Professional Documents
Culture Documents
Design, Analysis and Optimization of Thin Walled Semi-Monocoque Wing Structures Using Different Structural Idealization in The Preliminary Design Phase
Design, Analysis and Optimization of Thin Walled Semi-Monocoque Wing Structures Using Different Structural Idealization in The Preliminary Design Phase
www.emeraldinsight.com/1757-9864.htm
IJSI
5,3
Design, analysis and
optimization of thin walled
semi-monocoque wing structures
214 using different structural
Received 29 December 2013
Revised 26 March 2014 idealization in the preliminary
Accepted 29 May 2014
design phase
Downloaded by United Arab Emirates University At 04:45 22 June 2016 (PT)
Odeh Dababneh
Department of Aerospace Engineering, Cranfield University, Bedford, UK, and
Altan Kayran
Department of Aerospace Engineering, Middle East Technical University,
Ankara, Turkey
Abstract
Purpose – In modeling an aircraft wing, structural idealizations are often employed in hand
calculations to simplify the structural analysis. In real applications of structural design, analysis and
optimization, finite element methods are used because of the complexity of the geometry, combined
and complex loading conditions. The purpose of this paper is to give a comprehensive study on the
effect of using different structural idealizations on the design, analysis and optimization of thin walled
semi-monocoque wing structures in the preliminary design phase.
Design/methodology/approach – In the design part of the paper, wing structures are designed by
employing two different structural idealizations that are typically used in the preliminary design
phase. In the structural analysis part, finite element analysis of one of the designed wing configurations is
performed using six different one and two dimensional finite element pairs which are typically used to
model the sub-elements of semi-monocoque wing structures. Finally in the optimization part, wing
structure is optimized for minimum weight by using finite element models which have the same six
different finite element pairs used in the analysis phase.
Findings – Based on the results presented in the paper, it is concluded that with the simplified
methods, preliminary sizing of the wing configurations can be performed with enough confidence as
long as the simplified method based designs are also optimized iteratively, which is what is practiced in
the design phase of this study.
Originality/value – This research aims at investigating the effect of using different one and two
dimensional element pairs on the final analyzed and optimized configurations of the wing structure,
and conclusions are inferred with regard to the sensitivity of the optimized wing configurations with
respect to the choice of different element types in the finite element model.
Keywords Finite element analysis, Aerospace structural design, Structural optimization,
Wing torque box
Paper type Research paper
1. Introduction
International Journal of Structural Most of the thin walled structural members of aerospace vehicles have multi-cell box
Integrity beam configurations made of semi-monocoque construction. Reinforced thin walled
Vol. 5 No. 3, 2014
pp. 214-226 structural parts of aerospace vehicles are regions where significant weight savings can
© Emerald Group Publishing Limited be achieved, if structural optimization tools are used effectively in the preliminary
1757-9864
DOI 10.1108/IJSI-12-2013-0050 design phase of the structural design process.
In modeling reinforced thin walled aerospace sub-structures, structural idealizations Design, analysis
are often employed in simplified methods. Structural idealizations utilized in simplified
methods include certain assumptions with regard to the load carrying capability of thin
and optimization
walled shell structures such as wing skins, ribs, spar webs and reinforcing members
such as spar caps, stringers, etc. Typical structural idealizations that are used in the
simplified structural analysis of reinforced thin walled shell structures are based on
the assumption that thin walled shell members mainly carry either shear loads only or 215
combined shear and axial loads, whereas reinforcements mainly carry axial loads.
