حاله

materials
Article
High-Cycle, Low-Cycle, Extremely Low-Cycle Fatigue
and Monotonic Fracture Behaviors of Low-Carbon
Steel and Its Welded Joint
Younghune Kim 1 and Woonbong Hwang 2, *
1 Graduate Institute of Ferrous Technology, Pohang University of Science and Technology, 77 Cheongam-Ro,
Nam-Gu, Pohang 37673, Korea; gogman@postech.ac.kr
2 Department of Mechanical Engineering, Pohang University of Science and Technology, 77 Cheongam-Ro,
Nam-Gu, Pohang 37673, Korea
* Correspondence: whwang@postech.ac.kr

Received: 14 November 2019; Accepted: 6 December 2019; Published: 9 December 2019
Abstract: Low-carbon steels are commonly used in welded steel structures and are exposed to various
fatigue conditions, depending on the application. We demonstrate that the various transitions in
the fracture mode during fatigue testing can be distinguished by their different cyclic response
curves and microstructural features after fracture. Fractography, surface damage micrographs,
and microstructural evolution clearly indicated the transition of the fracture modes from high-cycle
to low-cycle, extremely low-cycle fatigue, and monotonic behavior. The high-cycle fatigue mode
showed initial cyclic softening, followed by cyclic stabilization, and showed inclusion-induced crack
initiation at fish-eyes, while the low-cycle fatigue mode showed initial cyclic hardening followed by
cyclic stabilization, where fractography images showed obvious striations. In addition, the extremely
low-cycle fatigue mode showed no cyclic stabilization after initial cyclic hardening, which was
characterized by quasi-cleavage fractures with a few micro-dimples and transgranular cracking, while
the monotonic fracture mode predominantly showed micro-dimples.
Keywords: low-carbon steel; fatigue modes; extremely low-cycle fatigue; fatigue test; fatigue transition
1. Introduction
Low-carbon steels generally contain 0.05 to 0.2 wt.% (extensively up to 0.3 wt.%) carbon along with
other alloying elements, such as manganese and silicon. Such steels are generally used for structural
applications, where the strength and ductility can be optimized via thermo-mechanical controlled
rolling (TMCR) or heat-treatment processes. For practical application, steel parts are usually joined by
welding, which can introduce residual stresses and heterogeneous microstructures [1,2] that can result
in locally inferior mechanical properties compared to the base material. For example, the fatigue life of
welds is shorter than that of the base material due to welding defects [3,4] or heat-affected zone (HAZ)
softening [5,6], where fatigue fracture occurs in these regions. Low-carbon steels are readily joined
with several common fusion welding processes such as plasma arc welding (PAW) and gas tungsten
arc welding (GTAW), but some studies [7–9] reported that friction stir welding process as an solid
state welding showed superior fatigue strength due to the synergetic effect of microstructure, superior
tensile properties and favorable residual stress, which inhibit the growth of cracks compared to other
joints. Welded steel structures are subjected to cyclic loading conditions from high-cycle fatigue (HCF)
to low-cycle fatigue (LCF), or even to extremely low-cycle fatigue (ELCF), whereas uniaxial loading
results in monotonic fracture (MF).
The relevant fatigue fracture modes can be identified by analyzing the fatigue life and stress levels,
as illustrated by Figure 1 [10].
Materials 2019, 12, 4111; doi:10.3390/ma12244111 www.mdpi.com/journal/materials

However, little attention has been paid to the consecutive fatigue transitions of low-carbon steel
depending on cyclic response curves and fractography in addition to strain (stress) levels and fatigue
life.
Moreover, for most structural applications, the different fatigue modes and fatigue lives of the
base material and welded joints at various strain (stress) amplitudes have not been thoroughly
Materials 2019, 12, 4111 2 of 23
elucidated, even though this is fundamental information for defining engineering safety factors.
log 𝜎
Fatigue limit
MF ELCF LCF HCF

log 𝑁
Schematics of classified fatigue modes [10]; abbreviations of fatigue modes are used as MF
Figure 1. Schematics
for monotonic fracture, ELCF for extremely low-cycle fatigue, LCF for low-cycle fatigue, and HCF for
high-cycle fatigue.
The elastoplasticity
The theories arethe
present study compared generally
fatigue assumed
behavioronof the basis of the steel
a low-carbon decomposition of theusing
with its welds total
strain into elastic and plastic components so as to treat the elastic and plastic parts
cyclic response curves and fractography over a range of strain amplitudes in order to describe the of the strains
separately [11,12].
transitions between the HCF, LCF, ELCF, and MF modes, which is critical for designing durable
Previous
structures using studies focused on differentiating the classical HCF and LCF modes, and ELCF from LCF.
this material.
The HCF mode is stress controlled and associated with relatively low stress levels that are relevant
to Materials
2. predominantly elastic components [13]. In this mode, the fatigue life is >104 cycles. In contrast,
and Methods
the LCF mode is strain controlled and related to large strain amplitudes or high stress levels that
2.1. Test Materials
produce and and
both plastic elastic components at each cycle [14], where fatigue lives are <104 cycles.
Welding
The ELCF mode is between the MF and LCF modes and is controlled by the ductility and cyclic strain;
The test material was a 4 mm-thick low-carbon steel sheet, which was a thermo-mechanical
initial damage under tensile conditions is followed by ductility exhaustion during cyclic loading until
controlled rolled coiled tubing (CT) steel [4] with the chemical composition shown in Table 1.
crack propagation occurs [15,16]. The fatigue life of ELCF is very short (<100 cycles). In other studies,
Thermo-mechanical controlled rolling is a controlled process, extending from slab reheating, via
specific fatigue modes and stress levels of stainless steels, copper alloys, aluminum alloys, and Ti alloys
rolling, up to and including (accelerated) cooling and, where necessary, tempering. Rolling is
as well as low-carbon steels have been extensively investigated [14–24].
However, little attention has been paid to the consecutive fatigue transitions of low-carbon
steel depending on cyclic response curves and fractography in addition to strain (stress) levels and
fatigue life.
Moreover, for most structural applications, the different fatigue modes and fatigue lives of the base
material and welded joints at various strain (stress) amplitudes have not been thoroughly elucidated,
even though this is fundamental information for defining engineering safety factors.
The present study compared the fatigue behavior of a low-carbon steel with its welds using cyclic
response curves and fractography over a range of strain amplitudes in order to describe the transitions
between the HCF, LCF, ELCF, and MF modes, which is critical for designing durable structures using
this material.
2. Materials and Methods
2.1. Test Materials and Welding

The test material was a 4 mm-thick low-carbon steel sheet, which was a thermo-mechanical
controlled rolled coiled tubing (CT) steel [4] with the chemical composition shown in Table 1.
Thermo-mechanical controlled rolling is a controlled process, extending from slab reheating,
via rolling, up to and including (accelerated) cooling and, where necessary, tempering. Rolling
is performed in accordance with a defined rolling-pass schedule, with finish rolling in the
non-recrystallised austenite [25,26]. The essential effect of thermo-mechanical controlled rolling
2019, 12,
Materials 2019, 12, 4111
x FOR PEER REVIEW 33 of 24
of 23
performed in accordance with a defined rolling-pass schedule, with finish rolling in the non-
is the attainment
recrystallised of the fine
austenite grain,The
[25,26]. this essential
being the effect
necessary basis for the combination
of thermo-mechanical of high
controlled mechanical
rolling is the
strength and toughness with only low alloying-element contents. For instance, Gladman
attainment of the fine grain, this being the necessary basis for the combination of high mechanical [27] mentioned
astrength
precipitation strengthening
and toughness withuponly
to 240low
MPaalloying-element
containing 0.09 wt.% Nb onFor
contents. the instance,
assumption of complete
Gladman [27]
NbC precipitation at the nanoscale. Funakawa et al. [28] reported the increased
mentioned a precipitation strengthening up to 240 MPa containing 0.09 wt.% Nb on the assumption yielding strength
up to 300 MPa
of complete NbC due to the precipitation
precipitation of nanometer
at the nanoscale. Funakawa size (Ti,Mo)C
et al. [28] carbides. CT increased
reported the steel is one of the
yielding
thermo-mechanical
strength up to 300 MPa controlled
due to steels used in oil field
the precipitation industriessize
of nanometer for (Ti,Mo)C
applications such asCT
carbides. work
steelstrings,
is one
drill, siphon, and velocity strings as well as sucker rod systems [29].
of the thermo-mechanical controlled steels used in oil field industries for applications such as work
strings, drill, siphon, and velocity strings as well as sucker rod systems [29].
Table 1. Chemical
In the welding experiments, composition
a plasma of coiled
arc welding tubing was
process (CT) adopted
steel (wt.%).
to weld the low-carbon
steel sheets. A single Cpass was Mn conducted
Si with
P aS currentOther
of 250 A at a Elements
Alloying travel speed of 250 mm/min
using a Powwel 500XP machine (Powwel Co., Ltd., Seoul, South Korea). An AWS A5.18:ER70S-6 filler
0.12 0.87 0.36 0.01 0.001 Cr, Ni, Mo, Nb, Ti
wire was fed at a rate of 1500 mm/min, and 100% argon shielding gas was supplied at a flow rate of 1.5
L/min on the front surface and at 20 L/min to the back surface. After welding, a stress relief heat
In the(SRHT)
treatment weldingwasexperiments,
conductedaatplasma
600 °Carc
for welding
15 min toprocess
relievewas adopted
residual to weld
stresses [30].the low-carbon
steel sheets. A single pass was conducted with a current of 250 A at a travel speed of 250 mm/min
using a Powwel 500XP machine
Table (Powwel
1. Chemical Co., Ltd.,ofSeoul,
composition coiled South
tubing Korea). An
(CT) steel AWS A5.18:ER70S-6 filler
(wt.%).
wire was fed at a rate of 1500 mm/min, and 100% argon shielding gas was supplied at a flow rate of
C Mn Si P S Other Alloying Elements
1.5 L/min on the front surface and at 20 L/min to the back surface. After welding, a stress relief heat
0.12 0.87 0.36 0.01 0.001
treatment (SRHT) was conducted at 600 ◦ C for 15 min to relieve Cr, Ni, Mo,stresses
residual Nb, Ti [30].
2.2. Tensile and

and Fatigue
Fatigue Testing
Testing
Tensile and
Tensile and fatigue
fatigue tests
tests of
of the
the base
base material
material (BM) (BM) andand welded
welded joints
joints (WJ)
(WJ) specimens
specimens were
performed using aa 100100 kN
kN MTS
MTS closed-loop
closed-loop servo-hydraulic
servo-hydraulic materials
materials testing
testing machine.
machine. The tensile
tensile
and fatigue test specimens were machined from steel sheets cut longitudinally to the rolling direction.
longitudinally the rolling direction.
study, the
In the present study, the BM
BM specimens
specimens werewere composed
composed only only of
of the
the base
base metal that retains the TMCR
microstructure. The WJ specimen were composed of the weld metal where the portions of the base base
metal has
metal and filler metal has been
been completely
completely meltedmelted during
during welding,
welding, the HAZ HAZ where
where mechanical
mechanical
properties or microstructure have been altered from the base metal by the heat of welding, welding, and the
base metal. The geometry of the fatigue specimens is shown in Figure Figure 2. 45°◦ bias
2. A 45 bias welding
welding design
design
was adopted
was adopted [31], which
which allows a smoother transition between the mechanical properties of the base
material and welds, while increasing the strength and decreasing the ductility ductility by distributing
distributing the
welding
welding stresses over a larger area than a 90°
over a larger area than a 90 bias. ◦ bias. After machining, a smooth surface finish was
was
obtained by polishing with emery papers of increasing fineness (320 down to 1000 grit). Tensile tests
were conducted
were conductedat ataaconstant
constantstrain
strainrate
rateofof6.7
6.7××10 10−3−3 s−1
−1. Fatigue tests were conducted in an identical
. Fatigue tests were conducted
manner for both the BM and WJ specimens at 0.05 Hz to 0.033 Hz under total strain control for total
strains of
strains of ±0.5%,
±0.5%,±1.0%,
±1.0%,±1.5%,
±1.5%, ±2.0%,
±2.0%, ±2.5%,
±2.5%,andand±3.0% ±3.0%using a triangular
using strain
a triangular waveform
strain withwith
waveform zero
mean strain at a constant strain rate of 4 × 10 −3 s−1. A
−3 12-mm
−1 extensometer was
zero mean strain at a constant strain rate of 4 × 10 s . A 12-mm extensometer was used with the used with the 16-mm-
gauge-length specimens.
16-mm-gauge-length Henceforth,
specimens. the abbreviations
Henceforth, the abbreviationsBX and BXWXandare
WXused for the
are used for BM andand
the BM WJ
specimens, respectively, where X is the strain; for example, B0.5 is the BM specimen
WJ specimens, respectively, where X is the strain; for example, B0.5 is the BM specimen tested at 0.5% tested at 0.5%
nominal strain.
nominal strain.
within ±
Figure 2. Axial fatigue specimen geometry; all units are in millimeters of tolerance within ± 0.05 mm,
with
with the
the stippled
stippled line
line indicating
indicating the
the welding
welding zone.
zone.
Materials 2019, 12, 4111 4 of 23
Materials 2019, 12, x FOR PEER REVIEW 4 of 24
2.3. Microscopy
Before observing
Before observingthe themicrostructure,
microstructure, immersion
immersion etching
etching waswas performed
performed using
using a standard
a standard 2%
2% nital
nital solution
solution (2%acid
(2% nitric nitric acid in followed
in ethanol) ethanol) by followed byethanol
rinsing in rinsingandin drying
ethanolinand drying
hot air. in hot air.
Microstructural
Microstructural characterization was performed on mounted cross-sections
characterization was performed on mounted cross-sections of each sample type. Wet grindingof each sample type. was
Wet
grinding was
performed performed
using usingup
SiC papers SiC
to papers up to
2000 grit, 2000 grit,
followed byfollowed by fineusing
fine polishing polishing usingand
diamond diamond
silica
and silica suspensions.
suspensions. Metallographic
Metallographic studies werestudies were using
performed performed using an
an Olympus Olympus
AX70 (OlympusAX70 (Olympus
Corporation,
Corporation,
Tokyo, Tokyo, Japan)
Japan) inverted inverted
metallurgical metallurgical
microscope. microscope. FM-700
The Future-Tech The Future-Tech FM-700
Vickers hardness Vickers
tester was
hardnesswith
applied tester was of
a load applied
9.81 Nwith
and aa load
dwelloftime
9.81of N 10
and a dwell
s to time of
the points, 10 s to the
regularly 0.3 points, regularly
mm distanced 0.3
from
mmcenter
the distanced
of thefrom
weldthe center
metal. In of the weld
addition, metal. In addition,
high-magnification high-magnification
fractographs fractographs
of ultrasonically cleanedof
ultrasonically cleaned samples were conducted using a JEOL JSM-6610LV
samples were conducted using a JEOL JSM-6610LV scanning electron microscope (SEM, JEOL Ltd., scanning electron
microscope
Tokyo, Japan)(SEM,
usingJEOL Ltd., Tokyo,
a secondary Japan)
electron using a secondary electron detector.
detector.
3. Results and Discussion
3.1. Microstructures
3.1. Microstructures
The
The microstructures
microstructures of of the
the base
base metal
metal and
and weld
weld metal
metal are
are shown
shown in in Figure
Figure 3.
3. The
The base
base metal
metal had
had
aa rolled fine polygonal ferrite (PF)-discontinuous pearlite band (P) microstructure with
rolled fine polygonal ferrite (PF)-discontinuous pearlite band (P) microstructure with the average the average
grain
grain size of 4.75
size of 4.75 µm in diameter
µm in diameter (Figure
(Figure 3a,b),
3a,b), which
which are
are typical
typical microstructures
microstructures in in low-carbon
low-carbon steels,
steels,
and
and aa tensile
tensile strength
strength of
of 600
600 MPa.
MPa. The
The weld
weld metal
metal showed
showed coarse
coarse granular
granular bainite
bainite with
with the
the average
average
grain
grain size of 33.8 µm in diameter (Figure 3c), but the structure was finer than that produced by
size of 33.8 µm in diameter (Figure 3c), but the structure was finer than that produced by gas
gas
tungsten
tungsten arcarcwelding
weldingprocess,
process,because
becauseplasma
plasma arcarc
welding
weldingprocess provides
process focused
provides heatheat
focused transfer that
transfer
results in a lower
that results heat heat
in a lower inputinput
leading to improved
leading to improvedpenetration in theinsmaller
penetration heat-affected
the smaller zonezone
heat-affected [32].
Previous studies observed fatigue fractures at the softened HAZ [5,6]; however, the
[32]. Previous studies observed fatigue fractures at the softened HAZ [5,6]; however, the inferior inferior coarse
WJ microstructures
coarse with lowwith
WJ microstructures hardness
low produced
hardness using a welding
produced using consumable with low strength
a welding consumable with orlowa
process with high heat input can cause fatigue fracture [3,4].
strength or a process with high heat input can cause fatigue fracture [3,4].
Figure 3. Microstructures
Microstructuresofof(a)(a)
base material,
base (b) (b)
material, high-magnification image
high-magnification of base
image material,
of base and
material,
microstructures of (c) weld metal.
and microstructures of (c) weld metal.
3.2. Mechanical Properties

