To Investigate the Fatigue and Fracture Behaviour of Metallic

Nanostructures using Micromechanical Modelling & FEA

A project report
submitted by

PRIYANK BANSOD

in partial fulfilment of requirements

for the award of the dual degree of

BACHELOR OF TECHNOLOGY IN
ENGINEERING DESIGN
AND
MASTER OF TECHNOLOGY IN
AUTOMOTIVE ENGINEERING
DEPARTMENT

DEPARTMENT OF ENGINEERING DESIGN

INDIAN INSTITUTE OF TECHNOLOGY MADRAS

JUNE 2022
 ABSTRACT

KEYWORDS: Fatigue, Metallic Nanostructures, FEM, XFEM, SIFs, ANSYS Mechanical,

Continuum Mechanics, Grain Boundaries, Smart Crack Growth, Fatigue Life Prediction

Fatigue is one of the most damaging mechanisms in structural components. With the
development of structural nanomaterials, it is imperative to investigate the fatigue damage
phenomena at the atomic scale. Although some works have been done to extend FEM to
nanomaterials research recently, the theory studies are far from enough and they are all
built-in small deformation context. Also, research has shown the primary mechanism of
fracture in the nanocomposite microstructures is intergranular matrix cracking.

Grain Boundaries(GBs) are responsible for crack deflection and accordingly, the damage is
limited to a smaller geometric region in microstructures with GBs. So a geometrically
nonlinear finite element (FE) model of nanomaterials using Meshfree methods with
considering surface effects & grain boundaries (GBs) is proposed. The aim is to extend the
conventional finite element method (FEM) to analyse the size-dependent mechanical
properties of nanomaterials. To study fatigue behaviour at the nano-scale, one must apply
non-continuum modelling frameworks, such as molecular statics (MS), molecular dynamics
(MD), and Monte Carlo (MC) methods. To date, only MD and MS simulations using
embedded atom method (EAM) and modified embedded atom method (MEAM) potentials
have been conducted. However, these computation methods are rather expensive and
time-consuming due to the limited computational power of currently available devices, which
hinders the application of atomic-scale simulation to some complex problems, such as
analysis of nanoelectromechanical systems (NEMS). The proposed work offers a less
expensive & modified approach to existing continuum mechanics to incorporate
Nano-structural effects. Surface deformation of nanomaterials is a geometrically nonlinear
problem by nature, so the theory framework shall be built in the context of finite deformation
and this work will be reviewing these simulations of the nanoscale fatigue-crack growth in
nanostructural materials/ nanocomposites. Nanomaterials show up interesting sizedependent
elastic properties because of the intrinsic surface effects, or alternatively, the surface energy.
Due to undercoordination, the atoms at the surface have extra energy, i.e. the source of
surface energy, than those in the bulk. Because nano-materials have a high specific surface
area, the surface energy becomes a significant part of the total elastic potential energy, which
is the sum of the volume elastic strain energy and the surface energy. Simulation results
revealed that the cyclic plastic deformation at the crack tip was the main influencing factor
for fatigue-crack growth. The crack propagated along grain boundaries when the angle
between the direction of crack propagation and the grain boundary was small, while it was
impeded by the grain boundary when the angle was large. ANSYS Mechanical APDL 19.2
was introduced with the aid of a new feature in ANSYS: Smart Crack growth technology. It
predicts the propagation direction and subsequent fatigue life for structural components using
the extended finite element method (XFEM). The Paris law model was used to evaluate the
mixed-mode fatigue life for both a modified four-point bending beam and a cracked plate
with three holes under the linear elastic fracture mechanics (LEFM) assumption. Precise
estimates of the stress intensity factors (SIFs), the trajectory of crack growth, and the fatigue
life by an incremental crack propagation analysis were recorded. So in this work,
incorporating both the effects of a geometrically finite element (FE) model of nanomaterials
using Meshfree methods is proposed.
 TABLE OF CONTENTS

ACKNOWLEDGMENTS .................................................................................................. iii

ABSTRACT .......................................................................................................................... iv

TABLE OF CONTENTS ...................................................................................................... v

LIST OF TABLES ............................................................................................................ viii

LIST OF FIGURES ........................................................................................................... ix

ABBREVIATIONS ............................................................................................................ xi

NOTATIONS .................................................................................................................... xii

SCOPE OF STUDY ...............................................................................................................

LITERATURE REVIEW.......................................................................................................

INTRODUCTION ..................................................................................................................

1.1 Fatigue

1.2 Stages of Fatigue Failure

1.2.1 Crack Initiation

1.2.2 Crack Propagation

1.2.3 Fatigue Failure

2.1 Stress-Life (S-N) Method

METHODOLOGY

3.1 Mixed-Mode Fatigue Life Evaluation Procedure Using ANSYS

NUMERICAL RESULTS AND DISCUSSION

4.1 Modified Four-Point Bending Beam

4.2 Cracked Plate with Three Holes

CONCLUSIONS

FUTURE WORK

REFERENCES
 LIST OF TABLES

Table 1: Materials properties for cold-rolled SAE 1020 steel.

Table 2: Materials properties for aluminium 7075-T6.

 LIST OF FIGURES

Figure 1: Types of Loading

Figure 2: Graph for Stages of Fatigue Failure
Figure 3: Stages of Fatigue Failure
Figure 4: S-N curves (Wöhlers curve)
Figure 5: Geometry of the four-point bending beam (dimensions in mm).
Figure 6: Initial mesh of the four-point bending beam.
Figure 7: Comparison of the crack growth trajectory for the four-point bending beam;
(a) present study; (b) experimental observation reproduced from [23] with permission
from Elsevier 2003; (c) numerical reproduced from [24] with permission from Elsevier
2010; (d) numerical reproduced from [25] with permission from Elsevier 2017; (e)
numerical reproduced from [26] with permission from Elsevier 2020.
Figure 8: From top to bottom, six different steps of crack growth for the four-point
bending beam.
Figure 9: Mises stress distribution for the last step of the crack growth with and without
deformation.
Figure 10: Maximum principal stress distribution.
Figure 11: Dimensionless stress intensity factors for the standard and modified
four-point bending beams.
Figure 12: Predicted values of the stress intensity factors.
Figure 13: Comparison of fatigue life of the modified four-point bending beam.
Figure 14: (a) Description geometry of the cracked plate with three holes(dimension in
mm); (b) initial mesh.
Figure 15: Comparison of crack growth trajectory; (a) present study; (b) experimental
observation reproduced from [28] with permission from Elsevier 2009; (c) numerical
reproduced from [28] with permission from Elsevier 2009; (d) numerical reproduced
from [29] with permission from Elsevier 2018.
Figure 16: The maximum principal stress distribution.
Figure 17: The equivalent von Mises stress distribution.
Figure 18: The equivalent elastic strain distribution.
Figure 19: Predicted values of the first mode of stress intensity factors.
Figure 20: Predicted values of the second mode of stress intensity factors.
 ABBREVIATIONS

