You are on page 1of 14

Engineering Structures 254 (2022) 113875

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Fatigue performance evaluation of steel-UHPC composite orthotropic deck


in a long-span cable-stayed bridge under in-service traffic
Shiqiang Qin a, b, Jiabin Zhang a, b, Chunlei Huang a, b, Liqiang Gao b, c, Yi Bao d, *
a
School of Civil Engineering and Architecture, Wuhan University of Technology, Wuhan 430070, Hubei China
b
State Key Laboratory for Health and Safety of Bridge Structures, Wuhan 430034, Hubei, China
c
China Railway Major Bridge Engineering Group Co., Ltd., Wuhan 430034 Hubei, China
d
Department of Civil, Environmental and Ocean Engineering, Stevens Institute of Technology, Hoboken, NJ 07030, United States

A R T I C L E I N F O A B S T R A C T

Keywords: Ultra-high-performance concrete (UHPC) emerges as an advanced material that is promising to improve the
Fatigue assessment fatigue performance of bridge decks. This study evaluates the fatigue performance of steel-UHPC composite deck
In-service traffic load in the Junshan Yangtze River Bridge, which is a long-span cable-stayed bridge, under in-service traffic loads. The
Orthotropic bridge deck
bridge was instrumented with strain gauges that provided in-situ strain data at fatigue-prone details. The strain
Steel-UHPC composite deck
Stress monitoring
data of fatigue-prone details were used to evaluate the maximum and effective stress ranges and fatigue life of
steel-UHPC deck. A finite element model was established to analyze stress distributions and validated by the in-
situ strain monitoring data. The results showed that the UHPC layer significantly reduced the maximum stress
ranges and the effective stress ranges of fatigue-prone details and increased fatigue life. Attention should be paid
to rib splice joints in bridge inspection and maintenance because its maximum stress range exceeded the
constant-amplitude fatigue limit by 40%.

1. Introduction the interface between steel deck and pavement is limited [8]. (3) The
pavement is prone to cracking. Once cracked, the contribution of
Orthotropic steel deck (OSD) is widely used in medium- and long- pavement to the stiffness of bridge deck is highly reduced, and, thus, the
span bridges owing to its light weight, high load-bearing capacity, and stress ranges at fatigue-prone details are increased, compromising the
strong spanning capacity [1,2]. An OSD is mainly composed of steel fatigue performance [16]. To increase the stiffness of deck system,
deck, rib, and floor beam, which are welded to bear loads together [3]. steel–concrete composite decks were proposed. Compared with con­
The deck thickness of OSD in early bridges was thin (about 10 ~ 12 ventional OSD, a reinforced concrete layer was cast on top of the steel
mm), and the welding quality was limited [4,5], making the welded deck plate, and steel headed studs welded on the steel deck plate were
joints of OSD susceptible to fatigue cracks. For example, shortly after used to integrate the OSD and reinforced concrete layer [17]. However,
construction, different degrees of cracking occurred on the OSD of the the crack resistance and tensile strength of conventional concrete are
Haseltal Bridge and the Sinntal Bridge in Germany, the Severn Bridge in low [18]. Once concrete cracks, steel bars are exposed to the environ­
the United Kingdom [6], the Throgs Neck Bridge in the United States ment and have accelerated corrosion, thus reducing the durability.
[7], the Tianjin Haihe Bridge in China [8], and the Jiangyin Yangtze There are many factors that can cause concrete cracking such as
River Bridge in China [9]. Fatigue life is a major concern of the dura­ shrinkage, overloading, and temperature changes [19]. It is essential to
bility of OSD in bridge applications. enhance the crack resistance and tensile strengths of concrete because
Many measures have been proposed to improve the fatigue resistance they are important to achieve and maintain the stiffness of bridge decks.
of OSD in bridges. Representative methods include drilling holes to ar­ In recent years, various high-performance fiber-reinforced cementitious
rest cracks [10], adding steel plates [11], and using fiber reinforcement composites have been developed to deliver high crack resistance, tensile
[12,13]. However, the effects of those measures are limited [14]. The strength, and long-term durability. Representative materials include
limited effects are attributed to three main reasons: (1) The stiffness of strain-hardening cementitious composite (SHCC) and ultra-high-
OSD is limited, resulting in high stresses [15]. (2) The bond strength of performance concrete (UHPC) [20,21]. Based on micromechanics,

* Corresponding author.
E-mail address: yi.bao@stevens.edu (Y. Bao).

https://doi.org/10.1016/j.engstruct.2022.113875
Received 29 June 2021; Received in revised form 16 December 2021; Accepted 9 January 2022
Available online 19 January 2022
0141-0296/© 2022 Elsevier Ltd. All rights reserved.
S. Qin et al. Engineering Structures 254 (2022) 113875

SHCC was developed by mechanistically tuning cementitious matrix, Representative previous research on the fatigue behaviors of OSD
fibers, and fiber–matrix interface, and features exceptional ductility. enhanced by UHPC is summarized in Table 1. Most previous research
Based on particle packing theory, UHPC was designed by maximizing focused on the deck of suspension bridges and performed through finite
the particle packing density. Compared with conventional high- element analysis or laboratory tests. The bridge test reported in refer­
performance fiber-reinforced concrete, UHPC features self- ence [32] was subjected to static loading of the bridge. The bridge test
consolidation and high mechanical strengths. Typically, UHPC has a reported in reference [33] was subjected to low-speed vehicle. To date,
water-to-binder ratio lower than 0.25 and a compressive strength at 28 there is lack of research on the fatigue performance of steel-UHPC
days higher than 120 MPa under standard curing conditions [22]. Due to composite decks of realistic cable-stayed bridges under in-service
the outstanding mechanical properties and durability, both SHCC and traffic loads. Another observation from the previous research is that
UHPC have been used to replace conventional concrete in steel–concrete improvement of the fatigue resistance of the deck systems was different.
composite decks [23] and develop innovative orthotropic bridge decks The discrepancy in improvement was likely associated with the different
to significantly enhance the fatigue performance [24]. geometry and load conditions of the deck systems. For example, when
Fatigue tests have been performed to investigate the effects of UHPC the thicknesses of the steel deck and UHPC layer were respectively 12
on fatigue performance of OSD for girder bridges and suspension mm and 50 mm, different results of the stress reduction at the fatigue-
bridges. Abdelbaset et al. conducted full-scale fatigue tests of OSD prone details were reported in references [16,29,36]. However, when
specimens with and without UHPC layer and found that the hot spot the same geometry was applied to the deck systems [16,36], inconsistent
stresses of fatigue-prone details were decreased by up to 83% because results were still observed. In a nutshell, it is still unclear how UHPC
the UHPC layer increased the stiffness and reduced the stress ranges of affects the fatigue life of bridge decks under realistic traffic loads
the bridge deck [25]. Unterweger et al. investigated the fatigue life of [35,38].
steel-UHPC composite deck, considering overload and precipitation ef­ To fill the above gaps, this research investigates the fatigue perfor­
fects, and found that use of an 80-mm-thick UHPC layer could increase mance of a steel-UHPC composite deck of a long-span cable-stayed
the service life of the OSD by at least 50 years [26]. Yuan et al. used bridge using in-situ strain monitoring data for the fatigue-prone zones of
UHPC to strengthen a cracked OSD of a suspension bridge through a full- the steel-UHPC composite deck under in-service traffic loads. Strain
scale model test [18]. In addition to fatigue tests, finite element models sensors were deployed on the steel-UHPC composite decks of the Jun­
have been developed to predict fatigue performance of OSD with UHPC. shan Yangtze River Bridge, a cable-stayed bridges with a main span of
Zhu et al. proposed a finite element model to investigate the stresses of 460 m. The strain distributions at fatigue-prone details were monitored
the fatigue-prone details of OSD in a girder bridge, and found that the and used to calculate the maximum and effective stress ranges that were
stress range on the deck side of the rib-to-deck weld was reduced by 71% in turn utilized to evaluate the fatigue life of fatigue-prone details. An
when a 45-mm-thick UHPC layer was used [27]. Jiang et al. performed a interesting feature of the Junshan Yangtze River Bridge is that the bridge
finite element analysis for fatigue-prone details of a steel-UHPC com­ uses a conventional OSD on the upstream side and a steel-UHPC com­
posite bridge deck of a suspension bridge and reported that the fatigue posite deck on the downstream side. In other words, the bridge provides
resistance could be highly enhanced by the UHPC layer [28]. Zhang a unique case for comparative research on the fatigue behaviors of the
et al. developed a finite element model for a steel-UHPC composite conventional OSD and steel-UHPC deck in the same bridge. This
bridge deck of a suspension bridge and found that the use of UHPC could research is expected to provide new insights into the enhancement of the
reduce the stress ranges of fatigue-prone details by up to 87% [29]. The fatigue resistance of different fatigue-prone details of long-span cable-
above fatigue tests and finite element studies have been verified by field stayed highway bridges under in-service traffic loads by utilizing in-situ
tests on the residual fatigue life of bridges [30]. Cao et al. reported that bridge monitoring data.
use of UHPC in OSD of a bridge reduced the peak tensile stresses of the The remainder of the paper is organized as follows: Section 2 in­
deck side and rib wall side of rib-to-deck welds by 92% and 80%, troduces the main structure and instrumentation of the Junshan Yangtze
respectively [31]. Su et al. claimed that the use of steel-UHPC composite River Bridge. Section 3 compares the strains/stresses measured from
bridge deck reduced the deflections by 23% for ribs and 27% for deck fatigue-prone details of the conventional OSD and steel-UHPC compos­
[32]. Pei et al. indicated that steel-UHPC composite deck could reduce ite deck. Section 4 calculates the effective stress range and the residual
the stress ranges of fatigue-prone details of suspension bridges [33]. Cui fatigue life for each fatigue-prone detail. Section 5 presents the finite
et al. studied the fatigue life of OSD under in-service traffic and welded element analysis for the maximum stress range in the UHPC layer of the
residual stress, and found that the fatigue life was dependent on the steel-UHPC composite deck under design loads, and identifies the
traffic [34]. Deng et al. evaluated the fatigue life of OSD considering the fatigue-prone zones that need special attention in bridge inspection and
road surface condition, and concluded that the fatigue life was over­ maintenance. Section 6 summarizes the conclusions and
estimated by up to 110% without considering road surface condition recommendations.
[35].

