Available online at www.sciencedirect.

com
Available online at www.sciencedirect.com

ScienceDirect
ScienceDirect
Structural
Available
Available Integrity
online
online atProcedia
at 00 (2018) 000–000
www.sciencedirect.com
www.sciencedirect.com
Structural Integrity Procedia 00 (2018) 000–000 www.elsevier.com/locate/procedia

ScienceDirect
www.elsevier.com/locate/procedia
ScienceDirect
Procedia Structural
Structural IntegrityIntegrity
Procedia1300(2018)
(2016)444–449
000–000
www.elsevier.com/locate/procedia

ECF22 - Loading and Environmental effects on Structural Integrity

ECF22 - Loading and Environmental effects on Structural Integrity
Design of Wing Spar Cross Section for Optimum Fatigue Life
Design of Wing Spar Cross Section for Optimum Fatigue Life
XV Portuguese
Khalid Conference
Eldwaib a on Fracture,
*, Aleksandar GrbovićPCF
a 2016, 10-12 February 2016, Paço de
, Gordana Kastratovićb, Mustafa Arcos, Portugal
Aldarwish a
Khalid Eldwaiba*, Aleksandar Grbovića, Gordana Kastratovićb, Mustafa Aldarwisha
Thermo-mechanical modeling ofof aof Belgrade,
high pressure turbine
Faculty of Mechanical Engineering, University of Belgrade, Kraljice Marije 16, 11120 Belgrade 35, Serbia
blade of an
a

Faculty
Faculty bof Mechanical
a
of Transport Engineering,
and Traffic University
Engineering, Belgrade,
University KraljiceVojvode
Marije Stepe
16, 11120 Belgrade
305 11000 35, Serbia
Belgrade, Serbia

airplane gas turbine engine

Faculty of Transport and Traffic Engineering, University of Belgrade, Vojvode Stepe 305 11000 Belgrade, Serbia
b