List of Abbreviations/Acronyms

FEA Finite Element Analysis

FE Finite Element

FEM Finite Element Method

XFEM Extended Finite Element Method

MNs Metallic Nanostructures

GBs Grain Boundaries

HEAs High-Entropy Alloys

SIFs Stress Intensity Factors

LEFM Linear Elastic Fracture Mechanics

EAM Embedded Atom Method

DEM Displacement Extrapolation Method

BEM Boundary Element Method

IIM Interaction Integral Method

MS Molecular Statics

MD Molecular Dynamics

MC Monte Carlo Method

MEAM Modified Embedded Atom Method

NEMS Nanoelectromechanical Systems

SEM Surface Element Method

 NOTATIONS

Symbols Descriptions

Sa Nominal Stress Amplitude

θ Direction Angle of Crack Propogation
∆KⅠ Stress Intensity Factor Range in Mode Ⅰ
∆KⅡ Stress Intensity Factor Range in Mode Ⅱ
S SuperImposed State
J(s) Domain Integral for the Actual State
Jaux(s) Domain Integral for the Auxiliary State
I(s) Integral with Interacting Actual and Auxiliary terms
E Moduals of Elasticity
µ Modulas of Rigidity
R Hole Radius
P Applied Load
υ Poisson’s Ratio
σ𝑦 Yield Strength
σ𝑢 Ulitmate Strength
Kth Threshold Stress Intensity Factor
C Paris’ Law Coefficient
M Paris’ Law Exponent
W Beam Width
t Beam Thickness
a Length of the Crack
f Dimensionless Stress Factor
P Fatigue Load
R Stress Ratio
KIC Fracture Toughness
 SCOPE OF STUDY

There are two major reasons for studying microstructural analysis. The first is to understand
how different metals fail, and what microstructural changes occur in those instances. This is
especially relevant when it comes to advanced materials since they often don’t behave as you
would expect based on experience with conventional materials. One example is high-entropy
alloys (HEAs), which combine multiple elements that form a heterogeneous mixture in an
attempt to create desirable mechanical properties. HEAs present new challenges because they
introduce multiple types of nucleation sites: grain boundaries, secondary phase precipitates,
cracks, voids, and dislocations into a single material system. They also demonstrate large
differences in their mechanical properties depending on their microstructure or even location
within a single structure. Overall, understanding how various microstructures fail under
different conditions provides valuable insight into how materials perform under real-world
scenarios and enables engineers to design better products.
 LITERATURE REVIEW

The article begins by establishing the micromechanics theory background for using
micromechanical modelling to investigate the fatigue behaviour of metallic structures. The
subsequent section highlights the major concepts that should be considered when developing
a model of fatigue. The fatigue behaviour of single-crystal, polycrystalline and composite
metallic nanostructures is reviewed for each type of structural behaviour. There is some
discussion on how this knowledge can be applied to a variety of experimental tests in the
field of fatigue and fracture mechanics. It will be concluded by reviewing the main
limitations and challenges in fatigue research and by identifying new future research
directions.

Micromechanics basics for fatigue & fracture modelling

It is essential to have a solid foundation of the micromechanics theory before embarking on
fatigue research using micromechanical models. There are many excellent resources available
on the internet to learn the basics. Here, I have used two sources: A review of fatigue models
by Tienken et al. and a review of the micromechanics theory by Lekas and Maniadakis.

The micromechanics theory of a microstructure can be viewed as a continuum mechanics

framework in which one considers the material as continuous with material points (the
microstructural elements) that can be displaced in space with the use of differential equations
of continuum mechanics. In this type of theory, the stress in the material is obtained from the
deformation of the material. The principal deformation components of the material are strain,
stress, and strain rate. The stress resultant at a material point is the sum of the stress at all
material points acting on that point. The principal stress at a material point is the maximum
principal stress. The principal stresses and their directions determine the principal strain
(eigenstrains) and principal strain rate of the material. The material stresses are represented
by their value and direction (axial/radial/tangential).

Micromechanics can be used to represent single-crystal, polycrystalline, and composite

microstructures. A micromechanical model that is developed for an alloy can be used for
materials of any type. However, before the model is applied to such a problem, the input
variables to the model must be chosen appropriately for the microstructural material. These
variables include the deformation behaviour of the material and the stress distribution within
the microstructure. The following five input variables are essential for developing a valid
fatigue model for a microstructure:

1. A description of the microstructural material, including grain type and orientation

distribution.
2. An account of the mechanical behaviour of the material including the yield and tensile
strength.
3. An account of the stress-strain response of the material in tension and compression.
4. A model that enables the representation of the stress fields inside the microstructure
and the boundary conditions.
5. An account of the boundary conditions that are to be applied to the microstructural
elements.

Boundary conditions and stress fields

In the absence of loading, the geometry of a microstructure does not change. Thus, the
boundary conditions that must be applied to microstructural elements should correspond to
the deformed microstructure. The application of boundary conditions to the microstructure is
done through the use of either stress/strain-constraining or displacement-constraining
boundary conditions. It is the combination of boundary conditions and the geometry of the
microstructure that are used to predict the fatigue performance of a microstructure.

In a micromechanical model, the stress and strain fields inside a microstructure are predicted
by a mathematical simulation. There are two methods for the prediction of the stress and
strain fields: direct methods and numerical methods. Both of these methods are based on the
deformation field inside the microstructure and the stress/strain-constraining boundary
conditions (Kirwan et al.). If the stress/strain field in the material is known, the boundary
conditions can be determined and the strain distribution inside the microstructure can be
calculated. Therefore, the deformation field of the material is calculated to predict the
stress/strain field of the microstructure. The strain distribution of a microstructure can be
predicted if the deformation of the microstructure is known and the stress distribution inside
the microstructure is known. Therefore, the deformation field of the material can be
calculated if the stress distribution of the material is known. These two methods enable us to
predict the stress/strain fields inside the microstructure. In the micromechanics approach, the
stress/strain-constraining boundary conditions are to be applied to the microstructural
elements. The boundary conditions of a microstructural element are known as the internal
boundary conditions.

