Sustainability 14 07253

sustainability
Article
Fatigue Crack Propagation and Life Analysis of Stud
Connectors in Steel-Concrete Composite Structures
Da Wang 1,2 , Benkun Tan 2, * , Shengtao Xiang 2 and Xie Wang 2
1 School of Civil Engineering, Central South University of Forestry and Technology, Changsha 410004, China;
wangda@csust.edu.cn
2 School of Civil Engineering, Changsha University of Science & Technology, Changsha 410114, China;
xst@stu.csust.edu.cn (S.X.); 20102020364@stu.csust.edu.cn (X.W.)
* Correspondence: tanbenkun@stu.csust.edu.cn; Tel.: +86-173-7572-7911
Abstract: To investigate the fatigue performance of the stud connectors of steel-concrete structures,
fatigue crack propagation analysis and fatigue life calculation were carried out. Firstly, the finite
element model with the initial crack based on linear elastic fracture mechanics (LEFM) was estab-
lished, and the parameter analysis of the stress intensity factors (SIFs) of the studs and cracks with
different geometric sizes was performed. Then, the propagation with mixed-type fatigue crack and
I-type fatigue crack of the stud were calculated, and the variation of effective SIFs with the fatigue
crack depth was analyzed. Finally, the flow chart of stud fatigue life evaluation which considers
crack initiation and stable propagation was presented, and the short stud of steel-UHPC composite
structures was taken as an example and verified. The calculation results show that the fatigue crack
propagation type and the initial crack have an obvious influence on the fatigue life of the stud. It has
acceptable accuracy that the fatigue life of short stud in UHPC simulated by considering the crack
initiation. The critical damage parameters are greatly affected by the fatigue stress amplitude, and
the initiation life of fatigue crack can account for more than 90% of the total fatigue life. This paper
Citation: Wang, D.; Tan, B.; Xiang, S.;
can provide a reference for evaluating the fatigue performance of studs in steel-concrete composite
Wang, X. Fatigue Crack Propagation
structures. Accurate evaluation of the fatigue life of stud connectors conforms to the concept of
and Life Analysis of Stud Connectors
sustainable development.
in Steel-Concrete Composite
Structures. Sustainability 2022, 14,
Keywords: stud connector; fracture mechanics; stress intensity factor; fatigue life; finite element
7253. https://doi.org/10.3390/
su14127253
analysis
Academic Editors: Qinghua Zhang,

Chuang Cui and Yi Bao
Received: 14 May 2022 1. Introduction

Accepted: 10 June 2022 The fatigue of metal materials is an important factor affecting the service life and
Published: 14 June 2022 operation safety of steel bridge structures [1]. On account of the influence of material
Publisher’s Note: MDPI stays neutral
properties and welding quality, the initiation and propagation of microscopic cracks under
with regard to jurisdictional claims in cyclic loading is an intractable problem. The engineering practice shows that it is difficult
published maps and institutional affil- to detect the microscopic cracks with naked eyes at the beginning of fatigue cracking [2].
iations. August Wöhler put forward the concept of stress amplitude-fatigue frequency curve [3]. In
the subsequent fatigue design, the nominal stress amplitude of the structure is controlled in
a reasonable range [4–7], and the different fatigue strengths of welded joints were classified
by the standard [8]. The essence of limiting the stress amplitude is to make the stress
Copyright: © 2022 by the authors. amplitude less than the threshold of microcrack propagation.
Licensee MDPI, Basel, Switzerland. The steel-concrete composite structure is widely applied in bridge construction because
This article is an open access article of its superior performance and convenient construction [9]. Researchers have carried out
distributed under the terms and extensive studies on its performance. Whitworth et al. [10] used neural network technology
conditions of the Creative Commons
to optimize the embodied energy of composite beams. Similarly, Xiang et al. [11] proposed
Attribution (CC BY) license (https://
a life-cycle simulation method of temperature field and temperature effect of steel-concrete
creativecommons.org/licenses/by/
composite bridge deck system through the BP-LSTM algorithm. Gunes et al. [1] investigated
4.0/).
Sustainability 2022, 14, 7253. https://doi.org/10.3390/su14127253 https://www.mdpi.com/journal/sustainability

Sustainability 2022, 14, 7253 2 of 19
the fatigue behavior of welded joints of steel-concrete composite beams with full depth
transverse stiffeners through repeated loading tests. Welded joints are one of the important
parts of steel-concrete composite structures with various numbers and styles, and the
welding details with eccentric load or residual stress are prone to fatigue cracking [12].
Recent accident studies show that deformation fatigue caused by welding defects between
studs and steel beams is one of the causes of the collapse of metro overpasses in Mexico
City [13]. The stud connector is an important component of steel-composite structures
which are welded to steel plates and encased in cast concrete. However, it is difficult to
carry out evaluate fatigue cracks without destructive detection owing to the coverage of
concrete [14].
As mentioned above, for the fatigue assessment of the stud connectors, it is to predict
the fatigue life under different conditions by fitting the nominal shear stress amplitude
and fatigue action times. However, in the operation of the bridge, it is located multi-axis
composite stress mode that the stud not only bears shear stress but also bears a certain
tension, especially for the short studs in the steel-UHPC composite structure. Therefore,
the traditional S-N method has certain limitations in evaluating the fatigue performance
of stud connectors with different types and application scenarios. Under the background
of the new materials and new structures in the increasing application of steel-concrete
composite structures [15,16], the formulas obtained by experimental fitting are difficult to
apply to different types and materials of studs. Lee et al. [17] found that the fatigue life
of large diameter studs which with diameters over 30 mm was lower than the predicted
value of existing specifications through the fatigue test. Xu et al. [18] proved that the rubber
sleeve would reduce the shear stiffness of the stud and lead to the reduction of fatigue
life. Cao et al. [19] conducted a push-out test on the short stud in steel-UHPC composite
structure and found that it had slightly superior fatigue performance compared with the
stud in ordinary concrete. In recent years, with the development of the finite element
method, the linear elastic fracture mechanics (LEFM) method has been widely used in
fatigue performance evaluation in the engineering field. Yang [20] analyzed the mechanical
properties and asymmetric fracture conditions of the column top plate in shallowly buried
coal seam by using the fracture mechanics method. Wang et al. [21] analyzed the fatigue
performance of the studs in the composite beam based on the LFEM method. However,
the initiation life of fatigue cracks has a significant effect on the fatigue performance of
the studs, and the research on the studs in composite structures based on the fracture
mechanics method remains to be further studied.
This study performed the fatigue crack propagation analysis and fatigue life evalu-
ation of stud connectors. Firstly, a finite element model of studs with initial cracks was
established based on LEFM. Then, the influence of different size parameters of the stud
and initial crack on SIFs was studied. The fatigue crack propagation of stud connectors
was analyzed, and the characteristics of three-dimensional fatigue crack and the variation
of SIFs were investigated. A hypothesis of I-type fatigue crack propagation was proposed
and compared with mixed-type fatigue crack propagation. Based on the above analysis,
a fatigue assessment flow chart was proposed which considers the fatigue initiation life
and stable propagation life. The critical damage plane method was used to calculate the
fatigue vulnerable element and the corresponding critical plane of the stud for obtaining
the fatigue crack initiation life of the stud. The stable propagation life of the stud was
obtained by substituting the simplified SIFs into the Paris formula. Finally, the total fatigue
life of short studs in UHPC was obtained and the calculation results were compared with
the test results and those of the specification.
2. Fatigue Propagation Analysis of Stud Connectors

2.1. Analysis Method of LEFM
As one of the most important components in the connection between steel and concrete,
a large number of experiments have been carried out on the fatigue performance of stud
connectors. At present, researches on the fatigue performance of stud connectors follow the
Sustainability 2022, 13, x 3 of 20
stud connectors. At present, researches on the fatigue performance of stud connectors fol-
low
S-N the
curveS-Nmethod
curve methodof metaloffatiguemetal fatigue
analysis, analysis,
and it isand it is determine
determine the relationship
the relationship between
between fatigue stress amplitude and fatigue life in
fatigue stress amplitude and fatigue life in the form of logarithmic formula the form of logarithmic formula
by fitting by a
fitting a large number of experiment data. However,
large number of experiment data. However, the application of the S-N curve method the application of the S-N curve
method has certain
has certain limitations limitations
whichwhich can evaluate
can only only evaluate the whole
the whole life life of materials
of materials and anddodo not
not distinguish the initial crack initiation from the macroscopic
distinguish the initial crack initiation from the macroscopic crack propagation process. On crack propagation process.
On
thethe other
other hand,
hand, forfor studconnectors,
stud connectors,the thedata
data ofof fatigue
fatigue action
actiontimes timesand andnominal
nominal stress
stress
amplitude are obtained by the push-out test or beam test, which
amplitude are obtained by the push-out test or beam test, which is required to involve is required to involve thethe
structural level rather than the metal material of the stud itself. Engineering
structural level rather than the metal material of the stud itself. Engineering projects show projects show
that
that although
althoughthe theS-N S-Ncurve
curvemethod
methodisisconservative
conservativeininstructural
structural design,
design, structural
structural fatigue
fatigue
fracture still occupies the majority of structural failure cases,
fracture still occupies the majority of structural failure cases, which is detrimental which is detrimental totothethe
sustainable development of bridge engineering. The LEFM
sustainable development of bridge engineering. The LEFM method, a mechanical analysis method, a mechanical analysis
method
method used usedto tostudy
studythe theconditions
conditionsfor forinitiation
initiationofofcrackcrackpropagation
propagation andandthethelawlawof of
crack propagation in cracked members has continuously developed
crack propagation in cracked members has continuously developed in recent years. At the in recent years. At the
level
level of standardization, the
of standardization, theshipping
shippingindustry
industryhas hascurrently
currently required
required an an
LNG LNGTypeType B
B cargo
cargo compartment to execute cracking expansion analysis
compartment to execute cracking expansion analysis based on the LEFM method [22]. At based on the LEFM method
[22]. At the application
the application level of level of the engineering
the engineering structure, structure, engineering
engineering critical critical
assessmentassessment
(ECA) is
(ECA) is applied to analyze existing cracks to support maintenance
applied to analyze existing cracks to support maintenance decision-making and formulate decision-making and
formulate inspection
inspection schemes [23]. schemes [23].
Based
Basedon onthetheabove
abovedescription
descriptionand anddiscussion,
discussion,the thefatigue
fatiguecrackcrack propagation
propagation ofofstud
stud
connectors was analyzed by the finite element method
connectors was analyzed by the finite element method of LEFM in this study, and of LEFM in this study, and thethe
variation
variation of ofSIFs
SIFsduring
duringthe thepropagation
propagationprocess processwas wasstudied.
studied.The The M-integral
M-integral method
method waswas
adopted to calculate the SIFs [24], which were co-simulated
adopted to calculate the SIFs [24], which were co-simulated by fracture mechanics softwareby fracture mechanics soft-
ware
FrancFranc
3D and 3D and
general generalfinitefinite element
element software
software Abaqus.
Abaqus. Themethod
The methodofofco-simulation
co-simulation is
is
shown in Figure 1, and the specific operation process is as follows: First,the
shown in Figure 1, and the specific operation process is as follows: First, thestructure
structure
was
was modeled by Abaqus software version 6.20 (crack-free FE model), and thematerial
modeled by Abaqus software version 6.20 (crack-free FE model), and the material
properties,
properties, boundary
boundaryconditions,conditions,and andcontact
contactrelationships
relationshipswere weredefined.
defined.Then, Then, fatigue
fatigue
cracks
cracks were
were introduced
introducedinto intothetheinitial
initialdefect
defectpositions
positionsofofstud studconnectors
connectors asas
shown
shown in in
Figure
Figure 11 and
andthe thesemi-elliptic
semi-ellipticcracked crackedFE FEmodel
modelwas wasestablished.
established.
Figure 1. Flow chart of co-simulation.