Table 1
Comparison of representative research on the effect of UHPC on fatigue of OSD.
Ref. Year Bridge type Steel deck UHPC layer Research method Stress reduction of fatigue-prone details by UHPC

Detail 1 Detail 2 Detail 3 Detail 4 Detail 5 Detail 6

[37] 2015 Suspension 12 mm 45 mm FEA 82% 51% 27% 22% 21% 28%
[29] 2016 Suspension 12 mm 50 mm FEA 86% 87% 49% 70% 50% 34%
[36] 2017 Suspension 12 mm 50 mm FEA 86% 52% 30% 38% 24% 23%
[28] 2017 Suspension 12 mm 45 mm FEA 83% NA 48% 66% 25% NA
[33] 2018 Suspension 14 mm 50 mm Bridge test 75% NA 35% 66% 18% 66%
[16] 2018 Cable-stayed 12 mm 50 mm FEA 87% 51% 27% 37% 13% 36%
[27] 2019 Continuous 16 mm 45 mm FEA 71% 26% 21% 34% 21% 25%
[18] 2019 Suspension 12 mm 60 mm Lab test 91% 74% NA NA NA 43%
[32] 2020 Continuous 16 mm 50 mm Bridge test 75% 42% NA 26% 4% NA
[39] 2021 Continuous 16 mm 50 mm FEA 63% 42% 51% 50% NA NA

Note: “Steel deck” refers to the thickness of the original steel deck of the OSD. “UHPC layer” is the thickness of the added UHPC. “FEA” stands for finite element
analysis. The fatigue-prone details 1 to 6 refer to deck side of rib-to-deck weld, rib side of rib-to-deck weld, floor beam side of rib-to-floor beam weld, rib beam side of
rib-to-floor beam weld, floor beam cutout, and rib splice, respectively.