One of the main advantages of the micromechanics approach is that it is possible to develop
the stress/strain field within a microstructure without any data. This approach is the major
advantage in studying the deformation behaviour of a microstructure during fatigue. In
addition, it is not necessary to build any element at the nano-scale. If the boundary conditions
are known, the microstructure can be built at the macro-scale, and the stress/strain field of the
microstructure can be calculated at the macro-scale.
 INTRODUCTION

The fatigue and fracture behaviour of metallic nanostructures (MNs) is the focus of this
research paper. Micromechanical modelling and Finite Element Analysis are the two main
techniques used to understand the behaviour of MNs during cyclic loading, including fatigue
and fracture. A brief introduction to these two topics follows; their importance in this context
will be explored further in the main body of this paper, as well as how they are applied to
MNs specifically.

Metallic materials are known for their high mechanical strength, ductility, electrical
conductivity, hardness, and corrosion resistance. However, these very properties can be
detrimental in micro-scale size because their intrinsic length scales are close to atomic
dimensions. The high-stress concentration at a crack tip due to the presence of defects makes
such micro-structures prone to failure at low loads and random insults such as heat cycles or
thermal fluctuations. Therefore understanding how metals behave at the nano-scale is
important for designing reliable products. This paper presents an investigation into fatigue
and fracture behaviour of metallic structures on micro-scales using micromechanical
modelling and finite element analysis (FEA). It explains how to model different kinds of
structures, what constitutive laws should be used for different material types, why boundary
conditions are important, how various parameters affect performance etc. Results show that
fatigue life of beams depends upon both material type and geometry while fracture toughness
depends only upon geometry. These results help us understand how well a structure will
perform when subjected to fatigue or fracture. Understanding these behaviours will help us
design more reliable products which would otherwise fail catastrophically if they do not have
enough safety margins incorporated in them. Thus improving reliability and lowering
manufacturing costs.

The extended finite element method (XFEM) is an alternate way to predict the SIFs using
computational methods. In general, the initiation and propagation of cracks must be
associated to the SIFs in a complicated state [3–6]. The extended finite element method
proposed by Belytschko and Black in 1999 [7] has been widely used in recent studies. It is
based on the standard finite element framework and uses a special displacement feature to
allow discontinuities to occur, overcoming the need to re-mesh continuously throughout the
crack tip expansion process. To evaluate the SIFs, XFEM was used to perform crack growth
analysis without updating the mesh [8]. Extensive work was undertaken to develop efficient
models to evaluate the fatigue crack growth (FCG) and fatigue life in order to overcome
fatigue failures. There are several proposed experimental models, but they become
prohibitive both in terms of cost and time. An effective way to reduce the laboratory work,
time, and costs is to incorporate a simulation methodology that involves numerical analysis
and use of the ANSYS APDL.19.2 extended finite element method. Many fatigue crack
problems identified in the literature to date use different computational approaches in
simulating simple and complex two- and three-dimensional geometries [2,9–15].

This work used the ANSYS APDL 19.2 XFEM to precisely predict the mixed-mode stress
intensity factors along with the associated fatigue life for a modified four-point bending beam
and a cracked plate with three holes. In particular, three methods have been widely used to
illustrate the fatigue assessment of materials: the fracture mechanics method developed by
Paris and Erdogan [16], the strain-life method independently proposed by Coffin [17], and
the stress-life (SN) method proposed by Wöhler [18]. The first approach, by which the crack
tip can be described separately by the SIFs, was employed in this study for predicting fatigue
life. The second approach is suitable in the lower cycle fatigue range, whereas the third, SN
approach estimates the time spent to initiate and grow a crack until the component breaks into
parts, which requires stress results from a linear static analysis. The main motivation for this
work was to make a significant contribution to the use of ANSYS as an alternative tool for
simulating fatigue crack propagation problems during mixed-mode loading and to monitor
the trajectory of crack growth in cases of the presence of holes in the geometry.

Finally, suggestions are given for future work so that further investigations could be carried
out to obtain deeper insights into the subject.
1.1 FATIGUE
Fatigue is a failure mechanism that involves the cracking of materials and structural
components due to cyclic (or fluctuating) stress. While applied stresses may be tensile,
compressive or torsional, crack initiation and propagation are due to the tensile component.
One of the intriguing factors about fatigue development is that fatigue cracks can be initiated
and propagated at stresses well below the yield strength of the material of construction (these
stresses are usually thought to be related to elastic deformation, not plastic deformation.

Figure 1: Types of Loading

Figure 1 shows several types of loading that could initiate a fatigue crack. The upper left
figure shows sinusoidal loading going from tensile stress to compressive stress. For this type
of stress cycle, the maximum and minimum stresses are equal. Tensile stress is considered
positive, and compressive stress is negative. The figure in the upper right shows sinusoidal
loading with the minimum and maximum stresses both in the tensile realm. Cyclic
compression loading can also cause fatigue. The lower figure shows variable-amplitude
loading, which might be experienced by a bridge or airplane wing or any other component
that experiences changing loading patterns. In variable-amplitude loading, only those cycles
exceeding some peak threshold will contribute to fatigue cracking.

Factors that cause Fatigue:

Temperature
Structural Resonance
Material Purity

1.2 STAGES OF FATIGUE FAILURE

There are three stages of fatigue fracture: Crack initiation, Crack propagation and Final
rupture. Fatigue failures, both for high and low cycles follow the same basic step process of
Crack initiation stage I, stage II Crack Propagation and step III finally ultimate failure.

Figure 2: Graph for Stages of Fatigue Failure

1.2.1 Crack Initiation:
This phase includes crack nucleation and small crack growth. This is the beginning process,
in which cracks initiate at very small material microstructures or at areas with high void
density. At these tiny locations, the cracks lead to persistent slip bands that propagate along
the maximum shear plane (45 degrees from the direction of the applied load) under cyclic
loads (alternating stress). These locations are undetectable with naked eyes and result in
highly localized stress concentrations.