Figure 1. Flow chart of co-simulation.
After
Afterthe
theinitial
initialcracks were
cracks wereintroduced,
introduced,the mesh of theofsub-model
the mesh needs to
the sub-model be redi-
needs to be
vided, which used a variety of types of elements as shown in Figure 2. To be
redivided, which used a variety of types of elements as shown in Figure 2. To be specific,specific, the
tetrahedral element
the tetrahedral was was
element adopted to divide
adopted the global
to divide mesh.mesh.
the global The elements of theofcrack
The elements tip
the crack
were composed of three-ring, the 15-node singular element, and the 20-node
tip were composed of three-ring, the 15-node singular element, and the 20-node hexahedral hexahedral
element
element were
were adopted
adopted to to define
define the
the inner
inner ring
ring and
andouter
outerring,
ring,respectively.
respectively.The
Thepentahedral
pentahe-
dral
element was adopted between the outer ring and the global mesh for transition.After
element was adopted between the outer ring and the global mesh for transition. After
the above steps were completed, the calculations of the finite element were performed.
The program automatically imported the calculation results into Franc 3D through Python
the above steps were completed, the calculations of the finite element were performed.
The program automatically imported the calculation results into Franc 3D through Python
script
script files,
files, and
and the
the stress–strain
stress–strain field
field and
and displacement
displacement calculation
calculation results
results of
of the
the crack
crack tip
tip
were
wereprocessed,
processed, threethreetypes
types of
of SIFs
SIFs (open,
(open, slide,
slide, and
and tear)
tear) were
were obtained,
obtained, respectively.
respectively. By By
defining
definingthe thefatigue
fatiguecrack
crackpropagation
propagation parameters
parameters and thethe
and propagation
propagation step, redefining
step, the
redefining
crack surface,
the crack and and
surface, dividing the mesh,
dividing the crack
the mesh, can be
the crack canpropagated. The Abaqus
be propagated. The Abaqus software
soft-
was used for calculation again, and the cycle was followed until the set critical
ware was used for calculation again, and the cycle was followed until the set critical size size was
reached.
was reached.Due toDuetheto
large
the computational workload
large computational of the experimental
workload model, model,
of the experimental the region
the
to be performed
region for crackfor
to be performed propagation analysis was
crack propagation defined
analysis was asdefined
a sub-model, the sub-model
as a sub-model, the
was meshedwas
sub-model separately
meshed with tetrahedron
separately element and
with tetrahedron then assembled
element with the original
and then assembled with the
model.
originalThe establishment
model. of a sub-model
The establishment greatly improves
of a sub-model the computational
greatly improves efficiency ef-
the computational of
fatigue crack propagation analysis.
ficiency of fatigue crack propagation analysis.
Figure 2.
Figure 2. Mesh
Mesh division
division of
of crack
crack tip.
tip.
2.2.
2.2. Finite
Finite Element
Element Modeling
Modeling
The
The push-out test
push-out test is
is usually
usually adopted
adopted to
to analyze
analyze the
the static
static and
and fatigue
fatigue properties
properties ofof
stud
stud connectors
connectors inin steel-concrete
steel-concrete composite
composite structures.
structures. InIn this
this study,
study, the
the specimens
specimens were
were
designed
designed according
according to to the
the standard
standard ofof Eurocode
Eurocode 44 [5,25],
[5,25], which
which isis composed
composed ofof I-steel
I-steel
plates, stud connectors, concrete plates, and structural steel bars. Two studs are arranged
plates, stud connectors, concrete plates, and structural steel bars. Two studs are arranged
on each side of the steel plate by arc welding, 4 in total. The transverse spacing of the studs
on each side of the steel plate by arc welding, 4 in total. The transverse spacing of the studs
is 80 mm, and the detailed size and layout of the specimen are shown in Figure 3. To study
is 80 mm, and the detailed size and layout of the specimen are shown in Figure 3. To study
the fracture mechanical properties of different sizes of studs, three stud parameters with
the fracture mechanical properties of different sizes of studs, three stud parameters with
different diameters and four different lengths were set. The detailed dimensions of the
different diameters and four different lengths were set. The detailed dimensions of the
stud and welded collar refer to the specification of the cylindrical head stud [26]. Only a
stud and welded collar refer to the specification of the cylindrical head stud [26]. Only a
quarter of the models were built based on the symmetry of the specimens. The C3D8R
quarter of the models were built based on the symmetry of the specimens. The C3D8R
element was used for concrete, steel beams, and studs. The truss element was used to
element was used for concrete, steel beams, and studs. The truss element was used to
define steel bars. The relation definition of the contact between the stud and concrete is
define steel bars. The relation definition of the contact between the stud and concrete is
important for calculation. In this paper, the contact between the two materials was set as
important for calculation. In this paper, the contact between the two materials was set as
face-to-face contact to simulate the combined action between those two materials. The
face-to-face
tangential contact was
direction to simulate the combined
set as a penalty functionaction between
(the friction those two
coefficient wasmaterials. The
taken as 0.4),
tangential direction was set as a penalty function
and the normal direction was set as hard contact. (the friction coefficient was taken as 0.4),
and the normal
Material direction was
parameters set as hard
for concrete, contact.
I-steel, stud connectors, and reinforcement in the
model are defined separately as follows, and the type of material is noted in parentheses.
The yield strength and ultimate strength of the stud connector (ML-15) are 442 MPa and
525 MPa, respectively. The yield strength and ultimate strength of the I-steel (Q345) are
352 MPa and 495 MPa, respectively, and the yield strength of the structural reinforcement
bar (HPB300) is 365 MPa. The elastic modulus of all steel was set at 2.06 × 105 MPa. The
plastic damage model is adopted for the material properties of concrete (C50), and the
parameters and calculation formulas of uniaxial tension and compression are given in the
literature [27]. The uniaxial tension parameters include the parameter value αt of concrete
uniaxial tensile stress–strain curve drop pair. The representative value f t,r of uniaxial tensile
strength of concrete; The peak tensile strain εt,r of concrete corresponds to the representative
value of uniaxial tensile strength. The uniaxial compression parameters include elastic
modulus Ec of concrete, parameter value αc of the descending section of the stress–strain
curve of concrete under uniaxial compression, and representative value f c,r of uniaxial
compressive strength of concrete. The peak tensile strain εc,r of concrete corresponds to

the representative value of uniaxial compressive strength. Table 1 shows the values of the
above parameters.
(a) (b) (c)
Figure3.3.Specimen
Figure Specimenlayout
layoutand
andgeometric
geometricsize.
size.(a)
(a)Front
Frontview;
view;(b)
(b)side
sideview;
view;(c)
(c)vertical
verticalview.
view.
Table Material
1. Materialparameters
properties offor concrete, I-steel, stud connectors, and reinforcement in the
concrete.
model are defined separately as follows, and the type of material is noted in parentheses.
TheStrength
yield strengthαtand ultimate f t,r strengthεtof the
,r/ 10
−4 stud connector
αc (ML-15)f c,r are 442 MPa
εc,r /10−3and
Grade
525 MPa, respectively. The yield strength and ultimate strength of the I-steel (Q345) are
C50 and 4951.95
352 MPa MPa, respectively,2.5 and the 1.07yield strength 2.48 of the structural
50 1.92
reinforcement
bar (HPB300) is 365 MPa. The elastic modulus of all steel was set at 2.06 × 10 MPa. The 5
plastic
2.3. damage
Analysis of SIFs model is adopted for the material properties of concrete (C50), and the
parameters
The fatigue calculation
and fracture modes formulas of uniaxial
of stud connectors tension and compression
mainly include the following are given threein the
literature
types. One[27].
is theThe uniaxial
damage tensioncollar
to welded parameters
causedincludeby welding the parameter
quality. The value
second αt ofisconcrete
fatigue
uniaxial tensile stress–strain curve drop pair. The representative
crack until failure in the stud shank is produced in the case of good welding quality. The value ft,r of uniaxial ten-
sile strength of concrete; The peak tensile strain ε of concrete
third is the fatigue damage caused by initial defects in the interface between the stud shank
t,r corresponds to the repre-
sentative
and welded valuecollarof uniaxial tensile strength.
due to welding quality. The The uniaxial
third type compression
of initial defectparameters
is moreinclude
likely
elastic modulus E of concrete, parameter value α of
to be caused. For the third case, the initial crack was introduced, and the fatigue
c c the descending section of the stress–
crack
strain curve was
propagation of concrete
analyzed under
in this uniaxial
study. compression, and representative value fc,r of uni-
axialIncompressive
the British strength
Standards of Guide
concrete. BSThe
7910 peak
[23],tensile straincrack
the initial εc,r of is
concrete
definedcorresponds
as a semi-
to the representative
elliptical shape, its geometricvalue of uniaxial
parameters compressive
are defined strength.
by long Table 1 showslength
half-degree the valuesc and of
the above
short parameters.
half-axis depth a, and the ratio between a and c was defined as the shape ratio of
crack. Before the fatigue crack propagation analysis, the influence of the initial crack and
Tablewith
stud 1. Material
different properties
sizes onoftheconcrete.
SIF was calculated. As mentioned above, there were three
different diameters (13 mm, 16 mm, and 19 mm), four different lengths (50 mm, 70−3mm,
Strength Grade αt ft,r εt,r/10−4 αc fc,r εc,r/10
90 mm, and 110 mm), and five different shape ratios of initial crack (0.2, 0.4, 0.6, 0.8 and 1.0)
C50 1.95 2.5 1.07 2.48 50 1.92
were selected to analyze. The value of displacement loading was taken as 0.1 mm, and the
initial crack depth a was set as 0.15 mm.
2.3. Analysis
A series of of SIFs
calculations were performed through co-simulation presented in Figure 1.
It is found that thefracture
The fatigue SIFs of modes
studs with of stud connectors
different diameters mainly include
under differentthe following
shape ratios three
of
types. One is the damage to welded collar caused by welding
initial crack are complex types dominated by the type I (open) SIFs, and the shape ratios quality. The second is fa-
tigue
of crack
initial until
crack notfailure in the the
only affect studsize shank is produced
of the SIFs but also in the
thecase of good welding
distribution quality.
characteristics.
The thirdSIF
Effective is the
(Kefffatigue
) is useddamage
to evaluate caused theby
SIFs initial defects fatigue
of complex in the interface
crack, which between the stud
is calculated
shank
by and welded
Equation (1), where collar
the vdue to welding
represents quality.
Poisson’s ratio The third
and wastype
takenofasinitial defect4isshows
0.3. Figure more
likely
the to be caused.
distribution of SIFsForwith
the an third
a/ccase,
ratio the
of 0.2 initial
and crack was introduced,
1, respectively. and the fatigue
Two characteristics are
crack propagation
presented. One is thewas analyzed
convex in this study.
distribution with the maximum value at the center of the crack
tip asIn the British
shown in Figure Standards
4a, andGuide the otherBS 7910
is the[23],concave the initial crack iswith
distribution defined as a semi-
the maximum
value at the
elliptical two endpoints
shape, its geometric as shown
parameters in Figure 4b. It indicates
are defined by long that the driving
half-degree length force of
c and
fatigue crack propagation
short half-axis depth a, and at thethecrackratio tip is different
between a anddepending
c was defined on the asinitial
the shape defect. ratio of
crack. Before the fatigue crack propagation analysis, the influence of the initial crack and
stud with different sizes on the SIF was calculated. As mentioned above, there were three
different diameters (13 mm, 16 mm, and 19 mm), four different lengths (50 mm, 70 mm,
90 mm, and 110 mm), and five different shape ratios of initial crack (0.2, 0.4, 0.6, 0.8 and
shows the distribution of SIFs with an a/c ratio of 0.2 and 1, respectively. Two characteris-
tics are presented. One is the convex distribution with the maximum value at the center
of the crack tip as shown in Figure 4a, and the other is the concave distribution with the
maximum value at the two endpoints as shown in Figure 4b. It indicates that the driving
Sustainability 2022, 14, 7253 force of fatigue crack propagation at the crack tip is different depending on the initial
6 of 19
defect.
(a) (b)
Complex SIFs
Figure 4. Complex
Figure SIFs distribution along with crack tips. (a) a/c
distribution along a/c = 0.2;
0.2;(b)
(b)a/c
a/c==1.0.
1.0.
To facilitate
To the analysis
facilitate the of the
analysis of the influence
influence of of different
different parameters,
parameters, the the comparison
comparison of of
K combining the stud length, diameter, and crack shape ratio
effcombining the stud length, diameter, and crack shape ratio is shown in Figure 5, which
Keff is shown in Figure 5, which
are compared to
are compared to the
the values
values of
of KKeff /K
eff/K eff ,max
eff,max , the, the Keff ,max
Keff,max represents
represents the the extreme
extreme value
value in each
in each set
setdata.
of of data. It be
It can canseen
be seen
fromfrom the figure
the figure that the thatKthe Keff at
eff at the
the midpoint
midpoint of thefront
of the crack crackisfront is
larger
larger than the endpoint value for a crack with a small shape
than the endpoint value for a crack with a small shape (long and narrow crack). The in- (long and narrow crack). The
influence of crack shape ratio on different sizes of the stud is obvious, and the Keff of the
fluence of crack shape ratio on different sizes of the stud is obvious, and the Keff of the
initial crack is almost not affected by the length of the stud.
initial crack is almost not affected by the length of the stud.
= 2I( K )
1/ 1/2
2
( )

