Composite Structures 228 (2019) 111367

Contents lists available at ScienceDirect

Composite Structures
journal homepage: www.elsevier.com/locate/compstruct

Seismic retrofitting of rectangular bridge piers using ultra-high performance T

fiber reinforced concrete jackets
⁎
Teng Tong, Siqi Yuan, Weiding Zhuo, Zhiqi He, Zhao Liu
Key Laboratory of Concrete and Prestressed Concrete Structures of Ministry of Education, School of Civil Engineering, Southeast University, Nanjing 210018, PR China

A R T I C LE I N FO A B S T R A C T

Keywords: An innovative seismic retrofitting technique for “as-built” reinforced concrete (RC) bridge piers was proposed,
Retrofitting by using ultra-high performance fiber reinforced concrete (UHPFRC) jackets. One “as-built” and two UHPFRC
UHPFRC jackets jacket-retrofitted rectangular cross-section RC piers were fabricated and cyclically loaded. Damage evolutions,
Cyclic loading skeleton curves, strength and stiffness degradations, ductility, self-centering and energy dissipation capacities
Fiber-based finite element
were derived and analyzed. The two jackets mitigated the concrete damage and enhanced the self-centering
capacity of a RC pier. The 850 mm-height jacket significantly increased the strength, but ductility and cumu-
lative energy dissipation capacity. On the other aspect, the 400 mm-height jacket exhibited superior perfor-
mances in ductility and cumulative energy dissipation capacity. Three strengthening mechanisms were brought
by a UHPFRC jacket, and they were cross-section enlargement effect, gap opening effect, and passive confine-
ment effect. Among them, formulas for the confinement effect are analytically obtained. The fiber-based finite
element (FE) model for the UHPFRC jacket-retrofitted RC pier is developed within OpenSees framework. The
simulation agrees well with the experimental results and can benefit further researches involved in the seismic
retrofitting with a UHPFRC jacket.

1. Introduction 1.1. RC jackets

A large portion of small- and medium- span length reinforced con- RC jackets are easy to be implemented, through enlarging the cross-
crete (RC) bridges in seismic prone zones were completed several section, and enhancing the confinement by additional close-spaced
decades ago. Deterioration of a structure’s mechanical properties, hoops or a spiral of small pitch [31,20]. Concrete strength of a RC jacket
seismic hazard re-evaluation, and substitution of the performance- is generally greater or at least equal to that of the shaft. Moreover, the
based seismic design for the force-based design, require many RC jacket thickness is generally greater than 100 mm, leading to the re-
bridges in service to be retrofitted [34,21]. The term “retrofit” herein duced space and apparent dead mass.
refers to the upgradation of a bridge that was inadequately designed or
detailed to meet the current seismic requirements. 1.2. Steel jackets
Bridge piers are the vulnerable component to dissipate ground en-
ergy input, failures of which lead to catastrophic disasters [2,3]. A Repair of a bridge pier with a steel jacket involves installing the
multitude of bridge piers completed prior to the 1970s in the U.S. were jacket by in-filed welding parts along the length, and filling the gap
not well designed, mainly with inadequately lap splice longitudinal between the jacket and pier with cement-based grout [29,30,12,11,13].
rebars inside the plastic hinge zone [32,7,10]. Bridge piers completed in Generally, a 50 mm space between the steel jacket and the supporting
China prior to the 1980s are now suffering from the low flexural and (e.g. footing or cap beam) is set, to avoid the possibility of excessive
shear reinforcement ratios [36], due to the nation’s limited economy. flexural enhancement and therefore the shift of plastic hinge zone.
Providing the targeted flexural strength, ensuring proper ductility, and Steel jackets are not recommended for a pier with the rectangle
precluding brittle shear or splice failure modes, therefore, are the main cross-section [31]. Little restraint of lateral dilation is provided by the
objectives for the seismic bridge retrofitting [35]. Several retrofitting bending of the jacket at the core. Different from a newly-built pier
techniques have been proposed, as: whose concrete can be poured directly into the steel jacket, the gap
between in-filed welding steel jacket and the shaft for a retrofitted pier

⁎
Corresponding author.
E-mail address: mr.liuzhao@seu.edu.cn (Z. Liu).