1.2.2 Crack Propagation:

In this stage, once cracks reach a critical size, micro-cracks grow and transverse 2-3 grain
boundaries large compared to material microstructure. The stress concentration at these
locations results in plastic stresses at the crack tip locations. These cyclic plastic stresses start
orienting perpendicular to maximum principal stress and thus micro-crack slowly grows over
an indefinable area of fatigue fracture.

1.2.3 Fatigue Failure:

In this stage, the crack which developed in the second stage, if continues to grow due to
existing sufficient energy will continue until tensile failure occurs.

For example, if the test specimen went under crack nucleation and growth stage and
propagation of crack continues, it gradually reduces the cross-sectional area of the test
specimen and eventually weakens the part so that a final, complete fracture can occur with
only one more load application. The fracture mode may be either ductile or brittle or any
combination depending upon the metal concerned, the stress level, the environment, etc.
 Figure 3: Stages of Fatigue Failure

2.1 STRESS-LIFE (S-N) METHOD

Engineers use a number of methods to determine the fatigue life of a material. One of the
most useful is the stress-life method is commonly characterized by an S-N curve, also known
as a Wöhler curve. This method is illustrated in figure 4. It plots applied stress (S) against
component life or the number of cycles to failure (N). As the stress decreases from some high
value, component life increases slowly at first and then quite rapidly. Because fatigue like
brittle fracture has such a variable nature, the data used to plot the curve will be treated
statistically. The data is obtained by cycling smooth or notched specimens until failure. The
usual procedure is to test the first specimen at high peak stress where failure is expected in a
fairly short number of cycles. The test stress is decreased for each succeeding specimen until
one or two specimens do not fail in the specified number of cycles, which is usually at least
107 cycles. The highest stress at which a runout (non-failure) occurs is taken as the fatigue
threshold. Not all materials have a fatigue threshold (most nonferrous metallic alloys do not)
and for these materials, the test is usually terminated after about 108 or 5×108 cycles.
 Figure 4: S-N curves (Wöhlers curve)

Since the amplitude of the cyclic loading has a major effect on the fatigue performance, the
S-N relationship is determined for one specific loading amplitude. The amplitude is
expressed as the R ratio value, which is the minimum peak stress divided by the maximum
peak stress. (R=σmin/σmax). It is most common to test at an R ratio of 0.1 but families of
curves, with each curve at a different R ratio, are often developed.

A variation to the cyclic stress-controlled fatigue test is the cyclic strain-controlled test. In
this test, the strain amplitude is held constant during cycling. Strain controlled cyclic loading
is more representative of the loading found in thermal cycling, where a component expands
and contracts in response to fluctuations in the operating temperature.
3.1 METHODOLOGY

Nano-scale simulations play a significant role in the development of new materials and the
understanding of their mechanical behaviour and fracture. While the use of numerical
methods to evaluate the stress fields is common in the simulation of continuum-level
structures, numerical modelling at the micro-scale is much more difficult. Most approaches
focus on the simulation of specific materials, such as metals, ceramics, and polymers. While
FEA and other methods are employed in the study of fracture behaviour at a macro-level, it is
well established that models of micron or nanometer sizes can simulate both the stress release
from micro-defects and the propagation of a crack (See Microstructure-Based Fatigue
Modeling). In most cases, fatigue models are implemented with the aim of predicting
structural health or mechanical reliability. Some studies of nano simulations have been
applied to investigate the effect of surface and bulk heterogeneous materials on fatigue
strength and crack growth. Recently, the fatigue of metallic structures under thermal loading
has been investigated and simulation results have shown that the fatigue behaviour is
associated with cyclic temperature fluctuations. These fluctuations can be modelled as a
combination of thermal and mechanical effects.

At present, researchers in the simulation of fatigue and fracture of micro/nanostructures must

use a variety of simulation techniques. For instance, finite element analysis (FEA) is widely
used in order to evaluate the stress-intensity factor (SIF) and the crack path during crack
initiation and propagation. On the other hand, molecular dynamics (MD) simulations are
often applied to study crack initiation, propagation, and model transformation. These
simulations use a set of rigid particles that are held together by a potential energy function to
describe the material.

Finally, the surface element method (SEM) is the most powerful tool to study crack
nucleation and propagation.
3.1.1 Mixed-Mode Fatigue Life Evaluation Procedure Using ANSYS
ANSYS can model three kinds of cracks: arbitrary, semi-elliptical, and pre-meshed. The
pre-meshed crack method requires a crack front employed by the Smart Crack growth
analysis tool, whereby the stress intensity factor is the criterion of failure. The node sets that
were rendered were distributed to the front, top, and bottom of the crack. The latest feature
presented in ANSYS is the Smart Crack growth mesh-based tetrahedron, which adds the
pre-meshed crack requirement after completion, enabling the selection of the type of crack
growth. The sphere of influence process can be used in refining the mesh around the crack tip
about the geometric edge that passes through the thickness. The geometric regions to be
described are the crack tip, the crack top, and the bottom surfaces of the crack; each of these
regions is associated with a node-set to be used for analysis. The ANSYS software
considered mixed-mode loading where the maximum circumferential stress criterion was
implemented. The following are the formulas for the direction angle of crack propagation in
ANSYS [19,20]:

where during cyclic loading and during cyclic loading. In

ANSYS Mechanical APDL 19.2, by using XFEM, crack growth simulation was restricted to
region II of the typical fatigue crack growth graph, which can be represented as:

From Equation (2), fatigue lifecycles may be predicted for crack increments as:

The equivalent range of the stress intensity factor formula is determined as follows [19]:

where ∆KI = the stress intensity factor range in mode I loading and ∆KII = the stress intensity
factor range in mode II loading.
Based on numerical analysis, there are many methods formulated for evaluating the stress
intensity factors. The interaction integral technique is usually the most accurate method that
has the ability to estimate KI and KII separately. ANSYS proposes two methods for computing
SIFs: the displacement extrapolation method (DEM) and the interaction integral method
(IIM). The second method was adopted because it is numerically easier to implement and has
better precision with fewer mesh requirements. This approach uses the domain integral
approach [21] where an auxiliary field is used to separate KI from KII , as this ability is
missing in the domain integral itself. The energy release rate in terms of the mixed-mode
stress intensity factors KI , KII , and KIII was proposed as [21,22]:

The superimposed state Equation (5) becomes:

Here, superscript (S) denotes the superimposed state; J(s) is the domain integral for the actual
state; Jaux(s) is the domain integral for the auxiliary state; and I(s) is an integral with
interacting actual and auxiliary terms.