3.2. Mechanical Properties
The
The tensile
tensilecurves
curvesininFigure
Figure4 show
4 showthatthat
the BM
the specimen had a yield
BM specimen had astrength and tensile
yield strength andstrength
tensile
up to 10% lower than those of the WJ specimen. However, the ductility of the
strength up to 10% lower than those of the WJ specimen. However, the ductility of the BM specimen BM specimen was
6.63
was times that ofthat
6.63 times the of
WJthe
specimen, as measured
WJ specimen, using Miranda’s
as measured equationequation
using Miranda’s [33], where
[33],the ductility
where the
was calculated from the ratio of the maximum displacement (D ) to the corresponding
ductility was calculated from the ratio of the maximum displacement (Du) to the corresponding
u displacement at
the onset of yielding
displacement at the onset (D ). At low strain ranges, materials of higher tensile strength generally
y of yielding (Dy). At low strain ranges, materials of higher tensile strength have
higher crack-growth
generally have higherresistance.
crack-growth By contrast,
resistance.at high strain ranges,
By contrast, at highmaterials of higher
strain ranges, ductility
materials have
of higher
higher crack-initiation resistance. Since strength level and ductility are usually
ductility have higher crack-initiation resistance. Since strength level and ductility are usually inversely related,
fatigue
inverselyresistance
related,involves
fatigue aresistance
tradeoff among strength
involves and ductility
a tradeoff among on strength
the assumption of high fracture
and ductility on the
toughness without defects [34].
assumption of high fracture toughness without defects [34].
Previous studies showed that the fatigue properties are strongly associated with the mechanical
properties of the material [35,36] and suggested that HCF resistance requires a high monotonic tensile
strength [35], where the stress-based Basquin law is often used to estimate the HCF life. In contrast,
LCF resistance requires a high monotonic ductility [36] and the strain-based Coffin–Manson law is
ductility, although the relative proportions are still being debated [37–39]. In early studies [15,40], the
ELCF mode was characterized as a rapid reduction and exhaustion of the residual ductility at very
large strain amplitudes, resulting in a short fatigue life (<100 cycles). As a result of this approach of
including the ductility when considering fatigue failure in the ELCF mode, it is important to elucidate
the fatigue
Materials 2019,resistance
12, 4111 of the WJ specimen in order to determine whether a tolerable level of strain
5 of is
23
exceeded during practical operation.
800 BM
WJ
600
Stress (MPa)
400
200
0 5 10 15 20 25 30 35
Strain (%)
Figure 4.
Figure Tensile curves
4. Tensile curves of base material
of base material(BM)
(BM) and
and welded
welded joint
joint (WJ)specimens.
(WJ)specimens.
Previous
The cyclicstudies showed that
stress response the fatigue
behavior properties
is determined by are
the strongly
mechanical associated
properties with of the mechanical
a material and
properties
should be of the material
considered when[35,36] and suggested
estimating the fatiguethat HCF resistance requires
life. Comprehensive a high[41,42]
studies monotonic havetensile
been
strength [35],
conducted where the
to identify thestress-based
major parametersBasquin law isfatigue
affecting often used to estimate
life. First, the HCFtolife.
it is necessary In contrast,
separate cyclic
strains into elastic and plastic components to elucidate their strain–stress behaviors under law
LCF resistance requires a high monotonic ductility [36] and the strain-based Coffin–Manson cyclicis
often used
loading. to predict
The the LCF
total strain life. However,
amplitude (Δεt) isELCF
defined resistance requires
by Equation both
(1), tensile
where Δεstrength and ductility,
p is the plastic strain
although the relative proportions are still being debated [37–39].
amplitude, Δεe is the elastic strain amplitude, E is the elastic modulus (Young modulus), and In early studies [15,40], the ELCF
Δσ is
mode
the wasamplitude.
stress characterized as a rapid reduction and exhaustion of the residual ductility at very large
strain amplitudes, resulting in a short fatigue life (<100 cycles). As a result of this approach of including
Δ𝜀 𝛥𝜀 Δ𝜀 𝛥𝜀 𝛥𝜎
the ductility when considering fatigue failure = in the + ELCF = mode, + it is important to elucidate the fatigue (1)
2 2 2 2 2𝐸
resistance of the WJ specimen in order to determine whether a tolerable level of strain is exceeded
In practical
during addition,operation.
the elastic strain amplitude can be obtained from extracting the elastic strain
components
The cyclicfrom theresponse
stress total strain amplitudes,
behavior as shown
is determined by thein mechanical
Equation (2), where σ′
properties off aismaterial
the fatigue
and
strength coefficient, 2N f is the number of reversals to failure, and b is the fatigue strength exponent.
should be considered when estimating the fatigue life. Comprehensive studies [41,42] have been
conducted to identify the major parameters E
Δ𝜀 affecting fatigue life. First, it is necessary to separate
= 𝜎 (2𝑁 ) (2)
cyclic strains into elastic and plastic components 2 to elucidate their strain–stress behaviors under cyclic
loading. The total is
This equation strain
similaramplitude (∆εt )equation
to Basquin’s is defined [43]byandEquation
has been (1), where
shown to∆ε is the plastic
fitpmaterials strain
exceeding
amplitude, ∆ε
the fatigue limit is the elastic strain amplitude, E is the elastic modulus (Young
e well. It can be deduced that high σ′f and low b values increase the fatigue strength. modulus), and ∆σ is the
stress
The amplitude.
plastic strain components are described using the Coffin–Manson equation [44], as shown in
∆εt ∆εp ∆εe ∆εp ∆σ
Equation (3), where, ε′f is the fatigue ductility = coefficient,
+ = and+c is the fatigue ductility exponent. (1)
2 2 2 2 2E
Δ𝜀
In addition, the elastic strain amplitude = can𝜀′ be
(2𝑁obtained
) from extracting the elastic strain (3)
2
components from the total strain amplitudes, as shown in Equation (2), where σ0 f is the fatigue strength
It is also
coefficient, 2Nclear that high ε′f and low c values lead to an increase in the fatigue life. Manson et al.
f is the number of reversals to failure, and b is the fatigue strength exponent.
[45] estimated the fatigue resistance by superposition of plastic and elastic components to give the
fatigue strength at a specific strain amplitude. ∆εe E Thus,0 Equations
b (1)–(3) can be combined to obtain an
= σ f 2N f (2)
expression between the total strain amplitude 2 (Δεt), plastic strain amplitude (Δεp) and elastic strain
amplitude (Δεe) as aisfunction
This equation similar to ofBasquin’s
the number of reversals
equation [43] andto failure
has been (2N f).
shown to fit materials exceeding
the fatigue limit well. It can be deduced that high σ0 f and low b values increase the fatigue strength.
The plastic strain components are described using the Coffin–Manson equation [44], as shown in
Equation (3), where, ε0 f is the fatigue ductility coefficient, and c is the fatigue ductility exponent.
∆εp c
= ε0 f 2N f (3)
2
Materials 2019, 12, 4111 6 of 23
It is also clear that high ε0 f and low c values lead to an increase in the fatigue life. Manson et al. [45]
estimated the fatigue resistance by superposition of plastic and elastic components to give the fatigue
strength at a specific strain amplitude. Thus, Equations (1)–(3) can be combined to obtain an expression
between the total strain amplitude (∆εt ), plastic strain amplitude (∆εp ) and elastic strain amplitude
(∆εe ) as a function of the number of reversals to failure (2Nf ).
∆εtΔ𝜀 ∆ε𝛥𝜀
p 𝛥𝜀e
∆ε c σ0 f
𝜎′ b
= = ++ = ε𝜀′0 f (2𝑁
= 2N f ) + (2𝑁2N f) (4)
2 2 22 22 𝐸E
In
In addition,
addition, the
the hardening
hardening coefficient
coefficient (K´)
(K0 ) and
and cyclic
cyclic hardening
hardening exponent
exponent (n´)
(n0 ) are
are described
described by
by
Ramberg–Osgood
Ramberg–Osgood equation:
equation:
/
Δ𝜀 ∆ε𝛥𝜀 Δ𝜀 𝛥𝜎 1/n0 𝛥𝜎
∆εt = p +∆εe = ∆σ + ∆σ (5)
2= 2 + 2 = 2𝐾0 + 2𝐸 (5)
2 2 2 2K 2E
The following equations for the material parameters can be derived from Equations (4) and (5).
The following equations for the material parameters can be derived from Equations (4) and (5).
𝑏
𝑛 = (6)
b𝑐
n0 = (6)
c
𝜎
𝐾 = σ0 (7)
0 (𝜀f )
K = n0 (7)
A linear Coffin–Manson relationship for the BM ε0f specimen was apparent, as plotted in Figure 5
on a log–log scale. The material parameters for the fatigue resistance (e.g., σ′f, b, ε′f, c, n´ and K´) were
A linear
obtained; Coffin–Manson
ε′f and relationship
c from the slope for theof
and exponent BM thespecimen was apparent,
plastic component as plotted(4),
of Equation in σ′
Figure 5 on
f and low
a log–log scale. The material parameters for the fatigue resistance
b from the slope and exponent of the elastic component of Equation (4), n´ from (e.g., σ 0 , b, ε0 , c, n0 and K0 ) were
f Equation
f (6), and K´
obtained;
from ε0 f and
Equation (7)c were
from subsequently
the slope and exponent
obtained, of andthelisted
plastic
in component
Table 2. of Equation (4), σ0 f and low b
fromTypically,
the slope and exponent of the elastic component of Equation 0 from Equation (6), and K0
(4), nspecimen
the slopes of the plastic strain amplitudes of the BM showed a drastic
from Equation
decrease (7) weretosubsequently
as compared obtained,
those of the elastic strainand listed in Table 2.
amplitudes.
-1
10
-2
Strain amplitude, Δε
10
-3
10
Total (BM)
Elastic (BM)
Inelastic (BM)
-4
10
0 1 2 3 4 5 6
10 10 10 10 10 10 10
Reversals to failure, 2Nf (cycles)
Figure
Figure 5.
5. Fatigue
Fatigue strain-life
strain-life curve
curve obtained
obtained by
by superposition
superposition of
of elastic
elastic and
and plastic
plastic strains.
strains.
Table 2.
Table Fatigue parameters
2. Fatigue parameters from
from equations.
equations.
σf𝝈𝒇 εf 𝜺𝒇
0 0
b𝒃 𝒄 n 𝒏 K 𝑲
0 0
Material
Material c
BM
BM 1098 −0.1050
1098 −0.1050 0.8232
0.8232 −0.6360
−0.63600.1633
0.16331125.6
1125.6
3.3. Fatigue Resistance at Welded Joint

The relationship of stable cyclic stress–strain for origin state material can be described by
Equation (5) and the cyclic stress–strain response curves of the BM and WJ specimens are shown in
Figure 6. Since the local stress–strain approach was developed for low-cycle fatigue, the prediction
Materials 2019, 12, 4111 7 of 23
Typically, the slopes of the plastic strain amplitudes of the BM specimen showed a drastic decrease
as compared to those of the elastic strain amplitudes.
3.3. Fatigue Resistance at Welded Joint

The relationship of stable cyclic stress–strain for origin state material can be described by
Equation (5) and the cyclic stress–strain response curves of the BM and WJ specimens are shown in
Figure 6. Since the local stress–strain approach was developed for low-cycle fatigue, the prediction
curves of fatigue lives in a life region share good agreement with experimental results. On the one
hand, the stress values for the WJ specimen at all strain amplitudes were commonly higher than7those
Materials 2019, 12, x FOR PEER REVIEW of 24
for the BM specimen. We propose that different plastic deformation occurred in the WJ specimen
because
because thethe WJ
WJ specimen
specimenhashasinhomogeneous
inhomogeneouscomplex
complexmicrostructure oror
microstructure strength mismatching
strength andand
mismatching its
strain ranges
its strain rangesshould be modified
should to compare
be modified with
to compare the the
with BMBM
specimen.
specimen.
800
600
Stress amplitude, Δσ (MPa)
400
BM, exp
WJ, exp
200 BM, pred
curve
WJ, pred
curve
0
0.000 0.005 0.010 0.015 0.020 0.025 0.030
Total strain amplitude, Δε
Figure 6. Cyclic stress–strain curves in the BM and WJ specimens.
Earlier studies have