These idealizations are widely used in standard textbooks of aerospace structural
design and analysis (Megson, 1990; Peery, 1950; Bruhn, 1973; Niu, 1997, 1999; Howe,
2004). Because of the complexity of the geometry and loading conditions of aerospace
sub-structures, finite element methods are often used in almost all phases of the
structural design and analysis. The use of finite element method necessitates the
Downloaded by United Arab Emirates University At 04:45 22 June 2016 (PT)
profile with a chord length of 1.524 m and aspect ratio of 6, and the wing material is
Al 2024 T3. The design is performed to meet the minimum requirements set forth in the
simplified design load criteria in appendix A of FAR Part 23 (FAA Federal Aviation
Regulations (FARS, CFR 14)). In the simplified method of analysis, the external
aerodynamic load is calculated as a line lift and pitching moment acting span-wise at
the 25 percent chord length. The span-wise distribution of the lift and the pitching
moment are calculated by, ESDU 95010 (1995) and the limit loads are calculated at the
corner points of the airspeed-load factor (V-N) diagram established in accordance with
appendix A of FAR 23. Internal loads are then calculated as sectional bending moment,
pitching moment and shear force acting at the 25 percent chord-line. Figure 1 gives the
line sketch of the wing section that shows the spar and the stringer locations and skin
and web thicknesses which are taken as the design variables. In the design process,
front spar location is allowed to vary between 20 and 25 percent of the chord length,
rear spar location is allowed to vary between 65 and 75 percent of the chord length, and
stringers are allowed to vary between 30 and 50 percent of the chord length.
To simplify the overall model, the trailing edge part behind the rear spar is not
considered in the design process. In addition, only a single stringer is considered on the
upper and lower skin between the front and the rear spar, and rib positions are taken
at constant intervals along the span of the wing, as shown in Figure 2.
In the design process, sheet thicknesses and spar flange and stringer areas are kept
constant in each bay, and they are allowed to change discretely at the rib stations.
Design of the wing structure is performed based on the following criteria for the
structural idealizations employed in the study.
Design criteria for structural idealization 1 (skins and webs carry shear load only and
spar flanges and stringers carry axial stress only). Bending stresses in the spar flanges
X
Figure 1.
Spar and stringer
locations and skin and Front Spar Web
web thickness definitions
Lower Mid-Skin Right Lower Skin
Rib 1
Design, analysis
Rib 2
and optimization
Rib 3
Rib 4
Bay 1 Rib 5
Rib 6
217
Bay 2
Bay 3 Rib 7
Bay 4
Y Bay 5
Figure 2.
Downloaded by United Arab Emirates University At 04:45 22 June 2016 (PT)
X
Bay 6 Rear Spar Rib locations along
Z
Stiffeners the span of the wing
Front Spar
and stiffeners are calculated using the unsymmetric beam bending formulation. Shear
stresses in the wing skins and spar webs are calculated based on the constant wing
skin/spar web shear flow assumption which is in accordance with the first structural
idealization:
• maximum shear stresses in the skins and webs of each bay should be less than
the shear stress allowable;
• maximum axial stress in the spar flange and stringers should be less than the
stress allowable; and
• local shear buckling of the wing skins and spar webs in each bay should be
prevented.
Design criteria for structural idealization 2 (skins and webs carry shear and axial
load and spar flanges and stringers carry axial stress only). Bending stresses in the spar
flanges, stiffeners and wing skins and spar webs are again calculated using the
unsymmetric beam bending formulation. Shear stresses in the wing skins and spar
webs are calculated based on the variable wing skin/spar web shear flow approach. To
account for variable shear flow, distinct shear flows are defined at the entrance and exit
of each spar flange and skin stiffener, as shown in Figure 3:
• maximum Von Mises stresses in the skins and webs of each bay should be less
than the stress allowable;
• maximum axial stress in the spar flange and stringers should be less than the
stress allowable;
q22
q55
Figure 3.
q5,s5,t5 Shear flow definitions
X q2,s2,t2
used in the second
q44 structural idealization
q11 q6,s6,t6 q6 q4,s4,t4
IJSI • combined tension and shear buckling of the lower wing skins should be prevented;
5,3 • combined compression and shear buckling of the upper wing skins should be
prevented; and
• combined bending and shear buckling of the spar webs should be prevented.