Ke f f K=eff K +I2 K+2I K 2
+
I II +
K 2 2
K /
I I III
I /
( 11−− vv ) (1)
0.88 0.91 0.94 0.97 1.00 1.03 0.4 0.5 0.6 0.7 0.8 0.9 1
1.1 1.2
End points of
End points of
1.0 crack front 1.0 crack front
0.9
Keff / Keff,max
Keff / Keff,max
0.8
0.8
0.6
0.7
0.4
0.6
Midpoint of
0.5 crack front 0.2 Midpoint of
0.2 0.2
0.4
50
0.4 crack front 13
0.6 70 0.6 16
a/c 0.8 90 a/c 0.8 d
h
1.0 110 1.0 19
(a) (b)
Figure
Figure 5. Comparison
5. Comparison ofeffK/K
of K eff/K underdifferent
,max under
eff eff,max differentsizes
sizesofofstud
studand
andcrack.
crack.(a)(a)Ratio
Ratioofof stud
stud height
height
and
and crack
crack shape;
shape; (b)(b) ratio
ratio of ofstudstud diameter
diameter andand crack
crack shape.
shape.
2.4. Analysis
2.4. of of
Analysis Fatigue Crack
Fatigue Propagation
Crack Propagation
According to the above analysis,
According to the above analysis, the finite
the element
finite element simulation
simulationofoffatigue
fatigue crack
crackprop-
prop-
agation for the stud was performed. The model with a diameter of 13 mm
agation for the stud was performed. The model with a diameter of 13 mm and a length of and a length
of 70
70 mm
mm inin the
the above
above analysis
analysis was
wasused
usedforforfatigue
fatiguecrack
crackpropagation
propagationanalysis,
analysis,andandthethe
nominal stress amplitude of the stud was 120 MPa with the upper limit of
nominal stress amplitude of the stud was 120 MPa with the upper limit of stress, which stress, which
waswastaken
takenasas150
150MPa.
MPa.The
Themethod
methodin inFigure
Figure 11 and
and instructions on the
instructions on the official
official website
website of
of the
thesoftware
softwareFranc
Franc3D3D was used for the calculation [28]. The propagation angle
was used for the calculation [28]. The propagation angle of a mixed- of a
mixed-type
type fatiguefatigue
crack crack was calculated
was calculated by theby the maximum
maximum circumferential
circumferential tensile tensile stress
stress criterion
criterion (MTS), and the calculation formula is shown in Equation (2). It is worth noting that
(MTS), and the calculation formula is shown in Equation (2). It is worth noting that a series
of fracture mechanics parameters must be defined before fatigue crack propagation cal-
culations are performed. In the simulation of fatigue crack propagation, it is generally
considered that the initial crack belongs to an engineering detectable crack and is no more
than 0.5 mm, previous studies indicate that the value of a0 in LEFM theory should not be
a series of fracture mechanics parameters must be defined before fatigue crack propagation
calculations are performed. In the simulation of fatigue crack propagation, it is generally
considered that the initial crack belongs to an engineering detectable crack and is no more
than 0.5 mm, previous studies indicate that the value of a0 in LEFM theory should not
be less than 0.1 mm. By defining the fatigue crack as the most unfavorable case with
a = c = 0.5 mm, Zhang et al. [29] analyzed the fatigue crack propagation in the double-sided
welding details of steel bridge decks effectively. In this paper, this typical shape of the
initial crack was introduced for calculation. In addition, other parameter settings of fracture
mechanics in Zhang’s study were referred to. The fatigue crack propagation is determined
by the threshold of SIFs (∆Kth ). The crack propagates when the amplitude of effective SIF is
greater than the threshold (∆K > ∆Kth ). Otherwise, the expansion stops, and the threshold
was taken as 63 MPa·mm1/2 . The critical value of SIFs is the fracture toughness (KIC ) of
the material. The crack propagation is in the stage of rapid expansion when the maximum
value of SIF (Kmax ) is close to the KIC , and until the Kmax at the crack tip reaches the fracture
toughness, and the material appears unstable fracture. The Paris formula is usually used to
describe the law of fatigue crack propagation. However, there is a lack of data related to the
fatigue properties of the stud material. C and m are both material parameters characterizing
crack propagation characteristics. Referring to the study performed by Xu et al. [30], C and
m were set to 4.74 × 10−14 MPa·mm1/2 and 3, respectively.
 v !
KI 2
u
KI 1u
θ MTS = 2arctan ± t +8 (2)
4K I I 4 KI I
The propagation path and spatial state diagram of the crack surface as shown in
Figure 6 were obtained. It can be seen from the figure that the fatigue crack has a fast
the length
speed alongdirection
the lengthatdirection
the initial stage
at the of propagation,
initial and eventually
stage of propagation, develops
and eventually into al-
develops
mostalmost
into a straight line which
a straight is consistent
line which with the
is consistent withtest
theresults conducted
test results by Oehlers
conducted et al.
by Oehlers
[31].
et Similar
al. [31]. to thetoexperimental
Similar the experimentalresults, the fatigue
results, cracks
the fatigue obtained
cracks by finite
obtained element
by finite sim-
element
ulation alsoalso
simulation tend to develop
tend to developtoward
toward thethe
steel plate.
steel TheThe
plate. available experimental
available experimentalresults in-
results
dicate the dip angle of fatigue crack initiation at the
indicate the dip angle of fatigue crack initiation at the interface interface between the stud shank
stud shank andand
theweld
the weldcollar
collar[31,32].
[31,32]. For
For the
the convenience
convenience of of calculation,
calculation,thethe following
followinganalysis
analysisassumes
assumes
thatthe
that theinitial
initialfatigue
fatiguecrack
crackofof
thethe stud
stud extends
extends onlyonly along
along the the surface
surface perpendicular
perpendicular to
to the
the stud
stud shank.
shank.
Figure6.6.The
Figure Thespatial
spatialshape
shapeof
ofcrack
crackpropagation.
propagation.
Based
Basedon
onthe
theabove
abovediscussion
discussionandandassumptions,
assumptions,the thecalculation
calculationresults
resultsofofSIFs
SIFsbased
based
on
on maximum circumferential tensile stress criterion (3D) and plane extension (2D)
maximum circumferential tensile stress criterion (3D) and plane extension (2D) areare
compared,
compared, asas shown
shown in inFigure
Figure7.7. The
The essence
essence of
of these
these two
twomethods
methods isismixed-type
mixed-type crackcrack
propagation
propagation and
and I-typed
I-typed crackcrackpropagation.
propagation. ForFor the
the 3D
3D extended
extended model, theKKI I always
model, the always
plays
plays a dominant role in the extended process, while the contribution of the other
a dominant role in the extended process, while the contribution of the other twotwo
SIFs
SIFs types
types to the KKeffeffamplitude
to the amplitudeis is almost
almost negligible.
negligible. TheTheKI isKchanging
I is changing rapidly
rapidly at theatcrack
the
propagation initial stage, and as the crack expands the SIFs of the open type are stable. In
the end, KI tends to decrease at the shear capacity critical position of the stud. Similarly
for the model in which cracks were assumed to propagate along the plane, the KI is dom-
inant in the early stage of crack propagation and tends to be stable. However, when the
Based on the above discussion and assumptions, the calculation results of SIFs based
on maximum circumferential tensile stress criterion (3D) and plane extension (2D) are
compared, as shown in Figure 7. The essence of these two methods is mixed-type crack
Sustainability 2022, 14, 7253 propagation and I-typed crack propagation. For the 3D extended model, the KI always 8 of 19
plays a dominant role in the extended process, while the contribution of the other two
SIFs types to the Keff amplitude is almost negligible. The KI is changing rapidly at the crack
propagation initial stage, and as the crack expands the SIFs of the open type are stable. In
crack propagation initial stage, and as the crack expands the SIFs of the open type are
the end, KI tends to decrease at the shear capacity critical position of the stud. Similarly
stable. In the end, KI tends to decrease at the shear capacity critical position of the stud.
for the model in which cracks were assumed to propagate along the plane, the KI is dom-
Similarly for the model in which cracks were assumed to propagate along the plane, the
inant
K in the early stage of crack propagation and tends to be stable. However, when the
I is dominant in the early stage of crack propagation and tends to be stable. However,
midpoint of the fatigue
when the midpoint crack
of the tip extended
fatigue to the center
crack tip extended of the
to the stud
center of section,
the studthe KII began
section, the
to surpass the KI and the increase rate was fast. Meanwhile, the amplitude of the Keff also
KII began to surpass the KI and the increase rate was fast. Meanwhile, the amplitude of the
keptalso
K increasing gradually.
kept increasing gradually.
eff
750 800
Stress intensity factors (MPa·mm1/2)
Stress intensity factors (MPa·mm1/2)