Earlier studies have shownshown thatthat the
the fatigue
fatigue strength
strength of of welded
welded joints
joints isis different
different from
from that
that in
in
homogeneous
homogeneous base base materials
materials due due toto the
the welding
welding defects
defects [46]
[46] oror weld
weld toetoe [47],
[47], and
and the
the high
high level
level of
of
residual stresses [48], complex microstructure, or strength
residual stresses [48], complex microstructure, or strength mismatching [49,50]. mismatching [49,50].
It should be
It should be noted
noted thatthat the
the existence
existence of of welding
welding defects
defects such
such asas porosities,
porosities, lacklack of
of fusion,
fusion, oror lack
lack
of
of penetration
penetration may mayintroduce
introducecracks cracksororcrack-like
crack-likedefects,
defects,which
which were
were found
found to be areas
to be preferable
areas for
preferable
fatigue crack initiation. In addition, the weld toe acts as the geometric stress
for fatigue crack initiation. In addition, the weld toe acts as the geometric stress concentrations and concentrations and fatigue
cracks
fatigueprimarily appear throughout
cracks primarily the weld the
appear throughout toe.weld
In thetoe.
present
In thestudy,
presentultrasonic
study, non-destructive
ultrasonic non-
testing was conducted for the welds to eliminate the welding defects
destructive testing was conducted for the welds to eliminate the welding defects that can that can cause the test cause
resultsthe
to
become corrupted. In addition, the weld toe was removed by milling the
test results to become corrupted. In addition, the weld toe was removed by milling the surface of the surface of the welding seam
to eliminate
welding seam thetostress
eliminateconcentration
the stress caused by geometric
concentration causeddiscontinuity.
by geometric discontinuity.
Residual
Residual stresses introduced by the welding process are detrimental
stresses introduced by the welding process are detrimental and and may
may reduce
reduce the fatigue
the fatigue
strength by accelerating fatigue crack initiation and growth. In addition,
strength by accelerating fatigue crack initiation and growth. In addition, the severe non-uniform the severe non-uniform
temperature distribution in
temperature distribution in the
the HAZ
HAZ around
around the the weld
weld can
can cause
cause residual
residual thermal
thermal and and mechanical
mechanical
stresses that greatly influence the microstructural evolution in this region
stresses that greatly influence the microstructural evolution in this region [1,2]. In other studies [1,2]. In other studies [51,52],
however, eliminating residual stresses did not lead to an improvement
[51,52], however, eliminating residual stresses did not lead to an improvement in the low-cycle in the low-cycle fatigue strength
and its strength
fatigue effect is notandtoo criticalisand
its effect can critical
not too be ignoredand canas significant
be ignoredplasticity can occur
as significant fromcan
plasticity external
occur
loading that relaxes the residual stresses. In the present study, a SRHT was conducted at 600 ◦ C for
from external loading that relaxes the residual stresses. In the present study, a SRHT was conducted
15 min°C
at 600 to for
remove
15 mineven the minor
to remove effects
even of residual
the minor effectsstresses.
of residual stresses.
For the smooth welded joint, mechanical inhomogeneity (complex microstructure or strength
mismatching) can be the main reason that leads to the decrease of the low-cycle fatigue strength,
although the stress concentrations and residual stresses were eliminated. In this test, at all the strain
amplitudes, fatigue crack initiation and propagation occurred at the heat-affected zones, where the
plasticity and ductility were weak. Although experimental tests were used to obtain material
mechanical properties, it is difficult to clarify these parameters owing to size limitation of the weld
zone. Because hardness is easier than other mechanical properties to determine and it does not
Materials 2019, 12, 4111 8 of 23
For the smooth welded joint, mechanical inhomogeneity (complex microstructure or strength
mismatching) can be the main reason that leads to the decrease of the low-cycle fatigue strength,
although the stress concentrations and residual stresses were eliminated. In this test, at all the
strain amplitudes, fatigue crack initiation and propagation occurred at the heat-affected zones, where
the plasticity and ductility were weak. Although experimental tests were used to obtain material
mechanical properties, it is difficult to clarify these parameters owing to size limitation of the weld
zone. Because hardness is easier than other mechanical properties to determine and it does not require
much space or material, many studies [53–56] used hardness distribution to determine the welded
joint material properties as proportional to the base material properties, and have found a simple linear
relation between the yield strength and hardness, including the following proportional form:
Hvi
Ki0 = Ki0 , i = 1, 2, , , n (8)
HvBM
where K0 i , K0 BM , Hvi and HvBM are the cyclic strength coefficient in the ith zone, the cyclic strength
coefficient in the base material, the Vickers hardness in the ith zone, and the Vickers hardness value in
the base material, respectively.
The fatigue parameters b, c of different zones are set as constant, and the fatigue strength coefficient
σ0 f is in proportion to hardness, that is
Hvi
σ0f i = σ0f BM , i = 1, 2, , , n (9)
HvBM
where σ0 f , σ0 BM are the fatigue strength coefficient in the ith zone and the base material, respectively.
For K0 , it can be obtained by Equation (7). Here ε0 f is treated as constant because K0 , σ0 f are in proportion
to hardness value. The hardness distribution obtained for the welded joints is shown in Figure 7. In this
study, it assumed all specimens have the same hardness distribution in the heat-affected zone and strain
localization occurred. Hence there are total four HAZs near the weld metal, including Zone I, II, III, IV,
and the Vickers hardness values decrease gradually and recover in this zone, from weld metal to base
metal. The cyclic stress–strain parameters and strain-life curve parameters determined from hardness
and base material parameters according to Equations (8) and (9) are shown in Table 3. The cyclic
stress–strain curves and empirical data of the base material are shown in Figure 8a. The predicted
fatigue lives in a life region share good agreement with experimental results. For the welded joint,
the cyclic stress–strain curve at Zone III was used where the fatigue fracture occurred, and the strain
range for experimental stress amplitude was adjusted as plotted in Figure 8b since the nominal strain
cannot be directly matched to the local strain. After this rearrangement, fatigue strain-life curve
obtained by superposition of the BM and WJ specimens was plotted in Figure 9 on a log–log scale.
All WJ specimens were in higher strain ranges than the BM specimens.
Table 3. Cyclic stress–stain and fatigue parameters with hardness distribution.

0
σf
0 0
Zone Hardness(Hv) E b c K n
Base material 202 203.0 1098.0 −0.1050 −0.6360 1125.6 0.1633
Weld metal 243 203.0 1322.3 −0.1050 −0.6360 1355.6 0.1633
HAZ-I 221 203.0 1203.8 −0.1050 −0.6360 1234.0 0.1633
HAZ-II 198 203.0 1075.1 −0.1050 −0.6360 1102.2 0.1633
HAZ-III 188 203.0 1020.9 −0.1050 −0.6360 1046.5 0.1633
HAZ-IV 200 203.0 1088.2 −0.1050 −0.6360 1115.6 0.1633
experimental results. For the welded joint, the cyclic stress–strain curve at Zone III was used where
the fatigue fracture occurred, and the strain range for experimental stress amplitude was adjusted as
plotted in Figure 8b since the nominal strain cannot be directly matched to the local strain. After this
rearrangement, fatigue strain-life curve obtained by superposition of the BM and WJ specimens was
plotted 2019,
Materials in Figure
12, 41119 on a log–log scale. All WJ specimens were in higher strain ranges than the BM
9 of 23
specimens.
300
Weld metal HAZ Base metal
I II III IV
Hardness (Hv) 250
200
0 2 4 6 8 10 12
Distance from weld metal center (mm)
Materials 2019,
Materials 2019, 12,
12, xx FOR
FOR PEER
PEER REVIEW
REVIEW 99 of
of 24
24
Figure 7. Microhardness profile of WJ specimen; arrow indicates fatigue fracture spot.
Figure 7. Microhardness profile of WJ specimen; arrow indicates fatigue fracture spot.
1000 1000
aa bb
1000 1000
BM,exp
exp WJ,exp
WJ,exp
BM, WJ,pred
Table 3. Cyclic stress–stainBM,
and
BM, pred
predfatigue parameters with hardness distribution. WJ,pred
800 800
800 800
Hardness
Zone E 𝝈𝒇 𝒃 𝒄 𝑲 𝒏
(Hv)
(MPa)
(MPa)
600 600
Stress,σσ(MPa)
Stress,σσ(MPa)
600 600
Base material 202 203.0 1098.0 −0.1050 −0.6360 1125.6 0.1633

Weld metal 243 203.0 1322.3 400 −0.1050 −0.6360 1355.6 0.1633
Stress,
Stress,
400 400
400
HAZ-I 221 203.0 1203.8 −0.1050 −0.6360 1234.0 0.1633

200 200
200 HAZ-II 198 203.0 1075.1 200 −0.1050 −0.6360 1102.2 0.1633
HAZ-III 188 203.0 1020.9 −0.1050 −0.6360 1046.5 0.1633
0 0
00.00
0.00HAZ-IV
0.02
0.02
0.04
0.04
0.06
0.06 200 0.10
0.08
0.08
0.10 203.0
0.12
0.12
0.14
0.14 1088.2 00.00
0.00−0.1050
0.02
0.02 0.04 −0.6360
0.04 0.06
0.06
0.08
0.08 1115.6
0.10
0.10 0.12 0.1633
0.12 0.14
0.14
Strain,εε Strain,εε
Strain,
Strain,
Figure 8.
Figure
Figure 8. Cyclic
8. Cyclic stress–strain
Cyclic stress–strain curves
stress–strain curves in the
curves in
in the (a)
the (a) BM
(a) BM and
BM and (b)
and (b) WJ
(b) WJ specimens.
WJ specimens.
specimens.
-1
10-1
BM
BM
10 WJ
WJ
-2
10-2
10
Strain,εε
Strain,
-3
10-3
10
-4
10-4
10 0 1 2 3 4 5 6
10100 101
10 102
10 103
10 104
10 105
10 106
10
Reversals to
Reversals to Failure,
Failure, 2N
2Nf (cycles)
(cycles)
f
Figure 9.
Figure
Figure 9. Fatigue
9. Fatigue strain-life
Fatigue strain-life curve
strain-life curve obtained
curve obtained by superposition
obtained by
by superposition of
superposition of BM
of BM and
BM and WJ
and WJ specimens.
WJ specimens.
specimens.
3.4. Cyclic Stress

3.4. Cyclic
Cyclic Stress Response
Response
3.4. Stress Response
Figure 10
Figure 10 shows
10 shows the
shows the cyclic
the cyclic stress
cyclic stress response
stress response curves
response curves of
curves of the
of the BM
the BM and WJ
BM and
and WJ specimens
specimens atat different
different strain
strain
Figure WJ specimens at different strain
amplitudes.
amplitudes. Different
Different features
features were
were observed
observed in
in the
the cyclic
cyclic stress
stress response
response curves
curves of
of the
the BM
BM and
and WJ
WJ
amplitudes. Different features were observed in the cyclic stress response curves of the BM and WJ
specimens, although
specimens, although early
although early fatigue
early fatigue failure
fatigue failurecommonly
failure commonly occurred
commonly occurred with the
occurred with increase
with the in the
the increasestrain
increase in amplitude.
in the
the strain
strain
specimens,
amplitude. For example, initial cyclic softening was observed for the B0.5 sample, while
amplitude. For example, initial cyclic softening was observed for the B0.5 sample, while initial cyclic initial cyclic
hardening was observed for B1.0–3.0 (intermediate and large strains), as shown in Figure
hardening was observed for B1.0–3.0 (intermediate and large strains), as shown in Figure 10a,b. Note 10a,b. Note
that for
that for the
the BM
BM specimen,
specimen, initial
initial cyclic
cyclic hardening
hardening followed
followed byby cyclic
cyclic stabilization
stabilization was
was widely
widely
observed, although it became ambiguous at higher strains. In addition, a rapid drop
observed, although it became ambiguous at higher strains. In addition, a rapid drop in cyclic stress in cyclic stress
before fracture was observed, which was primarily due to the formation of macroscopic cracks and
Materials 2019, 12, 4111 10 of 23
For example, initial cyclic softening was observed for the B0.5 sample, while initial cyclic hardening
was observed for B1.0–3.0 (intermediate and large strains), as shown in Figure 10a,b. Note that for the
BM specimen, initial cyclic hardening followed by cyclic stabilization was widely observed, although
it became ambiguous at higher strains. In addition, a rapid drop in cyclic stress before fracture was
observed, which was primarily due to the formation of macroscopic cracks and their unstable extension
to fracture. In contrast, initial cyclic hardening was generally observed for specimens W0.5–2.5,
showing cyclic hardening at high stress amplitudes with increasing strain, while no cyclic behavior
was observed for W3.0 (Figure 10c,d). It should also be noted that, for specimens W0.5–2.5, initial
cyclic hardening followed by cyclic softening readily occurred without cyclic stabilization, while there
was no cyclic
Materials 2019, 12,feature for W3.0.
x FOR PEER REVIEW 10 of 24
a B0.5 B1.0 b B0.5 B1.0

600 B1.5 B2.0 B1.5 B2.0
560
B2.5 B3.0 B2.5 B3.0
500
Stress, σ (MPa)
Stress, σ (MPa)
400
300
490
200
100
420
0
0 1000 2000 3000 4000 0 400 800

Number of Cycles, Nf Number of cycles, Nf
c W0.5 W1.0 d W0.5 W1.0

600 W1.5 W2.0 W1.5 W2.0
W2.5 W3.0 W2.5 W3.0
500 660
Stress, σ (MPa)
Stress, σ (MPa)
400
300
200 630
100
0 1000 2000 3000 4000 0 20