During the design process, for strength checks, stresses are calculated at the inboard
218 edge of each bay, whereas for local buckling checks, average stresses are calculated in
the skin and in the spar web panels in each bay. Both continuous and discrete choice of
sheet thicknesses and flange/stringer areas are used. Iterations are performed to reach
to the minimum mass of the wing structure while satisfying all stress and local
buckling constraints. In the discrete approach, standard sheet thicknesses and flange/
stringer areas are used in the iterative solution, and selections are made from lists of
Downloaded by United Arab Emirates University At 04:45 22 June 2016 (PT)
standard sheet thickness and flange/stringer area (Bruhn, 1973; MIL‐HDBK‐5H, 1998).
The initial iterations showed that for both idealizations, best spar and stringer
locations are at 25 percent chord length for the front spar, at 70 percent chord length for
the rear spar, at 50 percent chord length for the upper stringer and at 46 percent chord
length for the lower stringer. These values do not give absolute minimum weights for
each idealization, but they are considered to be the best ones when both structural
idealizations are considered. Results summarized in Table I shows that the use of the
second structural idealization results in approximately 10 kg lighter mass in the final
configuration compared to the use of first idealization.
Figure 4 shows the mass breakdown of the wing configurations designed by
employing both structural idealizations for the minimum maneuvering speed flight
78.60%
64.33%
Figure 4.
Mass breakdown of the
wing configurations 21.65%
designed by using 11.39%
structural idealizations 7.43% 5.80% 6.05% 4.21%
1 and 2
Skins Spar webs Flange Areas Ribs
condition. From Figure 4, it can be seen that because wing skins and spar webs are also Design, analysis
allowed to carry axial load in the second idealization, total skin and spar web mass
obtained by using the second idealization is higher than the total skin and spar web
and optimization
mass obtained by employing the first idealization.
2.2 Finite element analysis of the wing structure using different one and two dimensional
finite element combinations
219
The main objective of the finite element analysis study is to investigate the effect of
using different finite element types on the analysis results of a wing torque box which
is designed using the second structural idealization, based on the external aerodynamic
load calculated at the minimum maneuvering speed and maximum positive load factor.
Finite element analysis of the wing structure is performed by MSC Nastran (Schaeffer,
Downloaded by United Arab Emirates University At 04:45 22 June 2016 (PT)
2001) using the line lift and the pitching moment load used in the simplified method of
analysis. Figure 5 shows the distributed lift and the pitching moment acting through
the lower flange of the front spar, which is located at the 25 percent chord-line, for the
coarsest mesh model which has single element between the rib stations.
Table II summarizes the element pairs that are used to model the one dimensional
and the two dimensional members of the wing structure. Shell and membrane elements
with – R extension are the so-called revised elements of Nastran which also have
drilling degrees of freedom. It should be noted that the distributed line lift and pitching
moment loading necessitates the use of revised membrane elements in the wing ribs,
because with the standard membrane elements in the wing ribs, the distributed pitching
moment cannot be handled accurately.
3.8+002
3.7+002
8.1+003
3.6+002
8.0+003
3.4+002
7.9+003
2.9+002
7.7+003
6.7+003
1.9+002 Figure 5.
7.2–002 Distributed line
5.7+003
lift (N/m) and pitching
Y 5.7–001 moment (N.m/m) acting
X on the wing structure
Z
Y
Figure 6. X
Coarsest mesh of the wing Z
structure-107 elements
Y
Figure 7. X
Finest mesh of the wing
structure-3,097 elements Z
Fine mesh results given in Table IV confirm that Von Mises stresses predicted by the
simplified method are consistently higher than the Von Mises stresses predicted the by
the fine mesh finite element models. One main effect which is not considered in the simplified
beam model is the axial stresses generated in the flanges and skins due to torsion. The effect
of neglecting the torsion on the axial stresses is clearly seen in Table IV, from the increasing
ratio of the Von Mises stresses determined by the finite element solution to the Von Mises
stresses calculated by the simplified method toward the root of the wing.