Abstract
Abstract P. Brandãoa, V. Infanteb, A.M. Deusc*
Aircraft structure is the most obvious example where functional requirements demand light weight and strong structures. Shape
Aircraft
and sizing
a
structure
Department is ofthe
optimization most obvious
Mechanical
are being example
Engineering,
increasingly wherenowadays
Instituto
used functional
Superior forrequirements
Técnico, Universidade
designing demand light
de Lisboa,
lightweight weight
Av. and
Rovisco
structural strong
Pais, structures.
1, 1049-001
components. Shape
Lisboa,
The aim of this
and
paper sizing
is tooptimization are being of
present optimization increasingly used nowadays
I-section integral Portugal
wing sparfor made
designing lightweight
of aluminum structural
2024-T3. Thecomponents. The aim
efficient design, of this
based on
b
paperIDMEC,
optimum is to Department
present
fatigue
of Mechanical
life,optimization
was achieved
Engineering,
of using
I-section
Extended
Institutowing
integral Superior
sparTécnico,
Finite Element made
Method
Universidade
of aluminum
(XFEM)
de2024-T3.
Lisboa, Av.The
and its
Rovisco
ability to
Pais, 1,
efficient
simulate
1049-001
design,
crack
Lisboa,
based
growth on
in
Portugal
optimum
complex
c fatigue
geometry. life,
The was achieved
computations using
were Extended
carried Finite
out in Element
Morfeo/Crack Methodfor (XFEM)
Abaqus and its
software ability
which to simulate
relies on
CeFEMA, Department of Mechanical Engineering, Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais, 1, 1049-001 Lisboa, the crack growth
implementation in
complex
of XFEM. geometry. The computations
Shape optimization of the were carried
aircraft wingout sparin beam
Morfeo/Crack for Abaqus
was conducted
Portugal software which
by comparing relies crack
the fatigue on thegrowth
implementation
lives for
of XFEM.cross
different Shape optimization
section shapes, butof the aircraft
constant wing
cross spar beam
section area ofwastheconducted by comparing
spar. The analysis revealedthethat
fatigue
XFEM crack growth lives
is efficient for
tool for
different cross section shapes, but constant cross section area of the spar. The analysis revealed
complex three-dimensional configurations optimization where extended fatigue life is one of the most important objectives. that XFEM is efficient tool for
complex
Abstract three-dimensional configurations optimization where extended fatigue life is one of the most important objectives.
© 2018 The Authors. Published by Elsevier B.V.
© During
© 2018
2018The The Authors.
their
Authors. Published
operation, by Elsevier
modern
Published by B.V. B.V.
aircraft engine components are subjected to increasingly demanding operating conditions,
Peer-review
Peer-review under
under responsibility
responsibility of of Elsevier
the the
ECF22 ECF22 organizers.
organizers.
especially under
Peer-review the high pressure turbine
responsibility of the(HPT)
ECF22 blades. Such conditions cause these parts to undergo different types of time-dependent
organizers.
degradation,
Keywords: one
aircraft of which
structure; is creep. A
optimization; model Finite
Extended using Element
the finite element
Method; method
fatigue life. (FEM) was developed, in order to be able to predict
the creep
Keywords: behaviour
aircraft structure;ofoptimization;
HPT blades. FlightFinite
Extended data Element
recordsMethod;
(FDR)fatigue
for a life.
specific aircraft, provided by a commercial aviation
company, were used to obtain thermal and mechanical data for three different flight cycles. In order to create the 3D model
needed for the FEM analysis, a HPT blade scrap was scanned, and its chemical composition and material properties were
1. obtained.
Introduction The data that was gathered was fed into the FEM model and different simulations were run, first with a simplified 3D
1. Introduction
rectangular block shape, in order to better establish the model, and then with the real 3D mesh obtained from the blade scrap. The
The shape
overall expectedoptimization
behaviour inistermsbeing of increasingly
displacement was usedobserved,
to design lightweight
in particular structural
at the components.
trailing edge of the blade.It Therefore
is also being
such a
used The
model shape
to can
develop optimization
localized
be useful is being
shape
in the goal increasingly
strategies
of predicting used the
to restore
turbine blade to design
life, given alightweight
operational structural
set ofavailability
FDR data. components.
of ageing structural It is also being
components.
used to develop
However, localizedto shape
it is important strategies
note that all newtoaircraft
restoredesign
the operational availability
and structural of ageing
changes made structuralaircraft
to in-service components.
require
a damage tolerance analysis as outlined in the US Joint Services Structural Guidelines JSSG-2006 (1988),require
However,
© 2016 it
The is important
Authors. to note
Published bythat all
Elseviernew aircraft
B.V. design and structural changes made to in-service aircraft which
Peer-review
astates
damage all under
that tolerance responsibility
analysis as
safety-of-flight of the structures
Scientific
outlined
critical in theCommittee
US Joint
should beofServices
PCF 2016.
designed Structural Guidelines
using a damage JSSG-2006
tolerance analysis.(1988), which
The purpose
states that all safety-of-flight critical structures should be designed using a damage tolerance analysis.
of this requirement is to ensure that any cracks present in the structure will not cause loss of the structure for some The purpose
ofKeywords: High Pressure
this requirement Turbine
is to ensureBlade;
thatCreep; Finite Element
any cracks Method;
present in the3Dstructure
Model; Simulation.
will not cause loss of the structure for some

* Corresponding author.
* E-mail
Corresponding
address:author.
kedwaib@yahoo.com
E-mail address: kedwaib@yahoo.com
2452-3216 © 2018 The Authors. Published by Elsevier B.V.
Peer-review underThe
2452-3216 © 2018 responsibility of theby
Authors. Published ECF22 organizers.
Elsevier B.V.
Peer-review underauthor.
* Corresponding responsibility
Tel.: +351of218419991.
the ECF22 organizers.
E-mail address: amd@tecnico.ulisboa.pt

2452-3216 © 2016 The Authors. Published by Elsevier B.V.