Number of cycles, Nf Number of cycles, Nf
10. Cyclic
Figure 10. Cyclic stress
stressresponse
response(CSR)
(CSR)curves of of
curves BMBMandand
WJWJspecimens; CSRCSR
specimens; curves of (a)ofBM,
curves (a) (b)
BM,high
(b)
magnification of BM,
high magnification of(c)
BM,WJ,(c)and
WJ,(d)
andhigh
(d) magnification of WJof
high magnification specimens.
WJ specimens.
The
The stress–strain
stress–strainhysteresis
hysteresisloops
loopsofofthe theBMBMspecimens
specimensatat different
differentstrains areare
strains shown
shown in Figure
in Figure11.
It is important to note that the shape of the hysteresis loops changed from
11. It is important to note that the shape of the hysteresis loops changed from fusiform to semi- fusiform to semi-rectangular
with increasing
rectangular withstrain; hence,strain;
increasing the fraction
hence,of the
the elastic of
fraction component
the elasticwas higher than
component wasthat of the
higher plastic
than that
component. High strains resulted in large plastic strains that produced semi-rectangular
of the plastic component. High strains resulted in large plastic strains that produced semi-rectangular hysteresis
curves.
hysteresis In addition,
curves. Inataddition,
low cyclicatload
lownumbers,
cyclic load both maximum
numbers, bothtensile and compressive
maximum tensile andstresses at the
compressive
Y axis slightly
stresses at theincreased with increasing
Y axis slightly increased cycle number
with (i.e., ancycle
increasing apparent cyclic(i.e.,
number hardening phenomenon
an apparent cyclic
occurred), while at the half life cycles, maximum stresses decreased
hardening phenomenon occurred), while at the half life cycles, maximum stresses decreased due due to cyclic stabilization after
to
initial cyclic hardening.
cyclic stabilization after In addition,
initial cyclic the stress–strain
hardening. hysteresis
In addition, theloops showedhysteresis
stress–strain obvious serration flows
loops showed
at large strain
obvious amplitudes
serration flows at (B2.0 to B3.0),
large strainwhich was not
amplitudes observed
(B2.0 for the
to B3.0), lowest
which wasstrain amplitudes
not observed for(B0.5
the
to B1.0), although a rather attenuated serration flow was observed for B1.5.
lowest strain amplitudes (B0.5 to B1.0), although a rather attenuated serration flow was observed forPeriodic arrest and release
of dislocations
B1.5. through
Periodic arrest and dislocation
release of pile-ups
dislocationsand dislocation cell wall structures
through dislocation pile-ups and have been suggested
dislocation as
cell wall
the main triggers for serrated plastic flow [57], which will be discussed in
structures have been suggested as the main triggers for serrated plastic flow [57], which will be detail later considering the
TEM images
discussed in (Section 3.9).considering
detail later The stress–strain
the TEM hysteresis
images loops of the
(Section 3.9).WJ specimen
The at different
stress–strain strains
hysteresis are
loops
shown
of the WJ in Figure 12. These
specimen curves showed
at different strains area similar
shown tendency
in Figureto those of thecurves
12. These BM specimens,
showed changing
a similar
from fusiform to semi-rectangular type with increasing strain amplitude.
tendency to those of the BM specimens, changing from fusiform to semi-rectangular type However, the stress levels at
with
increasing strain amplitude. However, the stress levels at all strains in the WJ specimens were higher
than those in the BM specimens due to the WJ specimens being exposed to a higher stress (as
previously illustrated in Figure 8).
Materials 2019, 12, 4111 11 of 23
all strains in the WJ specimens were higher than those in the BM specimens due to the WJ specimens
Materials
being 2019, 12, x FOR
exposed PEER REVIEW 11 of 24
Materials 2019, 12,to a higher
x FOR stress (as previously illustrated in Figure 8).
PEER REVIEW 11 of 24
800
a 800
b 800
c
800
a 800
b 800
c
400 400 400
400 400 400
σ (MPa)
σ (MPa)
σ (MPa)
σ (MPa)
σ (MPa)
σ (MPa)
0 0 0
0 0 0
Stress,
Stress,
Stress,
Stress,
Stress,
Stress,
-400 Cycle1 -400 -400
-400 Cycle1
Cycle2 -400 -400
Cycle2
Cycle5
-800 Cycle5
Half Life Cycle -800 -800
-800 Half Life Cycle -800 -800
-0.006 -0.004 -0.002 0.000 0.002 0.004 0.006 -0.010 -0.005 0.000 0.005 0.010 -0.015 -0.010 -0.005 0.000 0.005 0.010 0.015
-0.006 -0.004 -0.002 Strain,
0.000 ε 0.002 0.004 0.006 -0.010 -0.005 0.000 ε
Strain, 0.005 0.010 -0.015 -0.010 -0.005 Strain,
0.000 ε 0.005 0.010 0.015
Strain, ε Strain, ε Strain, ε
800 d 800 e 800 f

800 d 800 e 800 f
400 400 400
400 400 400
σ (MPa)
σ (MPa)
σ (MPa)
σ (MPa)
σ (MPa)
σ (MPa)
0 0 0
0 0 0
Stress,
Stress,
Stress,
Stress,
Stress,
Stress,
-400 -400 -400
-400 -400 -400
-800 -800 -800

-800 -800 -800
-0.02 -0.01 0.00 0.01 0.02 -0.03 -0.02 -0.01 0.00 0.01 0.02 0.03 -0.03 -0.02 -0.01 0.00 0.01 0.02 0.03
-0.02 -0.01 0.00 ε
Strain, 0.01 0.02 -0.03 -0.02 -0.01 Strain,
0.00 ε 0.01 0.02 0.03 -0.03 -0.02 -0.01 Strain,
0.00 ε 0.01 0.02 0.03
Figure 11. Hysteresis loops

Hysteresis loops of
loopsof the
ofthe BM
theBM specimen
BMspecimen
specimenatatat(a)(a)
0.5%, (b) 1.0%,
0.5%, (c) 1.5%, (d) 2.0%, (e) 2.5%, and
Figure
Figure 11.11. Hysteresis (a) 0.5%, (b)(b) 1.0%,
1.0%, (c) (c) 1.5%,
1.5%, (d) (d) 2.0%,
2.0%, (e) 2.5%,
(e) 2.5%, and
(f)
and 3.0% strain
(f) 3.0% amplitudes;
strain arrows
amplitudes; indicate
arrows the
indicate serrations.
(f) 3.0% strain amplitudes; arrows indicate the the serrations.
serrations.
800
a 800
b 800
c
800
a 800
b 800
c
400 400 400
400 400 400
σ (MPa)
σ (MPa)
σ (MPa)
σ (MPa)
σ (MPa)
σ (MPa)
0 0 0
0 0 0
Stress,
Stress,
Stress,
Stress,
Stress,
Stress,
-400 Cycle1 -400 -400

-400 Cycle1
Cycle2 -400 -400
Cycle2
Cycle5
Cycle5
Half Life Cycle
-800 Half Life Cycle -800 -800
-800 -800 -800
-0.006 -0.004 -0.002 0.000 0.002 0.004 0.006 -0.010 -0.005 0.000 0.005 0.010 -0.015 -0.010 -0.005 0.000 0.005 0.010 0.015
-0.006 -0.004 -0.002 0.000 ε 0.002
Strain, 0.004 0.006 -0.010 -0.005 0.000 ε
Strain, 0.005 0.010 -0.015 -0.010 -0.005 Strain,
0.000 ε0.005 0.010 0.015
800
d 800
e 800
f
800
d 800
e 800
f
400 400 400
400 400 400
σ (MPa)
σ (MPa)
σ (MPa)
σ (MPa)
σ (MPa)
σ (MPa)
0 0 0
0 0 0
Stress,
Stress,
Stress,
Stress,
Stress,
Stress,
-400 -400 -400

-400 -400 -400
-800 -800 -800

-800 -800 -800
-0.02 -0.01 0.00 0.01 0.02 -0.03 -0.02 -0.01 0.00 0.01 0.02 0.03 -0.03 -0.02 -0.01 0.00 0.01 0.02 0.03
-0.02 -0.01 0.00 0.01 0.02 -0.03 -0.02 -0.01 0.00 ε 0.01
Strain, 0.02 0.03 -0.03 -0.02 -0.01 0.00 ε 0.01
Strain, 0.02 0.03
Strain, ε
Figure 12. Hysteresis loops

Hysteresis loops of
loopsof the
ofthe WJ
theWJ specimen
WJspecimen
specimenatatat(a)(a)
0.5%, (b) 1.0%,
0.5%, (c) 1.5%, (d) 2.0%, (e) 2.5%, and
Figure
Figure 12.12. Hysteresis (a) 0.5%, (b)(b) 1.0%,
1.0%, (c) (c) 1.5%,
1.5%, (d) (d) 2.0%,
2.0%, (e) 2.5%,
(e) 2.5%, and
(f) 3.0%
and (f) strain
3.0% amplitudes;
strain arrows
amplitudes; indicate
arrows the the
indicate serrations.
serrations.
(f) 3.0% strain amplitudes; arrows indicate the serrations.
Thestress
The
The stressamplitude
stress amplitudeasas
amplitude asa aafunction
functionofof
function offraction
fraction
fraction of
ofof cycles
cycles
cycles normalized
normalized
normalized to
to to the
thethe fatigue
fatigue
fatigue life
lifelife is depicted
depicted
is depicted
is in
in Figure
Figure
in Figure 13.
13. 13.
TheThe
The initial
initial cyclic
cyclic
initial cyclic hardening
hardening
hardening or
oror softening
softening
softening stage
stage
stage was
was relatively
wasrelatively short,
relativelyshort, generally<5%
generally
short,generally <5%
<5%of of the
ofthe
the
total
total fatigue
fatigue life.
life. A
A rapid
rapid drop
drop in
in the
the cyclic
cyclic stress
stress amplitude
amplitude was
was observed
observed
total fatigue life. A rapid drop in the cyclic stress amplitude was observed for all curves just before for
for all
all curves
curves just
just before
before
fractureatat
fracture
fracture atthe
thefinal
the final
final cyclic
cyclic stage,
stage,
cyclic stage, mostly
mostly
mostly within
within
within 10%
10%10% of the
of the
of the fatigue
fatigue
fatigue life.
life. life. It is
is possible
possible
It is possible
It thatfatigue
that the
that the fatigue
the fatigue
lives
lives predominantly
were
lives were predominantly
were predominantly determined
determined
determinedby thebyby the constant
constant
the constant stress amplitudes
stress amplitudes
stress amplitudes in the
the intermediate
in the intermediate
in intermediate stage,
stage, rather
stage,
rather
than at than
the at the
stages stages
of of
initial initial
cyclic cyclic
hardeninghardening (softening)
(softening) or the or
final the
rather than at the stages of initial cyclic hardening (softening) or the final rapid drop. final
rapid rapid
drop. drop.
Materials 2019, 12, x4111
FOR PEER REVIEW 12
12of
of 24
23
a b
600 a 600 b
600 600
σ (MPa)
σ (MPa)
400 400
σ (MPa)
σ (MPa)
400 400
Stress,
Stress,
B0.5 W0.5
Stress,
Stress,
200 B1.0
B0.5 200 W1.0
W0.5
200 B1.5
B1.0 200 W1.5
W1.0
B2.0
B1.5 W2.0
W1.5
B2.5
B2.0 W2.5
W2.0
0 B3.0
B2.5 0 W3.0
W2.5
0 B3.0 0 W3.0
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
0.0Fraction
0.2of Cycles
0.4 relative
0.6to Fatigue
0.8 Life 1.0 0.0Fraction
0.2of Cycles
0.4 relative
0.6to Fatigue
0.8 Life 1.0
Fraction of Cycles relative to Fatigue Life Fraction of Cycles relative to Fatigue Life
Figure 13. Stress amplitude as a function of fraction of cyclic number to fatigue life in (a) BM and (b)
Figure
Figure
WJ 13. Stress
13. Stress amplitude
specimens. amplitudeas
asaafunction
functionofoffraction
fractionofof
cyclic number
cyclic to to
number fatigue lifelife
fatigue in (a) BMBM
in (a) andand
(b)
WJ specimens.
(b) WJ specimens.
3.5. Cyclic Hardening/Softening Behavior
3.5. Cyclic Hardening/Softening
3.5. Cyclic Hardening/Softening Behavior
Behavior
The ratios of cyclic hardening (RH) and cyclic softening (RS) can be expressed as follows [58]:
The
The ratios
ratios of
of cyclic
cyclic hardening
hardening (RH)
(RH) and
and cyclic
cyclic softening
softening (RS)
(RS) cancan be
be expressed
expressedas
asfollows
follows[58]:
[58]:
𝑅𝐻 = (𝛥𝜎 − 𝛥𝜎 )/𝛥𝜎 (10)
RH𝑅𝐻==(∆σ
(𝛥𝜎 − 𝛥𝜎 )/𝛥𝜎
max − ∆σmax ) /∆σ f irst (10)
(10)
𝑅𝑆 =(𝛥𝜎 − 𝛥𝜎 )/𝛥𝜎 (11)
RS𝑅𝑆== ∆σ (𝛥𝜎 −− ∆σ𝛥𝜎 /∆σ )/𝛥𝜎 (11)
(11)
max hal f max
where, Δσmax, Δσfirst, and Δσhalf are the stress amplitudes at the maximum, first cycle, and half cycle,
where, ∆σΔσmax,as Δσ first, and Δσhalf are the stress amplitudes at the maximum, first cycle, and half cycle,
where, max , ∆σ
respectively, first , and
shown ∆σhalf are
in Figure 14.the
Thestress amplitudes
BM specimens at the maximum,
exhibited a high RS first
at lowcycle, and
strain half cycle,
amplitudes
respectively,
respectively,
(B0.5), while as as shown
RSshown in Figure
in Figure
dramatically 14. The
14. Theand
reduced, BM
BMRH specimens
specimens
increased exhibited
exhibited a high RS at
a high RSstrain
with increasing low strain
at lowamplitudes,amplitudes
strain amplitudes
where
(B0.5),
(B0.5),
the while
while
ratios RS dramatically
RS dramatically
intersected reduced,
at ±1.1% reduced,
strain. The and
andWJRH RH increased
increased
specimens with increasing
with
showed increasing strain
a high RSstrain amplitudes,
at lowamplitudes, where
where
strain amplitudes
the ratios
the ratios
(W0.5), intersected
intersected
while RS decreasedat ±1.1% strain.
at ±1.1%tostrain. The WJ
The WJ
a similar levelspecimens
specimens
as RH atshowedshowed a
intermediate high RS at low strain
RS at low strain
a high amplitudes amplitudes
amplitudes
(W1.0–2.5) and
(W0.5), while
(W0.5), while
reached zero for RS
RS W3.0.decreased
decreased RH and to
to aRS a similar
similar level
are two level as RH at
as RH at parameters
important intermediate
intermediateinamplitudes amplitudes
(W1.0–2.5)
characterizing (W1.0–2.5) and
and reached
cyclic deformation
reached
behaviors zero
zero for W3.0. for W3.0.
RH and RS
of materials. RH
That and
areis,
two RS are two
important
with high RH important
parameters parameters in
in characterizing
or RS, fatigue characterizing
cyclic by
life can be longer cyclic
deformation deformation
behaviors
cyclic hardening or
behaviors
of materials. of materials.
That is, withThat is,
high with
RH orhigh
RS, RH or
fatigue RS,
lifefatigue
can be life can
longer
softening behaviors depending on materials or strain amplitudes. The difference between RH and RS be
by longer
cyclic by cyclic
hardening hardening
or softeningor
softening
behaviors
and behaviors
depending
the relative depending
on materials
fluctuation on materials
with or or strain amplitudes.
strainamplitude
strain amplitudes. forThe WJThe
thedifference difference
between
specimen between
were RHand
RHrelatively
and RS and RS
the
small
and the
compared relative
to those of
relative fluctuation fluctuation
withthestrain with
BM specimen;
amplitudestrain
hence,amplitude
for theweWJ propose for
specimen the
thatwereWJ
the WJ specimen
specimen
relatively were relatively
hadcompared
small small
lower hardening
to those
compared
(softening) to those of the BM specimen; hence, we propose that the WJ specimen
ability. hence, we propose that the WJ specimen had lower hardening (softening) ability.
of the BM specimen; had lower hardening
(softening) ability.
a b
a 0.20 RH b0.20 RH
0.20 RRSH 0.20 RRSH
ratio
ratio
0.16 RS 0.16 RS
ratio
ratio
0.16 0.16
Softening
Softening
0.12 0.12
Softening
Softening
0.12 0.12
0.08 0.08
Hardening/
Hardening/
0.08 0.08
Hardening/
Hardening/
0.04 0.04
0.04 0.04
0.00 0.00
0.00 0.00
0.5 1.0 1.5 2.0 2.5 3.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5 1.0Strain 1.5 2.0Δε (%) 2.5
amplitude, 3.0 0.5 1.0Strain 1.5 2.0Δε (%) 2.5
amplitude, 3.0
Strain amplitude, Δε (%) Strain amplitude, Δε (%)
Figure
Figure 14. Hardening
Hardening (RH)
(RH) and
and softening
softening (RS)
(RS) ratio
ratio in
in (a) BM and (b) WJ specimens.
Figure 14. Hardening (RH) and softening (RS) ratio in (a) BM and (b) WJ specimens.
Analyzing the
Analyzing the differences
differencesbetween
betweenthe thecyclic
cyclicstrain
strainand
andmonotonic
monotonic strain
strainresponse
response is important
is important as
it can Analyzing
give the
informationdifferences
about thebetween
fatigue the cyclic
behavior strain
of the and monotonic
material. Figure strain
15 response
compares
as it can give information about the fatigue behavior of the material. Figure 15 compares the stress– the is important
stress–strain
as it can
curves
strain ofgive
curves information
the of
cyclic cyclic about
the stress and the
andfatigue
stressmonotonic behavior
tension.
monotonic of the
For
tension. the
For material.
BM BM Figure
thespecimens, 15 cyclic
the
specimens, comparesstress
the cyclic theatstress–
small
stress at
strain
strains curves
(B0.5) of
was the cyclic
slightly stress
lower and
thanmonotonic
the tension.
monotonic For
stresses the
at BM
the specimens,
same strain,
small strains (B0.5) was slightly lower than the monotonic stresses at the same strain, while the values the cyclic
while the stress at
values
small strains
increased with(B0.5) was
increasing slightly
strain lower than
(B1.0–3.0). the
We monotonic
propose stresses
that the at
CSR the same
below the
increased with increasing strain (B1.0–3.0). We propose that the CSR below the elastic strain showedstrain, while
elastic strain the values
showed
increased
cyclic
cyclic with increasing
stabilization
stabilization strainto
subsequent
subsequent to(B1.0–3.0).
initial We
an initial
an propose
cyclic
cyclic thatas
softening
softening theit
as itCSR
wasbelow
was belowthe
below theelastic
the stress strain
stress showed
of monotonic
of monotonic
cyclic stabilization
tension. On the other subsequent
hand, thetoCSR an initial
undercyclic softening
plastic as it was
strain showed belowcyclic
shorter the stress of monotonic
stabilization after
tension. On the other hand, the CSR under plastic strain showed shorter cyclic stabilization after
Materials 2019, 12, 4111 13 of 23