For the different finite element models listed in Table II, axial stresses calculated at
the centers of bays 2-5 on the upper spar cap of the front spar are compared in Tables V
and VI for the coarsest and finest mesh finite element models, respectively. Tables V
and VI also show the compressive axial stresses determined by the simplified method
using the structural idealization 2.
2.3 Structural optimization of the wing structure using different one and two dimensional
finite element combinations
Structural optimization of the wing structure is performed using the design, sensitivity and
optimization module of MSC Nastran (Vanderplaats R&D, Inc.) by both the coarsest and
the finest mesh finite element models using the gradient based method. For the coarse
mesh case, six different models given in Table II are used in the property optimization,
whereas for the fine mesh case finite element models with membrane elements are not
used. Optimization is performed in two stages. In the first stage, continuous optimization
problem is solved until convergence is achieved. In the second stage, discrete optimum
solution is generated by referencing the standard sheet thicknesses and flange/stringer
areas which are defined as the design variables.
Structural optimization problem is defined as:
The objective function: Minimize the total mass of the wing structure which is subject to
distributed line lift and pitching moment loading.
Stress constraints:
• maximum Von Mises stresses in the skins and webs of each bay should be less
than the stress allowable; and
• maximum axial stress in the spar flange and stringers should be less than the
stress allowable.
• thicknesses of skin and web panels and ribs are forced to decrease bay by bay
from the root to the tip of the wing in a discrete fashion; and
• spar cap and stiffeners areas are forced to decrease bay by bay from the wing root
to the tip of the wing in a discrete fashion.
Design variables:
• For the finite element models with flanges and stringers modeled by the rod
elements, 92 design variables are used. Design variables are taken as the thick-
nesses of wing skins, spar webs, ribs and spar flange and stringer areas.
• For the finite element models which have flanges and stringers modeled with
beam elements, 128 design variables are used. In the beam modeling of spar
flanges and stringers, it is assumed that beams have rectangular cross-sections,
and width and height of beam cross-sections are also taken as design variables.
Wing torque box configurations given in Table I are optimized for minimum weight
with all the design constraints included and for both coarsest and finest mesh models.
Tables VII and VIII show the masses of the optimized wing structures including the
masses of the wing determined using simplified method of analysis based on
unsymmetric beam theory. Since the round-up method is used in the discrete optimization,
optimized masses determined by the discrete variable optimization are higher than the
optimized masses determined by the continuous optimization.
As an example, for the fine mesh rod-shell finite element model, Figure 8 shows the
variation of the mass of the wing with respect to design cycle which refers to the
Model Initial mass (kg) Continuous optimization (kg) Discrete optimization (kg)
masses with all the constraints included. This result is an indication that deflection and
local buckling constraints have significant effect in driving the optimization problem
toward the minimum mass configuration. Based on the results of optimized mass
configurations given in Tables VII-IX, the following conclusions can be drawn:
• Optimized masses obtained with the fine mesh finite element models are higher
than the optimized masses obtained with the coarse mesh finite element models.
Model Initial mass (kg) Continuous optimization (kg) Discrete optimization (kg)
120.0
100.0
80.0
Wing Mass (Kg)
60.0
40.0
20.0
Figure 8.
Mass of the wing versus
0.0
design cycle for the fine
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
mesh rod-shell model
Design Cycle
6.35-003 Design, analysis
Thickness Scalar Plot
5.95-003
and optimization
5.56-003
5.16-003
4.76-003
4.37-003
3.97-003 225
3.57-003
3.18-003
2.78-003
Figure 9.
2.38-003
Thickness scalar plots of
1.99-003
lower skin, spar web and
Downloaded by United Arab Emirates University At 04:45 22 June 2016 (PT)
It should be noted that although the stresses at the centers of the domains of the
bays are lower for the fine mesh models, the maximum stresses in the domains of
each bay are higher in the fine mesh finite element models. Therefore, with the
fine mesh finite element models, optimized masses turn out to be higher than
the optimized masses obtained by the use of coarse mesh finite element models in
the optimization process.