2
S. Qin et al. Engineering Structures 254 (2022) 113875

2. Field bridge monitoring measuring 30 mm in thickness was placed on top of the UHPC. The
adopted UHPC was consistent with the UHPC mixture in reference [42].
2.1. Bridge description In conventional steel–concrete composite sections, the thickness of
concrete layer was usually thicker than 150 mm, and the length of
Fig. 1 shows the structure and main dimensions of the Junshan headed shear studs was longer than 100 mm. For example, the concrete
Yangtze River Bridge, which is an important highway bridge located in thickness was 150 mm, and the length of the headed studs was 105 mm
Hubei Province, China. The main bridge is a semi-floating cable-stayed in reference [43]. Due to the higher compressive strength of UHPC than
bridge with two towers and two planes of stay cables, as shown in Fig. 1 conventional concrete, the thickness of UHPC was greatly reduced in the
(a). The span layout of the bridge is 48 + 204 + 460 + 204 + 48 m, with steel-UHPC composite sections. In this project, the thickness of UHPC
the main span measuring 460 m in length, as shown in Fig. 1(b). There was 55 mm, as shown in Fig. 5. Due to the thin thickness of UHPC, short
are a total of 36 pairs of cables on one tower. The cable spacing on the headed studs were used, and the length was 35 mm, respectively. At the
steel box girder is 17 m. A typical cross section of the bridge is shown in position where there was a steel plate, the length of the headed studs was
Fig. 1(c). The width of the bridge deck is 38.8 m, with six traffic lanes 27 mm to achieve flush studs over the deck.
including three lanes for the downstream and three lanes for the up­ The performance of the short headed studs was evaluated through
stream. The thickness of deck plate ranges from 12 mm to 16 mm. The U- experiments, as elaborated in reference [40]. The steel-UHPC composite
ribs have a thickness of 6 mm and a depth of 260 mm. The spacing section was applied to a deck specimen, and the deck specimen was
between adjacent ribs is 300 mm. The floor beam where stay cables are tested under mechanical loads until failure. When the specimen failed,
anchored are 20 mm in thickness, other floor beams are 10 mm, and the the failure mode and damages were examined. The examination results
spacing between two floors is 3 m. indicated that there was no interface slippage between the concrete and
The Junshan Yangtze River Bridge represents a pioneer project of steel, suggesting that the short headed studs enabled the deck specimen
utilizing UHPC in OSD, as shown in Fig. 1(c) and Fig. 1(d). The bridge to achieve adequate shear transfer at the steel-UHPC interface. The shear
uses a conventional OSD with a 75-mm-thick asphalt concrete layer (See performance was related to the high mechanical strengths of the UHPC
Fig. 1(d)) on the upstream side and a steel-UHPC composite deck (See and studs. The fatigue resistance of the specimen was also investigated
Fig. 1(d)) on the downstream side, which provides a unique deck system in reference [40]. After the specimen was loaded under fatigue loads for
that can be used to compare the performance of conventional OSD and 10 million cycles, the steel-UHPC interface did not show any slippage.
steel-UHPC composite deck.
2.4. Field tests
2.2. Fatigue damages in conventional OSD
To evaluate the performance of steel-UHPC bridge deck, the strains
The Junshan Yangtze River Bridge was opened to traffic in 2001. In at fatigue-prone details were measured using strain gauges under
regular bridge inspection, fatigue cracks were observed from the rib-to- random traffic flow. The deployment of strain gauges was determined
deck welded joints, rib-to-floor-beam welded joints, floor beam cutout, through a preliminary finite element analysis of the stress andstrain
and rib splice welded joints. Fig. 2 depicts five representative types of distributions at the fatigue-prone zones of bridge decks. The finite
cracks, designated as Crack 1 to Crack 5. Crack 1 was located in the rib- element model is introduced in Section 5. The strain gauges were
to-deck welded joint, with four sub-types of cracks designated as Crack deployed at the thinnest section I-I (Fig. 1(b)). From the two wind
1–1 to Crack 1–4 [34]. Crack 2 and Crack 3 were located in the rib-to- fairing (Fig. 1(c), red dashed box) sides to the center along the bridge
floor beam welded joint, where the cracks initiate from the toe and/or width (i.e., transverse direction), the ribs are continuously designated as
root of the welded joint at its lower end and propagates continuously in rib 1 to rib 19, as shown in Fig. 1(c). The strain gauges were installed on
the floor beam and rib wall. Crack 4 and Crack 5 were located in the ribs 7, 8, 9, 11, 12, 14, 15, 18, and 19 marked by red dashed circles in
floor beam cutout and the rib splice weld. Many cracks penetrated Fig. 1(c). These ribs were under the heavy lanes. Fig. 6 shows the layout
through the plate thickness and were found from the top flange, as and designation of the strain gauges. T-01 ~ T08 were installed on the
shown in Fig. 3. top plate to measure the strains of rib-to-deck welded joint at the deck
Once generated, the fatigue cracks can grow and cause many issues, side. U-09 ~ U-16 were installed on the U-ribs wall side to measure the
such as the degradation of deck stiffness, pavement cracking, and strains of the rib-to-deck welded joint at the rib wall side. U21 ~ U24
corrosion of girders, which can threaten bridge safety [38]. The main were installed on the U-ribs to measure the strains of the rib-to-floor
defects of deck pavements are shown in Fig. 4. beam weld at the wall side. F-25 ~ F-32 were installed on the floor
beam to measure the strains of the rib-to-floor beam weld and floor
2.3. Steel-UHPC composite deck beam cutout at the floor beam side. The strain gauges were 6 mm away
from the toe or free edge to avoid stress concentration effects [3].
To improve the fatigue performance, the OSD deck was retrofitted by For comparison, the upstream half of the bridge, where the con­
applying a UHPC layer on top of the steel deck to form a steel-UHPC ventional OSD was kept, was instrumented with the same strain gauges
composite section, as shown in Fig. 5. Through-the-thickness cracks with an identical layout. The designation of the strain gauges was the
were found from the top flange of the OSD. The cracks would cause same as that on the steel-UHPC composite deck. Fig. 7 shows the
reflective cracks in the UHPC if the UHPC was in direct contact with the representative deployment of the strain gauges for different fatigue-
cracked flange. To avoid reflective cracks, steel strips measuring 80 mm prone details. The adopted strain gauges were based on electrical
in width and 8 mm in thickness were used to isolate the UHPC layer from resistance, which had a manufacturer-specified measurement accuracy
the steel deck. The effect of the steel plates was investigated and re­ of 1 µε. The strains of the fatigue-prone details were continuously
ported in reference [40]. The yielding strength of the steel plate was 345 monitored using a data acquisition system with a sampling frequency of
MPa. Shear studs were welded on the steel deck plate, which were 60 Hz for a duration of 7 days.
spaced at 200 mm along the bridge (longitudinal direction) and 150 mm
along the bridge width (transverse direction). Each stud measured 35 2.5. Traffic flow information
mm in length and 13 mm in diameter, except the studs welded on the
steel strips, which measured 27 mm in length so that the top surfaces of The investigated bridge was instrumented with a weigh-in-motion
all the studs were flushed. The thickness of the UHPC layer was 55 mm, (WIM) system to monitor the number and weight of vehicles, as
which was reinforced using a mesh of steel bars made using HRB400 shown in Fig. 1(b). During the field test, the axle loads of trucks were
steel with a nominal yielding strength of 400 MPa [41]. Asphalt concrete measured by the WIM system. Fig. 8 shows the statistical results of the

3
S. Qin et al. Engineering Structures 254 (2022) 113875

(a)

48 204 460 204 48

Cable 1~18 Cable 1~18


163.5
WIM

29+17×12+19=252 19+17×12+14+17×12+19=460
ᴵ 29+17×12+19=252

Note: WIM (Weigh-in-motion system)


(b)

Upstream Conventional OSD Steel-UHPC composite deck Downstream


1750 3250 3750 3750 3750 3500 3750 3750 3750 3250 1750
Lane 1 Lane 2 Lane 3 Lane 3' Lane 2' Lane 1'

2640
3000

Wind fairing
4270 30260 4270
38800

(c)

Asphalt concrete layer


UHPC layer
Asphalt concrete 300 Steel plate
layer
75

55
12~16

12~16

Steel panel Rib Steel panel Rib


260

180
Floor beam Floor beam

600 600
Conventional OSD Steel-UHPC composite deck
(d)
Fig. 1. Depiction of the Junshan Yangtze River Bridge: (a) a photo, (b) an elevation view (unit: m), (c) a cross section (unit: mm), and (d) the conventional OSD and
steel-UHPC composite deck (unit: mm). In (c), Lane 1 and Lane 1′ are heavy lanes, Lane 2 and Lane 2′ are normal lanes. Lane 3 and Lane 3′ are passing lanes.

4
S. Qin et al. Engineering Structures 254 (2022) 113875

results of stresses at the rib-to-deck welded joints of the OSD and steel-
UHPC deck.
Crack1
The deck side and the rib wall side of the rib-to-deck welded joints
1-1 1-2 are subjected to alternating tensile and compressive stresses, depending
Crack2 on the transverse location of the wheel loads [3]. Tensile stresses
Crack3
Crack5 Crack4 dominate the deck side of the rib-to-deck welded joints, while
1-4 1-3 compressive stresses dominate the rib wall side. The compressive and
tensile stresses of steel-UHPC composite deck are significantly lower
than those of the OSD. At the deck side of the rib-to-deck welded joints,
the maximum compressive stress is reduced by 73% (from 16.1 MPa to
Fig. 2. Depiction of the five main categories of fatigue cracks at fatigue-prone 4.4 MPa), and the maximum tensile stress is reduced by 29% (from 16.1
details of the OSD. MPa to 11.4 MPa). At the rib wall side of the rib-to-deck welded joints,
the maximum tensile stress is reduced by 33% (from 15.9 MPa to 10.8
MPa), and the maximum compressive stress is reduced by 46% (from
31.7 MPa to 17.0 MPa). The reduction of stresses can be attributed to the
Crack 1 enhancement of the stiffness of the bridge deck due to the use of UHPC.