https://doi.org/10.1016/j.compstruct.2019.111367
Received 12 September 2018; Received in revised form 17 July 2019; Accepted 4 September 2019
Available online 07 September 2019
0263-8223/ © 2019 Elsevier Ltd. All rights reserved.
T. Tong, et al. Composite Structures 228 (2019) 111367

blurs the confinement effect. Moreover, the corrosion of a steel jacket Table 1
limits its large-scale employment in environment-sensitive regions. Design parameters for the piers.
Pier Section (m) Longitudinal Stirrup N
1.3. FRP wrappings bar (HRB400) (HRB300)
Shaft UHPFRC jacket
FRP wrappings are usually bonded to the shaft with epoxy. Because
UR 0.45 × 0.50 – Eight C16 ϕ8 @80 mm 0.08 fc′ Ag
of greater tensile strength and tensile stiffness of FRP fibers, lesser R400 0.45 × 0.50 0.50 × 0.55 at two ends;
thickness is required for a FRP wrapping [1,18,19]. R850 0.45 × 0.50 0.50 × 0.55 ϕ8 @
FRP wrappings are suitable for the circular piers. To obtain sa- 160 mm at
tisfactory confinement effects for a rectangular pier with FRP wrap- middle.

pings, the concrete bolsters or other means of section shape modifica-

Note: N is external axial load; fc′ is cylindrical strength of concrete; and Ag is
tions are indispensable to form a continuously curved surface.
cross-section area of the shaft.
Nevertheless, the rupture of the FRP fibers generally occurs at the
corners of the pier, followed with the abrupt strength loss and the
6000 kN MTS electrohydraulic servo machine. The height-to-width
brittle failure mode [38]. On the other aspect, epoxy bond degradation
ratio for pier UR was around 2650/500 = 5.3, and the flexural-domi-
and fire-sensitive feature are also two factors influencing the wrap-
nated failure mode was expected.
ping’s long-term reliability.
The fabricated piers with the low flexural reinforcement ratio ρl
Alongside, more efforts are put into the quick rehabilitation of
characterized the typical RC pier completed prior to the 1980s in China.
bridge networks after an earthquake. As a consequence, reducing ex-
The ρl was 0.71% for the “as-built” piers. Nevertheless, the quantity
cessive lateral residual drift is becoming another concern for bridge
already increases to around (1–2)% for newly-built RC piers in China.
seismic design [17]. Bridge piers surviving in the Hyogo-Ken NanBu
Eight C16 HRB400 rebars served as the flexural rebars in the piers.
earthquake in 1995 were demolished, not due to the unacceptable
The 3.6 m long flexural rebars were continuous along the cap, the shaft
strength loss, but the greater than 1.75% residual drift [15]. Anchoring
and the footing. The purpose of continuous rebars was to mimic the
at the experience, the allowable residual drift was set as 1% for
longitudinal rebars with reliable mechanical connection or with a suf-
highway bridges in Japan [14].
ficient splice length. To circumvent the possible brittle bond failure
With the progress of cementitious-based materials, ultra-high per-
between longitudinal rebars and concrete, the authors provided a suf-
formance fiber reinforced concrete (UHPFRC) presents extraordinary
ficient anchorage length (0.65 m) in the footing. For transverse re-
mechanical properties, in terms of greater compressive and tensile
inforcements, ϕ8 HRB300 stirrups spaced at a distance of 80 mm at the
strengths, excellent toughness, energy dissipation capacity, etc
shaft’s two ends, and 160 mm at the middle, see Fig. 1.