• Results of the optimization study shows that optimized wing masses, determined
by the use of the different finite element models in the optimization process, are
very close to each other with only slight favorable overall mass on behalf of
models which have spar flanges and stringers meshed with beam elements.
• The mass of the wing configuration designed by the simplified method using the
second structural idealization is very close to the optimized masses determined
by the use of fine mesh finite element models in the optimization process.
From Table VII, it is seen that mass of the wing configuration designed by the
simplified method using the second structural idealization has more than 10 kg
mass penalty compared to the optimized masses determined by the coarsest
mesh finite element models.
3. Conclusion
The effect of using different structural idealizations on the design, analysis and
optimization of thin walled semi- monocoque wing structures in the preliminary design
IJSI phase is studied. Simplified hand calculation based design study by employing two
different structural idealizations showed that second structural idealization, which
5,3 assumes that skin panels and spar webs also carry axial stress besides the shear stress,
results in approximately 10 kg less mass in the final configuration compared to the use
of first idealization during the design analysis. Comparison of the stress results
obtained by the finite element analysis and simplified methods showed that simplified
226 method based on unsymmetric beam theory can be best simulated by the coarse mesh
finite element models. Results of the optimization study showed that optimized wing
masses, determined by the use of the different finite element models in the optimization
process, are very close to each other with only slight favorable overall mass on behalf
of models which have spar flanges and stringers meshed with beam elements.
Based on the results presented in this study, it is concluded that with the simplified
methods, preliminary sizing of the wing configurations can be performed with enough
Downloaded by United Arab Emirates University At 04:45 22 June 2016 (PT)
confidence as long as the simplified method based designs are also optimized
iteratively, which is what is practiced in the design phase of this study.
References
Bruhn, E.F. (1973), Analysis and Design of Flight Vehicle Structures, Tri-State Offset Company.
Butler, R. (1998), “The optimization of wing structures”, Aircraft Engineering and Aerospace
Technology, Vol. 70 No. 1, pp. 4-8.
ESDU 95010 (1995), “Computer program for estimation of span-wise loading of wings with
camber and twist in subsonic attached flow”, available at: www.esdu.com
FAA Federal Aviation Regulations (FARS, CFR 14), “FARS PART 23 Appendix A, A.23.1
General”, available at: www.flightsimaviation.com/data/FARS/part_23-appA.html
Howe, D. (2004), Aircraft Loading and Structural Layout, AIAA Publication, VA.
Megson, T.H.G. (1990), Aircraft Structures for Engineering Students, 2nd ed., John Wiley and
Sons, New York, NY.
MIL‐HDBK‐5H (1998), Military Handbook Metallic Materials and Elements for Aerospace Vehicle
Structures, Department of Defense.
MSC Inc., MSC Nastran (2004), Design Sensitivity and Optimization User’s Guide, MSC Inc., MSC Nastran.
Niu, M.C.Y. (1997), Airframe Stress Analysis and Sizing, 2nd ed., Hong Kong Conmilit Press Ltd,
Hong Kong.
Niu, M.C.Y. (1999), Airframe Structural Design, 2nd ed., Hong Kong Conmilit Press Ltd,
Hong Kong.
Peery, D.J. (1950), Aircraft Structures, McGraw Hill, 1st ed., New York, NY.
Schaeffer, H.G. (2001), MSC.Nastran Primer for Linear Static Analysis, MSC Software
Corporation, Santa Ana, CA.
Vanderplaats R&D Inc., DOT Optimization Software, available at: www.vrand.com/contact.html
(accessed March 15, 2009).
Vaughan, R.E. and Daniel M.F. (2004), “The correct use of finite element models for stress analysis of
aircraft”, Annual Forum Proceedings, American Helicopter Society, Vol. 60, Part 1, pp. 140-192.
Corresponding author
Odeh Dababneh can be contacted at: odeh.da@gmail.com