3.2. Rib-to-floor-beam welded joints

Due to the different stiffness of the rib and floor beam, the de­
Crack 2 Crack 3 Crack 4 formations of the ribs and floor beam under vehicle loads are different
[44]. Typically, the ribs can rotate regarding to the floor beam and
produce a large stress at rib-to-floor-beam welded joints. Usually, the
Crack 5 Crack 6 fatigue cracks initiate from the lower end of a welded joint and propa­
gate upwards along the welded joint. Fig. 10 plots the stresses measured
from the OSD and the steel-UHPC composite deck. At the floor beam
side, the tensile stresses dominate under the random traffic loads. At the
Fig. 3. Penetrating cracks on steel panel. rib wall side, the welded joints are subjected to alternating tensile and
compressive stresses, and, overall, the maximum tensile stresses are
traffic flow on the upstream and downstream, respectively. Compared slightly larger than the maximum compressive stresses. The stresses in
with the conventional OSD, the deck system with the steel-UHPC com­ the steel-UHPC composite deck are lower than the stresses in the OSD
posite section was subjected to heavier traffic in terms of the number without UHPC. At the floor beam side, the maximum tensile stress is
and weight of vehicles. reduced by 77% (from 34.0 MPa to 7.7 MPa). At the rib wall side, the
maximum tensile stress is reduced by 43% (from 27.7 MPa to 15.8 MPa),
and the maximum compressive stress is reduced by 66% (from 13.3 MPa
3. Results and analysis of stress monitoring
to 4.6 MPa). According to reference [45], the reduction of the stresses
indicates that the U-ribs of the steel-UHPC composite deck have less
The fatigue performance of OSD is mainly reflected by the maximum
torsional deformation with respect to the floor beam, and, thus, the
stress range and the effective stress range of each fatigue-prone detail.
steel-UHPC composite deck is expected to have a longer fatigue life at
The stress ranges and the number of stress cycles were determined based
the rib-to-floor-beam welded joints. Again, the reduction of stresses can
on the measurement results from the deployed strain gauges. This sec­
be attributed to the enhancement of the stiffness of the bridge deck due
tion presents the stress monitoring results for rib-to-deck welded joints,
to the use of UHPC.
rib-to-floor-beam welded joints, floor beam cutout, and rib splice wel­
ded joints.
3.3. Floor beam cutout

3.1. Rib-to-deck welded joints The curved cutout in the floor beam can generally reduce the
constraint of the floor beam for the U-ribs and the stress caused by the
The rib-to-deck welded joints are subjected to gaps and weld residual out-of-plane deformation of the floor beam [3]. However, the cutout can
stress generated in welding. When vehicle loads are applied to the bridge also reduce the stiffness of the floor beam, thus increasing the stresses
deck, stresses will be produced at the welded joints, which tend to cause around the cutout. Under in-service vehicle loads, fatigue cracks tend to
fatigue cracks from the weld toe or weld root, and the crack can prop­ occur at the edge of cutout. Fig. 11 shows the stresses measured from the
agate to the deck and the rib wall [44]. Fig. 9 shows representative floor beam cutout details of the OSD and steel-UHPC composite deck.

Fig. 4. Pavement damages on the conventional OSD: (a) reticular crack, and (b) upheaval.

5
S. Qin et al. Engineering Structures 254 (2022) 113875

Transverse
150 Longitudinal
Asphalt concrete layer
80 UHPC layer
Longitudinal Transvers
100 30 steel bar Ф12
120 55 Longitudinal
100 steel bar Ф10
Steel plate
Shear connector Shear connector
50 Steel panel
Ф13×27 Ф13×35

37.5

Before welding After welding


22

Shear connector
35 27 32
Ф13×35

13
22 Rib

Shear connector 27
19 24
Ф13×27

13
Fig. 5. Illustration of the main components of the steel-UHPC composite deck (unit: mm).

Rib T-01 Rib T-02 Rib T-03 Rib T-04 Rib


U-10 U-11 U-12
#7 U-09 #8 F-26 #9 #11 #12
U-21 U-22
F-25 F-27 U-17 F-28 U-18
(a) (b)

T-05 T-06 T-07 Rib T-08 Rib


U-13 Rib U-14 Rib U-16
F-30 U-15
#14 #15 U-24 #18 #19
U-23
F-29 F-31 U-19 F-32 U-20
(c) (d)

Fig. 6. Layout and designation of strain gauges in section I-I: (a) ribs 7 to 9, (b) ribs 11 and 12, (c) ribs 14 and 15, and (d) ribs 18 and 19.

The floor beam cutout is mainly subjected to compressive stresses. The deck are lower than those in the OSD. The maximum tensile stress is
stresses in the steel-UHPC composite deck are lower than those in the reduced by 29% (from 46.4 MPa to 33.0 MPa), and the maximum
OSD. The maximum compressive stress is reduced by 13% (from 73.4 compressive stress is reduced by 19% (from 14.9 MPa to 12.1 MPa).
MPa to 61.8 MPa), indicating that the steel-UHPC composite deck can
reduce the stress range at the floor beam cutout. 4. Evaluation of fatigue life

Based on the stress monitoring data, the effective stress range and the
3.4. Rib splice welded joints
residual fatigue life of each fatigue-prone detail are analyzed to evaluate
the fatigue performance of the steel-UHPC composite deck in this
Under in-service vehicle loads, the U-ribs are subjected to positive
section.
moment, meaning that the bottom of the U-ribs are subjected to tension
[3]. Therefore, fatigue cracks can be generated at the rib splice welded
joints. Fig. 12 shows the stresses measured from at the rib splice welded 4.1. Effective stress range
joints of the OSD and steel-UHPC composite deck. The rib splice welded
joints are subjected to alternating tensile and compressive stresses, and Based on the results of stresses measured from the fatigue-prone
the tensile stress dominates. The stresses in the steel-UHPC composite details, the rainflow counting method can be used to determine the

6
S. Qin et al. Engineering Structures 254 (2022) 113875

Fig. 7. Deployment of strain gauges at fatigue-prone details: (a) floor beam cutout, (b) rib-to-deck weld, and (c) rib splice weld.

Summary: 8512 vehicles Summary: 9945 vehicles


Axle weight
less than 2 tons

es
Axle weight

cl
hi
less than 2 tons

ve
%
96
3102 vehicles

24
23
2897 vehicles 36% 5686 vehicles
34% 57%
1863 vehicles
Axle weight Axle weight 19%
greater than 4 tons 2513 vehicles greater than 4 tons
Axle weight
30%
2~4tons

Axle weight
2~4tons

(a) (b)

Axle weight
Axle weight Axle weight Axle weight
less than 2 tons
less than 2 tons greater than 4 tons
es
greater than 4 tons
cl
hi

1338 vehicles
ve
%
86

25%
19
13

2448 vehicles 3387 vehicles


45% 48%
1653 vehicles 2341 vehicles
30% 33%
Axle weight
Axle weight 2~4tons
2~4tons
Summary: 5439 vehicles Summary: 6426 vehicles

(c) (d)
Fig. 8. Statistical results of the traffic flow on the bridge: (a) the normal lane on the OSD, (b) the normal lane on the steel-UHPC composite deck, (c) the heavy lane
on the OSD, and (d) the heavy lane on the steel-UHPC composite deck.

stress spectrum [46]. Then, the Miner’s linear cumulative damage rule composite deck, the effective stress range in each fatigue-prone detail
[47] is applied to calculate the effective stress range: is significantly reduced. For example, the effective stress range at floor
∑ m m1 beam side of the rib-to-floor beam welded joints is reduced by 66%
ni s
Sr,eff = ( ∑ i ) (1) (from 13.3 MPa to 4.5 MPa). Again, the reduction of effective stress
ni range can be attributed to the enhancement of bending stiffness of the
where Sr,eff is the effective stress range; ni is the number of loading bridge deck due to the addition of the UHPC layer. The deflection of the
cycles at the stress range of Si, and m is the inverse of the slope of S-N deck and the rotation of U-ribs regarding to floor beam are reduced.
curve (m = 3, according to AASHTO LRFD [47]). Therefore, the effective stress ranges of the fatigue-prone details of the
Fig. 13 shows the results of the effective stress range of each fatigue- steel-UHPC composite deck are lower than those of the OSD.
prone detail of the steel-UHPC composite deck. With the steel-UHPC

7
S. Qin et al. Engineering Structures 254 (2022) 113875

20 20
D-02 D-02
#8 #8

10 10

Stress (MPa)

Stress (MPa)
0 0

-10 -10

-20 -20
0 10 20 30 40 50 60 0 10 20 30 40 50 60
t (min) t (min)
(a) (b)
20 20
10 10
Stress (MPa)

Stress (MPa)
0 0
-10 -10
-20 -20
U-10 U-10
-30 #8 -30 #8

-40 -40
0 10 20 30 40 50 60 0 10 20 30 40 50 60
t (min) t (min)
(c) (d)
Fig. 9. Results of stresses measured from the rib-to-deck welded joints: (a) deck side of the OSD, (b) deck side of the steel-UHPC composite deck, (c) rib wall side of
the OSD, and (d) rib wall side of the steel-UHPC composite deck.