Peer-review under responsibility of the Scientific Committee of PCF 2016.
2452-3216  2018 The Authors. Published by Elsevier B.V.
Peer-review under responsibility of the ECF22 organizers.
10.1016/j.prostr.2018.12.074
 Khalid Eldwaib et al. / Procedia Structural Integrity 13 (2018) 444–449 445
2 Author name / Structural Integrity Procedia 00 (2018) 000–000

predetermined period of in-service operation. Indeed, it is essential that structures should be designed such that at no
time in its operational life will the residual strength of the structure fall beneath limit load, according to Jones R. et
al. (2004).
Several types of wing spar structure have been studied and optimized in the literature (Ajith V. S. et al. (2017)).
The efficient design was achieved by the use of strength of material approach. Girennavar M., et al. (2017)
considered a wing spar as a beam with discrete loads at different stations. The design was carried out as per the
external bending moment at each station. Weight optimization of the spar was carried out by introducing lightening
cut-outs in the web region. Attempt has been made by Datta, D., and Deb, K. (2006), to design the optimum cross-
sections for load-carrying structures, using a multi-objective evolutionary algorithm, for simultaneously maximizing
moment of inertias and minimizing the cross-sectional area. According to Jones R. et al. (2004) when designing a
fatigue optimized structure, it was essential not only to reduce the peak stress but also to ensure that the critical
crack lengths associated with cracks at the critical design feature were not reduced. An approach to the optimization
of the thin-walled cantilever open section beams subjected to the bending and to the constrained torsion was
considered by Anđelić, N., and Milošević-Mitić, V. (2007). To reduce the induced undesired stresses, a load
carrying beam like wing spar should has its related area moment of inertia as large as possible. However, an increase
in such moment of inertia comes with an increase in the transverse cross-sectional area and hence, the weight of the
spar. Therefore, the maximization of moment of inertia should not take place at the cost of the excessive weight of
the spar (Datta, D., and Deb, K. (2006)).
The aim of research presented in this work was to design an optimized shape and size of a wing spar cross
section based on fatigue life obtained for fatigue crack propagation phase. The task of achieving the optimal design
was carried out by maximizing moment of inertia at constant cross-section area of the wing spar. This was done
through a several case studies.

2. Optimization methodology

Under service loading, characterized by many load cycles, fatigue cracks initiate from the most severe stress
concentrators (i.e. riveted holes) on the differential wing spar. This has been observed in experimental work of
Petrašinović, D. et al. (2012), in which fatigue life was determined for 2024-T3 spar of light aircraft. Fatigue cracks
appeared in the bottom caps (flanges) and then grew in a direction perpendicular to the spar web until total caps’
failure. Failure like this can lead to catastrophic consequences during the flight if the crack is not detected and cap
repaired. Integral structures are more resistive to crack appearance since they don’t have many stress concentrators,
but their shape must be carefully chosen to provide reasonable fatigue life after the crack initiation.
Shape optimization of the integral spar was conducted by comparing the fatigue crack growth life for three
different cross section shapes and same cross section area of the spar beam (idea was to keep constant mass of the
spar). The first analyzed shape was I-section (used for modelling main integral spar – case A) with the same
dimensions as the differential spar (riveted structure) used in work of Petrašinović, D. et al. (2012). The second
shape was a channel-section (U-section spar – case B) and, finally, the third shape was I-section with intermediate
cap (I-section with a cap spar– case C). All three shapes with dimensions are shown in Fig. 1, 2 and 3.
First two shapes are well known and are frequently used in spar design, but I-section with a cap is somehow
unusual and rarely used before. This shape has additional, intermediate cap at an assumed height from the bottom
cap. In the event of fatigue crack appearance at the bottom cap, this cap may fail but the top cap, web and the
intermediate cap should remain intact and spar could carry designed load. In this paper, case C is represented
through three subcases, all with different dimensions (size optimization was objective, too). These subcases were
named C1, C2 and C3. The overall dimensions of the models C1, C2 and C3 are shown in Table 1.

3. Dimensions of analyzed models and applied displacements

In order to applied adequate displacements at free ends of spar beam models, moment of inertia (𝐼𝐼) for each
cross section was extracted from CATIA v5. Then, the required displacement (Δ) for each model was calculated
using the well-known equation:
446 Khalid Eldwaib et al. / Procedia Structural Integrity 13 (2018) 444–449
Author name / Structural Integrity Procedia 00 (2018) 000–000 3

𝑃𝑃𝐿𝐿3
∆= 3𝐸𝐸𝐸𝐸 (1)

where P is the load applied on the spar (same for all models); E is the modulus of elasticity (same for all models,
material 2024-T3); L is length of the spar (same for all models) and I is the moment of inertia of chosen cross
section (values shown in Table 2).
3 b
5
1. 0 1 a
1 1.
6 1
6
b a
4. 18. 3
4 4. 3
2
3