tension. On the other hand, the CSR under plastic strain showed shorter cyclic stabilization after initial
initialhardening
cyclic cyclic hardening with increasing
with increasing strain as strain
it was as
in it
thewas in the
plastic plastic deformation
deformation range. for
range. However, However,
the WJ
for the WJcyclic
specimen, specimen,
stressescyclic stresses higher
were slightly were slightly higher than
than monotonic monotonic
stresses stresses
at all strains, at allinstrains,
resulting initial
resulting
strain in initial strain hardening.
hardening.
a 800 b
600 600
Stress, σ (MPa)
Stress, σ (MPa)
400 400
200
BM,Tensile 200
BM,Tensile
WJ,Tensile WJ,Tensile
0 BM, Fatigue BM, Fatigue
WJ, Fatigue WJ, Fatigue
0
-5 0 5 10 15 20 25 30 35 0 1 2 3
Strain, ε (%) Strain, ε (%)
Figure 15. (a)

Figure 15. (a) Comparison
Comparison of
of the
the cyclic
cyclic stress–strain
stress–strain curve
curve and
and the
the monotonic
monotonic tensile
tensile stress–strain
stress–strain
curves
curves and
and(b)(b)high
highmagnification
magnificationofof00to
to3.5%
3.5%strain
strainrange.
range.
3.6.
3.6. Classification
Classification of
of Fatigue
Fatigue Fracture
Fracture Mode
Mode
Classification of the fatigue fracture
Classification of the fatigue fracture mode was based
mode on theon
was based fatigue life and life
the fatigue stress levels,
and modified
stress levels,
from Figurefrom
modified 1. InFigure
this study,
1. Infour
thismodes
study, were
fourderived
modesdepending
were derivedon the materials and
depending strain
on the amplitudes,
materials and
i.e., HCF
strain for B0.5, LCF
amplitudes, i.e.,for
HCF B1.0–3.0 and
for B0.5, W0.5,
LCF for ELCF forand
B1.0–3.0 W1.0–2.5, and MF
W0.5, ELCF forfor W3.0. An
W1.0–2.5, andoverall
MF for scheme
W3.0.
for
Andetermining
overall scheme the corresponding
for determining modes can be estimated
the corresponding from can
modes Figures 10 and 15.from
be estimated The B0.5 specimen
Figure 10 and
was
Figure 15. The B0.5 specimen was under the elastic limit of the BM specimen (Figure 15b), whichHCF
under the elastic limit of the BM specimen (Figure 15b), which was in the stress-controlled was
mode which showed initial cyclic softening followed by cyclic stabilization
in the stress-controlled HCF mode which showed initial cyclic softening followed by cyclic (Figure 10b). In addition,
the B1.0–B3.0 (Figure
stabilization and W0.5 specimens
10b). were the
In addition, overB1.0–B3.0
the elasticand
limit of the
W0.5 BM and WJ
specimens werespecimens,
over therespectively,
elastic limit
which were classified as strain-controlled LCF (cyclic hardening
of the BM and WJ specimens, respectively, which were classified as strain-controlledfollowed by cyclic stabilization or
LCF (cyclic
quasi-stabilization). However, the W0.5–2.5 specimens were over the elastic limit
hardening followed by cyclic stabilization or quasi-stabilization). However, the W0.5–2.5 specimens of the WJ specimens
(Figure
were over15b)theand werelimit
elastic macroscopically in the stress
of the WJ specimens concentration
(Figure 15b) and werestatemacroscopically
represented by initial
in the cyclic
stress
hardening, followed by cyclic softening and no stabilization (ELCF),
concentration state represented by initial cyclic hardening, followed by cyclic softening which is evidence of ductility
and no
exhaustion.
stabilizationSample
(ELCF), W3.0
whichshowed no cyclicofdamage
is evidence behavior,
ductility but instead
exhaustion. Sample showed
W3.0 features
showed similar to
no cyclic
monotonic tensile conditions. This classification becomes more apparent
damage behavior, but instead showed features similar to monotonic tensile conditions. This when comparing the fracture
features discussed
classification in themore
becomes next apparent
section. when comparing the fracture features discussed in the next
section.
3.7. Fracture Features of Base Material after Fatigue Tests
Fatigue Features
3.7. Fracture fracture of
features from fractography
Base Material After Fatigue images
Tests are good indicators for identifying fracture
modes. In the HCF mode, incipient cracks determine the fatigue life [19,20] and defining the transition
Fatigue fracture features from fractography images are good indicators for identifying fracture
between crack initiation and propagation is still controversial [59]. Meanwhile, the microscopic initial
modes. In the HCF mode, incipient cracks determine the fatigue life [19,20] and defining the transition
crack length and stress intensity at the crack tip are crucial for HCF resistance [21,60]. In addition,
between crack initiation and propagation is still controversial [59]. Meanwhile, the microscopic initial
HCF cracks originate from a sub-surface inclusion with a diameter of several tens of micrometers [61],
crack length and stress intensity at the crack tip are crucial for HCF resistance [21,60]. In addition,
and a typical “fish-eye” fracture morphology is often observed. The strong residual compressive
HCF cracks originate from a sub-surface inclusion with a diameter of several tens of micrometers
stress field has significant improvement in fatigue strength due to a lattice distortion, which causes the
[61], and a typical “fish-eye” fracture morphology is often observed. The strong residual compressive
primary crack nucleation site (usually located at some defects such as a non-metallic inclusion) in HCF
stress field has significant improvement in fatigue strength due to a lattice distortion, which causes
regions. The crack initiation begins by decohesion of the inclusion-matrix interface, and then the crack
the primary crack nucleation site (usually located at some defects such as a non-metallic inclusion) in
propagates forming a distinct fracture pattern known as the “fisheye”. A distinct crack propagation
HCF regions. The crack initiation begins by decohesion of the inclusion-matrix interface, and then
(fisheye) can be observed on the fracture surface. Approximately, the inclusion as crack nucleus in the
the crack propagates forming a distinct fracture pattern known as the “fisheye”. A distinct crack
HCF regime is located within fisheye. The existence of inclusion and fisheye is an essential feature of
propagation (fisheye) can be observed on the fracture surface. Approximately, the inclusion as crack
the interior failure in the HCF regime [62–64].
nucleus in the HCF regime is located within fisheye. The existence of inclusion and fisheye is an
essential feature of the interior failure in the HCF regime [62–64].
Figure 16 shows representative images of fractured surfaces for each fracture mode observed in
our experiments. In the HCF mode, a “fisheye” fracture morphology due to fatigue crack initiation
content of CaO accelerates the dephosphorization process in presence of FeO by dissolving the
phosphates.
In the LCF mode, Forsyth et al. [67] differentiated two types of striations, Type A (ductile
striations), which consist of light and dark bands, and Type B (brittle striations), which consist of
river-like patterns
Materials 2019, 12, 4111 with limited ductility. They also showed optical fractography images of 14 ofthe
23
transition from ductile to brittle striations on a grain boundary facet of Al-5%Mg-4%Zn alloy [68]. In
the present study, in the LCF mode of the BM specimen, striations with narrow spacings were
Figure 16
conspicuous shows16d),
(Figure representative
which mayimages be theof fractured
typical Typesurfaces for each fracture mode observed in
A striations.
our experiments. In the HCF mode, a “fisheye” fracture
To elucidate the fatigue damage features for each fracture mode, morphology due tothe
fatigue crack
surface initiation
damage of at
thea
sub-surfacewas
specimens CaOobserved
compositeafterwithcyclic
a diameter of 10
loading, asµm in theinB0.5
shown specimen
Figure 17. In was
the observed
HCF mode (Figure
after 16a,b).
cyclic
Distinct fracture striations were observed at the fisheye fracture (Figure 16c).
loading tests, plastic flows induced by plastic deformations along the shear stress direction In addition, such inclusions
were
originate from
developed dephosphorization
(Figure 17a). In the LCFduring
mode of steel
themaking. In conventional
BM specimen, steel making process,
surface micro-fissures and cracks Iron ore
were
raw materials
observed, which contain
implied impurities
that mostsuch as phosphorus
surface grains of theand BMsulphur,
specimenwhich have numerous
contained been known to have
intrusions
detrimental impacts on the final steel properties. Accordingly, dephosphorization
and extrusions (Figure 17b); hence, a homogeneous deformation morphology was observed due to process is conducted
by adding
the uniform CaO/SiO 2 , which
shear stress increasesduring
distribution the slag basicity
the fatigue[65,66]. Thus,indicated
test, which higher content of CaO
that Stage accelerates
I crack growth
the dephosphorization process in presence
was the predominant fatigue crack propagation mechanism. of FeO by dissolving the phosphates.
Figure 16. Fracture surfaces according to fracture modes; fracture surfaces in (a) low-magnification
image of B0.5, the corresponding high-magnification image in area indicated by the (b) white box and
whitecircle
(c) white circlein in
(a),(a), (d) low-cycle
(d) low-cycle fatigue
fatigue (LCF) (LCF)
of B2.0of B2.0 specimens
specimens *The direction
* The loading loading isdirection is
identically
identically normal
normal to the images. to the images.
In the LCF mode, Forsyth et al. [67] differentiated two types of striations, Type A (ductile striations),
which consist of light and dark bands, and Type B (brittle striations), which consist of river-like patterns
with limited ductility. They also showed optical fractography images of the transition from ductile to
brittle striations on a grain boundary facet of Al-5%Mg-4%Zn alloy [68]. In the present study, in the
LCF mode of the BM specimen, striations with narrow spacings were conspicuous (Figure 16d), which
may be the typical Type A striations.
To elucidate the fatigue damage features for each fracture mode, the surface damage of the
specimens was observed after cyclic loading, as shown in Figure 17. In the HCF mode after cyclic
loading tests, plastic flows induced by plastic deformations along the shear stress direction were
developed (Figure 17a). In the LCF mode of the BM specimen, surface micro-fissures and cracks were
observed, which implied that most surface grains of the BM specimen contained numerous intrusions
and extrusions (Figure 17b); hence, a homogeneous deformation morphology was observed due to the
Materials 2019, 12, 4111 15 of 23
uniform shear stress distribution during the fatigue test, which indicated that Stage I crack growth was
the predominant fatigue crack propagation mechanism.
Figure 17.
Figure 17.Surface
Surface damage
damage morphology
morphology according
according to themodes;
to the fracture fracture modes;
surface surface
damage damage
morphologies
morphologies in (a) high-cycle fatigue (HCF) of B0.5 and (b) low-cycle fatigue (LCF) of B2.0
in (a) high-cycle fatigue (HCF) of B0.5 and (b) low-cycle fatigue (LCF) of B2.0 specimens. * The loading
specimens.*The
direction loading
is identically direction
applied is identically
to the images. applied to the images.
3.8. Fracture Features