4.2. Residual fatigue life Fig. 15(a) shows a part of the global model. In the global model, the
boundary conditions of the model were consistent with the real bridge.
According to AASHTO LRFD, the residual fatigue life can be calcu­ The box bridge girder was modelled using plate elements Shell63, which
lated: has both bending and membrane capabilities and permits both in-plane
and normal loads. Each node of Shell63 element has six degrees of
A
Y= m
∑ (2) freedom. The UHPC deck was modelled using solid elements Solid45,
Sr,eff ⋅365⋅ ni
which has eight nodes, and each node has three degrees of freedom.
∑ Each element has six faces, and each face has four lines. Stress stiffening
where Y is the residual fatigue life in years; ni is the measured average
and large deflection can be considered in both Shell63 and Solid45 el­
number of daily stress cycles; and A is a constant depending on the detail
ements. The real bridge has asphalt concrete pavement, which is ignored
category specified by AASHTO LRFD [47]. The fatigue life of the fatigue-
in the global model, because the temperature of pavement is high due to
prone detail is “infinite”, when the maximum stress range is lower than
the black color and the direct exposure to sunlight, and the high tem­
the constant-amplitude fatigue limit (CAFL) [48].
perature can highly reduce the stiffness of pavement. This is consistent
Fig. 14 compares the maximum stress range with the CAFL for each
with references [49,50]. With the global model, the internal forces and
fatigue-prone detail of the steel-UHPC composite deck. Except the rib
moments of the steel-UHPC composite deck can be determined and used
splice welded joint, the other investigated fatigue-prone details have
as the boundary conditions of the sub-model. The position of the sub-
“infinite” fatigue life because their maximum stress ranges are lower
model is marked in the global model, which is a region around rib 12
than the corresponding CAFL values. According to Eq. (2), the fatigue
in the wheel path of the heavy lane. This research focuses on the stress
life of rib splice welded joint is calculated as 100 years.
distributions in the sub-model, so solid elements Solid185 and refined
mesh sizes are adopted. Compared with Solid45 elements, Solid185 el­
5. Finite element analysis
ements have different element formulations to avoid volumetric locking.
A mesh size convergence analysis was performed to determine the
5.1. Finite element model
appropriate mesh size. In the sub-model, the mesh size of the fatigue-
prone details was 2 mm, as shown in Fig. 15(b).
To analyze the detailed stress distributions around the fatigue-prone
Since no delamination was observed from the bridge at the interfaces
details of bridge decks, a global model of the bridge and a local model of
of the steel-UHPC deck system, the adjacent layers of the steel deck,
steel-UHPC composite deck were established using software ANSYS
UHPC, and pavement were bonded by coupling the nodes. In the finite
[27], as shown in Fig. 15. The local model is part of the global model, so
element model, it was assumed that all materials had linear elastic
it can be designated as sub-model. The global model is used to determine
properties [33,40]. Inelastic material properties were not considered
the boundary conditions for the sub-model, and the sub-model is used to
since the bridge was not damaged under the normal operation loads. The
analyze the detailed stress distributions because the weld details can be
adopted material properties were consistent with the real bridge. The
simulated with refined mesh sizes.
density, modulus of elasticity, and Poisson’s ratio of the steel were 7850

8
S. Qin et al. Engineering Structures 254 (2022) 113875

40 40
#14 #14
H-30 H-30
30 30

Stress (MPa)

Stress (MPa)
20 20

10 10

0 0

-10 -10
0 10 20 30 40 50 60 0 10 20 30 40 50 60
t (min) t (min)
(a) (b)
40 40
#9 #9
30 U-22 30 U-22
Stress (MPa)

Stress (MPa)
20 20
10 10
0 0
-10 -10
-20 -20
0 10 20 30 40 50 60 0 10 20 30 40 50 60
t (min) t (min)
(c) (d)
Fig. 10. Results of stresses measured from the rib-to-floor-beam welded joints: (a) the floor beam side of OSD, (b) the floor beam side of steel-UHPC composite deck,
(c) the rib wall side of OSD, and (d) the rib wall side of steel-UHPC composite deck.

60 60
40 #14 40 #14
H-31 H-31
20 20
Stress (MPa)

Stress (MPa)

0 0
-20 -20
-40 -40
-60 -60
-80 -80
0 10 20 30 40 50 60 0 10 20 30 40 50 60
t (min) t (min)

(a) (b)
Fig. 11. Results of stresses measured from the floor beam cutout: (a) OSD, and (b) steel-UHPC composite deck.

kg/m3, 206 GPa, and 0.3, respectively. The density, modulus of elas­ lanes during travelling, and the error can be 30 cm at the left and right
ticity, and Poisson’s ratio the UHPC were 2400 kg/m3, 42.6 GPa, and sides of the design lanes. To reasonably consider the transverse distri­
0.2, respectively. The density, modulus of elasticity, and Poisson’s ratio bution of vehicle wheel loads, five loading cases were considered, as
of the asphalt concrete were 2350 kg/m3, 3 GPa, and 0.2, respectively. shown in Fig. 17 (a). In the five loading cases, except one loading case
that corresponds to loading at the center line of the traffic lane (Case 3),
the other four loading cases have errors: Case 1 (30 cm, left), Case 2 (15
5.2. Loading cases
cm, left), Case 3 (0 cm), Case 4 (15 cm, right), and Case 5 (30 cm, right).
Movement of wheel loads on the deck is simulated by the step by step
Because the longitudinal influence line of OSD is short, biaxial loads
loading cases in the longitudinal direction of the bridge. The center of
is utilized in this study. The truck load is shown in Fig. 16. The front and
front and rear axle start at right above floor beam 2 and move in the
rear axles are loaded with two wheels, and the wheel load action area is
right above floor beam 4. There are a total of 155 loading cases
600 mm × 200 mm. The axle weight was 140 kN. The wheel axle spacing
considered for each transverse loading case, along with the space in­
is 1.4 m in the longitudinal direction and 1.85 m in the transverse
terval of 0.2 m.
direction.
According to BS5400 [51], vehicles do not exactly follow the design

9
S. Qin et al. Engineering Structures 254 (2022) 113875

60 60
50 #9 50 #9
40 U-17 40 U-17

Stress (MPa)

Stress (MPa)
30 30
20 20
10 10
0 0
-10 -10
-20 -20
0 10 20 30 40 50 60 0 10 20 30 40 50 60
t (min) t (min)
(a) (b)
Fig. 12. Results of stresses measured from the rib splice welded joints: (a) OSD, and (b) steel-UHPC composite deck.

23% higher than the measurement results, because the stiffness and load
28 distribution effect of pavement were not considered in the finite element
OSD
model, as stated in section 5.1. The errors of the other fatigue details
24 Steel-UHPC
were less than 15%, indicating that the finite element model provided
19% reasonable predictions of stresses. The other main error sources included
20
23% the simplification of the support conditions and idealization of materials
Stress (MPa)

16 in the finite element model.