ΔKI ΔKII ΔKIII ΔKeff ΔKI ΔKII ΔKIII ΔKeff
550
500
350
Midpoint growth path 200

150
−-50 −-100
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9
Crack propagation length (mm) Crack propagation length (mm)
(a) (b)
Figure7.7.Crack
Figure propagation
Crack length-SIFs
propagation curve. (a)
length-SIFs MTS extended
curve. model (3D);
(a) MTS extended (b) planar
model extended
(3D); (b) planar
model (2D).
extended model (2D).
The
The ratio
ratioof ∆K
ofΔK III and
III andΔK∆K II to I∆K
II to ΔK with
I withthe change
the changeof crack propagation
of crack propagationdepthdepthis sum- is
marized
summarized in Figure 8a. It8a.
in Figure directly shows
It directly the dominant
shows the dominantrelationship of complex
relationship SIFs with
of complex SIFs
different propagation
with different propagationmodels and initial
models crackcrack
and initial sizes,sizes,
and the andcase where
the case a0 = 0.2
where a0 =and0.2 aand
0/c0
=a01.0 was also calculated. As can be seen from the figure that
/c0 = 1.0 was also calculated. As can be seen from the figure that when the expansion when the expansion depth is
close
depthtoisthe size to
close of the stud
size radius,
of the stud the ΔK radius, the ∆K
II/ΔKI both exceeds
II /∆K1.0 under
I both the plane
exceeds 1.0 expansion
under the
model in the casemodel
plane expansion of a0 = in0.2the
andcase a0 =of 0.5.
a0The= 0.2effective
and a0 = SIF amplitudes
0.5. of the
The effective SIFtwo different
amplitudes
extended
of the twomodels
different were compared,
extended models as shown in Figure 8b.
were compared, It is found
as shown that the
in Figure 8b. ΔK along
It iseff found
the the ∆K
thatdepth along
direction
eff the
differs depth
little direction
in the early differs
stage little
of in
fatigue the early
crack stage of
propagation. fatigue
However, crack
propagation. However,
the ΔKeff of the 2D model increases the ∆K of the 2D model increases gradually
eff gradually when the crack propagation depth is greater when the crack
propagation
than the radius depth is greater
of the stud, which than couldthe radius
lead toof athe stud,difference
certain which could lead
in the to a certain
calculation of
difference
fatigue lifein the calculation
values of the assumed of fatigue life values
propagation of the assumed propagation model.
model.
1.9 750
Stress intensity factor (MPa·mm1/2)
ΔKII/ΔKI ΔKIII/ΔKI ΔKeff

a=0.5 (3D) a=0.5 (3D)
Stress intensity factor ratio
650 a=0.5 (3D)

1.4 a=0.5 (2D) a=0.5 (2D) a=0.5 (2D)
a=0.2 (2D) a=0.2 (2D) a=0.2 (2D)
550
0.9
450
0.4
350
−-0.1 250
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9
Crack propagation length (mm) Crack propagation length (mm)
(a) (b)
Figure
Figure 8.8.Variation
Variation curves
curves of SIFs.
of SIFs. (a) ∆K(a) ΔKII/ΔKI and ΔKIII/ΔKI; (b) ΔKeff.
II /∆KI and ∆KIII /∆KI ; (b) ∆Keff .
To clarify the difference between

To between the
the calculation
calculation results
results of
of mixed-type
mixed-type crack
crack propa-
propa-
gation and
gation and I-type crack propagation, the difference in effective
effective SIF
SIF amplitude
amplitude (∆K eff, diff))
(ΔKeff,diff
between the two propagation models
between with aa00 == 0.5
models with 0.5 mm
mm and
and the
the curve
curve of
of crack
crack depth
depth versus
versus
the number of cycles is presented in Figure 9. Referring to Figure 9a, the ΔKeff of the mixed-
type crack propagation model is larger than that of the I-type crack propagation model
before the fatigue crack depth reaches the stud radius. After that, the situation was re-
versed. In the final form of the extension, the ΔKeff of the I-type crack propagation model
the number of cycles is presented in Figure 9. Referring to Figure 9a, the ∆Keff of the
mixed-type crack propagation model is larger than that of the I-type crack propagation
model before the fatigue crack depth reaches the stud radius. After that, the situation
was reversed. In the final form of the extension, the ∆Keff of the I-type crack propagation
model is 32.2% higher than the mixed-type crack propagation model. The fatigue life
can be obtained by accumulative action times of each crack propagation step with crack
depth, as shown in Figure 9b. It can be found that the calculation result of the I-type crack
propagation model is 7.2% higher than that of the mixed-type crack propagation model.
The fatigue life of the model with an initial crack depth of 0.2 is 3.14% higher than that of
the model with an initial crack depth of 0.5 under the same I-type crack propagation model.
From the above analysis, it can be concluded that both the initial crack and the fatigue
propagation mode have a significant influence on the fatigue life of the stud connector.
300
Stress intensity factor (MPa·mm1/2)
ΔKeff, diff
200
The radius of the stud

100
−-100
0 1 2 3 4 5 6 7 8 9
Crack propagation length (mm)
(a) (b)
Figure 9.
Figure 9. TheThedifference betweenmixed-type
differencebetween mixed-type crack propagationand
crack propagation andI-type
I-type crack
crack propaga-
propagation.
tion. (a) ΔK eff, diff; (b) crack depth versus the number of cycles.
(a) ∆Keff,diff ; (b) crack depth versus the number of cycles.
3. Simplified
Simplified Calculation
Calculation Method Method forfor Fatigue
Fatigue Life
Life ofof Stud
Stud Connectors
Connectors
Although
Although the the fatigue
fatiguecrackcrackpropagation
propagationpattern
patternofofthe thestud
studwas
was obtained
obtained according
accordingto
the method
to the method above, sincesince
above, the stud
the connector is wrapped
stud connector in concrete,
is wrapped and the contact
in concrete, and thebetween
contact
different
between materials with the complex
different materials with the element,
complex the non-convergence
element, of the calculation
the non-convergence was
of the calcu-
quite frequent, which lead to a situation that it is still difficult for the
lation was quite frequent, which lead to a situation that it is still difficult for the fatigue to fatigue to crack
propagation calculation.
crack propagation Meanwhile,
calculation. the fatigue
Meanwhile, crack propagation
the fatigue parameters
crack propagation for fatigue
parameters for
life calculation of the stud material still need to be determined by a lot
fatigue life calculation of the stud material still need to be determined by a lot of experi- of experiments. To
evaluate the fatigue life of stud effectively, instead of only calculating the
ments. To evaluate the fatigue life of stud effectively, instead of only calculating the num- number of fatigue
actions from the
ber of fatigue nominal
actions from stress
the amplitude
nominal stressaccording to the according
amplitude S-N curve,toa relatively simplea
the S-N curve,
calculation method needs to be developed. Based on the above calculation,
relatively simple calculation method needs to be developed. Based on the above calcula- for fatigue crack
propagation
tion, for fatigue crack propagation at the interface of stud shank and welded collar,SIFs
at the interface of stud shank and welded collar, the variation trend of the
under mixed-type
variation crackunder
trend of SIFs propagation
mixed-typeis notcrack
different from thatisof
propagation I-type
not crackfrom
different propagation,
that of I-
in which
type crackthepropagation,
differences mainlyin which exist
theindifferences
the post-cracking
mainlystage.
exist Therefore, the study on
in the post-cracking the
stage.
fatigue life of rubber-sleeved studs conducted by Xu et al. [30] was referred to in the study,
Therefore, the study on the fatigue life of rubber-sleeved studs conducted by Xu et al. [30]
and the fatigue crack propagation was simplified. Based on the classic S-N curve method,
was referred to in the study, and the fatigue crack propagation was simplified. Based on
fatigue crack initiation life and macroscopic crack propagation life were considered, and
the classic S-N curve method, fatigue crack initiation life and macroscopic crack propaga-
the fatigue performance of the stud can be evaluated by the simplified fatigue calculation
tion life were considered, and the fatigue performance of the stud can be evaluated by the
method of fracture mechanics. On the whole, the fatigue crack propagation process of stud
simplified fatigue calculation method of fracture mechanics. On the whole, the fatigue
connectors can be divided into the initial crack initiation stage (stage I), the fatigue crack
crack propagation process of stud connectors can be divided into the initial crack initiation
stable propagation stage (stage II), and the crack unstable propagation stage (stage III).
stage (stage I), the fatigue crack stable propagation stage (stage II), and the crack unstable
Stage III develops rapidly which is the last stage of fatigue failure, and it is acceptable not
propagation stage (stage III). Stage III develops rapidly which is the last stage of fatigue
to be counted in the overall fatigue life. Therefore, the fatigue life (Nf ) of the stud in this
failure, and it is acceptable not to be counted in the overall fatigue life. Therefore, the
research was obtained by calculating the sum of the initiation life (Nstage I ) and the stable
fatigue life (Nf) of the stud in this research was obtained by calculating the sum of the
propagation life (Nstage II ) of crack. The Smith–Watson–Topper (SWT) critical plane damage
initiation[33]
method lifefor
(Nstage I) and the stable propagation life (Nstage II) of crack. The Smith–Watson–
multiaxial fatigue was applied to calculate the crack initiation life Nstage I ,
Topper (SWT) critical plane damage method [33] for multiaxial fatigue was applied to cal-
culate the crack initiation life Nstage I, and the stable extension life Nstage II was calculated by the
LEFM method. The flow chart of stud fatigue life assessment is presented in Figure 10.
The stud shear connector is subjected to shear force, bending moment, and axial load
under the action of fatigue load in the concrete deck, and it is in the multi-axial stress–
point of the element, it is usually located in the stress–strain concentrated area, for mem-
bers subjected to complex stresses, and the fatigue crack propagates in the most unfavor-
able plane of the element.
Sustainability 2022, 14, 7253 Δε 10 of 19
D= σ max (3)
2
where Δε and σmax represent the normal strain amplitude and maximum normal stress of
and the stable extension life Nstage II was calculated by the LEFM method. The flow chart of
each fatigue load cycle, respectively, and at the critical plane, the damage parameter
stud fatigue life assessment is presented in Figure 10.
reaches its maximum value.
Establishing finite element model
Static calculation and determination of

fatigue concern area
Stress and strain values of each integral point

in the fatigue concern area The SIFs is calculated by fracture
mechanics formula or finite
element method
Stress and strain of each element at their
various critical planes
The fatigue crack propagation

parameters were determined by
Maximum and minimum Maximum stress of of experiments
strains of each element in each element in each
each plane plane
Paris formula
The maximum SWT parameter of each
element
SWT damage parameters of fatigue concern Macroscopic fatigue crack

area propagation life Nstage II of stud
connectors
Fatigue crack initiation life (Nstage I) of stud
connectors
Fatigue life value (Nf) of stud connector
Figure 10. Flow chart of stud fatigue life assessment.