[8,24,28,6]. Therefore, UHPFRC emerges as an attractive alternative
After the fabrication of the three “as-built” piers, two of them were
option for the seismic bridge retrofitting. By replacing normal concrete
“retrofitted” with 0.05 m thickness UHPFRC jackets. As shown in
around the lapped rebars in the splice region, Dagenais et al. [7] found
Table 1, pier R400 denotes the retrofitted pier with a 400 mm-height
that UHPFRC was effective in eliminating the brittle bond failure which
jacket, whereas pier R850 denotes the one with a 850 mm-height jacket.
frequently occurred in a rectangular pier with cross-sectional aspect
It is acknowledged that the height of a jacket affects a retrofitted pier’s
ratio exceeding 2.0. Xu et al. [40] cyclically loaded a UHPFRC pier and
seismic responses. The plastic hinge zone of pier UR was around
pointed out that it superseded a conventional concrete pier in ductility.
Lp = 393 mm, calculated by the equation 7.6.2.1–1 in the Seismic De-
Yang et al. [41] combined UHPFRC and posttension techniques in a
sign Criteria 1.7 [34]. The height of pier R400’s jacket approximated Lp ,
precast segmental pier and achieved agreed self-centering capacity and
whereas the height of pier R850’s jacket was deliberatively selected as
mitigated concrete damage.
≈2 Lp. For a typical RC jacket, the thickness is generally greater than
In this study, a new seismic retrofitting technique is proposed, with
100 mm. Anchoring at the enhanced compressive and tensile strengths
the employment of UHPFRC jackets. Their retrofitting effects on a RC
of UHPFRC, 50 mm thickness was selected for the UHPFRC jackets (See
pier’s flexural capacity are investigated, whereas the effects on the
Fig. 2).
shear capacity are beyond the scope. To this end, three “as-built” RC
Roughness treatment of normal concrete was performed to achieve
rectangular piers were fabricated, with large height-to-width ratio and
an agreed interfacial bond between the UHPFRC jacket and the RC
low flexural reinforcement ratio. Among them, one pier served as the
shaft, see Fig. 3a. Before casting the jacket (see Fig. 3b), the debris was
prototype one, and the other two were “retrofitted” with UHPFRC
removed though watering the roughened surface. Attention was paid
jackets. Cyclic pseudo-static loading was applied and the piers’ hys-
that keeping the roughened surface moisture was necessary when
teretic behaviors were compared. Strengthening mechanisms of a
casting the jacket.
UHPFRC jacket on a RC pier are also discussed. Passive confinement
UHPFRC generally requires a high-demanding curing condition. The
effect of a UHPFRC jacket on the shaft is analytically evaluated, which
initial curing should be maintained at 20 °C for 12–48 h, followed by
results from the constraint of normal concrete’s dilation [39]. At the
the high-temperature steam curing at 90 °C for another 24–72 h [42].
end, the fiber-based finite element (FE) model with the OpenSees fra-
The above curing condition is easy to be realized for a precast concrete
mework is constructed, specified for a UHPFRC jacket-retrofitted RC
member, but a cast-in-place (CIP) concrete member. For the easy of
pier.
retrofitting, the curing of the UHPFRC jackets under natural environ-
ment was adopted, see Fig. 3c and d. The released curing condition
2. Experimental protocol
undoubtedly leaded to the unignorable but still acceptable early-age
strength loss, to some extent.
2.1. Specimen design and fabrication