50% 66% Similar to the finite element analysis of the steel-UHPC composite
12 37% 19% deck, a finite element model was established using the same method for
the original OSD. The load distribution effect of the asphalt pavement
8
was considered because the load distribution effect played an important
4 role for OSDs, as indicated in references [8,38]. The presence of an
asphalt layer distributed the loads from vehicle wheels [8,38], in the
0 determination of the stresses at fatigue-prone details such as the welded
Deck Rib Floor beam Rib Cutout Rib splice
joints of OSD.
(a) (b) (c)

Fig. 13. Results of the effective stress ranges at: (a) the rib-to-deck welded 5.4. Simulation results and discussions
joints, (b) rib-to-floor-beam welded joints, and (c) cutout and rib splice wel­
ded joints. 5.4.1. Deflections of bridge decks
Deflection is an important index that reflects the flexural stiffness of
5.3. Validation of the finite element model bridge decks. Fig. 18 shows the maximum deflections of OSD and steel-
UHPC composite deck under the same design vehicle loads in the
The finite element analysis results of stresses were compared with different loading cases. Compared with the OSD, the maximum deflec­
the measurement results to validate the finite element model. The tion of the steel-UHPC composite deck is reduced from 1.3 mm to 0.9
measurement results were reported in reference [52]. The finite element mm (by 31%). The reduction of the deflection can be attributed to the
analysis was performed using the same loading cases as those of the increase of the flexural stiffness due to the use of UHPC in the steel-
measurement. Table 2 shows the finite element analysis results and the UHPC composite deck.
measurement results. Overall, the errors were in a reasonable range. The
predicted stress ranges at the deck side of rib-to-deck welded joints were

180 Steel-UHPC Sr,eff= 6.6 MPa


CAFL Fatigue life ∞
164.1
160 68.4 68.4 68.4 68.4 Sr,eff= 7.0 MPa
61.8 Fatigue life ∞
60 55.6
Stress (MPa)

47.8 43.8 Sr,eff= 8.8 MPa


40 Fatigue life ∞
26.9
20 18.0
Sr,eff= 8.8 MPa
10.9
Fatigue life ∞ Sr,eff= 16.5 MPa
0 Fatigue life ∞
Cutout Rib splice Deck Rib Floor beam Rib Sr,eff= 13.5 MPa
Fatigue life 100 years
(Rib-to-deck) (Rib-to-floor beam)
(a) (b)
Fig. 14. Results of residual fatigue life of the fatigue-prone details of the steel-UHPC deck: (a) comparison of the maximum stress range and CAFL, and (b) dis­
tribution of residual fatigue life.

10
S. Qin et al. Engineering Structures 254 (2022) 113875

of stress extraction points have been shown in Fig. 15(b) and Fig. 17(a).
The maximum stress ranges of the six different fatigue-prone details of
rib 12 can be extracted for both the conventional OSD and the steel-
Y
UHPC composite deck, as shown in Fig. 20. The dash lines are used to
Sub-model area group the different fatigue-prone details according to their positions in
Z X
the decks.
Rib #12 Floor beam 5 Due to the use of the UHPC deck, the maximum stress ranges of the
fatigue-prone details are reduced. The reduction percentage of the stress
Floor beam 3
(a) range of the rib 12 is 39% for the deck side of the rib-to-deck welded
Floor beam 1 joints and 54% for the rib wall side of the rib-to-deck welded joints. The
stress range reduction of the rib 12 is 46% for the floor beam side of the
rib-to-floor beam welded joints and 39% for the rib wall side of the rib-
Floor beam
to-floor beam welded joints. The stress range reduction of the rib 12 is
Rib 12% for the floor beam cutout and 31% for the rib splice welded joints.
The results are consistent with the results of the effective stress range in
6mm
Fig. 13, except for the rib-to-floor beam welded joints. The crack is close
Rib
to the strain gauges on UHPC side, which may result in higher stresses at
this location. Therefore, the stress range reductions of finite element
model are larger than the measured results at this location. It should be
pointed out that the steel-UHPC composite deck has a limited
6mm
improvement for the fatigue performance of the floor beam cutout. The
6mm
stress range reduction of floor beam cutout is less than 20%.

5.4.4. Influence of overload


6mm

Overloaded vehicles were identified from the information collected


Rib
by the weight-in-motion system. To consider the influence of overload,
(b)
the stress ranges of the fatigue-prone details at steel-UHPC composite
Fig. 15. Depiction of the established finite element models: (a) the global deck are analyzed, under 1.5 times the design load, as shown in Fig. 21.
model, and (b) the meshed sub-model. Under 1.5 times the design load, the maximum stress range of rib splice
welded joints is 66.6 MPa. The value exceeded their corresponding CAFL
value by 42%. However, the stress ranges of the other fatigue-prone
140kN 140kN details are less than the corresponding CAFL values, thus satisfying
1400 the requirement of “infinite” life. In summary, the stress range in the rib
splice weld is greater than CAFL under 1.5 times the design load, which
do not meet the design requirements. Under the action of cyclic over­
600

loaded trucks, there is a risk of fatigue cracking in this area. During


future bridge maintenance, special attention should be paid to the rib
200
1850

splice welded joints.

6. Conclusions

5800 1400 X
This paper investigates the fatigue performance of a steel-UHPC
composite deck of a long-span cable-stayed bridge under in-service
Z traffic loads. Strain distributions around fatigue-prone details were
measured in the field and used to evaluate the residual fatigue life. A
Fig. 16. Layout and magnitude of the truck load adopted in the finite element finite element model was established to analyze the stress distributions
model (unit: mm). of the steel-UHPC composite deck. Based on the above investigations,
the following conclusions are drawn:
5.4.2. Stress distributions in UHPC
Fig. 19 shows the maximum tensile stresses of the UHPC layer of the (1). Compared with the OSD, the steel-UHPC composite deck reduced
steel-UHPC composite deck under the design vehicle loads in the the maximum stress ranges and the effective stress ranges for the
different loading cases. The maximum tensile stress of the transverse fatigue-prone details of bridge deck, thus increasing the fatigue
direction is lower than that of the longitudinal direction under the life. According to the evaluation of the residual fatigue life using
investigated loading cases. Along the longitudinal direction of the field monitoring data of stresses, the residual fatigue life of the rib
bridge, the maximum tensile stress is 7.3 MPa in Case 1, which is lower splice welded joint is 100 years, and the other fatigue-prone de­
than the nominal tensile stress tolerance value (25.4 MPa) [53], indi­ tails satisfy the requirements of infinite life. Special attention
cating that the crack resistance of the steel-UHPC composite bridge deck should be paid to the rib splice welded joints in the bridge in­
satisfies the service requirements, under the design load. spection and maintenance.
(2). Due to the use of the UHPC, the maximum deflection of the steel
5.4.3. Stress distributions in steel plates panel under the design load was reduced by 31% (from 1.3 mm to
The stress ranges in the fatigue-prone details are investigated using 0.9 mm). The reduction of the deflection indicated enhancement
the finite element model. The analysis focuses on the right-hand wheel of section stiffness. The stiffness enhancement was attributed to
load action area, which is the left of rib 12. The location and the number

11
S. Qin et al. Engineering Structures 254 (2022) 113875

Right-hand
wheel load
Case5

Case4 12-1
12-2 Rib #12
Case3
12-3 12-5
Case2 12-4
12-6
Location of stress extraction point
Case1

7 8 9 10 11 12

11 12 11 12 11 12 11 12 11 12
Case1 Case2 Case3 Case4 Case5

(a)

140kN 140kN 140kN 140kN


Moving direction

Sub-model area

Floor beam 1 Floor beam 3 Floor beam 5

(b)
Fig. 17. Illustration of loading cases: (a) transverse loading cases considering inaccurate wheel loading positions, and (b) longitudinal loading cases.