H a
10 6 4 10
b 5
0 1. 0 5 1. 0
5
0 0

a
b 2
25 2 a
18. b
.7 4
4 4
a
1. 1. b 1
6 6 1
Fig. 1 Cross section of main integral spar Fig. 2 Cross section of U spar Fig. 3 Dimensions of I-section with intermediate cap

Table 1 The overall dimensions for I-section with a cap spar (shown in Figure 3).

Case a1 b1 a2 b2 a3 b3 a4 b4 a5 b5 H
C1 1.6 41.0 1.6 21.0 13.4 4.2 23.4 4.2 58.4 1.0 100.0
C2 1.6 54.7 1.6 54.7 5.0 3.0 5.0 3.0 85.2 1.0 100.0
C3 1.6 53.6 1.6 53.6 5.0 3.0 5.0 3.0 90.2 1.0 105.0
# All dimensions in mm

Table 2 Moments of inertia and appropriate displacements for different spars

Spar category (type) No. Case No. I (mm4) Displacement (mm)

Main integral spar 1 Case A 654,900 3.000
U-section spar 2 Case B 558,300 3.519
3 Case C1 583,500 3.367
I-section with a cap spar 4 Case C2 662,000 2.968
5 Case C3 726,300 2.705

Value of displacement 3mm for main integral spar was chosen because exactly the same displacement was
observed in experimental work of Petrašinović, D. et al. (2012).

4. Results of fatigue crack growth simulations

All simulations of cracks’ growths were performed using Morfeo/Crack for Abaqus software. This software was
verified by many authors including Eldwaib, K. A. et al. (2017), and it proved to be useful tool in fatigue crack
propagation analyses. In the analyses of defined spar models, the fatigue crack initiation stage was ignored since the
emphasis was on the crack growth phase only. Two initial penny shaped cracks were inserted in all cases: one in the
right lower cap (flange) and another in the left lower cap next to fixed end of the spar beam, where the most severe
 Khalid Eldwaib et al. / Procedia Structural Integrity 13 (2018) 444–449 447
4 Author name / Structural Integrity Procedia 00 (2018) 000–000

stress concentration occurs. Two cracks were propagated simultaneously in cases A and C because in work of
Petrašinović, D. et al. (2012) two cracks appeared on lower caps during the experiment with aluminium spar. In the
case B one initial crack was used because U-section spar had only one lower cap. In this paper comparisons are
made for cracks initiated at the same spots (left lower cap). To determine the number of cycles of displacement that
will grow cracks to certain lengths, Paris law was integrated (Schijve, J. (2008)) using material coefficients 𝑚𝑚 = 3.2
and 𝐶𝐶 = 2.382 × 10−12 and stress ratio 𝑅𝑅 = 0.15. Values of fatigue life obtained in simulations were later used to
attain the optimized shape and size of the spar cross sections.
4.1 Comparison of fatigue lives (no. of cycles) for different cases
The variation of crack length “a” vs. number of cycles “N” of applied displacement is shown in Figure 4 for all
analysed cases. It can be seen that the shortest fatigue life was obtained in case B where moment of inertia of spar
beam cross section was lowest (558,300mm4) and the applied displacement was highest (3.519mm). There is almost
constant difference (approximately 100,000) between number of cycles in case A and case B, from cracks’ lengths 2
mm up to 21mm. Although the moment of inertia in case C1 is little bit higher than in case B and displacement is
approximately 12% higher than in case A (see Table 2), number of cycles in C1 is a bit less than number in A, with
the difference 5,000 – 20,000 cycles during the cracks’ growth. This is result of C1 geometry consistency and is a
direct consequence of reinforcement achieved by adding additional flange above the lower flange.
In case C2 number of cycles was little bit less than that in case A until cracks reached 6mm; after that, number
of cycles in case C2 started to be bigger and bigger and for length 20mm difference was about 94,000 cycles. This
was result of modifications in the size and position of intermediate flange (see Table 1) that led to stronger spar and
more fatigue resistant lower area of cross section. It is worth mentioning that there was no significant difference in
moments of inertia in cases C2 and A and, consequently, applied displacement were almost the same.
Finally, Figure 4 shows that the longest fatigue life occurred in case C3 for which number of cycles at a=21mm
is more than 1,000,000 cycles bigger than that in Case A. This fatigue life was obtained after height H of cross
section C2 was increased from 100mm to 105mm (see Table 1) along with the increase of dimension a 5 from
85.2mm to 90.2mm. Newly obtained cross section C3 had significantly larger moment of inertia and consequently
lower displacement (Δ=2.705mm) which – at the end of the day – led to the longest fatigue life.
22
20 main integral spar (case A)
18 U-section spar (case B)
16 spar with intermediate flange (case C1)
14
crack length (mm)