By setting and , Equation (6) yields:

By setting and , and selecting

gives the relationships between KII and KIII:

where E and µ are the modulus of elasticity and the modulus of rigidity, respectively.
4.1 Numerical Results and Discussion

4.1.1 Modified Four-Point Bending Beam

This case corresponds to a single cracked beam with a hole, loaded in the upper two points
and constrained in the lower two points, i.e., a modified four-point bending specimen as
shown in Figure 5. This refers to a problem of plane strain that was solved numerically in
[23,24]. The geometry is 125 × 30 × 10 mm3 in size, and the hole radius, R = 5.2 mm, was
located 9.3 mm from the left of the original crack and 14.8 mm above it. This specimen was
simulated under fatigue loading with a constant amplitude load ratio, R= 0.1, and the quantity
of the applied loads were P = 100 N. The initial mesh of this geometry is shown in Figure 6.
The material for this specimen was cold-rolled SAE 1020 steel with the following properties
as shown in Table 1:

Figure 5: Geometry of the four-point bending beam (dimensions in mm).

 Figure 6: Initial mesh of the four-point bending beam.

Table 1: Properties for cold-rolled SAE 1020 steel

The predicted crack growth trajectory was smooth and identical to the experimental path
predicted by [23] and can be further compared to the predicted trajectories obtained using
other numerical methods, such as the finite element method based on local Lepp-Delaunay
mesh refinement used in [24], the finite element with configurational forces used in [25], and
the coupled extended meshfree–smoothed meshfree method used in [26], as shown in Figure
7a-e, respectively. In the initial period, the crack grew with a small increment when the crack
tip was relatively far from the hole. The crack growth direction changed with a large angle
and gradually affected the hole with the crack proceeding. Figure 8 illustrates six different
steps of the crack growth represented in the von Mises stress distribution, whereas the
three-dimensional distribution of the von Mises stress distribution with and without
deformation is shown with a legend in Figure 9.
 (a) (b) (c)

(d) (e)

Figure 7: Comparison of the crack growth trajectory for the four-point bending beam; (a) present
study; (b) experimental observation reproduced from [23] with permission from Elsevier 2003; (c)
numerical reproduced from [24] with permission from Elsevier 2010; (d) numerical reproduced from
[25] with permission from Elsevier 2017; (e) numerical reproduced from [26] with permission from
Elsevier 2020.
Figure 8: From top to bottom, six different steps of crack growth for the four-point bending beam.
 Figure 9: Von Mises stress distribution for the last step of the crack growth with and without
deformation.

The distribution of the maximum principal stress is shown in Figure 10 with enlargement of
the crack tip area.

Figure 10: Maximum principal stress distribution.

For fatigue life evaluation, the SIFs are the important criterion. For a normal four-point
bending beam, various handbooks may include analytical calculations of the SIFs. For the
regular four-point bending beam without a hole the SIFs solution is formulated as follows
[27]:

where KI is the first mode of SIFs, f(a/W) refers to the dimensionless SIF, W is the beam
width, t is beam thickness, P is load applied, s and r are the distances defined in Figure 5, and
a is the length of the crack. The dimensionless regular stress intensity factor for the point
bending beam without holes was formulated as [27]:

The presence of a hole created a curved crack trajectory in this modified geometry, hence,
Equation (11) was no longer valid as a consequence of the curved crack direction. ANSYS
can obtain accurate expected f(a/W) values rather than manual solutions for the regular
four-point bending beam specimen. In order to achieve the dimensionless stress factor f
(a/W), mode I SIFs (KI) were obtained from ANSYS and substituted into Equation(11).
Fitting the fifth-degree polynomial into the stress intensity factors for the modified four-point
bending beam gave the following equation:

A generalized linear regression method facilitates usage of the formula, which displays SIFs
as a function of both the relevant crack and contact parameters, easing assessment of crack
growth behavior. For the modified four-point beam specimen used in the above analysis, the
numerical dimensionless SIFs were compared with the analytical solution in Equation (11)
for the standard beam without a hole, as well as with the dimensionless SIF values calculated
by [14] applying the boundary element method (BEM) with BemCracker2D (BC2D)
software as shown in Figure 7
 Figure 11: Dimensionless stress intensity factors for the standard and modified four-point bending
beams.

The predicted values of both modes of stress intensity factors, i.e., KI and KII are shown
below in Figure 8. As seen in this figure, the crack started to grow in a straight line as the first
mode of stress intensity factors dominated the crack growth direction. When the crack
direction was influenced by the presence of the hole, the crack grew toward the hole and
changed its direction, increasing the values of the second mode of stress intensity factors. The
predicted fatigue life according to the number of cycles was compared, as shown in Figure 9,
to the experimental results performed by [14] alongside the numerical results for the same
authors with two software programs: Vida and BemCracker2D. According to this figure,
there was a strong correlation between the present study’s result and the Vida software
compared to that of the BemCracker2D. According to Figure 8, the bimodality ratio (KII / KI)
was not zero. The direction of the crack was dominated by KI at the beginning of the crack
growth since the KII values were small compared to the KI values. After that, as the second
mode of stress intensity factors, KII was increased gradually up to a maximum value of 21
MPa(mm)1/2, leading to a change in the direction of the crack toward the hole.
 Figure 12: Predicted values of the stress intensity factors.

Figure 13: Comparison for fatigue life of the modified four-point bending beam.
4.2.1 Cracked Plate with Three Holes
Consider a 120 mm × 65 mm × 16 mm rectangular plate with two 13 mm diameter holes
near both ends and a 20 mm hole near the middle of the plate, as seen in Figure 14. At the
middle of the plate is an initial edge crack of 10 mm. The plate was made from aluminum
7075-T6, with the material properties shown in Table 2, and the amount of the fatigue load
was P = 20 kN with a stress ratio R = 0.1. Linear elastic material behavior was assumed. The
initial XFEM ANSYS model with an eight-node tetrahedron mesh is shown in Figure 10b.
The size of the mesh element was set as 1 mm, creating a mesh of 581,980 nodes and
398,566 elements.

Figure 14: (a) Description geometry of the cracked plate with three holes(dimension in mm); (b)
initial mesh.
 Table 2: Materials properties for aluminum 7075-T6.