Figure 10. Flow chart of stud fatigue life assessment.
After SWT parameters related to fatigue life are obtained, the fatigue crack initiation
The stud shear connector is subjected to shear force, bending moment, and axial load
life can be calculated by Equation (4) [34]. Referring to the study performed by Xu et al.
under the action of fatigue load in the concrete deck, and it is in the multi-axial stress–
[30],
strainthe parameters
state, which can σ f′also ′ in the formula were set to 350 MPa and 0.0715, respec-
andbeε fseen from the distribution characteristics of the complex
tively, while b and
SIFs. Therefore, theccritical
were set to −0.07
damage and method
plane −0.4, respectively.
of multiaxial fatigue life prediction was
used to calculate the most unfavorable
Δε (σ f′ )
plane direction, and the initiation life Nstage I of the
( stage I ) f f ( stage I )
2b b+c
stud. The SWT parameters are σ max = to
used N
analyze + σ ′ε ′ N
multiaxial stress problems as shown(4) in
2 E
Equation (3). The product in Equation (3) is a characterization of strain energy which is
considered that the critical plane direction is the direction in which the SWT parameter
value of a point reaches its maximum. In general, for the critical plane method, the position
of the critical plane is determined by the stress and strain parameters of the integral point
of the element, it is usually located in the stress–strain concentrated area, for members
subjected to complex stresses, and the fatigue crack propagates in the most unfavorable
plane of the element.
∆ε
D = σmax (3)
2
where ∆ε and σmax represent the normal strain amplitude and maximum normal stress
of each fatigue load cycle, respectively, and at the critical plane, the damage parameter
reaches its maximum value.
Sustainability 2022, 13, x After SWT parameters related to fatigue life are obtained, the fatigue crack initiation
life can be calculated by Equation (4) [34]. Referring to the study performed by Xu et al. [30],
the parameters σf0 and ε0f in the formula were set to 350 MPa and 0.0715, respectively, while
b and c were set to −0.07 and −0.4, respectively.
where b and c represent
∆ε
the
σf0
fatigue strength index and fatigue ductility
2b b+c
tively. σ f′ and ε2f′ σmax =
represent Nfatigue
stage I σf0 ε0f Nstagecoefficient
+strength I (4) ducti
and fatigue
E
respectively.
where b and c represent the fatigue strength index and fatigue ductility index, respectively.
σf0 and εAs
0 represent fatigue strength coefficient and fatigue ductility coefficient, respectively.
f mentioned above, the most unfavorable position needs to be dete
As mentioned above, the most unfavorable position needs to be determined before
fatigue life calculation and analysis, and then the maximum SWT parame
fatigue life calculation and analysis, and then the maximum SWT parameter and its cor-
responding
responding critical
critical plane direction
plane direction were To
were calculated. calculated.
be specific, To bethe
firstly, specific, firstly, t
stress dis-
bution ofofthe
tribution thestud
stud connector
connector in the in the steel-concrete
steel-concrete specimen wasspecimen was
calculated, andcalculated
the
most unfavorable position under fatigue load was preliminarily determined. Then, SWT
unfavorable position under fatigue load was preliminarily determined. T
parameters of the element at the most unfavorable position were evaluated. Finally, the
rameters
normal of the component
stress–strain element at of the mostwas
any plane unfavorable position
obtained through were evaluated.
the coordinate trans- F
mal stress–strain
formation matrix as showncomponent
in Figure 11.of any
The plane transformation
coordinate was obtained through
formula is shownthe coord
in Equation (5). In Figure 11, the coordinate system before transformation is xyz, and a new
mation matrix as shown in Figure 11. The coordinate transformation formu
coordinate system x0 y0 z0 is obtained after a coordinate transformation. Based on the above
Equation
calculation of (5). In Figure
the most 11,position
unfavorable the coordinate system before
and the corresponding transformation
critical plane of fatigue is
coordinate
crack, the fatiguesystem x’y’z’
initiation life ofis
theobtained after a coordinate
stud was calculated transformation.
according to Equation (4). Base
calculationσ of
= σthe 2 most 2 unfavorable 2 position and the corresponding critic
xx n x + σyy ny + σzz nz + 2τxy n x ny + 2τyz ny nz + 2τxz n x nz
tigue crack, (5)
ε =the
ε xx nfatigue
2 2initiation life of the stud was calculated according
2
x + ε yy ny + ε zz nz + 2γxy n x ny + 2γyz ny nz + 2γxz n x nz
t
Figure
Figure 11. 11. Schematic
Schematic diagramdiagram ofposition.
of any plane any plane position.
In Equation (5), σ and ε represent the normal stress and strain after the transformation
plane, nx , ny , and nz represent the cosine of the 2 angle between the post-transformation
σ = σ xx n 2
x + σ yy n y + σ zz n 2
z + 2τ xy
coordinate system and the pre-transformation coordinate system, respectively.
n x n y + 2τ yz nThe + 2τ xz nx nz
y nz calcula-
ε =ε xx nx2 +stress
tion formula is as follows. Firstly, the
ε nand y + ε zz nz + 2γ xy nx n y + 2γ yz n y nz + 2γ xz nx nz
2 strain2 components in Equation (5) were
extracted from the calculation results of theyyfinite element model, and the angle parameters
θ and ϕ in Equation (6) were successively evaluated from 0◦ to 180◦ which was separated by
In Equation (5), σ and ε represent the normal stress and strain afte
mation plane, nx, ny, and nz represent the cosine of the angle between the
mation coordinate system and the pre-transformation coordinate system
The calculation formula is as follows. Firstly, the stress and strain compone
10◦ . Then, the direction cosine value was calculated through Equation (6) and substituted
into Equation (5). Finally, the normal stress and strain after coordinate transformation
were obtained, which were used for the subsequent calculation of SWT parameters and
fatigue life.
n x = cos θ sin ϕ
ny = sin θ sin ϕ (6)
nz = cos ϕ
where θ is the angle between the x-axis and the plane normal vector projection on the xy
plane, and ϕ is the angle between the plane normal vector and the z-axis.
The LEFM method was used to calculate the stable extended fatigue life (Nstage II ) of
the stud. Based on the above finite element calculation of fracture mechanics, the SIF was
simplified and obtained by Equation (5), it was assumed that a linear correlation with the
fatigue crack depth, and then the fatigue life was calculated through Equation (6). The
values of C and m were taken as 4.74 × 10−14 and 3.0, respectively [32]. The initial crack a0
is generally obtained by non-destructive testing, and the initial crack a0 of the stud was set
as 2 mm in this study. √
K = Fσ πa (7)
Z a
f da
Nstage I I = (8)
a0 C (∆K )m
where F represents the geometric modification coefficient of fatigue crack, which was
calculated according to 1.12 in this study. σ represents the stress of the structure without
initial crack, a0 represents the initial fatigue crack depth, and af represents the crack depth
under fatigue failure.
The shear resistance of a stud under fatigue load is mainly determined by the shear
area of the stud shank. Before the initial fatigue crack appeared, the shear resistance area
was the circular section area A of the stud with diameter d, As the crack expanded under
fatigue load, the shear resistance area of the stud decreased continuously, and, finally,
decreased to the ultimate shear resistance area Af corresponding to the upper limit of the
fatigue load amplitude Pmax (Equation (9)). Therefore, the fatigue crack depth af in fatigue
failure can be calculated by Equation (10), and fu is the ultimate tensile strength of the
stud material.
Pmax
Af = (9)
fu
Af af
= 1− (10)
A d
4. Analysis of Calculation Examples

4.1. Model Establishment
The fatigue experiment of short stud connectors in a steel-UHPC composite structure
was performed by Cao et al. [19], and it was set as a calculated example to verify the
effectiveness of the above method. The finite element model was established regarding the
push-out test which was according to the standard of Eurocode 4 [5]. Specific parameter
settings are as follows. The UHPC deck was taken as 50 mm, I section steel with a thickness
of 12 mm. The spacing of ordinary steel bars is 50 mm × 55 mm. Four short studs with
diameters of 13 mm and 35 mm in height were arranged on each side of I-shaped steel,
with a horizontal and vertical spacing of 110 mm and 200 mm, respectively. The elastic
modulus of steel was set to 2.06 × 105 MPa. The yield strength and ultimate strength of
stud and I-shaped steels were 345 MPa and 430 MPa, respectively, the yield strength and
ultimate strength of steel were 335 MPa and 400 MPa, respectively, and the Poisson’s ratio
of steel was set to 0.3. The compressive strength of the UHPC was taken as 129.1 MPa,
and the initial elastic modulus was set to 42.6 GPa. It is similar to the finite element model
for fatigue crack progradation, the example model consists of a UHPC plate, short stud,
I-shaped steel plate, and structural steel bar. The modeling method was consistent with the
previous model and not described here, the difference lies in the definition of constitutive
relations for UHPC materials, which was explained here. The tension and compression
stress–strain curves of UHPC were referred to in Figure 12. The specific parameters were set
Sustainability 2022, 13, x
according to the experimental results in the literature [35–37], and the calculation formula 14 of 20
is shown in Figure 12. Based on the principle of energy equivalence, the damage factor D
of UHPC under tension and compression was defined by Equation (11). The parameter
meanings of the formula in the Figure 12 are as follows. Et0 represents strain at peak
represents
tension, crack width
ft represents parameter,
average stress pduring
represents
strainthe parameters
hardening, obtained by
εtp represents axial tension
ultimate strain
test fitting, fc represents the compressive strength, ξ
under tension, lc represents the extension distance measured by the specimen, wcompressive
represents the ratio of p represents
crack
strainwidth
value parameter, p represents
to compressive the parameters
peak strain, obtained
and ε0 represents by axial
peak straintension test fitting, fc
in compression.
represents the compressive strength, ξ represents the ratio of compressive strain value to
σ
compressive peak strain, and ε0 represents D =peak
1 − strain in compression. (11)
E0 ε
r
σ
D = 1− (11)
E0 ε
(a) (b)
12. Stress–strain curve of UHPC material.

