Available online at www.sciencedirect.

com
Available
Available online
online at at www.sciencedirect.com
www.sciencedirect.com
Available online at www.sciencedirect.com

ScienceDirect
ScienceDirect
Structural
Procedia
Structural Integrity
Structural Procedia
Integrity
Integrity Procedia 00(2016)
(2016)
2 (2016)
00 000–000
3073–3080
000–000
www.elsevier.com/locate/procedia
www.elsevier.com/locate/procedia
Structural Integrity Procedia 00 (2016) 000–000
www.elsevier.com/locate/procedia

21st European Conference on Fracture, ECF21, 20-24 June 2016, Catania, Italy
21st European Conference on Fracture, ECF21, 20-24 June 2016, Catania, Italy
Fatigue Crack Growth Under Variable Amplitude Loading Through
Fatigue Crack
XV Portuguese Growth
Conference on Under
Fracture, Variable
XFEM
PCF Amplitude
2016, 10-12 February 2016, LoadingPaço deThrough
Arcos, Portugal
XFEM
Thermo-mechanical Haydarmodeling
Dirika,b , of
a,b
a
Tuncay high pressure
Yalçinkaya a,∗
a,∗
turbine blade of an
Haydar Dirik , Tuncay Yalçinkaya
airplane gas (TAI),
turbine engine
a
Middle East Technical University, Department of Aerospace Engineering, Ankara 06800, Turkey
Turkish b
Aerospace Industries 06980 Ankara, Turkey
a
Middle East Technical University, Department of Aerospace Engineering, Ankara 06800, Turkey
b Turkish Aerospace Industries (TAI), 06980 Ankara, Turkey
P. Brandãoa, V. Infanteb, A.M. Deusc*
AbstractaDepartment of Mechanical Engineering, Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais, 1, 1049-001 Lisboa,
Abstract Portugal
Predicting
b fatigue crack growth (FCG) rate and path under variable amplitude loading (VAL) is a crucial issue in damage tolerant
IDMEC, Department of Mechanical Engineering, Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais, 1, 1049-001 Lisboa,
design commonly used in aerospace industry. The aim of the current Portugalstudy is to predict FCG life under VAL through Extended
Predicting fatigue crack growth (FCG) rate and path under variable amplitude loading (VAL) is a crucial issue in damage tolerant
Finite
c Element Method (XFEM) and to explicitly illustrate both FCG life and crack propagation. For this purpose an algorithm is
design commonly used in aerospace industry. The aim of the current studyUniversidade
CeFEMA, Department of Mechanical Engineering, Instituto Superior Técnico, de FCG
is to predict Lisboa, Av.under
life Rovisco Pais,
VAL 1, 1049-001
through Lisboa,
Extended
developed and integrated in ABAQUS software to analyze 3D crack Portugal
propagation under VAL using Modified Generalized Willen-
Finite Element Method (XFEM) and to explicitly illustrate both FCG life and crack propagation. For this purpose an algorithm is
borg (MGW) retardation model. The results are compared with NASGRO crack propagation software and experimental FCG test
developed and integrated in ABAQUS software to analyze 3D crack propagation under VAL using Modified Generalized Willen-
data on 7075-T6 aluminum alloy under various over load (OL) and over load-under load (OL-UL) conditions which exhibit a good
borg (MGW) retardation model. The results are compared with NASGRO crack propagation software and experimental FCG test
Abstract
agreement.
data on 7075-T6 aluminum alloy under various over load (OL) and over load-under load (OL-UL) conditions which exhibit a good
c 2016 The Authors. Published by Elsevier B.V.
agreement.
Copyright
During©their 2016 The Authors.modern
Published by Elsevier B.V.components
This is an
Peer-review underoperation,
responsibility aircraft
of the engine
Scientific Committee of open access
are
ECF21. article under
subjected the CC BY-NC-ND
to increasingly licenseoperating conditions,
demanding
c especially
2016 The Authors. Publishedturbine
the high pressure by Elsevier
(HPT)B.V.
(http://creativecommons.org/licenses/by-nc-nd/4.0/). blades. Such conditions cause these parts to undergo different types of time-dependent
Peer-review
Peer-review
Keywords:
degradation,under
underresponsibility
one
Fatigue of which
crack growth;isofvariable
responsibility the Scientific