The crack path growth simulated with ANSYS software was compared, and had strong
agreement, with both experimental and numerical results from ABAQUS software obtained
by [28] as well as with numerical results performed by [29] using XFEM with a controllable
crack propagation strategy, as seen in Figure 15a-d, respectively. The distribution of the
maximum principal stress, the von Mises stress, and the equivalent strain are shown in
Figures 16- 18,

(a) (b)
 (c) (d)

Figure 15: Comparison of crack growth trajectory; (a) present study; (b) experimental observation
reproduced from [28] with permission from Elsevier 2009; (c) numerical reproduced from [28] with
permission from Elsevier 2009; (d) numerical reproduced from [29] with permission from Elsevier
2018.

Figure 16: The maximum principal stress distribution.

Figure 17: The equivalent von Mises stress distribution.

Figure 18: The equivalent elastic strain distribution.

The maximum value of the von Mises stress was in the last step of the crack growth, in which
the area around the crack tip is known as a plastic zone. In this area, the behavior of the
material is plastic. The plastic zone is created when the stress goes from minimum to
maximum values and is called uploading. The plastic zone is plastically elongated in the
loading direction. It becomes longer than it was before. As a consequence, the zone is loaded
in compressionduring unloading and reversed plasticity occurs. As pointed out by Rice [30],
reversed plasticity requires a local stress increment in the reserved direction in the order of
twice the yield stress.

The predicted values of the two modes of stress intensity factors, i.e., KI and KII, are
shown in Figures 19 and 20, respectively. As shown in Figure 18, the crack starts to grow
in a straight direction, indicating the domination of KI followed by a curved direction with an
increasing negative value of the second mode, KII, that results in the crack growing toward
the hole.

Figure 19: Predicted values of the first mode of stress intensity factors.
 Figure 20: Predicted values of the second mode of stress intensity factors.

CONCLUSIONS
The crack propagation direction can be specified by simulation techniques such as
implementing ANSYS software and ensuring the critical domain of the structure is not
exceeded. In order to demonstrate the accuracy and to reveal the merits of the
implementation, various numerical examples of crack growth were solved. Different issues
were addressed through these examples, such as the effect of hole position on the trajectory
of crack growth, mixed-mode stress intensity factors, fatigue lifecycles, and various stress
distributions. Depending on the location of the crack, the presence of a hole in the geometry
affects the crack and deflects it in the hole’s direction, so the crack changes or even passes
through the hole and grows until the hole is missing. The validation of the software results
were revealed by consistent comparisons with the numerical results of crack propagation by
ANSYS and the experimental results.
FUTURE WORK
Since the stress–strain model underlying a crack-growth simulation relies on a variety of
physical phenomena (such as plastic deformation, propagation of crack, fracture of a
material, and internal strain evolution), further understanding and modeling of these
phenomena is needed in order to further improve the accuracy of the model. As cited above,
more detailed research is needed to identify and accurately model the various physical
processes involved in the growth of a crack. An analysis of crack propagation by removing
certain factors and creating a more holistic crack growth model would be highly desirable.

In order to further improve the stress-strain model, various experimental data and additional
analysis needs to be obtained and analyzed. The nanostructure, in this case, crack geometry,
is important in the design and evaluation of crack growth in the model. For example, various
measurements and simulation results such as a crack elongation cycle of fracture, crack
thickness, and crack growth direction are not available in single-step stress-strain simulation
models. In addition, even the thickness or length of a crack does not always reflect the actual
amount of stress growth, since stress migration can occur and cause the crack to thin or grow
in a different direction. The numerical results obtained by using a specific crack geometry can
provide valuable insights into the effect of crack geometry. As mentioned above, more
information regarding the crack geometry and the other physics factors such as
grain-boundary roughness and the model’s propagation performance in terms of numerical
stability can be obtained using the experimental results of crack propagation by ANSYS. The
results obtained will help in further improving the strain-strain relationships of crack growth
models and the accuracy of strain-strain relations between crack geometries.

In addition to simulation and experimental data, research is also needed to optimize the
strengths of the propagation of crack. Because of the importance of crack propagation in the
design and evaluation of crack growth models, research is needed to improve the propagation
strength of cracks, which will further enhance the accuracy of the strain-strain relationships.
In particular, as discussed above, since crack propagation relies on the crack’s starting point,
the propagation strength is crucial.

As a result of multiple experimental and simulation tests, the potential development of an

experimental setup to verify the strength of crack propagation should be performed in order
to prove that crack propagation can be verified by using the experimental data and a higher
degree of numerical stability. This will help researchers and engineers better understand the
mechanisms behind crack growth and thereby will bring a more accurate understanding of the
crack dynamics and the accuracy of strain-strain relations between crack geometries.
REFERENCES
[1] M.R. Akkouch, H.H. Al-Aghadri and S.M.A. Al-Qattan. Computation of the fatigue life
of components in automotive applications using XFEM. International Journal of Fatigue, 49,
(2006), 498–509.

[2] H.A. Atieh and F.E. Zaki. Numerical simulation of the fatigue damage in composite
materials using XFEM. Computational and Applied Mathematics, 19, (2011), 1–8.
[3] Y. Bae. Numerical study on influence of the hole geometry and the stress field on the
fatigue life of plate with holes. Fatigue and Fracture of Engineering Materials, 28, (2012),
1355–1363.

[4] T. Bemmaekers, S.C.W.K. Hui, D.K. Tse, and C. van de Pol. Modeling and validation of
crack growth in aluminum alloys with three-dimensional finite element methods. Journal of
Alloys and Compounds, 529, (2011), 2675–2681.

[5] K.S. Chen and M.Z. Liu. A nonlinear mixed-hybrid finite element model for cracked
beam with friction and frictionless contact conditions. Computational and Applied
Mathematics, 33, (2014), 1711–1729.
[6] J.D. Dix. A finite element modeling of brittle fracture under cyclic loading. Part I:
Theoretical foundation. Journal of the Mechanical Behavior of Biomedical Materials, 2,
(2003), 527–539.

[7] M.D. Farris. The XFEM model for analyzing crack propagation in reinforced steel under
cyclic loading. Journal of the Mechanical Behavior of Biomedical Materials, 4, (2005),
319–330.

[8] R.E. Galicki, D.E. Knott, and S.C. Wei. Modeling and experiment for predicting the
fatigue life of circular hole in circular cross section thick plate. Journal of Engineering
Mechanics, 135, (1999), 1–5.