Figure 12. material. (a) Tensile stress–strain
stress–strain curve; (b) compression
compression
curve.
stress–strain curve.
4.2. Verification of
4.2. Verification of Load-Slip
Load-Slip Curves
Curves
Firstly,
Firstly, the static performance of
the static performance of the
the stud
stud connector
connector obtained
obtained byby the
the finite
finite element
element
method
method was verified. Three push-out specimens (SAT-1~SAT-3) with the same parameters
was verified. Three push-out specimens (SAT-1~SAT-3) with the same parameters
were
were carried
carried out
out by
by Cao
Cao et
et al.
al. [19],
[19], and
and the
the slip
slip and
and ultimate
ultimate bearing
bearing capacity
capacity ofof the
the stud
stud
were recorded. The comparison between the load-slip curve calculated by
were recorded. The comparison between the load-slip curve calculated by the finite ele- the finite element
model and the test results is shown in Figure 13. It can be concluded from Figure 13 that the
ment model and the test results is shown in Figure 13. It can be concluded from Figure 13
results calculated by the finite element model are consistent with the measured load-slip
that the results calculated by the finite element model are consistent with the measured
curve, and the results are most similar to those of specimen SAT-2. The stud in the UHPC
load-slip curve, and the results are most similar to those of specimen SAT-2. The stud in
deck of the static performance is mainly divided into two stages, at the beginning of the
the UHPC deck of the static performance is mainly divided into two stages, at the begin-
load, a load of stud-slip can be regarded as a linear correlation, and then the stud material
ning of the load, a load of stud-slip can be regarded as a linear correlation, and then the
yields strength after the interface slip velocity gradually accelerated, until you reach the
stud material yields strength after the interface slip velocity gradually accelerated, until
limit strength stud shear failure occurs, these characteristics are also reflected in the finite
you reach the limit strength stud shear failure occurs, these characteristics are also re-
element calculation results. The yield load and ultimate load measured in the experiment
flected in the finite element calculation results. The yield load and ultimate load measured
are 400 kN and 496 kN, respectively. Comparatively, the yield load calculated by the finite
in the experiment are 400 kN and 496 kN, respectively. Comparatively, the yield load cal-
element method is relatively small which are 376 kN and 456 kN, respectively. The errors
culated by the finite element method is relatively small which are 376 kN and 456 kN,
between the finite element model and experiment are 6% and 2%, respectively, which is
respectively.
acceptable forThe errors between
the prediction of thethe finite
shear element
bearing modelofand
capacity the experiment are 6%
stud connectors. Theand 2%,
above
respectively, which is acceptable for the prediction of the shear bearing capacity
analysis shows that the calculation results of the finite element are in good agreement with of the
stud connectors. The above analysis shows that the calculation results of the
the experimental results. The verification of the static experiment model provides the basisfinite element
are the
for in good agreement
following with the
calculation, andexperimental
the fatigue life results.
valuesThe verification
calculated of the10static
in Figure experi-
are verified
ment model provides the basis
by the fatigue test results below. for the following calculation, and the fatigue life values
calculated in Figure 10 are verified by the fatigue test results below.
Sustainability 2022,14,
Sustainability2022, 13,7253
x 14 of
15 of 19
20
500
500
400
400
(kN)
300
Load(kN)
300
Load
200 SAT-1
200 SAT-1
SAT-2
SAT-2
100 SAT-3
100 SAT-3
FEM
FEM
0
00.0 0.2 0.4
0.6 0.8 1.0 1.2
0.0 0.2 0.4
0.6 0.8 1.0 1.2
Slip (mm)
Slip (mm)
Figure 13. Comparison of load slip curves of studs.
Figure13.
Figure 13.Comparison
Comparisonof
ofload
loadslip
slipcurves
curvesof
ofstuds.
studs.
4.3. Calculation of SWT Parameters

4.3.
4.3. Calculation
Calculation of ofSWT
SWTParameters
Parameters
In this study, fatigue life calculations and analyses of six specimens (N1~N6) were
In
In this study, fatiguelife
this study, fatigue lifecalculations
calculationsand andanalyses
analysesof ofsix
sixspecimens
specimens(N1~N6) (N1~N6) were were
performed. The
performed. The failure
failure modes
modes observed
observed in in stud
stud fatigue
fatigue teststests inin UHPC
UHPC are are summarized
summarized in in
performed. The failure modes observed in stud fatigue tests in UHPC are summarized in
the
the following two cases [19]. One case is that the fatigue crack source of the first condition
thefollowing
followingtwo twocases
cases[19].
[19]. One
One casecase is
is that
that the
the fatigue
fatigue crackcrack source
source of of the
the first
firstcondition
condition
appears at
appears at the welded
welded collar of of the studstud with the the steel plate.plate. With the the repeated action action of
appears at the the welded collar collar of thethe studwith with thesteel steel plate. With With therepeatedrepeated actionof of
the
the fatigue load, the crack gradually extends on the steel plate and forms a small concave
thefatigue
fatigueload,load,thethecrack
crackgradually
graduallyextends
extendson onthe thesteel
steelplate
plateand andformsformsaasmall smallconcave
concave
surface. In
surface. In the
thesecond
secondcase, case,the
thefatigue
fatigue crack
crack appears
appears at the interface of the studstudshank and
surface. In the second case, the fatigue crack appears at atthethe interface
interface of the
of the stud shank shank
and
welded
and collar,
welded which
collar, can
which be regarded as the cross-section of the stud was gradually pen-
welded collar, which can can be regarded
be regarded as the as cross-section
the cross-section of the ofstud
the stud
was was gradually
gradually pen-
etrated by the
penetrated crack
by the along
crack withwith
along the transverse expansion untiluntil
the shear limitlimit
of theofstud
etrated by the crack along with the the transverse
transverse expansion
expansion until the shear the shearlimit of the the
stud
was was
stud reached. As mentioned
reached. As mentioned above,above,thethe stress
stress at at
thethe intersection
intersection ofofthe
the stud
stud shank
shank and
and
was reached. As mentioned above, the stress at the intersection of the stud shank and
welded collar
welded collar is the mostmost unfavorable, which which is preliminarily determined determined as the the fatigue
welded collar isisthe the mostunfavorable,
unfavorable, whichisispreliminarily preliminarily determined as as thefatigue
fatigue
crack
crack location.
location. For
For simplicity,
simplicity, SWT SWT parameters
parameters for for
the the
entire entire
fatigue fatigue
detail detail
region region
describedde-
crack location. For simplicity, SWT parameters for the entire fatigue detail region de-
scribed
in in Figure
Figurein10Figurewere 1010 were
notwere not
calculated calculated here.
here. Therefore, Therefore, taking
taking specimen specimen
N1 as an N1 as an exam-
scribed not calculated here. Therefore, taking specimen N1example,
as an exam-the
ple, the calculation
calculation process process proposed
proposed in Figure in
10 Figure
was 10 was
listed as listed asThe
follows. follows.
SWT The SWT param-
parameters at the
ple, the calculation process proposed in Figure 10 was listed as follows. The SWT param-
eters at the
interface interface
between studbetween
shank studwelded
and shank were and welded
calculated, were calculated,
where the where
stress is the stress is
concentrated,
eters at the interface between stud shank and welded were calculated, ◦ . The
where the stress is
concentrated,
and the distance andbetween
the distance between
the angles onthe eachangles
planeon waseach setplane
to 10was set SWT
to 10°.parameter
The SWT
concentrated, and the distance between the angles on each plane was set to 10°. The SWT
parameter
result result
of each planeof each
of theplane
mostofunfavorable
the most unfavorable
element iselement shown is inshown
Figurein Figure
14. 14. The
The angle θ
parameter result of each plane of the most unfavorable element is shown in Figure 14. The
angle θ and φ of the most unfavorable plane of the maximum
and ϕ of the most unfavorable plane of the maximum parameter are calculated, which are parameter are calculated,
angle
◦ and θ and φ of the most unfavorable plane of the maximum parameter are calculated,
which
90 20◦90°
are and 20°, respectively.
, respectively. In subsequent In analysis,
subsequent analysis,
it was determinedit wasthat determined
the fatigue that the
crack
which are 90° and 20°, respectively. In subsequent analysis, it was determined that the
fatigue crack
developed fromdeveloped
the planefrom the plane corresponding
corresponding to this position.toThe thisdata
position.
obtained Theabovedata obtained
analysis
fatigue crack developed from the plane corresponding to this position. The data obtained
above
is listedanalysis is listed
as follows, as follows,
the SWT parameter the SWT value parameter
is 74,460,value is 74,460,
the normal the normal
vector vector
of the critical
above analysis is listed as follows, the SWT parameter value is 74,460, the normal vector
plane
of the for the global
critical plane forplane (−0.000272,
theisglobal plane is0.342, 0.940),0.342,
(−0.000272, the normal
0.940),strain
the normalamplitudestrain∆ε is
am-
of the critical plane for the global plane is (−0.000272, 0.342, 0.940), the normal strain am-
915 µε, the
plitude Δε maximum
is 915 με, the normal
maximumstress σ normal
max is 166.8
stress MPa,
σ max and
is the
166.8 normal
MPa, andstress
the amplitude
normal is
stress
plitude Δε is 915 με, the maximum normal stress σmax is 166.8 MPa, and the normal stress
135.1 MPa.
amplitude is 135.1 MPa.
amplitude is 135.1 MPa.
Figure 14.
Figure 14. SWT
SWT parameter
parameter diagram
diagram of
of stud
stud shank-weld
shank-weld collar
collar interface.
interface.
Figure 14. SWT parameter diagram of stud shank-weld collar interface.
4.4. Calculation of Fatigue Life

SWT parameter and normal stress amplitude ∆σ were obtained according to the above
calculation, the initiation life (Nstage I ) and the stable propagation life (Nstage II ) of crack can
be calculated by Equations (4) and (8). As shown in Table 2, the fatigue test results (Ne )
and calculation results (Nf ) of specimens N1~N5 are summarized, the relationship between
stress amplitude and fatigue life of stud is usually expressed by logarithmic relation. The
error between logNe and logNf ranges from 0.5% to 7.7%, and the results of logNe /logNf
range from 0.98 to 1.04, which shows that the calculated results have a good correlation
with the experimental data. By further calculation, the correlation coefficient between the
calculated results in this paper with the experimental results is 0.9955. The above analysis
shows that the method of predicting the fatigue life by considering the fatigue initiation
life and stable extension life of the stud is feasible. Among all the specimens, the predicted
results of N1 were closest to the experimental values.
Table 2. Fatigue life calculation results.
Number F max F min ∆F τ max τ min ∆τ Ne Nstage I Nstage II Nf logNe /logNf

N1 1 122 22 100 115 21 94 11,787,000 9,883,700 1,299,136 11,182,836 1.00
N2 151 27 124 143 26 117 1,130,000 1,206,000 398,725 1,604,725 0.98
N3 162 29 133 152 27 125 1,688,000 600,000 392,224 992,224 1.04
N4 175 31 143 165 30 135 441,000 243,000 224,643 467,643 0.99
N5 175 21 154 165 20 145 620,000 221,000 185,308 406,308 1.03
N6 1 122 22 100 115 21 94 / 12,590,000 892,000 13,482,000 /
1 N1 and N6 were consistent except that the material parameters of UHPC were set differently.
Statistical analysis can be carried out on the relationship of action times under different
fatigue load amplitudes through data fitting, and the S-N curve of short stud in steel-UHPC
composite structure can be established. The design specification [5–7] is also guided by
the S-N curve formula which is shown in Equation (12), and the parameter values in the
formulas of different specifications are comprehensively given in Table 3. The parameter
values obtained by the fatigue experiment results of specimens N1~N5 are also given in
Table 3, and its survival rate is 95% which ensures the safety of the prediction [19].
m0 log ∆τ + log N = C 0 (12)
where N represents the number of fatigue loads, ∆τ represents the amplitude of the nominal
shear stress on the stud, and C0 and m0 represent the parameters obtained by fitting the
experimental data (Table 3).
Table 3. Parameters of the S-N curve formula.
Source of Parameter Data m0 C0