Table 2 2.0
Comparison of stress ranges from the finite element model and in-situ sensors.
OSD
Fatigue-prone details Stress range (MPa) Error
1.6 Steel-UHPC
Finite element Measurement
Deflection (mm)

analysis 1.3 1.3 1.3 1.3


1.2
Deck side of rib-to-deck weld 11.2 8.6 23% 1.2
Rib side of rib-to-deck weld 16.2 16.2 0%
Floor beam side of rib-to-floor 24.3 25.2 4% 0.9 0.9 0.9 0.9 0.9
beam weld 0.8
Rib side of rib-to-floor beam weld 26.1 26.7 2%
Floor beam cutout 67.9 59.5 12%
Rib splice 16.4 18.8 13%
0.4

the presence of the UHPC that was integrated with the OSD and
0.0
formed a steel-UHPC composite section. Case 1 Case 2 Case 3 Case 4 Case 5
(3). Under the design load, the maximum tensile stress of the UHPC
layer is 7.3 MPa, which is less than its nominal flexural-tensile Loading case
stress tolerance (25.4 MPa), satisfying the service requirement
Fig. 18. The maximum deflections of the OSD and steel-UHPC composite deck
of crack resistance. Under 1.5 times the design load considering in different cases.
overloads, the maximum stresses of the rib splice welded joints
exceed its CAFL values, thus failing to satisfy the requirement of
infinite life. composite deck is less than 20%. The reduction of stresses by the
(4). The steel-UHPC composite deck has a limited improvement for use of UHPC layer for the deck side of rib-to-deck welded joints is
the fatigue performance of the floor beam cutout. Compared with lower than 40%, which is lower than the results (70%-90%) re­
the original OSD, the reduction of the effective stress range ported in previous research.
reduction of the floor beam cutout due to the steel-UHPC

12
S. Qin et al. Engineering Structures 254 (2022) 113875

10.0 CRediT authorship contribution statement


Case1
Case2 Shiqiang Qin: Conceptualization, Investigation, Methodology,

7.3
7.5 Case3 Writing – original draft. Jiabin Zhang: Investigation, Visualization,
Writing – review & editing. Chunlei Huang: Investigation, Writing –
Stress (MPa)

Case4
review & editing. Liqiang Gao: Investigation, Writing – review &

5.2
5.0
Case5

4.8
4.8
5.0 editing. Yi Bao: Conceptualization, Methodology, Writing – review &

3.7
editing.

3.3
2.9
2.7
2.4
2.5
Declaration of Competing Interest

0.0 The authors declare that they have no known competing financial
Longitudinal Transverse interests or personal relationships that could have appeared to influence
Direction of tensile stress the work reported in this paper.
Fig. 19. Simulation results of the maximum tensile stresses of the UHPC layer
in different cases. Acknowledgements

This research was financially supported by the Natural Science


90 Foundation of China (grant No. 51608408), and the Open Projects
OSD 12% Foundation of State Key Laboratory for Health and Safety of Bridge
75 Steel-UHPC Structures of China (grant No. BHSKL20-08-GF).
31%
Stress range (MPa)

60 References
39%
45 46% [1] Qian ZH, Abruzzese D. Fatigue failure of welded connections at orthotropic
bridges. Frattura ed integrità strutturale 2009;3(9):105–12.
54% [2] Xiao Z, Yamada K, Ya S, Zhao X. Stress analyses and fatigue evaluation of rib-to-
30 39% deck joints in steel orthotropic decks. Int J Fatigue 2008;30(8):1387–97.
[3] Connor, R.J., 2012. Manual for design, construction, and maintenance of
15 orthotropic steel deck bridges (No. FHWA-IF-12-027). United States. Federal
Highway Administration.
[4] Zhang Q, Liu Y, Bao Yi, Jia D, Bu Y, Li Q. Fatigue performance of orthotropic steel-
0 concrete composite deck with large-size longitudinal U-shaped ribs. Eng Struct
Deck Rib Floor beam Rib Cutout Rib splice 2017;150:864–74.
(a) (b) (c) [5] Liu Y, Zhang Q, Meng W, Bao Yi, Bu Y. Transverse fatigue behaviour of steel-UHPC
composite deck with large-size U-ribs. Eng Struct 2019;180:388–99.
Fig. 20. Simulation results of the maximum stress range of rib 12: (a) the rib- [6] Wolchuk R. Lessons from weld cracks in orthotropic decks on three European
bridges. J Struct Eng 1990;116(1):75–84.
to-deck welded joints, (b) rib-to-floor-beam welded joints, and (c) cutout and
[7] Haight R, Chang S, Kushmock R. In: Orthotropic deck rehabilitation at the Throgs
rib splice welded joints. Neck Bridge. Metropolis and Beyond; 2005. p. 1–10.
[8] Wang L, Su X, Ma Y, Deng M, Zhang J, Cai CS. Strengthening of steel decks for
180 Steel-UHPC cable-stayed bridge using ultra-high performance concrete: A case study. Adv
Struct Eng 2020;23(16):3373–84.
164.1 CAFL [9] Cao B-y, Ding Y-L, Song Y-S, Zhong W. Fatigue life evaluation for deck-rib welding
details of orthotropic steel deck integrating mean stress effects. J Bridge Eng 2019;
160 104.5 24(2):04018114. https://doi.org/10.1061/(ASCE)BE.1943-5592.0001344.
Stress range (MPa)

[10] Hassel, H.L., Hartman, A.S., Bennett, C.R., Matamoros, A.B. and Rolfe, S.T., 2010.
Distortion-induced fatigue in steel bridges: Causes, parameters, and fixes. In
80 68.4 68.4 68.4 68.4 Structures Congress 2010 (pp. 471-483).
66.6 [11] Chen Z, Li C, Ke L, Guo L, Song G. Fatigue cracks cutting repair and optimization of
60 cope holes in orthotropic steel decks. China J Highway Transport 2020:1–14 [In
47.8 46.1 Chinese].
40 35.0 [12] Ke L, Li C, He J, Lu Y, Liu Y. Fatigue evaluation and CFRP strengthening of
diaphragm cutouts in orthotropic steel decks. Steel Compos Struct 2021;39(4):
22.2 24.8 453–69.
20 [13] Liu J, Guo T, Feng D, Liu Z. Fatigue performance of rib-to-deck joints strengthened
with FRP angles. J Bridge Eng 2018;23(9):04018060. https://doi.org/10.1061/
0 (ASCE)BE.1943-5592.0001286.
Cutout Rib splice Deck Rib Floor beam Rib [14] Guo T, Liu J, Deng Y, Zhang Z. Fatigue performance of orthotropic steel decks with
FRP angles: field measurement and numerical analysis. J Perform Constr Facil
(Rib-to-deck) (Rib-to-floor beam) 2019;33(4):04019042. https://doi.org/10.1061/(ASCE)CF.1943-5509.0001308.
[15] Yao Bo, Cheng G, Wang X, Cheng C. Characterization of the stiffness of asphalt
Fig. 21. Simulation results of the maximum stress range of the fatigue-prone surfacing materials on orthotropic steel bridge decks using dynamic modulus test
details under 1.5 times the design load. The numbers on the columns repre­ and flexural beam test. Constr Build Mater 2013;44:200–6.
sent the stress ranges (unit: MPa). [16] Shao X, Qu W, Cao J, Yao Y. Static and fatigue properties of the steel-UHPC
lightweight composite bridge deck with large U ribs. J Constr Steel Res 2018;148:
491–507.
Based on this research, future research opportunities are identified. [17] Walter R, Olesen JF, Stang H, Vejrum T. Analysis of an orthotropic deck stiffened
with a cement-based overlay. J Bridge Eng 2007;12(3):350–63.
This research showed that the fatigue behaviors of steel-UHPC com­
[18] Yuan Y, Wu C, Jiang Xu. Experimental study on the fatigue behavior of the
posite decks under in-service traffic loads were different from the be­ orthotropic steel deck rehabilitated by UHPC overlay. J Constr Steel Res 2019;157:
haviors under static loads in previous finite element analysis and 1–9.
[19] Liu Y, Zhang Q, Bao Yi, Bu Y. Fatigue behavior of orthotropic composite deck
laboratory tests. Further research is needed to establish a holistic un­
integrating steel and engineered cementitious composite. Eng Struct 2020;220:
derstanding of the effects of UHPC layer on the fatigue resistance of OSD 111017. https://doi.org/10.1016/j.engstruct.2020.111017.
of bridges. It is important to investigate the life-cycle cost of steel-UHPC [20] Guo P, Meng W, Xu M, Li VC, Bao Y. Predicting mechanical properties of high-
composite decks for bridges. performance fiber-reinforced cementitious composites by integrating
micromechanics and machine learning. Materials 2021;14(12):3143.

