa, and and 12.6 per cent for 4-, 5-, and 6-spar arrangements,
the shear modulus is G12 ¼ 12 GPa with Poisson respectively, compared to the 2-spar wing structure.
ratio of n12 ¼ 0.4; at all the other wing structural It is of interest to investigate the influence of the
components (covers, spar webs, ribs, etc.) the stiffer spar flanges (MS2) to the mass breakdown of
original material system MS1 (E11 ¼ E22 ¼ 52 GPa, each individual wing components. In Table 3, the per-
G12 ¼ 26 GPa, and n12 ¼ 0.4), has been kept, as in the centage mass reduction for each wing is presented.
previous designs. The component comparison is performed between
The SL1 strain limit set has been kept for the inves- the solutions obtained using the combination of
tigation of the material influence to the potential MS1 þ MS2 and the original material system MS1.
mass benefits, such that results are directly compara- It is concluded from the results of Table 3, that the
ble to those of the basic sizing solution of section 4. spar flange mass has decreased substantially as
In Fig. 10, the detailed mass breakdown results expected; the reduction is between 19 per cent and
obtained using MS2 material for the spar flanges and 22.7 per cent for the multispar and 25.7 per cent for
MS1 material everywhere else are presented. It may be the conventional arrangement. In addition, the stiffer
observed by comparing the results of Figs 7 and 10, material of the spar flanges has caused a remarkable
that the application of spar flanges with higher stiff- saving to cover mass as well, in the multispar
ness in the main (axial) direction (through MS2), arrangements.
results to considerably lighter designs compared to It should be also observed that the 6-spar arrange-
the respective wings of MS1 material system. More ment is slightly lighter than the 5-spar arrangement in
specifically, comparing the original MS1 wing to the this case. The small mass penalty that exists between
combined MS1 þ MS2 wing, the 2-spar configura- the 6-spar and the 5-spar configurations of the orig-
tion becomes slightly lighter, i.e. 3.6 per cent, while inal solution is disappeared, as it is expected. In more
the 4-, 5-, and 6-spar arrangements become signifi- detail, the enhanced material properties of the
cantly lighter, i.e. 12.6 per cent, 12.1 per cent, and flanges give a significant mass saving that overcomes
13.6 per cent, respectively. Comparison of the the mass penalty that the arrangement of the spars at
MS1 þ MS2 wings between themselves indicates sig- the wing tip cause. Therefore, it can be observed that
nificant mass saving of 10.2 per cent, 11.5 per cent, by altering the material system properties to that of
MS2, the total wing mass is reduced.

7 CONCLUSIONS

An investigation of the possible advantages of a mul-

tispar main wing design comprising 4-, 5-, and 6-spar
arrangements in comparison to the conventional
structural arrangement (2-spars) has been per-
formed. A flexible parametric FE model that can be
easily and efficiently modified to produce different
structural multispar layouts of a lateral wing has
been developed. An optimization procedure has
been linked to the parametric FE model, resulting in
Fig. 10 Mass breakdown of multispar configurations the optimal sizing of all wing structural members by
using combination of material system MS1 satisfying all pre-set constraints. For all multispar
and MS2 configurations, a parametric study was also

Table 3 Mass reduction of the individual wing components resulting by adopting combined
MS1 þ MS2 instead of MS1 alone
Structural 2-spar 4-spar 5-spar 6-spar
wing component arrangement arrangement arrangement arrangement
Covers (%) 2.1 12.4 12.7 14.7
Spar flanges (%) 25.7 22.7 19.0 21.3
Spar webs (%) 0.8 0.0 1.8 1.6
Ribs (%) 0.0 0.8 0.0 0.0
Total mass 3.6 12.6 12.1 13.6
of the wing (%)