of the
creep. Committee
Scientific
Aamplitude
model using ofXFEM;
Committee
the
loading; ECF21.
of element
finite ECF21. method (FEM) was developed, in order to be able to predict
the creep behaviour of HPT blades. Flight data
Keywords: Fatigue crack growth; variable amplitude loading; XFEM; records (FDR) for a 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
obtained. 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
1. overall
Introduction
expected behaviour in terms of displacement was observed, in particular at the trailing edge of the blade. Therefore such a
The FCG
model can belife of structural
useful in the goal of components in aerospace
predicting turbine blade life,industry under
given a set constant
of FDR data. amplitude loading (CAL) can be
determined
The FCGwith lifegood accuracy components
of structural by using various methods existing
in aerospace industryinunder
literature. The early
constant pioneering
amplitude loadingwork of Paris
(CAL) canand
be
© 2016
Erdogan The Authors.
(1963) Published
introduced a by Elsevier
stress B.V.
intensity factor (SIF) based empirical relation (Paris equation)
determined with good accuracy by using various methods existing in literature. The early pioneering work of Paris and for FCG analysis
Peer-review
which was under responsibility
modified by considering of the Scientific
ofCommittee
mean of PCF(see
2016.
Erdogan (1963) introduced a stress the effect factor
intensity the (SIF) stress
based Walker
empirical (1970)).
relation Forman
(Paris et al. (1967)
equation) for FCG introduced
analysis
the parameter
which was of critical
modified by SIF for
consideringpredicting
the the
effect final
of thefracture
mean regime.
stress Hartman
(see Walker and Schijve
(1970)). (1970)
Forman et suggested
al. (1967) a modified
introduced
Keywords: High Pressure Turbine Blade; Creep; Finite Element Method; 3D Model; Simulation.
form of Paris of
the parameter law by adding
critical SIF fora predicting
parameter theof threshold SIFregime.
final fracture range and
HartmanElberand
(1971) modified
Schijve (1970)the Paris equation
suggested by
a modified
considering crack closure concept. The NASGRO equation is used in this study for FCG
form of Paris law by adding a parameter of threshold SIF range and Elber (1971) modified the Paris equation by rate determination, which
is based on the
considering crack
crack growth
closure equation
concept. The according
NASGRO to Forman
equation(1992)
is usedconsidering
in this studyplasticity-induced crack closurewhich
for FCG rate determination, using
the crack opening function introduced by Newman (1984). Through these enhancements
is based on the crack growth equation according to Forman (1992) considering plasticity-induced crack closure usingof Paris equation, crack
propagation under CAL can be predicted with great accuracy. However, structural components
the crack opening function introduced by Newman (1984). Through these enhancements of Paris equation, crack especially in aircrafts
are commonly
propagation undersubjected
CAL can to variable
be predictedamplitude loading
with great (VAL)However,
accuracy. during their service
structural life. The problem
components especiallyof in
predicting
aircrafts
FCG life under VAL is still challenging due to load sequence effect. Neglecting the effect
are commonly subjected to variable amplitude loading (VAL) during their service life. The problem of predicting of load cycle in FCG
FCG life under VAL is still challenging due to load sequence effect. Neglecting the effect of load cycle in FCG
* Corresponding author. Tel.: +351 218419991.
∗ E-mail address:
Corresponding amd@tecnico.ulisboa.pt
author. Tel.: +90-312-210-4258 ; fax: +90-312-210-4250.
∗ E-mail address: yalcinka@metu.edu.tr
Corresponding author. Tel.: +90-312-210-4258 ; fax: +90-312-210-4250.
2452-3216 © 2016 The Authors. Published by Elsevier B.V.
E-mail address:
2452-3216 c 2016 yalcinka@metu.edu.tr
The Authors. Published by Elsevier B.V.Committee of PCF 2016.
Peer-review