3.8. Fracture Features of
of Welded
Welded Joint
Joint after
AfterFatigue
FatigueTests
Tests
In
In the LCF regime with increasing increasing strain
strain or in materials with low fatigue resistance, the striations
became blurred and the spacings between them widened as a transition to the quasi-cleavage (QC)
fracture mode occurred.
occurred. On On thethe other
other hand,
hand, at at very large strains, or in materials with low LCF LCF
resistance
resistance [22,69], QC or monotonic fracture surfaces are
[22,69], QC or monotonic fracture surfaces areoften
oftenobserved.
observed.
In the LCF mode of the WJ specimen (W0.5), striations were blurred with primary transgranular
cracks
cracks (Figure
(Figure18a),
18a), which
whichmay be the
may be Type B striations.
the Type After initial
B striations. After cyclic
initialhardening in the WJin
cyclic hardening specimen
the WJ
(Figure
specimen
fatal crack10d), cyclica10d),
(Figure
(above stabilization
cyclic
critical length)at constant
occurs, atstrain
stabilization at amplitudes
constant
which point,strain was
all crack observed
amplitudes (a balance
was
energies converge tobetween
observed form cyclic
(a the
balance
fatal
softening
between and
cyclic dynamic
softening recovery
and processes
dynamic [70,71]),
recovery where
processes cyclic strain
[70,71]),
crack [77]. Generally, the strain amplitude and microstructural features are considered the key localization
where cyclic occurred,
strain resulting
localization
in frequentresulting
occurred,
parameters transgranular
determining cracking
in frequent
the LCF [72,73].
fractureInmode.
transgranular the ELCF
Themode
cracking ratio of
[72,73]. the
ofIn WJELCF
the
Stage Ispecimen
andmode (W2.0),
Stage ofIIthe aWJ
crackQC fracture
specimen
growth is
with
(W2.0),a few
a QCmicro-dimples
fracture with was
a fewobserved (Figure
micro-dimples 18b),
was which
observed differed
(Figure from
18b),
determined by the applied strain amplitude [78]. Stage I crack growth is dominant for smaller strains, the striations
which differed during
from the
the
LCF
which testresult
striations orduring
MFinduring
HCF tensile
the behavior
LCF testtests.
or MF Cleavage
[79,80]. On thefracture
during tensile with
tests.
contrary, micro-dimples
forCleavage has with
fracture
larger strains, the been micro-dimples
attributed
fraction of StagetoIIELCF
has
crack
at large
been
growth cyclic
attributed strains,
increases,to ELCF which
and the usually
at fracture
large cyclic leads to
strains,
behavior a short fatigue
which usually
transitions life of
to theleads
ELCF <100
tomode.cycles [15,74].
a shortInfatigue At large
life of
addition, the<100 strains,
MF cycles
mode
the WJ
[15,74]. specimen
At large (W3.0)
strains, showed
the WJ ductile
specimen fracture,
(W3.0) characterized
showed by
ductile micro-dimples
fracture,
was prevalent at sufficiently large strains, given that micro-dimples were solely observed after tensile (Figure
characterized 18c),
by which
micro-
was
tests.identical
dimples (Figureto the
18c),fracture
which surface observed
was identical after
to the tensilesurface
fracture testingobserved
(Figure 18d).
after tensile testing (Figure
18d).
a Tranof
The surface damage theu lar
sg ran WJ specimens was also b M observed
icro - after cyclic loading and tensile tests,
Tran sg ran u lar
cracki n g d im p les
as shown in Figure 19. In the LCF mode of the WJ specimen, various cracki n g well-developed cracks were
clearly observed, where most cracks were connected at a higher angle of 38–69° (Figure 19a) than
Q u asi-
those in the LCF mode of the BM
Striatio n sspecimen (17–52°). Meanwhile, cleavag ine the ELCF and MF modes of the
WJ specimen, plastic flows induced by plastic deformations along the shear stress direction were
developed (Figure 19b,c), which were P ro gidentical
ress to the surface morphology of the MF in the tensile test
(Figure 19d); these were analogous to those in the HCF mode, but plastic flows at higher angles of
x1000 10㎛ x1000 10㎛
30–72° in the ELCF and MF modes are more likely to deform than those at lower angles of 11–30° in
the HCF mode due cto large applied cyclic tensile and d compressive loads.
Based on these observations, we concluded that the fatigue fracture and surface damage
M icro -
changed depending on the materials M icro - d im p les
d im p lesand strain amplitudes.
The ratio of crack initiation and crack propagation over the fatigue life depends on the failure
mode. In the HCF mode, an incipient crack determined that the fatigue life and portions of crack
initiation take up to 90% of fatigue life depending on the initiated crack size from an inclusion,
material strength, and stress level [23,75].
x1000 10㎛ x1000 10㎛
In the case of the LCF mode, crack growth determines the fatigue life as the incipient cracks can
be readily
Figure formed
Figure 18. at large
18. Fracture strain
surfaces amplitudes
according [24]. modes;
fracture
to fracture The LCF
modes; phenomenon
fracture surfaces in
surfaces incan be divided
(a) low-cycle
(a) low-cycle into four
fatigue
fatigue
steps(LCF)
and of
(LCF) two crack-growth
ofW0.5,
W0.5, (b)extremely
(b) extremely stages in microstructural
low-cycle
low-cycle fatigue
fatigue (ELCF)
(ELCF) of evolution
of W2.0,
W2.0, and and [76].
(c) (c)Fatigue
monotonic
monotonic damage
fracture begins
fracture
(MF) (MF)
of in
W3.0of Step
1, and symptoms
W3.0
specimen, and of
specimen, (d) latent
andMF(d) cracks
in MF appear,
in tensile
tensile test. test.showing
*The unstable
*The loading
loading is cyclic
direction
direction hardening
is identically
identically normal orthe
normal
to softening.
to In Step 2,
the images.
images.
incipient cracks are formed on the planes of high shear stresses and gradually propagate inward; this
is called “Stage I crack-growth.” In Step 3, cracks on the planes of high tensile stresses propagate
dramatically perpendicular to the applied load; this is called “Stage II crack-growth.” In the final step,
fracture occurs in the remaining reduction area after full crack growth. Micro-cracks nucleate and
grow rapidly until they are arrested by grain boundaries; multiple cracks grow and interact until a
Q u asi-
Striatio n s cleavag e
P ro g ress
x1000 10㎛ x1000 10㎛

Materials 2019, 12, 4111 16 of 23
c d
The surface damage of the WJ specimens

M icro - was also observeddMimafter
icro -
p les cyclic loading and tensile tests,
d im p les
as shown in Figure 19. In the LCF mode of the WJ specimen, various well-developed cracks were
clearly observed, where most cracks were connected at a higher angle of 38–69◦ (Figure 19a) than
those in the LCF mode of the BM specimen (17–52◦ ). Meanwhile, in the ELCF and MF modes of
the WJ specimen, plastic flows induced by plastic deformations along the shear stress direction were
developed (Figure 19b,c), which were x1000 10㎛
identical to the surface morphology x1000 10㎛of the MF in the tensile test
(Figure 19d); these were analogous to those in the HCF mode, but plastic flows at higher angles of
Figure 18. Fracture surfaces according to fracture modes; fracture surfaces in (a) low-cycle fatigue
◦
30–72(LCF)
in the ELCF(b)
and MF modes are more likely to deform than ◦
of W0.5, extremely low-cycle fatigue (ELCF) of W2.0, andthose at lower fracture
(c) monotonic angles of 11–30
(MF) of in
the HCF
W3.0mode due and
specimen, to large applied
(d) MF cyclic
in tensile test.tensile and compressive
*The loading loads.
direction is identically normal to the images.
Figure 19.19.Surface
Surface damage
damage morphology
morphology according
according to themodes;
to the fracture fracture modes;
surface damage surface damage
morphologies
morphologies
in (a) low-cyclein (a) low-cycle
fatigue (LCF) of fatigue (LCF)
W0.5, (b) of W0.5,
extremely (b) extremely
low-cycle fatigue low-cycle fatigue
(ELCF) of W2.0, and(ELCF) of W2.0,
(c) monotonic
and (c) monotonic
fracture (MF) of W3.0fracture (MF) and
specimens, of W3.0 specimens,
(d) MF in tensileand
test.(d) MFloading
*The in tensile test. *The
direction loading direction
is identically applied
is identically
to the images.applied to the images.
Based on these Evolution

3.9. Microstructural observations,
Afterwe concluded
Fatigue Tests that the fatigue fracture and surface damage changed
depending on the materials and strain amplitudes.
Persistent slip bands (PSBs) are observed within the HCF and LCF modes, which evolve via
The ratio of crack initiation and crack propagation over the fatigue life depends on the failure
irreversible slips constituting intrusion and extrusion during cyclic tensile and compressive loads.
mode. In the HCF mode, an incipient crack determined that the fatigue life and portions of crack
initiation take up to 90% of fatigue life depending on the initiated crack size from an inclusion, material
strength, and stress level [23,75].
In the case of the LCF mode, crack growth determines the fatigue life as the incipient cracks can
be readily formed at large strain amplitudes [24]. The LCF phenomenon can be divided into four
steps and two crack-growth stages in microstructural evolution [76]. Fatigue damage begins in Step
1, and symptoms of latent cracks appear, showing unstable cyclic hardening or softening. In Step
2, incipient cracks are formed on the planes of high shear stresses and gradually propagate inward;
this is called “Stage I crack-growth.” In Step 3, cracks on the planes of high tensile stresses propagate
dramatically perpendicular to the applied load; this is called “Stage II crack-growth.” In the final step,
fracture occurs in the remaining reduction area after full crack growth. Micro-cracks nucleate and
grow rapidly until they are arrested by grain boundaries; multiple cracks grow and interact until a
fatal crack (above a critical length) occurs, at which point, all crack energies converge to form the
fatal crack [77]. Generally, the strain amplitude and microstructural features are considered the key
parameters determining the LCF fracture mode. The ratio of Stage I and Stage II crack growth is
determined by the applied strain amplitude [78]. Stage I crack growth is dominant for smaller strains,
which result in HCF behavior [79,80]. On the contrary, for larger strains, the fraction of Stage II crack
Materials 2019, 12, 4111 17 of 23
growth increases, and the fracture behavior transitions to the ELCF mode. In addition, the MF mode was
prevalent at sufficiently large strains, given that micro-dimples were solely observed after tensile tests.
3.9. Microstructural Evolution after Fatigue Tests

Persistent slip bands (PSBs) are observed within the HCF and LCF modes, which evolve via
irreversible slips constituting intrusion and extrusion during cyclic tensile and compressive loads.
PSBs are readily observed in face-centered cubic (FCC) materials as 12 close-packed slip systems
running in {111} octahedral planes and the <110> direction. Various studies of PSBs have been
conducted [81–83] owing to the ease of obtaining fractographic images and analyzing the dislocation
behavior, especially in austenitic twinning-induced-plasticity and transformation-induced-plasticity
steels [84,85]. However, body-centered cubic (BCC) materials are not close packed and the slip moves
in the most atom-dense {110} planes, which have a lower atomic packing factor (APFα = 0.68) than
that of FCC materials (APFγ = 0.74). Hence, a higher critical shear stress is required to initiate cracks
in BCC structures than in FCC ones because in the initial stage, fatigue cracks initiate and propagate at
the PSB–matrix interfaces under the shear stress in fracture Mode II, which acts parallel to the plane of
the crack and perpendicular to the crack front [86,87]. Most studies concerning BCC materials were
conducted using pure metals, as impurities introduced by alloying elements increase the complexity
of the dislocation behavior. However, the findings of these studies are relevant as the presence of
the small impurity concentrations does not completely alter the fundamental slip planes, but rather
influences the level of cross-slips. Plastic deformation in BCC materials is controlled by the motion of
screw dislocations in a periodic potential originating from their non-planar core structure. Studies of
the slip systems of pure iron [88,89] reported the movement of screw dislocations on the {110} planes,
inferring that edge dislocations can glide on {110} planes. Dislocation sources always move on {110}
planes and require the motion of screw dislocations. In Fe–3%Si [90], slip probably occurred on the
{110} plane, while other studies proposed {112} slip via edge dislocations [91], wavy {112} slip [92],
or both {110} and {112} slip [93].
In the present study, in the base material, elongated subgrains and walls prevailed and some
nano-voids at the dislocation lines or walls could be seen in the HCF mode (B0.5) (Figure 20a),
and dislocation pile-ups at the dislocation lines with micro-voids were present on the {110} planes
(Figure 20d). Owing to the largely elastic fatigue loadings in the HCF mode, it becomes easy to observe
the energy dissipated by crack growth and microstructural change at the nanoscale. However, less
dislocation tangles and veins were present in the LCF mode (B2.0), while incipient dislocation lines
were observed (Figure 20b), and the diffraction patterns of the tangles in the LCF mode (Figure 18e).
In the welded joint, heavy dislocation tangles and veins were observed along the strain-localized
zone in the ELCF mode (W2.0) (Figure 20c), and they lie on the {110} planes (Figure 20f), where several
dislocations and cross-slips can be easily activated during the deformation of the BCC material. High
dislocation accumulation at the local zones such as dislocation tangles occurs in the elastic and plastic
fatigue loadings.
3.10. Fatigue Features According to Fatigue Fractography and Surface Damage

The fatigue properties according to the classified fatigue fracture modes are illustrated in Figure 21.
In the present study, the fatigue-fracture features such as fatigue fractography and surface damage
depending on the fatigue fracture modes were observed on the BM and WJ specimens and classified
into HCF for B0.5, LCF for B1.0–3.0 and W0.5, ELCF for W1.0–2.5, and MF for W3.0. An overall scheme
for determining the corresponding modes can be estimated from the cyclic stress–strain curves and
cyclic stress response (Figure 15). The B0.5 specimen was under the elastic limit of the BM specimen
in the HCF mode. The B1.0–B3.0 and W0.5 specimens were over the elastic limit of the BM and WJ
specimens, respectively, in the LCF mode. However, the W0.5–2.5 specimens were over the elastic
limit of the WJ specimens in the ELCF mode. W3.0 showed no cyclic damage behavior, but instead
showed features similar to monotonic tensile conditions.
dislocation tangles and veins were present in the LCF mode (B2.0), while incipient dislocation lines
were observed (Figure 20b), and the diffraction patterns of the tangles in the LCF mode (Figure 18e).
In the welded joint, heavy dislocation tangles and veins were observed along the strain-localized
zone in the ELCF mode (W2.0) (Figure 20c), and they lie on the {110} planes (Figure 20f), where several
dislocations and cross-slips can be easily activated during the deformation of the BCC material. High
dislocation
Materials accumulation at the local zones such as dislocation tangles occurs in the elastic and plastic
2019, 12, 4111 18 of 23
fatigue loadings.
Figure 20. STEM