Eurocode 4 [5] 8 21.935
JSCE [6] 8.55 23.42
AASHTO [7] 10 26.15
N1~N5 [19] 8 22.3587
The stud fatigue life values in UHPC with different elastic modulus and compressive
strength were calculated, and the material was set as UHPC with the elastic modulus of
43 GPa and compressive strength of 170.9 MPa as described in the literature [35]. The SWT
damage value of the most unfavorable element of the stud was calculated as shown in
Figure 15. It can be seen from the figure that SWT damage patterns of different intensities
are similar in distribution, and the maximum SWT damage value was 71,810 which is
lower than the value of the N1 specimen with the compressive strength of 129.1 MPa. The
calculated fatigue initiation life value is 12.59 million times, and the stable extended life
value is 892,000 times after changing the material parameters of UHPC. Figure 16 presents
Sustainability 2022, 13, x 17
Sustainability 2022, 14, 7253 calculated fatigue initiation life value is 12.59 million times, and the stable 16 of 19
extended
value is 892,000 times after changing the material parameters of UHPC. Figure 16 pres
the ratio of SWT parameters of UHPC with different compressive strengths, and it is
the ratio of SWT parameters of UHPC with different compressive strengths, and it is that
only the values with θ = 90° and φ = 20° were compared for convenience. Based on
only the values with θ = 90◦ and ϕ = 20◦ were compared for convenience. Based on the
above discussion, it can be seen that there are certain differences in the fatigue pe
above discussion, it can be seen that there are certain differences in the fatigue performance
mance
of of concrete
studs in studs inwith
concrete with
different different
strengths, strengths,
and the ratio isand the uniform
almost ratio is almost uniform
except for a ex
for points,
few a few points, especially
especially when ϕ = when◦
20 . φ = 20°.
Figure SWT
Figure15.15. parameter
SWT diagram
parameter of studofinstud
diagram UHPCin with
UHPC different
with strengths.
different strengths.
2.0
θ =90°
The ratio of SWT parameters
φ =20°
1.5
1.0
0.5 Position of the maximum

SWT parameter
0.0
0 20 40 60 80 100 120 140 160 180
Angle value (°)
Figure16.16.
Figure SWTSWT parameter
parameter ratioratio of UHPC
of UHPC with different
with different compressive
compressive strength. strength.
The experiment results of specimens N1~N6 and the curve formula obtained by fitting
The experiment results of specimens N1~N6 and the curve formula obtained b
were drawn, as shown in Figure 17. The normalized curves in Table 3 and the results
ting wereindrawn,
calculated as shown
this paper were also inshown
Figurein17. The17
Figure normalized
for comparison.curves in be
It can Table
seen3that
and the re
calculated
the fatigue life inofthis paper
short studswere
in UHPC alsoisshown
outsideintheFigure 17 curve
predicted for comparison.
derived fromItfitting
can be seen
the experimental data. For the N1 specimen, the fatigue stress amplitude is 94 MPa, the from fi
the fatigue life of short studs in UHPC is outside the predicted curve derived
fatigue life value obtained
the experimental data. inForthethe
testN1is 11.787 millionthe
specimen, times, and the
fatigue result
stress calculated is
amplitude by94 MPa
Eurocode 4 [5] is 1.412 million times, the difference in fatigue life values
fatigue life value obtained in the test is 11.787 million times, and the result calculatewas caused by the
larger elastic modulus and constraint action of UHPC. According to the data in Table 2, it
Eurocode 4 [5] is 1.412 million times, the difference in fatigue life values was cause
can be found that the initial fatigue crack initiation life accounts for a large proportion of
thewhole
the largerfatigue
elasticlife,
modulus
up to 93%. and constraint
It accounts foraction
a largerof UHPC. According
proportion of the wholetofatigue
the data in T
2, itfor
life can be connectors
stud found thatthat theare
initial fatigue
subjected crackfatigue
to small initiation life
stress accountsasfor
amplitude a large
shown in propo
of the whole fatigue life, up to 93%. It accounts for a larger proportion of the whole fat
Figure 18. It is indicated that a considerable part of the time is in the initial fatigue crack
initiation stageconnectors
life for stud of the stud in theare
that actual bridge operation.
subjected However,
to small fatigue the fatigue
stress failureas show
amplitude
Figure 18. It is indicated that a considerable part of the time is in the initial fatigue c
initiation stage of the stud in the actual bridge operation. However, the fatigue fa
occurs relatively quickly after the fatigue crack expands to the stage of steady propaga
which is detrimental to the safety of the structure. Therefore, for the sustainable opera
Sustainability2022,
Sustainability 2022,13,
13,xx 18
18
occurs relatively quickly after the fatigue crack expands to the stage of steady propagation
which is detrimental to the safety of the structure. Therefore, for the sustainable operation
of steel-UHPC
of steel-UHPC composite
composite bridges,
bridges, ititisisparticularly
particularlyimportant
importantto todetect
detectthethe fatigue
fatigue cr
c
of steel-UHPC composite bridges, it is particularly important to detect the fatigue cracks in
in the
in
the the welding
welding
welding detailsdetails
details of
of studof stud connectors.
stud connectors.
connectors. Meanwhile, Meanwhile,
Meanwhile, the fatigue
the
the fatigue stress fatigue stress
stress
amplitude amplitude
amplitude
is the key ii
keyparameter
key parameter
parameter todetermine
to determine
to determine the
the
the fatigue fatigue
fatigue
crack crackpropagation
crack
propagation propagation lifeof
life ofstud
life of stud connectors.studconnectors.
connectors.
300
300
(MPa)
amplitude (MPa)
250
250 NNf f
200
200 NNee
stress amplitude
150
150
Nominal stress
100
100 Eurcode 44
Eurcode
JSCE
JSCE
Nominal
AASHTO
AASHTO
Cao et al. [19]
50
50
1044
10 1055
10 1066
10 1077
10 1088
10
Fatiguelife
Fatigue life
Figure
Figure
Figure 17.
17.
17.S-NS-N
S-N curve
curve ofshort
short
of short
curve of stud
stud stud
in inUHPC.
UHPC.
in UHPC.
12
12
NNee
10
10
NNStage
StageIIII
(×105))
5
NNStage
life (×10
88 StageI I
Fatigue life
66
Fatigue
44
22
00
N1
N1 N2
N2 N3
N3 N4
N4 N5
N5
SpecimenNumber
Specimen Number
Figure
Figure
Figure 18.
18.
18. Bar
Bar
Bar diagram
diagram
diagram offatigue
of
of the thefatigue
the fatigue life.
life. life.
5. Conclusions and Observations