Copyright ©under
under responsibility
2016 responsibility
The Authors. of
of by
Published
the Scientific
Elsevier B.V. This is an open access article under the CC BY-NC-ND license
Peer-review the Scientific Committee of ECF21.
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
2452-3216 c 2016 The Authors. Published by Elsevier B.V.
Peer reviewunder
Peer-review under responsibility
responsibility of Scientific
of the the Scientific Committee
Committee of ECF21.
of ECF21.
10.1016/j.prostr.2016.06.384
3074 Haydar Dirik et al. / Procedia Structural Integrity 2 (2016) 3073–3080
2 Haydar Dirik and Tuncay Yalçinkaya / Structural Integrity Procedia 00 (2016) 000–000

calculations under VAL leads to inaccurate life predictions. Therefore much attention has been paid to the FCG
analysis of structural components under VAL as well. Many models have been proposed to include the effects of OL
and UL in crack propagation analysis. The first category of these models is based on plastic zone correction theory
in the vicinity of crack tip such as Wheeler (1972) and Willenborg model (see Willenborg and Wood (1971)). The
second category is known as crack closure models (see e.g., Elber (1971), Newman (1981)). In addition to predicting
FCG life in terms of load cycle, predicting the crack growth path is also an crucial issue. To estimate both FCG
life and crack growth path, numerical methods such as Finite Element Method (FEM) can be used (see e.g.,Solanki
et al. (2003)) but the mesh should conform to the crack geometry and the mesh around the crack tip must be updated
whenever crack growth occurs, so the computational efficiency is low. In addition to the FEM, a new finite element
technique called XFEM is developed by Belytschko and Black (1999). The mathematical background behind XFEM
is partition of unity concept presented by Melenk and Babuska (1996). Using the concept of local partition of unity
makes it possible to enrich the finite element approximation space. This enrichment in XFEM allows to represent
discontinuities and singularities around the crack independent from mesh. The main advantage of XFEM compared
to conventional FEM is that, it is not necessary to generate a mesh that conforms to the crack boundaries to represent
the geometric discontinuity. By using this advantage, it has been successfully applied to the modeling of crack growth
and validated in various works (see e.g. Amiri et al. (2013), Baietto et al. (2013), Varfolomeev et al. (2014)). The
capability of the structure to withstand the crack propagation without catastrophic failure must be shown by testing to
fulfil certification rules in global aviation industry. Since testing facilities is expensive and laborious, there is always
need for reliable tools which is verified with reliable test results in damage tolerant design of structures for FCG life
prediction to minimize the number of tests.
In this study, an algorithm is developed and implemented in ABAQUS software to meet the need of reliable tool
for FCG simulations that makes crack propagation simulations and FCG life prediction under VAL by combining SIF
prediction capability through XFEM, NASGRO crack growth equation and MGW retardation model. Using NASGRO
equation in FCG analysis gives us the advantages of covering the effect of threshold SIF range and critical SIF
compared to Paris equation which couldn’t cover these effects. MGW model has good accuracy in covering the effect
of VAL and easy to implement (see Nasgro (2002)). Stationary crack modeling methodology is used in ABAQUS
and crack is propagated automatically by FORTRAN script. The estimated fatigue life is compared with experimental
data on 7075-T6 aluminum alloy which is presented in Porter (1972). Implementation of a retardation model to the
XFEM-based automatized crack growth procedure to account for load sequence effect is a new contribution to the
literature. The efficiency of the methodology is illustrated on a simple component test and its performance will be
tested on more complex industrial components in the near future.