13
S. Qin et al. Engineering Structures 254 (2022) 113875

[21] Meng W, Khayat KH. Mechanical properties of ultra-high-performance concrete [37] Ding N, Shao X. Study on fatigue performance of light-weighted composite bridge
enhanced with graphite nanoplatelets and carbon nanofibers. Compos B Eng 2016; deck. China Civil Eng J 2015;48(1):74–81 [In Chinese].
107:113–22. [38] Zhu Z, Yuan T, Xiang Ze, Huang Y, Zhou YE, Shao X. Behavior and fatigue
[22] Du J, Meng W, Khayat KH, Bao Yi, Guo P, Lyu Z, et al. New development of ultra- performance of details in an orthotropic steel bridge with UHPC-deck plate
high-performance concrete (UHPC). Compos B Eng 2021;224:109220. https://doi. composite system under in-service traffic flows. J Bridge Eng 2018;23(3):
org/10.1016/j.compositesb:2021.109220. 04017142. https://doi.org/10.1061/(ASCE)BE.1943-5592.0001167.
[23] Liu Y, Zhang Q, Bao Yi, Bu Y. Static and fatigue push-out tests of short headed shear [39] Feng Z, Li C, He J, Ke Lu, Lei Z, Vasdravellis G. Static and fatigue test on
studs embedded in Engineered Cementitious Composites (ECC). Eng Struct 2019; lightweight UHPC-OSD composite bridge deck system subjected to hogging
182:29–38. moment. Eng Struct 2021;241:112459. https://doi.org/10.1016/j.
[24] Qi J, Cheng Z, Zhou K, Zhu Y, Wang J, Bao Yi. Experimental and theoretical engstruct.2021.112459.
investigations of UHPC-NC composite slabs subjected to punching shear-flexural [40] Wang Y, Shao X, Chen J, Cao J, Deng S. UHPC-based strengthening technique for
failure. J Build Eng 2021;44:102662. https://doi.org/10.1016/j. orthotropic steel decks with significant fatigue cracking issues. J Constr Steel Res
jobe.2021.102662. 2021;176:106393. https://doi.org/10.1016/j.jcsr.2020.106393.
[25] Abdelbaset H, Cheng B, Tian L, Li H-T, Zhang Q-H. Reduce hot spot stresses in [41] GB/T 1499.2-2018. Steel for the Reinforcement of Concrete—Part 2:Hot Rolled Ribbed
welded connections of orthotropic steel bridge decks by using UHPC layer: Bars . 2018, Standardization Administration, China. [In Chinese].
Experimental and numerical investigation. Eng Struct 2020;220:110988. https:// [42] Meng W, Valipour M, Khayat KH. Optimization and performance of cost-effective
doi.org/10.1016/j.engstruct.2020.110988. ultra-high performance concrete. Mater Struct 2017;50(1):1–16.
[26] Unterweger H, Novak F. Strengthening of orthotropic steel decks using UHPC- [43] Jung F, Mansperger T. The ortho-composite-slab of the Elbebridge Wittenberge. In:
concrete instead of asphalt layer for additional at least 50 years in service. In In: IABSE symposium report. International Association for Bridge and Structural
Eurosteel 2017: 8th European Conference on Steel and Composite Structures. 2017. Engineering; 2014. p. 1186–91.
[27] Zhu Z, Xiang Ze, Zhou YE. Fatigue behavior of orthotropic steel bridge stiffened [44] Sim H-B, Uang C-M. Stress analyses and parametric study on full-scale fatigue tests
with ultra-high performance concrete layer. J Constr Steel Res 2019;157:132–42. of rib-to-deck welded joints in steel orthotropic decks. J Bridge Eng 2012;17(5):
[28] Jiang Xu, Yuan Y, Wu C, Luo C. Fatigue Life Assessment of Orthotropic Steel Deck 765–73.
with UHPC Pavement. J Eng 2017;2017:1–10. [45] Wolchuk R. Steel orthotropic decks: developments in the 1990s. Transp Res Rec
[29] Zhang S, Shao X, Cao J, Cui J, Hu J, Deng Lu. Fatigue performance of a lightweight 1999;1688(1):30–7.
composite bridge deck with open ribs. J Bridge Eng 2016;21(7):04016039. https:// [46] Downing S, Socie D. Simple rainflow counting algorithms. Int J Fatigue 1982;4(1):
doi.org/10.1061/(ASCE)BE.1943-5592.0000905. 31–40.
[30] Connor RJ, Fisher JW. Report on Field Measurements and Assessment of the I-64 [47] Miner MA. Cumulative Damage in Fatigue. ASME. J Appl Mech 1945;12(3):
Kanawha River Bridge at Dunbar. West; 2001. A159–64.
[31] Cao J, Shao X, Zhang Z, Zhao H. Retrofit of an orthotropic steel deck with compact [48] American Association of State Highway and Transportation Officials. AASHTO
reinforced reactive powder concrete. Struct Infrastruct Eng 2016;12(3):411–29. LRFD bridge design specifications. 2017, Washington, D.C., United States.
[32] Su L, Wang S, Gao Y, Liu J, Shao X. In Situ experimental study on the behavior of [49] Battista RC, Pfeil MS, Carvalho EML. Fatigue life estimates for a slender orthotropic
UHPC composite orthotropic steel bridge deck. Materials 2020;13(1):253. steel deck. J Constr Steel Res 2008;64(1):134–43.
[33] Pei B, Li L, Shao X, Wang L, Zeng Y. Field measurement and practical design of a [50] Cui C, Zhang Q, Hao H, Li J, Bu Y. Influence of asphalt pavement conditions on
lightweight composite bridge deck. J Constr Steel Res 2018;147:564–74. fatigue damage of orthotropic steel decks: parametric analysis. J Bridge Eng 2018;
[34] Cui C, Bu Y, Bao Yi, Zhang Q, Ye Z. Strain energy-based fatigue life evaluation of 23(12):04018093. https://doi.org/10.1061/(ASCE)BE.1943-5592.0001313.
deck-to-rib welded joints in OSD considering combined effects of stochastic traffic [51] BS5400. British Standard Institute, Steel, Concrete and Composite Bridges in Part 10:
load and welded residual stress. J Bridge Eng 2018;23(2):04017127. https://doi. Code of Practice for Fatigue, 2005. London (UK).
org/10.1061/(ASCE)BE.1943-5592.0001181. [52] Ling L, Tang L, Shua J, Wang Y, Li Y, Wang Y. Experimental study of steel-UHPC
[35] Deng Lu, Zou S, Wang W, Kong X. Fatigue performance evaluation for composite composite deck on moving vehicle loading. Highway 2020;66(1):100–7 [In
OSD using UHPC under dynamic vehicle loading. Eng Struct 2021;232:111831. Chinese].
https://doi.org/10.1016/j.engstruct.2020.111831. [53] GDJTG/T A01-2015. Technical Specification for Ultra-high Performance Light-
[36] Peng B, Shao X. Study on fatigue performance of lightweight composite bridge weighted Composite deck structure. 2015, Department of Transportation of
deck with closed ribs. China Civil Eng J 2017;50(4):89–96 [In Chinese]. Guangdong, China. [In Chinese].

14

You might also like