Figure 20. STEM images of of
images (a)(a)
HCF,
HCF,(b)
(b)LCF,
LCF,and
and (c)
(c) ELCF (magnification
ELCF (magnification of of ×40,000),
×40,000), andand
(d–f)(d–f)
TEMTEM
imagesimages2019,(magnification
(magnification
Materials 12, x FOR PEER of ×13,500).
of ×13,500).
REVIEW 19 of 24
log σ
Fatigue limit
Fracture MF ELCF LCF HCF

Mode
log 𝑁
Np , Ductility Mixed (Intermediate) Ni , YS
Corresponding B1.0-B3.0 B0.5

Domain W3.0 W1.0-W2.5 W0.5
*Monotonic
uniaxial
Determinant tensile
strain
control * Cyclic strain-based control * Cyclic stress-based control
*Strength * Mainly plastic

Components and and ductility *Plastic and elastic
ductility components components *Elastic components
*Quasi-cleavage *(Sub) Surface crack
Primary Fracture *Micro- with few micro- *Striations with spacings initiation induced by
Surface dimples dimples non-metallic inclusion
*High
Surface Damage angle
plastic *High angle *Micro-fissure and cracks
flow plastic flow *Multiple cracks *Low angle plastic flow
Microstructural *Dislocation tangles *Dislocation lines *Elongated subgrains and

Evolution and veins and veins walls with nano-voids
Figure
Figure 21. 21. Fatigue
Fatigue properties according
properties according totoclassified fatigue
classified modes.
fatigue modes.
4. Conclusions
Various strain amplitudes (Δε/2 = ± 0.5–3%) were imposed on BM and WJ specimens of a low-
carbon steel with F+P microstructures to observe the transition of the fatigue fracture mode. The
conclusions drawn from this study are as follows:
(i) In the base material, the HCF and LCF fracture modes were observed through cyclic stress
Materials 2019, 12, 4111 19 of 23
Fractured surfaces for each fracture mode were observed in Figure 16. In the HCF mode, a CaO
composite with a diameter of 10 µm was observed within a fisheye feature on the sub-surface of the
B0.5 specimen (Figure 16a–c). In the LCF mode, striations with spacings were commonly observed
(Figure 16d, Figure 18a). In the ELCF mode, a QC fracture with a few micro-dimples was observed
(Figure 18b). At very large strains, it showed monotonic fracture, characterized only by micro-dimples
(Figure 18c), which was identical to the fracture surface observed after static tensile testing (Figure 18d).
The surface damages were also observed after cyclic loading and tensile tests. In the HCF
mode, plastic flows induced by plastic deformations along the shear stress direction were developed
(Figure 17a). In the LCF mode, micro-cracks were commonly observed (Figure 17b, Figure 19a). In the
ELCF and MF modes, plastic flows induced by plastic deformations along the shear stress direction
were developed (Figure 19b,c) at high angles ranging between 30–72◦ , which were identical to that in
the static tensile test (Figure 19d).
For microstructural evolution, in the HCF mode, elongated subgrains and walls prevailed and
some nano-voids at the dislocation lines or walls could be seen (Figure 20a). However, in the LCF
mode, less dislocation tangles and veins could be seen, while incipient dislocation lines were observed
(Figure 20b). In addition, in the ELCF mode, heavy dislocation tangles and veins were observed along
the strain-localized zone (Figure 20c).
In the HCF mode, dislocation pile-ups at the dislocation lines with micro-voids were present on
the {110} planes (Figure 20d). The diffraction patterns of the tangles in the LCF mode (Figure 20e)
and the dislocation veins in the ELCF mode (Figure 20f) show that they lie on the {110} planes, where
several dislocations and cross-slips can be easily activated during the deformation of the BCC material.
4. Conclusions
Various strain amplitudes (∆ε/2 = ± 0.5–3%) were imposed on BM and WJ specimens of a
low-carbon steel with F+P microstructures to observe the transition of the fatigue fracture mode.
The conclusions drawn from this study are as follows:
(i) In the base material, the HCF and LCF fracture modes were observed through cyclic stress
responses and fractography. The cyclic stress response in the HCF mode showed initial cyclic softening,
followed by cyclic stabilization. The fractography showed inclusion-induced crack initiation at fish-eyes.
In the microstructural evolution, elongated subgrains and walls prevailed and some nano-voids at the
dislocation lines or walls could be seen.
(ii) The LCF fracture mode commonly was observed. The cyclic stress response showed initial
cyclic hardening followed by cyclic stabilization, where fractography images showed obvious striations
in the base material, while striations become blurred with transgranular cracking. In the microstructural
evolution, dislocation lines and tangles could be seen.
(iii) In the welded joint, the ELCF mode at large strains showed no cyclic stabilization after initial
cyclic hardening, which was characterized by QC fractures with a few micro-dimples and transgranular
cracking. In the microstructural evolution, dislocation veins as well as tangles could be seen.
(iv) In the welded joint, the MF mode at large strain (±3%) predominantly showed micro-dimples
(identical features to those observed after monotonic tensile loading).
Fractography, surface damage morphologies, and microstructural evolution showed evidence of
the transition of the fracture modes from LCF to ELCF and MF modes. These findings deepen the
understanding of fatigue fracture behavior of low-carbon steels, and are expected to contribute to the
use of this material in structures with improved durability and safety. For instance, structures might
consider the damage tolerance design, carrying the localized loadings imposed by welding in addition
to loads of its own axial stresses, and specifying the transition of the fatigue fracture mode depending
on cyclic responses and fatigue fractography including the strain and fatigue life.
Author Contributions: Conceptualization, formal analysis and validation, Y.K.; Writing—Review and Editing, W.H.
Funding: This research received no external funding.
Materials 2019, 12, 4111 20 of 23
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Lee, S.; Kim, B.C.; Kwon, D. Correlation of microstructure and fracture properties in weld heat-affected zones
of thermomechanically controlled processed steels. Metall. Trans. 1992, 23, 2803–2816. [CrossRef]
2. Teng, T.L.; Chang, P.H. A study of residual stresses in multi-pass girth-butt welded pipes. Int. J. Press.
Vessel. Pip. 1997, 74, 59–70. [CrossRef]
3. Luft, H.B. Development of welding procedure specification for girth welds in coiled tubing. In SPE/ICoTA
Coiled Tubing Roundtable; Society of Petroleum Engineers: Houston, TX, USA, 1999; pp. 1–16. [CrossRef]
4. Padron, T.; Aitken, B. CT100+ bias weld fatigue life estimations–Are adjustments required? In SPE/ICoTA
Coiled Tubing and Well Intervention Conference and Exhibition; SPE-179045-MS; Society of Petroleum Engineers:
Houston, TX, USA, 2016. [CrossRef]
5. Li, X.; Shi, K.; Peng, T.; Wang, H.D. Welding thermo cycle and HAZ softening of CT80 butt weld joint.
Adv. Mater. Res. 2011, 295, 1901–1904. [CrossRef]
6. Li, X.; Shi, K.; Liu, Y.; Wang, H. Research on coiled tubing butt welding process and properties of welded
joint. Trans. JWRI 2012, 21–22. Available online: https://ir.library.osaka-u.ac.jp/repo/ouka/all/23056/jwri_
wse2011_021.pdf (accessed on 27 July 2019).
7. Prasad, K.S.; Rao, C.S.; Rao, D.N. Critical Review of Fatigue Properties of Friction Stir Welds of Aluminum
Alloys. J. Mater. Metal. Eng. 2014, 4, 1–6.
8. Balasubramanian, V.; Lakshminarayanan, A.K.; Malarvizhi, S. Effect of Welding Processes on Fatigue
Behaviour of AISI 409M Grade Ferritic Stainless Steel Joints. Adv. Mater. Res. 2013, 794, 391–412. [CrossRef]
9. Devireddy, K.; Devuri, V.; Cheepu, M.M.; Kumar, B. Analysis of the Influence of Friction Stir Processing on
Gas Tungsten Arc Welding of 2024 Aluminum Alloy Weld Zone. Int. J. Mech. Prod. Eng. Res. Dev. 2018, 8,
243–252. [CrossRef]
10. Bleck, W.; Dahl, W.; Nonn, A.; Amlung, L.; Feldmann, M.; Schäfer, D. Numerical and experimental analyses
of damage behaviour of steel moment connection. Eng. Fract. Mech. 2009, 76, 1531–1547. [CrossRef]
11. Lee, Y.L.; Barkey, M.E. Metal Fatigue Analysis Handbook, Chapter 7-Fundamentals of Cyclic Plasticity Theories;
Butterworth-Heinemann: Waltham, MA, USA, 2012; pp. 253–297. [CrossRef]
12. Lubarda, V.A. Constitutive analysis of large elasto-plastic deformation based on the multiplicative
decomposition of deformation gradient. Int. J. Solids Struct. 1991, 27, 885–895. [CrossRef]
13. Manonukul, A.; Dunne, F.P.E. High- and low-cycle fatigue crack initiation using polycrystal plasticity. Proc. R.
Soc. Lond. 2004, 460, 1881–1903. [CrossRef]
14. Taeishi, K.; Hanji, T. Low cycle fatigue strength of butt-welded steel joint by means of new testing system
with image technique. Int. J. Fatigue 2004, 26, 1349–1356. [CrossRef]
15. Kunio, T.; Shimizu, M.; Ohtani, N.; Abe, T. Microstructural aspects of crack initiation and propagation.
In Extremely Low Cycle Fatigue; American Society for Testing and Materials: West Conshohocken, PA, USA,
1988; pp. 751–764. [CrossRef]
16. Kuroda, M. Extremely low cycle fatigue life prediction based on a new cumulative fatigue damage model.
Int. J. Fatigue 2002, 24, 699–703. [CrossRef]
17. Mughrabi, H. Specific features and mechanisms of fatigue in the ultrahigh cycle regime. Int. J. Fatigue
2006, 28, 1501–1508. [CrossRef]
18. Pereira, J.C.R.; de Jesus, A.M.P.; Fernandes, A.A.; Varelis, G. Monotonic, low-cycle fatigue, and ultralow-cycle
fatigue behaviors of the X52, X60, and X65 piping steel grades. J. Press. Vessel Technol. 2016, 138, 031403.
[CrossRef]
19. Magnin, T.; Coudreuse, L.; Lardon, J.M. A quantitative approach to fatigue damage evolution in fcc and bcc
stainless steel. Scr. Metall. 1985, 19, 1487–1490. [CrossRef]
20. Polák, J.; Liškutían, P. Nucleation and short crack growth in fatigued polycrystalline copper. Fatigue Fract.
Eng. Mater. Struct. 1990, 13, 119–133. [CrossRef]
21. Khireddine, D.; Khireddine, M.H. Low cycle fatigue behaviour of an aluminium alloy with small shearable
precipitates: Effect of surface coating. Int. J. Fatigue 2000, 22, 585–591. [CrossRef]
Materials 2019, 12, 4111 21 of 23
22. Shao, C.W.; Zhang, P.; Liu, R.; Zhang, Z.J.; Pang, J.C.; Zhang, Z.F. Low-cycle and extremely-low-cycle fatigue
behaviors of high-Mn austenitic TRIP/TWIP alloys: Property evaluation, damage mechanisms and life
prediction. Acta. Mater. 2016, 103, 781–795. [CrossRef]
23. Eylon, D.; Pierce, C.M. Effect of microstructure on notch fatigue properties of Ti–6Al–4V. Metall. Trans.
1976, 7, 111–121. [CrossRef]
24. Kamaya, M. Fatigue properties of 316 stainless steel and its failure due to internal cracks in low-cycle and
extremely low-cycle fatigue regimes. Int. J. Fatigue 2010, 32, 1081–1089. [CrossRef]
25. Shao, Z.; Wang, Z.; Li, Z.; Wang, S.; Wang, J. Effect of Thermomechanical Processing on the Microstructure
and Mechanical Properties of Low Carbon Steel. In Proceedings of the 5th International Conference on
Advanced Design and Manufacturing Engineering, Shenzhen, China, 19–20 September 2015. [CrossRef]
26. Endo, S.; Nakata, N. Development of Thermo-Mechanical Control Process (TMCP) and high performance
steel in JFE Steel. JFE Tech. Rep. 2015, 20, 1–7.
27. Gladman, T. Precipitation hardening in metals. Mater. Sci. Technol. 1999, 15, 30–36. [CrossRef]
28. Funakawa, Y.; Shiozaki, T.; Tomita, K.; Yamamoto, T.; Maeda, E. Development of High Strength Hot-Rolled
Sheet Steel Consisting of Ferrite and Nanometer-Sized Carbides. ISIJ Int. 2004, 44, 1945–1951. [CrossRef]
29. API Recommended Practice 5C8. Recommended Practice for Care, Maintenance and Inspection of Coiled Tubular
Product; American Petroleum Institute: Washington, DC, USA, 2010; pp. 1–4.
30. American Welding Society. Stress Relief Heat Treatment. In AWS D1.1: Structural Welding Code-Steel, 17th ed.;
American Welding Society: Miami, FL, USA, 2000; pp. 158–159.
31. Brown, P.A.; Van Arnam, D. Coiled Tubing Weld Cycle Life–Part I: Final Project Report; DEA-97; Coiled Tubing
Engineering Services: Conroe, TX, USA, 1995.
32. Lucas, W.; Howse, D. Activating Flux-Increasing the Performance and Productivity of the TIG and Plasma
Process. Weld. Met. Fabr. 1996, 64, 11–17.
33. Miranda, E.; Bertero, V. Evaluation of strength reduction factors earthquake-resistant design. J. Earthq. Spectra
1994, 10, 357–379. [CrossRef]
34. Manson, S.S.; Hirschberg, M.H. The role of ductility, tensile strength and fracture toughness in fatigue.
J. Frankl. Inst. 1970, 290, 539–548. [CrossRef]
35. Grosskreutz, J.C. Strengthening and fracture in fatigue (approaches for achieving high fatigue strength).
Metall. Trans. 1972, 3, 1255–1262. [CrossRef]
36. Mughrabi, H.; Höppel, H.W.; Kautz, M. Fatigue and microstructure of ultrafinegrained metals produced by
severe plastic deformation. Scr. Mater. 2004, 51, 807–812. [CrossRef]
37. Komotori, J.; Shimizu, M. Grain size effect in the low cycle fatigue of a steel under mean strain. Trans. Jpn.
Soc. Mech. Eng. 1989, 55, 401–408. [CrossRef]
38. Hatanaka, K.; Fujimitsu, T. Some considerations on cyclic stress-strain relation and low cycle fatigue life.
Jpn. Soc. Mech. Eng. 1984, 50, 291–300. [CrossRef]
39. Hatanaka, K. Cyclic stress-strain response and low cycle fatigue life in metallic materials. Jpn. Soc. Mech. Eng.
1984, 50, 831–838. [CrossRef]
40. Komotori, J.; Shimizu, M. Microstructural effect controlling exhaustion of ductility in extremely low-cycle
fatigue. Jpn. Soc. Mech. Eng. 1991, 57, 2879–2883. [CrossRef]
41. Suh, C.M.; Kitagawa, H. Crack growth behaviour of fatigue microcracks in low carbon steels. Fatigue Fract.
Eng. Mater. Struct. 1987, 9, 409–424. [CrossRef]
42. Brechet, Y.; Magnin, T.; Sornette, D. The Coffin-Manson law as a consequence of the statistical nature of the
LCF surface damage. Acta Metall. Mater. 1992, 40, 2281–2287. [CrossRef]
43. Basquin, O.H. The exponential law of endurance tests. Proc. ASTM 1910, 10, 625–630.
44. Coffin, L.F., Jr. A study of the effects of cyclic thermal stresses on a ductile metal. Trans. ASME 1954, 76,
931–950.
45. Manson, S.S.; Hirschberg, M.H. Fatigue Behaviour in Strain Cycling in the Low and Intermediate Cycle
Range. In Fatigue: An Interdisciplinary Approach; Syracuse University Press: Syracuse, NY, USA, 1964; p. 133,
ASIN B000K03HIM.
46. Kin, I.T.; Yamada, K.; Kainuma, S. Fatigue behavior of butt welded joints containing inclined lack
of-penetration. Struct Eng. Earthq. Eng. 2001, 18, 53–62. [CrossRef]
47. Pang, H. Analysis of weld toe profiles and weld toe cracks. Int. J. Fatigue 1993, 15, 31–36. [CrossRef]
Materials 2019, 12, 4111 22 of 23
48. Barsoum, Z. Residual Stress Analysis and Fatigue Assessment of Welded Steel Structures. Ph.D. Thesis, KTH
Royal Institute of Technology, Stockholm, Sweden, 2007.
49. Kim, J.H.; Oh, Y.J.; Hwang, I.S.; Kim, D.J.; Kim, J.T. Fracture behavior of heat-affected zone in low alloy steels.
J. Nucl. Mater. 2001, 299, 132–139. [CrossRef]
50. Zhang, H.Q.; Zhang, Y.H.; Li, L.H.; Ma, X.S. Influence of weld mismatching on fatigue crack growth behavior
of electron beam welded joints. Mater. Sci. Eng. 2002, 334, 141–146. [CrossRef]
51. Cordiano, H.V. Effect of Residual Stresses on the Low Cycle Fatigue Life of Large Scale Weldments in High
Strength Steel. J. Eng. Ind. 1970, 92, 86–92. [CrossRef]
52. Kudryavtsev, Y.; Kleiman, J. Fatigue of Welded Elements: Residual Stresses and Improvement Treatments.
Weld. World 2007, 51, 255.
53. Lee, H.; Kim, N.; Lee, T.S. Overload failure curve and fatigue behaviour of spot-welded specimens.
Eng. Fract. Mech. 2005, 72, 1203–1221. [CrossRef]
54. Chang, B.; Shi, Y.; Lu, L. Studies on the stress distribution and fatigue behaviour of weld-bonded lap shear
joints. J. Mater. Process. Technol. 2001, 108, 307–313. [CrossRef]
55. Cahoon, J.R.; Broughton, W.H.; Kutzak, A.R. The determination of yield strength form hardness measurements.
Metall. Trans. 1971, 2, 1979–1983.
56. Zuniga, S.M.; Sheppard, S.D. Determining the constitutive properties of the heat affected zone in a resistance
spot weld. Model. Simul. Mater. Sci. Eng. 1995, 3, 391–416. [CrossRef]
57. Rao, G.S.; Chakravartty, J.K.; Saibaba, N.; Mahobia, G.S.; Chattopadhyay, K.; Santhi Srinivas, N.C.
Disappearance and reappearance of serrated plastic flow under cyclic loading: A study of dislocation
substructures. Mater. Sci. Eng. 2014, 603, 114–120. [CrossRef]
58. Guo, P.C.; Qian, L.H.; Meng, J.Y.; Zhang, F.C.; Li, L.F. Low-cycle fatigue behavior of a high manganese
austenitic twin-induced plasticity steel. Mater. Sci. Eng. 2013, 584, 133–142. [CrossRef]
59. Ibrahim, M.F.E.; Miller, K.J. Determination of fatigue crack initiation life. Fatigue Fract. Eng. Mater. Struct.
1980, 2, 351–360. [CrossRef]
60. Kumar, A.; Torbet, C.J.; Jones, J.W.; Pollock, T. Nonlinear ultrasonics for in situ damage detection during
high frequency fatigue. J. Appl. Phys. 2009, 106, 024904. [CrossRef]
61. Wang, Q.Y.; Berard, J.Y.; Rathery, S.; Bathias, C. High cycle fatigue crack initiation and propagation behaviour
of high-strength spring steel wires. Fatigue Fract. Eng. Mater. Struct. 1999, 22, 673–677. [CrossRef]
62. Dvorak, I.; Hanak, J. Fatigue fracture initiation and propagation in nitrided parts. Proc. Fatigue 1999, 99,
481–486.
63. Pokluda, J.; Dvorak, I.; Horakova, H.; Major, S. Influence of plasma-nitriding surface layer on fatigue life of
steel specimens under push-pull and bending-torsion. Proc. Fatigue 2006, 6, A24.
64. Slamecka, K.; Pokluda, J.; Kianicova, M.; Major, S.; Dvorak, I. Quantitative fractography of fish-eye crack
formation under bending-torsion fatigue. Int. J. Fatigue 2010, 32, 921–928. [CrossRef]
65. Fruehan, R.J. The Making, Shaping and Treating of Steel: Steelmaking and Refining, 11th ed.; U.S.S. Co.: Pittsburgh,
PA, USA; A. S. for Metals: Novelty, OH, USA, 1998; Volume 2.
66. Tayeb, M.A. Phosphorus Control in DRI-EAF Steelmaking: Thermodynamics, Effect of Alumina, and Process
Modeling; Carnegie Mellon University: Pittsburgh, PA, USA, 2015.
67. Forsyth, P.J.E.; Stubbington, C.A.; Clark, D. Cleavage facets observed on fatigue-facture surfaces in an
aluminum alloy. J. Inst. Met. 1962, 90, 238–239.
68. Forsyth, P.J.E.; Sampson, E.G.F. Corrosion fatigue and stress corrosion cracking of an aluminium 5%
magnesium 4% zinc alloy totally immersed in 3% NaCl and other corrodents. R. Aircr. Estab. Tech. Rep.
1965, 65, 158.
69. Liu, R.; Zhang, Z.J.; Zhang, P.; Zhang, Z.F. Extremely-low-cycle fatigue behaviors of Cu and Cu–Al alloys:
Damage mechanisms and life prediction. Acta. Mater. 2015, 83, 341–356. [CrossRef]
70. Morrow, J.D. Cyclic plastic strain energy and fatigue of metals. In Internal Friction, Damping and Cyclic
Plasticity; ASTM International: West Conshohocken, PA, USA, 1965; Volume 378, pp. 45–87. [CrossRef]
71. Feltner, C.E.; Laird, C. Cyclic stress-strain response of fcc metals and alloys. I. Phenomenologial experiments.
Acta Metall. 1967, 15, 1621–1632. [CrossRef]
72. Basinski, Z.S.; Basinski, S.J. Fundamental aspects of low amplitude cyclic deformation in face-centred cubic
crystals. Prog. Mater. Sci. 1992, 36, 89–148. [CrossRef]
Materials 2019, 12, 4111 23 of 23
73. Brown, L.M. Cracks and extrusions caused by persistent slip bands. Philos. Mag. 2013, 93, 3809–3820.
[CrossRef]
74. Kang, L.; Ge, H.B. Predicting ductile crack initiation of steel bridge structures due to extremely low-cycle
fatigue using local and non-local models. J. Earthq. Eng. 2013, 17, 323–349. [CrossRef]
75. Baxter, W.J.; Wang, P.C. Finite element prediction of high cycle fatigue life aluminium alloys. Metall. Trans.
1990, 21, 1151–1159. [CrossRef]
76. Plumbridge, W.J.; Ryder, D.A. The Metallography of Fatigue. Metall. Rev. 1969, 14, 119–142. [CrossRef]
77. Tokaji, K.; Ogawa, T.; Ohya, K. The effect of grain size on small fatigue crack growth in pure titanium.
Int. J. Fatigue 1994, 16, 571–578. [CrossRef]
78. Miller, K.J.; Ibrahim, M.F.E. Damage accumulation during initiation and short crack growth regimes.
Fatigue Eng. Mater. Struct. 1981, 4, 263–277. [CrossRef]
79. Vasek, A.; Polák, J. Low cycle fatigue damage accumulation in ARMCO-iron. Fatigue Eng. Mater. Struct.
1991, 14, 193–204. [CrossRef]
80. Stolarz, J. Multicracking in low cycle fatigue-a surface phenomenon? Mater. Sci. Eng. 1997, 234, 861–864.
[CrossRef]
81. Weidner, A.; Amberger, D.; Pyczak, F.; Schönbauer, B.; Stanzl Tschegg, S.; Mughrabi, H. Fatigue damage in
copper polycrystals subjected to ultrahigh-cycle fatigue below the PSB threshold. Int. J. Fatigue 2010, 32,
872–878. [CrossRef]
82. Mughrabi, H. Cyclic slip irreversibilities and the evolution of fatigue damage. Metall. Mater. Trans. 2009, 40,
1257–1279. [CrossRef]
83. Mughrabi, H. Microstructural fatigue mechanisms: Cyclic slip irreversibility, crack initiation, non-linear
elastic damage analysis. Int. J. Fatigue 2013, 57, 2–8. [CrossRef]
84. Yang, H.K.; Doquet, V.; Zhang, Z.F. Fatigue crack growth in two TWIP steels with different stacking fault
energies. Int. J. Fatigue 2017, 98, 247–258. [CrossRef]
85. Hamada, A.S.; Karjalainen, L.P.; Puustinen, J. Fatigue behavior of high-Mn TWIP steels. Mater. Sci. Eng.
2009, 517, 68–77. [CrossRef]
86. Sun, C.T.; Jin, Z.H. The Elastic Stress Field around a Crack Tip. In Fracture Mechanics; Academic Press:
Waltham, MA, USA, 2012; pp. 25–50. ISBN 978-0-12-385001-0.
87. Rolfe, J.M.; Barsom, S.T. Fracture and fatigue control in structures. In Applications of Fracture Mechanics,
3rd ed.; ASTM: West Conshohocken, PA, USA, 1999; ISBN 978-0-75-067315-0.
88. Caillard, D. Kinetics of dislocations in pure Fe. Part, I. In situ straining experiments at room temperature.
Acta Mater. 2010, 58, 3493–3503. [CrossRef]
89. Caillard, D. Kinetics of dislocations in pure Fe. Part II. In situ straining experiments at low temperature.
Acta Mater. 2010, 58, 3504–3515. [CrossRef]
90. Erickson, J.S. Mobility of edge dislocations on {112} slip planes in 3.25% silicon iron. J. Appl. Phys. 1962, 33,
2499–2506. [CrossRef]
91. Noble, F.W.; Hull, D. Deformation of single crystals of iron 3% silicon. Philos. Mag. 1965, 12, 777–796.
[CrossRef]
92. Taoka, T.; Takeuchi, S.; Furubayashi, E. Slip systems and their critical shear stress in 3% silicon iron. J. Phys.
Soc. Jpn. 1964, 19, 701–711. [CrossRef]
93. Saka, H.; Imura, T. Direct measurement of mobility of edge and screw dislocations in 3% silicon-iron by high
voltage transmission electron microscopy. J. Phys. Soc. Jpn. 1972, 32, 702–716. [CrossRef]
© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).