2. Theoretical background

2.1. Crack growth and retardation law used in this study

In this study, NASGRO equation of FCG is used which takes into account the mean stress effect, the threshold
value of the SIF range and the fracture toughness of the material. NASGRO equation is expressed as

∆Kth p
da 1− f (1 − )
= C( ∆K) n ∆K (1)
dN 1−R ∆Kmax p
(1 − )
∆Kc

where N is the number of applied load cycles, a is the crack length, ∆K is the SIF range defined as Kmax − Kmin , ∆Kth
is the threshold SIF range, Kc is the fracture toughness of the material. C, n, p, and q are empirically derived material
constants f is the parameter for closure effect.
The loading history should be considered in determination in FCG analysis. Significant accelerations or retardation
can occur in crack growth rate due to OL and UL. Modified Generalized Willenborg Model (MGW) which is used
in this study origins from Willenborg model. Willenborg model (see Willenborg and Wood (1971) ) was updated
 Haydar Dirik et al. / Procedia Structural Integrity 2 (2016) 3073–3080 3075
Haydar Dirik and Tuncay Yalçinkaya / Structural Integrity Procedia 00 (2016) 000–000 3

by Gallagher (1974) by modifying the residual SIF used in this model with an empirical material constant. This
generalized model can only deal with OL effect. MGW model, used in this study takes into account the effect of UL
as well (see Brussat (1997)). The MGW model extends the Generalized Willenborg model by taking into account
the reduction of retardation effects due to UL. Generelized Willenborg model considers the effect of crack growth
retardation by using an effective SIF concept which is given as Kie f f = Ki − KR in which Ki is the typical SIF for the
ith cycle and KR is the residual SIF represented as


ai − aoL
KR = KRW = oL
Kmax 1− − Kmax,i (2)
r poL

oL
where ai current crack size, aoL crack size at the occurrence of the OL, r poL yield zone produced by the OL, Kmax
maximum SIF of the OL and Kmax,i maximum SIF for the current cycle. Gallagher and Hughes (1974) introduced an
empirical material constant into the calculation of residual SIF. They suggested that KR = ΦKRW where Φ is given by

Kmax,th
1−
Kmax,i
Φ= oL
(3)
S −1

where Kmax,th is the threshold SIF level associated with zero FCG rates and S oL is the OL (shut-off) ratio required
to cause crack arrest for the given material. In this model retardation effect is obtained by the change in the effective
stress ratio

ef f
Kmin,i Kmin,i − KR
Re f f = = (4)
ef f
Kmax,i Kmax,i − KR

Thus, for the ith load cycle, the crack growth increment ∆ai is given by

da
∆ai = = f (∆K, Re f f ) (5)
dN

MGW retardation model (see Brussat (1997)) uses a factor Φ = 2.523Φ0 /(1 + 3.5(0.25 − RU ))0.6 when RU < 0.25 and
Φ = 1 when RU ≥ 0. In here RU is the ratio of current UL stress to OL stress and Φ0 is a material dependent parameter
typically ranges from 0.2 to 0.8. In this study it is used as 0.4 as suggested by the NASGRO material database for
7075 T6.

2.2. Extended finite element method (XFEM)

The extended finite element method (XFEM) is based on a local enrichment of approximation space. The enrich-
ment is realized through the partition of unity concept. Let all nodes be represented by the set Γ, the nodes around the
crack tips and faces are represented by the sub set Γt and Γc respectively. Then the displacement approximation for
crack modeling in the XFEM has the form (see Belytschko and Black (1999)),

 4

     
Ni (x) α 
u x f em = Ni (x)(ui ) + Ni (x)H(x)(ai ) +  F α (x)b i  (6)
i∈Γ i∈Γc i∈Γt α=1
3076 Haydar Dirik et al. / Procedia Structural Integrity 2 (2016) 3073–3080
4 Haydar Dirik and Tuncay Yalçinkaya / Structural Integrity Procedia 00 (2016) 000–000

Where u x f em displacement vector, Ni is the nodal shape function, ui is the nodal displacement vector for non-enriched
nodes , H(x) is the Heaviside function which is +1 on one side of the discontinuity and −1 on other side of the
discontinuity, ai nodal enriched degree of freedom vector associated with Heaviside function, Fα (x) asymptotic crack
tip function, bαi nodal enriched degree of freedom vector associated with crack tip enrichment.