spar with intermediate flange (case C2)

12
10 spar with intermediate flange (case C3)
8
6
4
2
0
0 200000 400000 600000 800000 1000000 1200000 1400000
No. of cycles
Fig. 4 Comparison of the fatigue lives (number of load cycles) for different cases

4.2 Comparison of fatigue crack growth rates (𝑑𝑑𝑑𝑑/𝑑𝑑𝑑𝑑) for different cases
The slope of the crack growth curve 𝑑𝑑𝑑𝑑/𝑑𝑑𝑑𝑑 (better known as crack growth rate - CGR) is good indicator of how
damaged structure is going to behave under given loading conditions. Using the data obtained from Abaqus, i.e. the
coordinates of the nodes on the crack front and number of cycles evaluated for each (incremental) step of crack
propagation, d𝑎𝑎 ⁄d𝑁𝑁 for each case was estimated. For that purpose, equation (2) was used:
448 Khalid Eldwaib et al. / Procedia Structural Integrity 13 (2018) 444–449
Author name / Structural Integrity Procedia 00 (2018) 000–000 5

d𝑎𝑎 (𝑎𝑎𝑖𝑖+1 −𝑎𝑎𝑖𝑖 )

( ) = (2)
d𝑁𝑁 𝑖𝑖 𝑁𝑁𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐
d𝑎𝑎
where ( ) is crack growth rate per each propagation step; 𝑖𝑖 is step number, and 𝑁𝑁𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 is number of estimated
d𝑁𝑁 𝑖𝑖
load cycles per each step. Obtained results are presented in Figure 5 and Figure 6

4.5E-04
Main integral spar
4.0E-04
U-section spar
3.5E-04
Intermediate flange (1)
3.0E-04
da/dN (mm/cycles)

Intermediate flange (2)

2.5E-04
Intermediate flange (3)
2.0E-04
1.5E-04
1.0E-04
5.0E-05
0.0E+00
0 1 2 3 4 5 6 7 8 9 10 11 12
crack length (mm)
Fig. 5 Comparison of d𝑎𝑎⁄d𝑁𝑁 calculated per each step of crack propagation (crack length 0 – 12mm)

2.4E-03
2.3E-03 Main integral spar
2.2E-03
2.1E-03 U-section spar
2.0E-03
1.9E-03 Intermediate flange (1)
1.8E-03
1.7E-03
1.6E-03 Intermediate flange (2)
1.5E-03
1.4E-03 Intermediate flange (3)
da/dN (mm/cycles)

1.3E-03
1.2E-03
1.1E-03
9.5E-04
8.5E-04
7.5E-04
6.5E-04
5.5E-04
4.5E-04
3.5E-04
2.5E-04
1.5E-04
5.0E-05
12 13 14 15 16 17 18 19 20
crack length (mm)
Fig. 6 Comparison of d𝑎𝑎⁄d𝑁𝑁 calculated per each step of crack propagation (crack length 12 – 20mm)