Proc. IMechE Vol. 226 Part G: J. Aerospace Engineering

Downloaded from pig.sagepub.com at NANYANG TECH UNIV LIBRARY on May 23, 2015
96 D G Stamatelos and G N Labeas

performed, concerning different allowable strain aeroelasticity. Chin. J. Aeronaut., 2008, 21(5),
levels and different tailored material properties. 402–410.
The investigation resulted in the conclusion that a 5 Colson, B., Bruyneel, M., and Grihon, S.
Optimization methods for advanced design of air-
multispar configuration has certain advantages from
craft panels: a comparison. Optimiz. Eng., 2009,
the structural point of view, as it leads to mass savings in Press.
due to the positively changing of the wing panels 6 Aydin Komur, M. and Sonmez, M. Elastic buckling
aspect ratios, which result in higher critical buckling of rectangular plates under linearly varying in-plane
loads, as well as the enhanced bearing–bypass perfor- normal load with a circular cutout. Mech. Res.
mance; these weight savings may become very signif- Commun., 2008, 35(6), 361–371.
icant if the material system and the design philosophy 7 Bisagni, C. and Vescovini, R. Analytical formulation
of the wing are properly selected. Among all the for local buckling and post-buckling analysis of stiff-
examined configurations, the 6-spar wing with tai- ened laminated panels. Thin-Wall. Struct., 2009,
47(3), 318–334.
lored material properties was proven to be the lightest
8 Kim, C. and Yoon, I.-S. The stacking sequence opti-
design, which compared to the conventional 2-spar mization of stiffened laminated curved panels with
arrangement is by 12.6 per cent lighter. The results different loading and stiffener spacing. J. Mech. Sci.
are promising and encourage further investigations of Technol., 2006, 20(10), 1541–1547.
multispar wing configurations. 9 McCullers, L. A. and Naberhaus, J. D. Automated
structural design and analysis of advanced compos-
FUNDING ite wing models. Comput. Struct., 1973, 3(4),
925–935.
The results presented in this study are partially 10 Snell, M. B. and Barlothomew, P. The engineering
optimization of hybrid composite/metallic wing
obtained in the frame of the EU funded research proj-
boxes for buckling and strength constrain. Compos.
ect Advanced Low Cost Aircraft Structure ‘ALCAS’. Struct., 1987, 7(1), 21–58.
The authors wish to acknowledge EU for financial 11 Rothwell, A. Multi-level optimization of aircraft shell
support of this project, as well as AIRBUS-UK for structures. Thin Wall. Struct., 1991, 11(1-2), 82–103.
the specific support of this investigation. 12 Seresta, O. and Adams, D. B. Optimal design of com-
posite wing structures with blended laminates.
ß Authors 2011 Compos. Part B, 2007, 38(4), 469–480.
13 Sunger, E. Approximate calculation of multispar
REFERENCES cantilever and semicantilever wings with parallel
ribs under direct and indirect loading, 1932,
August.
1 Watson, A., Kennedy, D., and Williams, F. W. Design
14 Herbeck, L. and Wilmes, H. Design rules for a CFRP
of structures by a splitting method. Comput. Struct.,
outer wing. In Proceedings of the 23rd ICAS Congress
1999, 70(4), 377–386.
of Aeronautical Sciences 2002 (ICAS 2002, CD),
2 Aston, G. and Williams, F. W. Simplified methods for
pp. 1–10 (ICAS Congress, Toronto).
the buckling analysis of composite multi-spar wing
15 Chintapalli, S. and Elsayed, M. The development of
boxes. Compos. Struct., 1994, 28(2), 215–223.
a preliminary structural design optimization method
3 Baron, J. R. and William, J. G. Survivability of multi-
spar co-cured composite wing structure (1). In Pro- of an aircraft wing-box. Aerosp. Sci. Technol., 2010,
ceedings of the Structural Dynamics and Materials 14, 188–198.
Conference, 1996, pp. 53–61 (Collection of Technical 16 ANSYS user’s manual, Version 10.0, 2004.
Papers - AIAA/ASME/ASCE/AHS/ASC Structures). 17 European Integrated Project ‘Advanced low cost air-
4 Shao, K., Wu, Z., and Yang, C. Analysis and flexible craft structure’ – ALCAS, contract no. AIP4-CT-2005-
structural modelling for oscillating wing utilizing 516092.

Proc. IMechE Vol. 226 Part G: J. Aerospace Engineering

Downloaded from pig.sagepub.com at NANYANG TECH UNIV LIBRARY on May 23, 2015