2.3. Problem statement and finite element model description

Validation of the developed algorithm is conducted through experimental data comparison under different types of
VAL conditions. Porter (1972) presents experiments on center notched 7075-T6 aluminum alloy specimens with 305
mm width, 915 mm length, 4.1 mm thickness and the initial crack size (2a) is 12.7 mm. Material properties for fatigue
life estimation is summarized in Table 1.

Table 1. Material properties for Al 7075-T6.

Material Yield Strength [MPa] Modulus of Elasticity [GPa] C n p q
Al 7075-T6 520 69.6 9.86x10−12 2.9 0.5 1

ABAQUS commercial software with XFEM crack modelling capability is used for SIF evaluation and FCG simula-
tions through three-dimensional hexahedron elements (C3D8). Mesh sensitivity analysis is carried out by considering
different mesh sizes to excrat KI on varying crack length and the optimum mesh configuration is used for further FCG
analysis. Figure 1 represents the finite element model of the specimen. The model, which has two partitions for the
domains of crack tip 1 and 2, is indeed three-dimensional, but it is shown as two-dimensional for clarity. Unit load is
applied from both ends of the specimen.

Fig. 1. FEM model of specimen with initial crack.

2.3.1. Numerical procedure

In this study, a FORTRAN script which calls ABAQUS for each analysis step is used for extracting SIF and for
FCG analysis. The FCG algorithm is presented in Figure 2. The FORTRAN script starts with input data of the material,
initial crack geometry and location. Three-dimensional model is then built in ABAQUS along with loading conditions,
XFEM crack definition, mesh generation and, etc. Next, FORTRAN script reads first mode SIF values from ABAQUS
solution and makes an averaging on the contour values to determine the SIF value. FORTRAN script calculates the β
factor by using this SIF value obtained under unit load. The calculated β factor is used for the further SIF calculations
for the load cases defined in load spectrum. At each loop crack tip plastic radius is calculated and retardation check is
made. The calculated crack growth at each cycle is summed until the total crack growth reached a predefined value.
This predefined value can be set by user to shorten the analysis time. But, it should be a small enough for the correct
crack path determination. After the predefined crack growth increment is reached, the script updates the crack tip by
calculating the new crack tip coordinates. This analysis is carried out till calculated SIF values exceed the critical SIF
or net section yielding occurs on the part. Crack propagation is obtained by successive linear extensions of crack tip.
Direction of crack growth is determined by KII=0 criterion which assumes that the crack growth occurs in a direction
of second mode SIF is zero.
 Haydar Dirik et al. / Procedia Structural Integrity 2 (2016) 3073–3080 3077
Haydar Dirik and Tuncay Yalçinkaya / Structural Integrity Procedia 00 (2016) 000–000 5

Fig. 2. Fatigue Crack Growth Algorithm

Figure 3.a and 3.b represent one block of loading for single OL cycle and single OL-UL cycle respectively, where
the baseline constant amplitude loading spectrum was periodically interrupted by applying tensile OL and OL-UL.
All relevant data associated with the loading spectrum were collected from Porter data.

3. Results

In this section, the numerical results are presented in terms of number of load cycle versus crack length and also
compared with results from NASGRO software and test data for seven different loading spectrums. Results agree with
NASGRO and test data with some minor discrepancies related to material and test data. Figure 4 shows comparisons
of predicted FCG life curves by NASGRO, XFEM and Porter test data for CAL.
3078 Haydar Dirik et al. / Procedia Structural Integrity 2 (2016) 3073–3080

6 Haydar Dirik and Tuncay Yalçinkaya / Structural Integrity Procedia 00 (2016) 000–000

σol
m cycle

Stress [MPa ]
n cycle n cycle
Stress [MPa ]

σol σmax
m cycle
σmax σmin

σmin σul

a b
Fig. 3. Schematic representation of one block of loading spectrum used in analysis a) OL spectrum b) OL-UL spectrum.

80
Nasgro
XFEM
70
Test Data

60

σmin = 3.45 MPa

Crack Length [mm]

50 σmax = 68.95 MPa

n=1 CAL
40

30

20

10

0
0 1 2 3 4 5 6 7 8
Number of Cycle 4
x 10

Fig. 4. Comparison of NASGRO, XFEM and Porter test data for CAL.