حاله

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

حاله

Uploaded by

Copyright:

Available Formats

materials

Materials 2019, 12, 4111; doi:10.3390/ma12244111 www.mdpi.com/journal/materials

MF ELCF LCF HCF

2. Materials and Methods

2.1. Test Materials and Welding

2.2. Tensile and

3. Results and Discussion

3.2. Mechanical Properties

3.3. Fatigue Resistance at Welded Joint

3.3. Fatigue Resistance at Welded Joint

Figure 6. Cyclic stress–strain curves in the BM and WJ specimens.

Earlier studies have

Table 3. Cyclic stress–stain and fatigue parameters with hardness distribution.

Weld metal HAZ Base metal

Hardness (Hv) 250

Base material 202 203.0 1098.0 −0.1050 −0.6360 1125.6 0.1633

HAZ-I 221 203.0 1203.8 −0.1050 −0.6360 1234.0 0.1633

3.4. Cyclic Stress

a B0.5 B1.0 b B0.5 B1.0

0 1000 2000 3000 4000 0 400 800

c W0.5 W1.0 d W0.5 W1.0

0 1000 2000 3000 4000 0 20

800 d 800 e 800 f

-800 -800 -800

Figure 11. Hysteresis loops

-400 Cycle1 -400 -400

-400 -400 -400

-800 -800 -800

Figure 12. Hysteresis loops

Materials 2019, 12, x FOR PEER REVIEW 13 of 24

Figure 15. (a)

3.8. Fracture Features

x1000 10㎛ x1000 10㎛

The surface damage of the WJ specimens

Based on these Evolution

3.9. Microstructural Evolution after Fatigue Tests

3.10. Fatigue Features According to Fatigue Fractography and Surface Damage

Figure 20. STEM

Fracture MF ELCF LCF HCF

Np , Ductility Mixed (Intermediate) Ni , YS

Corresponding B1.0-B3.0 B0.5

*Strength * Mainly plastic

Microstructural *Dislocation tangles *Dislocation lines *Elongated subgrains and

Conflicts of Interest: The authors declare no conflict of interest.

You might also like

Strength Mainly plastic

Microstructural Dislocation tangles Dislocation lines *Elongated subgrains and