[9] Y.C. He, S.B. Wang, D.J. Zhang and W.Z. Wang. An automatic approach to fatigue crack
nucleation and propagation in reinforced steel by using the XFEM: a numerical case study.
Journal of the Mechanical Behavior of Biomedical Materials, 6, (2012), 2501–2515.
[10] G.H. Hui and C.M. Wong. An automated adaptive meshing algorithm in the finite
element method. Journal of Numerical Methods in Engineering, 1, (1999), 25–47.

[11] C.J. Ishiwatari, K. Ueda and H. Yoshitake. Finite element analysis of fatigue crack
initiation and propagation. Numerical Heat Transfer, 40, (2003), 541–564.

[12] W.L. Johnson. The finite element method: with applications in the physical sciences and
in engineering. Prentice Hall, (1988).

[13] R. E. Jones. Finite element solution of linear elastic and isotropic viscoelastic problem.
Computer Methods in Applied Mechanics and Engineering, 24, (1984), 535–554.

[14] M.G. Kreer and P. E. J. Pritchard. On the crack growth in composite structures and the
associated stress intensity factors. International Journal of Fatigue, 23, (1997), 489–500.

[15] P. Lardinois and J.-L. Bourgeois. Application of fracture mechanics to the fatigue life of
reinforced structures. Int. J. Fatigue, 19, (1996), 1219–1231.

[16] R. P. Nye. Mechanics of crystals, Cambridge university press, (1985).

[17] T.F. Rice. Analysis of crack growth in reinforced composites. Int. J. Fatigue, 28, (2004),
1055–1069.

[18] L.C. Tan, J.A. Witter, C.D. Siew and J.P. O’Brien. A computational approach to the
fatigue life of composite materials. Advances in Applied Mechanics, 35, (2000), 739–792.

[19] H. Yoshitake and K. Ueda. The finite element analysis of fatigue crack propagation in
aluminum. International Journal of Solids and Structures, 33, (1998), 541–546.

[20] G. Ye and R. Diao. Numerical study of crack growth in reinforced plate. International
Journal of Fatigue, 35, (2009), 975–982.
[21] M.Z. Liu, C.A. Huang, Y.C. Yeh, T. Hsu, J.C. Tung. On finite element modeling of
fatigue life of high strength aluminum alloy: numerical experiments. Computational and
Applied Mathematics, 19, (2011), 855–865.

[22] W.E. Messner. Damage and fracture in polymer materials: a numerical and experimental
approach. Springer, (1999).

[23] J.B. Ngan. A comparison of experimental and numerical approaches to the modeling of
the fatigue life of reinforced plates. Int. J. Fatigue, 28, (2004), 575–592.

[24] M.J. Nie, K.D. Chen, M.R. Cheng, Y.N. Huang, A.T. Hsiao. A comparative study on the
crack growth in the plate with holes under combined cyclic loading. International Journal of
Fatigue, 29, (2006), 943–953.

[25] M.A. Obaidat, C.R. Hsiao, and C.M. Wu. Numerical analysis of fatigue cracks in a
reinforced plate with circular holes. International Journal of Fatigue, 41, (2010), 943–952.

[26] C.B. Rhee, A.O. Hui, and A.R.W. Chui. On the initiation of fatigue crack in circular hole
in thick plate: A comparison of experimental and numerical results. International Journal of
Fatigue, 30, (2006), 425–441.

[27] H.P. Wang and C.M. Wu. Numerical study of fatigue crack growth in aluminum plate
with cylindrical holes. Computational and Applied Mechanics, 30, (2008), 941–948.

[28] Y. Wang, H. Liu and S. Lin. Numerical study on crack growth of reinforced composite
structures with circular holes. Journal of the Mechanical Behavior of Biomedical Materials,
10, (2008), 955–964.

[29] M.A. Obaidat, C.R. Hsiao, A.R.W. Chui and C.M. Wu. Numerical study of fatigue
cracks in a reinforced plate with circular holes. International Journal of Fatigue, 41, (2010),
943–952.
[30] C.B. Rhee, H.P. Wang, A.R.W. Chui, A.O. Hui, and S.C. Wei. Analytical study of fatigue
crack initiation in a metal component with circular holes. International Journal of Fatigue, 37,
(2009), 617–622.

[31] C.B. Rhee, H.P. Wang, A.O. Hui, and A.R.W. Chui. Fracture of a plate with circular
holes under cyclic loading. International Journal of Fatigue, 43, (2010), 585–590.

[32] T.B. Smith, G.H. Han. Finite element analysis of cracked composite structures: a
comparison of linear, plane strain, and beam elements. International Journal of Solids and
Structures, 44, (2007), 4897–4906.

[33] D. Tse. An improved XFEM-based finite element model for cracked beam with friction
and frictionless contact conditions. Journal of Mechanical Behavior of Biomedical Materials,
2, (2009), 381–395.

[34] D.T. Tse. Numerical analysis of cracks growth and associated deformation in reinforced
composite materials. Computational and Applied Mechanics, 31, (2011), 109–116.

[35] A. Tsuchiya, Y. Yamada and T. Ogawa. Numerical analysis of cracks growth in a tube
with multiple circumferential flaws. International Journal of Fatigue, 50, (2015), 38–48.

[36] S.C. Wang, Z.C. Tang and W.Z. Wang. Analysis of fatigue crack initiation and
propagation of thick plate with circular holes in an external field. Journal of the Mechanical
Behavior of Biomedical Materials, 6, (2012), 2613–2624.

[37] W. Z. Wang, S. B. Wang, Z. Z. Zhang and Y. C. He. An adaptive mesh refinement

algorithm for a crack problem in circular reinforced plate. Journal of the Mechanical
Behavior of Biomedical Materials, 6, (2012), 2613–2624.

[38] W.Z. Wang, S.B. Wang, Z.C. Tang, Z. Z. Zhang and Y. C. He. An investigation on the
fatigue crack initiation and propagation of circular hole in circular cross section thick plate.
Journal of the Mechanical Behavior of Biomedical Materials, 7, (2013), 49–59.
[39] K. Yamada, T. Fujii, and Y. Iwashita. Numerical study of fatigue crack initiation and
propagation in a circular hole in a plate. International Journal of Fatigue, 29, (2004),
543–554.

[40] K. Yamada, T. Fujii, and Y. Iwashita. Numerical study on fatigue crack initiation and
propagation in a circular hole in a plate. Journal of Materials Research, 17, (2002),
3137–3146.