80 70
Nasgro Nasgro
XFEM XFEM
70 60
Test Data Test Data

60
50
σmin = 3.45 MPa σmin = 3.45 MPa
Crack Length [mm]

Crack Length [mm]

50 σmax = 68.95 MPa σmax = 68.95 MPa

σol = 76.54 MPa 40 σol = 103.43 MPa
40 m=29, n=1 m=29, n=1
30
30

20
20

10 10

0 0
0 2 4 6 8 10 12 0 2 4 6 8 10 12 14 16 18
Number of Cycle 4
x 10 Number of Cycle 4
x 10

Fig. 5. Comparison of NASGRO, XFEM and Porter test data for different OL level.

Figure 5 shows comparisons of predicted values by NASGRO, XFEM and Porter test data under different OL ratio
for spectrum defined in Figure 3a, and Figure 6 illustrates comparisons under same OL ratio but for different OL
spacing for spectrum defined in Figure 3a. Figure 7 presents the results for OL-UL spectrum defined in Figure 3b
at different load ratios. Predicted values by XFEM are in good agreement with NASGRO and test data. Retardation
effects of OL can be seen by comparing Figure 4 and Figure 5. It is obvious that neglecting the effect of load sequence
in fatigue calculations under VAL can lead to completely invalid life predictions. Figure 8 shows simulation results
 Haydar Dirik et al. / Procedia Structural Integrity 2 (2016) 3073–3080 3079
Haydar Dirik and Tuncay Yalçinkaya / Structural Integrity Procedia 00 (2016) 000–000 7

70 60
Nasgro Nasgro
XFEM XFEM
60 Test Data Test Data
50

50
σmin = 3.45 MPa 40 σmin = 3.45 MPa
Crack Length [mm]

Crack Length [mm]

σmax = 68.95 MPa σmax = 68.95 MPa
40 σol = 103.43 MPa σol = 103.43 MPa
m=50, n=1 30 m=300, n=1
30

20
20

10
10

0 0
0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5 3 3.5 4
Number of Cycle 5
x 10 Number of Cycle 5
x 10

Fig. 6. Comparison of NASGRO, XFEM and Porter test data for different OL sequences.

45 35
Nasgro Nasgro
XFEM XFEM
40
Test Data Test Data
30

35

σmin = 51.72 MPa 25 σ = 51.72MPa

min
Crack Length [mm]

Crack Length [mm]

30 σmax = 103.43 MPa

σmax = 103.43 MPa
σol = 155.14 MPa σol = 155.14 MPa
25 σul = 31.03 MPa 20 σul = 5.17 MPa

m=50, n=1 m=50, n=1

20
15

15

10
10

5 5
0 2 4 6 8 10 12 0 2 4 6 8 10
Number of Cycle 4
x 10 Number of Cycle 4
x 10

Fig. 7. Comparison of NASGRO, XFEM and Porter test data for different OL-UL ratios.

Fig. 8. Crack growth simulation steps for spectrum defined in Fig 3.a (n = 1, m = 29) (σmin = 3.45, σmax = 68.95, σoL = 76.54)

of crack propagation at four different steps using the spectrum defined in Figure 3a ( n = 1, m = 29 ) and as expected
from KII=0 criteria crack propagates in straight ahead direction at almost zero degree.
3080 Haydar Dirik et al. / Procedia Structural Integrity 2 (2016) 3073–3080
8 Haydar Dirik and Tuncay Yalçinkaya / Structural Integrity Procedia 00 (2016) 000–000

4. Conclusion

In this study, the developed XFEM-based procedure for fatigue crack growth under VAL is verified with experi-
mental results on 7075 T6 aluminum center cracked specimen. The comparison of developed procedure reveals good
agreement with the analytical and experimental results. Implementation of a retardation model to the XFEM-based
automatized crack growth procedure to account for load sequence effect is a new contribution to the literature. Prop-
agating the crack in a stepwise manner gives more control and the analysis possibility to the user unlike ABAQUS
self-crack propagation simulation by using XFEM. Moreover, since it not needed to propagate crack at every load cy-
cle there is a substantial increase in the computational efficiency. The obtained accuracy of XFEM in predicting FCG
under VAL in an automated scheme gives the users the confidence to apply the methodology at industrial components
in the near future.