5.Conclusions
5. Conclusionsand
andObservations
Observations
In this study, the LEFM method was used to calculate the fatigue crack propagation of
Inconnectors
In
the stud thisstudy,
this study, thesteel-concrete
in the
the LEFMmethod
LEFM method wasused
was
composite used tocalculate
to calculate
structure, thefatigue
the fatiguecrack
and the characteristics crack propag
of SIFspropaga
of
werethe
of the stud
analyzed. connectors
stud connectors
Combinedin in the
thethe
with steel-concrete
steel-concrete composite
composite
simplified analysis structure,
methodstructure, and the
andcrack
of the fatigue the ofcharacterist
characteristi
the
SIFsthe
stud,
SIFs were
were analyzed.
fatigue Combined
life of the
analyzed. with
stud in thewith
Combined UHPC the
deck
the simplified analysis
was calculated.
simplified methodconclusions
The method
analysis research ofthe
of thefatigue
fatiguecra
cra
are
the as follows.
the stud,
stud, the
the fatigue
fatigue life
life of
of the
the stud
stud in
in the
the UHPC
UHPC deckdeck was
was calculated.
calculated. The
The research
research
clusions
(a) areas
It belongs
clusions are as follows.
to follows.
the complex crack of the stud connector which is dominated by the open
type SIF. The distribution of SIFs is different under the different crack shapes, and the
(a) ItItbelongs
(a) belongsto tothe
thecomplex
complexcrack crackof ofthe
thestud
studconnector
connectorwhichwhichisisdominated
dominatedby bythe
theo
extreme value of Keff is greatly affected by the depth and width of the crack, but not
type SIF.
type SIF. The
significantly The distribution
distribution
affected by the length of SIFs
of SIFs
of is
theis different under
different
stud. under the the different
different crack
crack shapes,
shapes
(b) Thethe extreme
the fatigue
extreme value
value
crack of
of K
surface K is greatly
is stud
ofeffthe
eff greatly affected
affected
connector by
can by the depth
the depth
be obtained and width of the
and width of the crack
by three-dimensional crack
notsignificantly
not
fatiguesignificantly affected
propagation.affected bythe
by thelength
It is semi-elliptical length ofthe
of
in the earlythe stud.
stud.
stage of the fatigue crack propaga-
tion,
(b) The
(b) and the
Thefatigue crack
fatiguecrack front
cracksurfacegradually
surfaceof ofthedevelops
thestud into
studconnectora straight
connectorcan canline in the laterby
beobtained
be obtained stage.
by The
three-dimens
three-dimensi
midpoint of the fatigue
fatiguepropagation. crack front grows
propagation.ItItisissemi-elliptical faster,
semi-ellipticalin and
inthethe fatigue
theearly
earlystagecrack
stageof surface
ofthe fatiguetocrack
tends
thefatigue crackpr
pr
fatigue
incline toward the I-beam.
gation, and
gation, and the
the crack
crack front
front gradually
gradually develops
develops intointo aa straight
straight line
line in
in the
the later
later st
s
The midpoint
The midpoint of of the
the fatigue
fatigue crack
crack front
front grows
grows faster,
faster, and
and the
the fatigue
fatigue crack
crack sur
su
tendsto
tends toincline
inclinetoward
towardthe theI-beam.
I-beam.
(c) There
(c) Thereisislittle
littledifference
differencebetween
betweenthe
theΔK ΔKeffeffof
ofthe
theI-type
I-typecrack
crackpropagation
propagationand andthth
themixed-type
the mixed-typecrackcrackpropagation
propagationin inthe
theearly
earlystage
stageofoffatigue
fatiguepropagation.
propagation.Howe How
(c) There is little difference between the ∆Keff of the I-type crack propagation and that of
the mixed-type crack propagation in the early stage of fatigue propagation. However,
the ∆Keff of the I-type crack propagation model increases gradually after the crack
propagates through half of the stud section, which leads to a certain difference in
the number of effects calculated by the two methods in the late fatigue propagation
period. The fatigue life of the I-type crack propagation model is 7.2% higher than that
of the mixed-type crack propagation model.
(d) The calculation method of crack initiation life is considered by simplifying SIF and
combining the SWT critical damage plane method. Compared with the experimental
values, it is proved that the calculated values of fatigue life of stud connectors provide
better predictive values. In addition to the stress amplitude of the traditional S-N
curve, different material sizes, material properties, and contact characteristics of
structures can also be taken into account by the calculation method in this paper
combined with finite element modeling. It can effectively evaluate the fatigue life of
stud connectors in steel-concrete composite structures.
(e) The proportion of crack initiation life in the fatigue life of stud connectors in UHPC is
different under different stress amplitudes, and the compressive strength of UHPC
also affects the fatigue life of stud connectors. In the whole process of fatigue failure,
the initiation life of fatigue crack accounts for a large proportion, up to more than
90%. The fatigue stress amplitude is the key parameter to determine the fatigue crack
propagation life of stud connectors.
Author Contributions: Conceptualization, D.W. and B.T.; Data curation, B.T. and S.X.; Funding
acquisition, D.W.; Investigation, S.X. and X.W.; Methodology, D.W. and B.T.; Project administration,
D.W. and B.T.; Resources, D.W.; Software, B.T., S.X., and X.W.; Supervision, D.W. and S.X.; Validation,
D.W. and B.T.; Writing—original draft, B.T. and X.W.; Writing—review & editing, D.W. and S.X. All
authors have read and agreed to the published version of the manuscript.
Funding: This research was funded by the National Natural Science Foundation of China (Grant no.
51878072), the Science and Technology Program of Hunan Province (Grant no. 2020RC4049), and the
Science Research Start Foundation for Bring in Talents of the Central South University of Forestry
and Technology (Grant no. ZK2021YJ036).
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Not applicable.
Conflicts of Interest: The authors declare that they have no known competing financial interests or
personal relationships that could appear to influence the work reported in this paper.
References
1. Gunes, B.; Ilki, A.; Kemal, O.N. Determination of monitoring parameters for fatigue behavior of steel-concrete composite bridge
girders with welded full depth transverse stiffeners. Sustainability 2019, 12, 283. [CrossRef]
2. Lim, Y.Y.; Soh, C.K. Electro-mechanical impedance (EMI)-based incipient crack monitoring and critical crack identification of
beam structures. Res. Nondestruct. Eval. 2014, 25, 82–98. [CrossRef]
3. Wöhler, A. Versuche zur Ermittlung der auf die Eisenbahnwagenachsen einwirkenden Kräfte und die Widerstandsfähigkeit der
Wagen-Achsen. Z. Bauwes. 1860, 10, 583–614.
4. Freudenthal, A.M.; Gumbel, E. Physical and statistical aspects of fatigue. Adv. Appl. Mech. 1956, 4, 117–158.
5. CEN. EN 1922-1-1 Eurocode 2: Design of Concrete Structures–Part 1-1: General Rules and Rules for Buildings; European Committee for
Standardization: Brussels, Belgium, 2005.
6. JSSC. Guidelines for Performance-Based Design of Steel-Concrete Hybrid Structures; Japan Society of Civil Engineers: Tokyo, Japan, 2002.
7. AASHTO. AASHTO LRFD Bridge Design Specifications; American Association of State Highway: Washington, DC, USA, 2012.
8. Huang, Y.; Zhang, Q.; Bao, Y.; Bu, Y. Fatigue assessment of longitudinal rib-to-crossbeam welded joints in orthotropic steel bridge
decks. J. Constr. Steel Res. 2019, 159, 53–66. [CrossRef]
9. Wang, D.; Liu, Y.; Liu, Y. 3D temperature gradient effect on a steel-concrete composite deck in a suspension bridge with field
monitoring data. Struct. Control Health Monit. 2018, 25, e2179. [CrossRef]
10. Whitworth, A.H.; Tsavdaridis, K.D. Genetic Algorithm for Embodied Energy Optimisation of Steel-Concrete Composite Beams.
Sustainability 2020, 12, 3102. [CrossRef]
11. Xiang, S.; Wang, D.; Yang, L.; Tan, B. Study on the life cycle simulation method of the temperature field and temperature effect of
a steel–concrete composite bridge deck system. Meas. Control 2021, 54, 1068–1081. [CrossRef]
12. Cheng, Z.; Zhang, Q.; Bao, Y.; Deng, P.; Wei, C.; Li, M. Flexural behavior of corrugated steel-UHPC composite bridge decks. Eng.
Struct. 2021, 246, 113066. [CrossRef]
13. Mexico City Metro Overpass Collapse. Available online: https://en.wikipedia.org/wiki/Mexico_City_Metro_overpass_collapse
(accessed on 12 May 2022).
14. Ovuoba, B.; Prinz, G.S. Investigation of residual fatigue life in shear studs of existing composite bridge girders following decades
of traffic loading. Eng. Struct. 2018, 161, 134–145. [CrossRef]
15. Ferreira, F.P.V.; Tsavdaridis, K.D.; Martins, C.H.; De Nardin, S. Steel-Concrete Composite Beams with Precast Hollow-Core Slabs:
A Sustainable Solution. Sustainability 2021, 13, 4230. [CrossRef]
16. Li, C.; Lei, Z.; Feng, Z.; He, W.; Tan, L. Research on static performance of lightweight STC-steel composite deck. J. Transp. Sci. Eng.
2021, 37, 26–33. (In Chinese) [CrossRef]
17. Lee, P.G.; Shim, C.S.; Chang, S.P. Static and fatigue behavior of large stud shear connectors for steel–concrete composite bridges. J.
Constr. Steel Res. 2005, 61, 1270–1285. [CrossRef]
18. Xu, X.; Zhou, X.; Liu, Y. Behavior of rubber-sleeved stud shear connectors under fatigue loading. Constr. Build. Mater. 2020,
244, 118386. [CrossRef]
19. Cao, J.; Shao, X.; Deng, L.; Gan, Y. Static and fatigue behavior of short-headed studs embedded in a thin ultrahigh-performance
concrete layer. J. Bridge Eng. 2017, 22, 04017005. [CrossRef]
20. Yang, D. Analysis of Fracture Mechanics Theory of the First Fracture Mechanism of Main Roof and Support Resistance with
Large Mining Height in a Shallow Coal Seam. Sustainability 2021, 13, 1678. [CrossRef]
21. Wang, Y.; Nie, J. Fatigue behavior of studs in a composite beam based on fracture mechanics. J. Tsinghua Univ. 2009, 49, 1467–1470.
(In Chinese) [CrossRef]
22. Wolfe, R.J.; Vanderveldt, H.; Henn, A.E. Some considerations of fracture mechanics applications in ships design, construction and
operation. Eng. Fract. Mech. 1975, 7, 565. [CrossRef]
23. BSI. BS 7910; 2013+A1:2015 Guide to Methods for Assessing the Acceptability of Flaws in Metallic Structures. BSI Standards
Limited: London, UK, 2015.
24. Wang, D.; Tan, B.; Wang, L.; Chen, F.; Xiang, S. Numerical Study on Stress Intensity Factors for Stud Connectors of Steel–Concrete
Connection. Int. J. Steel Struct. 2021, 21, 1775–1789. [CrossRef]
25. Wang, B.; Huang, Q.; Liu, X. Comparison of Static and Fatigue Behaviors between Stud and Perfobond Shear Connectors. KSCE J.
Civ. Eng. 2018, 23, 217–227. [CrossRef]
26. GB/T 10433-2002; Cheese Head Studs for Arc Stud Welding. China Standards Press: Beijing, China, 2002.
27. GB50010-2010; Code for Design of Concrete Structures. Ministry of Housing and Urban-Rural Development of the People’s
Republic of China (MOHURD): Beijing, China, 2011.
28. Fracture Analysis Consultants, Inc. Available online: http://fracanalysis.com/software.html (accessed on 7 March 2021).
29. Zhang, Q.; Guo, Y.; Li, J.; Yuan, D.; Bu, Y. Fatigue Crack Propagation Characteristics of Double-sided Welded Joints between Steel
Bridge Decks and Longitudinal Ribs. China J. Highw. Transp. 2019, 32, 49–56. [CrossRef]
30. Xu, X.; Zhou, X.; Liu, Y. Fatigue life prediction of rubber-sleeved stud shear connectors under shear load based on finite element
simulation. Eng. Struct. 2021, 227, 111449. [CrossRef]
31. Oehlers, D.J.; Seracino, R.; Yeo, M.F. Fatigue behaviour of composite steel and concrete beams with stud shear connections. Prog.
Struct. Eng. Mater. 2000, 2, 187–195. [CrossRef]
32. Hanswille, G.; Porsch, M.; Ustundag, C. Resistance of headed studs subjected to fatigue loading: Part I: Experimental study. J.
Constr. Steel Res. 2007, 63, 475–484. [CrossRef]
33. Smith, R.N.; Watson, P.; Topper, T.H. A stress-strain function for the fatigue of metals. J. Mater. 1970, 5, 767–778.
34. Kujawski, D. A deviatoric version of the SWT parameter. Int. J. Fatigue 2014, 67, 95–102. [CrossRef]
35. Zhang, Z.; Shao, X.; Li, W.; Zhu, P.; Chen, H. Axial tensile behavior test of ultra high performance concrete. China J. Highw. Transp.
2015, 28, 50. (In Chinese) [CrossRef]
36. Li, L.; Fan, X.; Shi, X.; Shao, X. Experimental study on flexural behavior of large-scale prestressed UHPC T-shaped beam. China
Civ. Eng. J. 2018, 51, 84–94. (In Chinese) [CrossRef]
37. Yang, J.; Fang, Z. Research on stress-strain relation of ultra high performance concrete. J. Concr. 2008, 7, 11–15. (In Chinese)

Sustainability 14 07253

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

Sustainability 14 07253

Uploaded by

Copyright:

Available Formats

sustainability

Academic Editors: Qinghua Zhang,

Received: 14 May 2022 1. Introduction

Sustainability 2022, 14, 7253. https://doi.org/10.3390/su14127253 https://www.mdpi.com/journal/sustainability

2. Fatigue Propagation Analysis of Stud Connectors

Sustainability 2022, 14, 7253 3 of 19

Figure 1. Flow chart of co-simulation.

Sustainability 2022, 14, 7253 4 of 19

Sustainability 2022, 13, x 5 of 20

(a) (b) (c)

Stress intensity factors (MPa·mm1/2)

Midpoint growth path 200

ΔKII/ΔKI ΔKIII/ΔKI ΔKeff

650 a=0.5 (3D)

To clarify the difference between

The radius of the stud

Establishing finite element model

Static calculation and determination of

Stress and strain values of each integral point

The fatigue crack propagation

SWT damage parameters of fatigue concern Macroscopic fatigue crack

Fatigue life value (Nf) of stud connector

Figure 10. Flow chart of stud fatigue life assessment.

4. Analysis of Calculation Examples

12. Stress–strain curve of UHPC material.

4.3. Calculation of SWT Parameters

4.4. Calculation of Fatigue Life

Table 2. Fatigue life calculation results.

Number F max F min ∆F τ max τ min ∆τ Ne Nstage I Nstage II Nf logNe /logNf

m0 log ∆τ + log N = C 0 (12)

Table 3. Parameters of the S-N curve formula.

Source of Parameter Data m0 C0

0.5 Position of the maximum

5. Conclusions and Observations

You might also like