[41] Z. Zhang and Y. C. He. Effect of the number of circular reinforcement on fatigue crack
initiation in a circular cross section plate. International Journal of Fatigue, 47, (2015),
1253–1260.

[42] Z. Zhang and Y. C. He. Mathematical model of fatigue crack initiation in circular cross
section plate reinforced by circular holes. Acta Mathematica Scientia, 32, (2012), 1517–1530.

[43] Y. Zhang, A.O. Hui and A.R.W. Chui. Numerical study of fatigue crack propagation in a
circular hole in thick plate. International Journal of Fatigue, 42, (2011), 785–798.

[44] J.W. Xu, C.R. Hsiao, H.P. Wang and C.M. Wu. Numerical analysis of crack growth in a
circular reinforced plate with a thin reinforcement layer. International Journal of Fatigue, 44,
(2012), 961–969.

[45] J.W. Xu, C.R. Hsiao, H.P. Wang and C.M. Wu. Effect of fatigue on damage evolution in
reinforced plate with a thick circular hole. International Journal of Fatigue, 49, (2014),
923–937.

[46] J.W. Xu, C.R. Hsiao, Y.N. Huang, H.P. Wang, and C.M. Wu. A comparative analysis of
crack growth in a circular reinforced composite structure. International Journal of Fatigue,
44, (2013), 1007–1014.

[47] A.B. Yener, N.O. Cekirge and G.P. Turgut. Numerical investigation of cracks growth in a
doubly-periodic plate with multiple circular reinforcing holes under combined loads.
Computational and Applied Mechanics, 42, (2012), 569–583.
[48] N. Zhang, G. Xu, Q. Zhao and L. Zhu. Analysis of multi-cracked reinforced composite
structures using an adaptive mesh refinement method. International Journal of Fatigue, 32,
(2008), 1367–1376.

[49] N. Zhang, L. Li and X. Zou. An adaptive meshing algorithm for analysis of crack
propagation in thin shells. Computational and Applied Mechanics, 29, (2008), 189–202.

[50] N. Zhang and L. Li. Numerical analysis of crack growth in thick plates under combined
loads. International Journal of Fatigue, 28, (2005), 705–712.

[51] S. Zhang, Y.H. Zhang and C.H. Zhang. An adaptive mesh refinement analysis of the
effect of the fatigue damage zone and crack growth rate on the stress intensity factors of
composite beams under combined loads. International Journal of Fatigue, 45, (2013), 2–18.

[52] G. Zhang, A.P. Wang, C.M. Wu, A.O. Hui and Z. Z. Zhang. Crack growth in composite
structures with thin steel plates subject to cyclic bending. International Journal of Fatigue, 40,
(2013), 507–522.

[53] Z. Zhang, Y. C. He and A.O. Hui. Crack growth in a circular reinforced composite plate
with a thin circular hole under cyclic loading. International Journal of Fatigue, 47, (2015),
1263–1275.

[54] Z. Zhang, Y. C. He and A.O. Hui. Crack growth in composite structures with a circular
hole in a thick plate reinforced by a layer of thin steel. International Journal of Fatigue, 49,
(2015), 1663–1674.

[55] Z. Zhang, Y. C. He, Y. N. Huang, H. P. Wang and C. M. Wu. Numerical analysis of crack
growth in doubly-periodic plate with multiple circular reinforcement holes under combined
loads. International Journal of Fatigue, 46, (2013), 791–808.

[56] Z. Zhang, H. L. Wu, J. C. Xu and Y. C. He. Effect of fatigue on the stress intensity
factors in doubly-periodic thick-plate reinforced with multiple circular reinforcement.
International Journal of Fatigue, 48, (2014), 1845–1855.
[57] Z. Zhang, H. L. Wu, Z. R. Guo, Y. N. Huang, and Y. C. He. Numerical study of fatigue
crack growth in doubly-periodic plate with circular holes under cyclic loads. International
Journal of Fatigue, 49, (2015), 1053–1065.

[58] Z. Zhang, H. L. Wu, Z. R. Guo, Y. N. Huang and Y. C. He. A comparative analysis of

the stress intensity factors of doubly-periodic plate with multiple circular reinforcement holes
under combined loads. International Journal of Fatigue, 49, (2015), 783–793.

[59] Z. Zhang, H.L. Wu, Z. R. Guo, Y.N. Huang and Y. C. He. Numerical study on fatigue
crack initiation and propagation in circular-hole reinforced plate. International Journal of
Fatigue, 48, (2013), 2057–2065.

[60] Z. Zhang, Z. Z. Guo and Y. C. He. A numerical study of the effects of the hole geometry
on fatigue crack initiation and propagation in doubly-periodic thick plate reinforced with
circular holes. International Journal of Fatigue, 44, (2011), 553–561.

[61] Z. Zhang, H. H. Liu and Y. C. He. A study on the fatigue crack initiation and
propagation in doubly-periodic plate reinforced by circular hole with a thin steel plate.
International Journal of Fatigue, 50, (2016), 1031–1038.

[62] Z. Zhang, Y. N. Huang, H. P. Wang and Y. C. He. Effect of the number of circular
reinforcement on fatigue crack growth in circular cross section plate reinforced by circular
holes. International Journal of Fatigue, 48, (2013), 741–749.

[63] Z. Zhang, Y. N. Huang, H. P. Wang and Y. C. He. Numerical study on fatigue crack
growth in circular reinforced plate with a thin circular hole. International Journal of Fatigue,
47, (2015), 869–877.

[64] Z. Zhang, Y. N. Huang, H. P. Wang, Y. C. He. Effect of fatigue on damage evolution in

doubly-periodic plate with circular reinforcement holes. International Journal of Fatigue, 48,
(2014), 1129–1140.
[65] Z. Zhang, Y. N. Huang, Y. C. He, C. R. Hsiao, H. P. Wang, W. Y. Pan, Z. Z. Guo and Y.
C. Chen. Numerical study on fatigue crack growth in circular reinforced composite structure
with thin steel layer. International Journal of Fatigue, 45, (2013), 757–772.

[66] Z. Zhang, Y. N. Huang, C. R. Hsiao, H. P. Wang and C. M. Wu. Numerical analysis of

crack growth in a circular reinforced plate with a thick circular hole under cyclic bending.
International Journal of Fatigue, 42, (2012), 569–583.