5. Acknowledgements

The authors greatly acknowledge Eyüp Evren Taşkınoğlu and Fatih Özbakış from Turkish Aerospace Industries
(TAI) for providing valuable inputs for this work.

References

Belytschko , T., Black, T., 1999. Elastic crack growth in finite elements with minimal remeshing. Int. J. Numer. Methods Engng. 45, 601–620.
Amiri, F., Anitescu, C., Arroyo, M., Bordas, S.P.A., Rabczuk, T., 2013. Xlme interpolants, a seamless bridge between xfem and enriched meshless
methods. Comput. Mech. 53, 45–57.
Baietto, M., Pierres, E., Gravouil, A., Berthel, B., Fouvry, S., Trolle, B., 2013. Fretting fatigue crack growth simulation based on a combined
experimental and {XFEM} strategy. Int J. Fract. 47, 31 – 43.
Brussat, T.R., 1997. Private communication .
Elber, W., 1971. The significance of fatigue crack closure. ASTM STR 486, 230–242.
Forman, R.G., Kearney, V.E., Engle, R.M., 1967. Numerical analysis of crack propagation in cyclic-loaded structures. J. Basic Eng. 89, 459–464.
Forman, R. G. Mettu, S.R., 1992. Behavior of surface and corner cracks subjected to tensile and bending loads in ti-6al-4v alloy. ASTM STP 1131
, 519–546.
Gallagher, J., Hughes, T., 1974. Influence of yield strength on overload affected fatigue crack growth behaviour in 4340 steel. AFFDL-TR-74-28 .
Gallagher, J.P., 1974. A generalized development of yield-zone models. AFFDL-TM-74-28 .
Hartman, A., Schijve, J., 1970. The effects of environment and load frequency on the crack propagation law for macro fatigue crack growth in
aluminium alloys. Eng. Fract. Mech. 1, 615 – 631.
Melenk, J., Babuska, I., 1996. The partition of unity finite element method: Basic theory and applications. Comput. Methods in Appl. Mech. Eng.
139, 289 – 314.
Nasgro, 2002. Fracture mechanics and fatigue crack growth analysis software (version 4.02), reference manual. NASA Johnson Space Center and
Southwest Research Institute .
Newman, J.C., 1981. A crack closure model for predicting fatigue crack growth under aircraft spectrum loading. ASTM-STP 748, 53–84.
Newman, J.C., 1984. A crack opening stress equation for fatigue crack growth. Int J. Fract. 24, 131–135.
Paris, P.C., Erdogan, F.A., 1963. A critical analysis of crack propagation. J. Basic Eng. 85, 528–533.
Porter, T.R., 1972. Method of analysis and prediction for variable amplitude fatigue crack growth. Eng. Fract. Mech. 4, 717 – 736.
Solanki, K., Daniewicz, S., Jr., J.N., 2003. Finite element modeling of plasticity-induced crack closure with emphasis on geometry and mesh
refinement effects. Eng. Fract. Mech. 70, 1475 – 1489.
Varfolomeev, I., Burdack, M., Moroz, S., Siegele, D., Kadau, K., 2014. Fatigue crack growth rates and paths in two planar specimens under mixed
mode loading. International Journal of Fatigue 58, 12 – 19.
Walker, E.K., 1970. The effect of stress ratio during crack propagation and fatigue for 2024-T3 and 7076-T6 aluminum. In: Effect of environment
and complex load history on fatigue life. ASTM STR 462. Philadelphia: American Society for Testing and Materials , 1–4.
Wheeler, O.E., 1972. Spectrum loading and crack growth. J. Basic Eng. 94, 181–186.
Willenborg, J. P. Engle, R.M., Wood, H.A., 1971. A crack growth retardation model using an effective stress concept. AFFDL-TM-71-1-FBR .