In Figures 5 and 6 significantly higher CGR values can be observed for cases A, B, and C1, compared to small
CGRs in cases C2 and C3. For small cracks’ lengths (Fig. 5), differences in CGR values are not significant, but after
3mm cases C2 and C3 become very different from the other three. Cases A, B and C1 have similar CGR values until
8 mm and then main integral spar values (case A) start to grow faster than the others. Differences become even
greater when cracks reach lengths higher than 12mm (Fig. 6), while the greatest difference can be seen at a=20mm:
CGR in case C3 is about 2 × 10−4 𝑚𝑚𝑚𝑚⁄𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 , though CGR in case A is almost ten times bigger ( 2.2 ×
10−3 𝑚𝑚𝑚𝑚⁄𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 ).
 Khalid Eldwaib et al. / Procedia Structural Integrity 13 (2018) 444–449 449
6 Author name / Structural Integrity Procedia 00 (2018) 000–000

It should be noted that in the cases where two cracks were propagated simultaneously (A, C1, C2, C3) behaviour of
the second crack (on the lower right cap) varied from case to case. For example, in case C2 second crack met first
(Figure 7) after 27 steps of propagation (cracks were propagated in steps of approximately 1mm in all five cases) and
newly formed crack continued to grow in vertical spar wall, while in case A deformation of the spar prevented the
growth of the second crack after 8 steps. All these findings prove that geometry of cross section significantly influences
fatigue life of damaged structure; therefore, this matter clearly needs to be given more attention in the future.

Fig. 7 Two cracks met on the bottom cap (case C2)

5. Conclusions

In aircraft design, cracks’ appearance on the wing parts is allowed (as defined in so-called fail-safe design
philosophy), but life until critical crack length must be evaluated with high confidence in order to prevent undesirable
consequences. That’s not an easy job because in most cases cracks appear on parts with complex geometry subjected to
variable loads during service and data on critical lengths are generally available only for simpler load/geometry
configurations. This is reason why emphasis – in a research presented here – was on developing XFEM based
computational method for successful crack growth life estimation. On the other hand, main goal was to obtain
optimized size and shape (for given mass of the spar) that will significantly extend fatigue life.
Five different shapes had been analysed (I-section, U-section and three variants of I-section with intermediate
cap (flange)) and it was found that the optimum spar with the longest crack growth life and the lowest crack growth
rate had I-section with intermediate cap shape (case C3). Fatigue life of this spar was nearly 1,400,000 cycles, while
in the next best case (C2) life was significantly shorter (450,000 cycles). Typical I-section had the lowest fatigue life
(250,000 cycles) and let’s not forget that this is spar shape used in more than 90% of the light airplanes nowadays.
As the results of this research suggest, to increase fatigue life of light aircraft spar with intermediate cap should
be used. It was proven that in the event of crack initiation at the bottom cap, where cracks usually appear, this cap
will fail sooner or later, but the top cap, web and the intermediate cap will remain intact and spar could carry
designed load much longer than any other configuration.

References

JSSG-2006, Joint Service Specification Guide, Aircraft Structures, Dept. of Defense, October 1988.
Ajith, V. S., Paramasivam R., Vidhya, K. (2017). Study of optimal design of spar beam for the wing of an aircraft. International Journal of
Engineering Development and Research, 5(3), 179-193.
Jones, R., Pitt, S., & Peng, D. (2004). Structural optimisation for light weight durable structures. In Structural Integrity and Fracture International
Conference (SIF'04) (pp. 171-178).
Datta, D., and Deb, K. (2006). Design of optimum cross-sections for load-carrying members using multi-objective evolutionary algorithms. Int. J.
of Systemics, Cybernetics and Informatics (IJSCI), 57-63.
Girennavar M., et al. (2017). Design, Analysis and Testing of Wing Spar for Optimum Weight. International Journal of Research and Scientific
Innovation (IJRSI), Volume IV, Issue VII, 104-112.
Anđelić, N., and Milošević-Mitić, V. (2007). An approach to the optimization of thin-walled cantilever open section beams. Theoretical and
applied mechanics, 34(4), 323-340.
Eldwaib, K. A., Grbovic, A., Kastratovic, G., Radu, D., Sedmak, S. (2017). Fatigue life estimation of CCT specimen using XFEM. Structural
Integrity and Life, 17(2), 151-156.
Petrašinović, D., Rašuo, B., Petrašinović, N. (2012). Extended finite element method (XFEM) applied to aircraft duralumin spar fatigue life
estimation. Technical Gazette, 19(3), 557-562.
Schijve, J., Fatigue of Structures and Materials, 2nd edition, Springer (2008).