Three 1:2 scaled “as-built” RC bridge piers were fabricated in the 2.2. Material properties
lab. After 28-day natural curing, two piers were “retrofitted” with
UHPFRC jackets. Table 1 lists the relevant design parameters for the The shaft, the top cap and the footing adopted C50 concrete. Based
tested piers. Pier UR denotes the “as-built” pier, owning a on the Chinese code GB/T50152 [22], the mean cubic
0.70 × 0.80 × 0.70 m cap, a 0.45 × 0.50 × 2.30 m shaft, and a (150 × 150 × 150 mm) compressive strength fcu ′ was 44.2 MPa (see
0.90 × 0.70 × 1.30 m footing, see Fig. 1. Horizontal pseudo-static Table 2), which was equivalent to the mean cylindrical compressive
cyclic loading was exerted at the midpoint of the cap, through a strength fc′ = 34.9 MPa.

2
T. Tong, et al. Composite Structures 228 (2019) 111367

Fig. 1. Geometry and reinforcement details for pier UR (unit: mm).

2.3. Test set up and loading scheme

Test setup is shown as the photo in Fig. 6. To represent the super-

structure’s self-weight in a real bridge, the two screw-thread bars were
pre-tensioned, with the bottom being fixed to the ground and the top
being tensioned through two hydraulic jacks. Cyclic lateral displace-
ment-control loading was applied at midpoint of the cap through an
MTS electrohydraulic servo machine. The reaction force and the hor-
izontal displacement at the loading point were precisely monitored by
an MTS experiment system. The pseudo-static loading scheme is illu-
strated in Fig. 6. Three repeated cycles were performed at each loading
phase. The 5 mm increment was set for loading phases with ≤50 mm
amplitude, and was subsequently increased to 10 mm until the termi-
nation of the test.
Fig. 2. Locations of the UHPFRC jackets.
The vertical axial load N = 0.08fc′ Ag (about 628 kN) was initially
applied to the pier through tensioning the two screw-thread bars. Of
Reinforcement steel (HRB400) with mean yielding strength special attention was that two issues were encountered, with the
fsy = 432 MPa and mean peak strength fsp = 576 MPa was used for the elongation and the inclination of the bars, see Fig. 7. At a large lateral
longitudinal rebars, see Table 3 and Fig. 4. Yielding strength of trans- displacement, the bars elongated and the tensile force would inevitably
verse reinforcements (HRB300) was around 320 MPa. increase, leading to the vertical axial load N ≫ 628 kN, which would
Proportions of the UHPFRC mixture are listed in Table 4. Com- deviate from a real bridge structure whose self-weight of the super-
mercial UHPFRC product was provided by Subote New Materials CO., structure remained unchanged. To eliminate the disturbance, oil pres-
LTD (Jiangsu, China). The 0.2 mm-dia. steel fibers randomly distributed sures of the two hydraulic jacks were manually controlled as approxi-
in the matrix and they took about 2% volume fraction. The following mately a constant value. Oil pressures were released at the large lateral
properties were measured: 2600 MPa tensile strength, 205 GPa elastic displacement and were increased at the small lateral displacement. The
modulus, 7840 kg/m3 density, and around 20 mm length. The mean recorded tensile force in the two bars varied within the range from
cubic (150 × 150 × 150 mm) compressive strength for UHPFRC, after 628 kN to 660 kN. Alongside, the inclination of thread bars leaded to a
28-day natural curing, was 129.7 MPa, which was equivalent to the trace of horizontal restoring force. The recorded lateral displacement-
mean cylindrical compressive strength 102.5 MPa. The elastic modulus force relationship from the MTS system was revised to eliminate the
for UHPFRC was around 40.1 GPa. disturbance, see the formula in Fig. 7.
Two 500 mm length custom-fabricated dog-bone specimens were For pier retrofitting in practice, the jacket is cast while the shaft is
prepared to measure the UHPFRC’s tensile strength, with the simultaneously loaded by the superstructure. In other words, the jacket
50 × 100 mm rectangular cross-section at the middle, see Fig. 5a. The would not carry the superstructure’s self-weight. In the laboratory, it
alignment of the test specimen was carefully checked before testing. was unrealistic to consistently load the shaft for more than 28 days, for
The pin-fixed ends were guaranteed to avoid secondary flexural stress the purpose of merely casting and curing of the UHPFRC jacket without
and to ensure the centric loading. The tested displacement-force curves the influences of the superstructure. Consequently, the composite cross-
are plotted as the solids lines in Fig. 5b. The measured UHPFRC’s tensile section of the shaft and the jacket together carried the N = 0.08 fc′ Ag
strength after 28-day natural curing was around 5.52 MPa. axial load. The load sequence would generate ignorable effects on
seismic responses of bridge piers, after the exploration by the software

3
T. Tong, et al. Composite Structures 228 (2019) 111367

(a) Roughening normal concrete surface; (b) Casting the jacket;

(c) Wetting of the jacket; (d) Curing of the jacket.

Fig. 3. Surface treatment of normal concrete and construction of the UHPFRC jacket.

Table 2 3.1. Pier UR

Compressive strengths (MPa) for C50 cubic samples.
Batch Sample
The hysteretic curve after modification (see Section 2.3) for pier UR
is shown in Fig. 8 and it shows quite symmetric for push/pull cycles.
1 2 3 4 5 6 Average Peak load was witnessed at δ = 2.26%, followed by a plateau until
δ = 3.02%. The maximum lateral strengths were +108.4 kN and
1 46.4 47.7 42.5 47.1 42.8 40.6 44.7
2 42.6 45.2 44.4 43.7 41.9 41.1 43.2
−112.7 kN, respectively. Pier UR endured 15 complete sets of three
3 43.2 47.5 43.9 44.9 43.1 40.9 44.8 push/pull loops before its lateral load-carrying capacity dropped to
80% of its peak value Fp , at δ = 3.77%. One longitudinal rebar failed in
tension at δ = 4.91%, due to cyclic fatigue.
AIS [7]. Damage evolutions of pier UR are plotted in Fig. 9 and summarized
in Table 5. The first hairline flexural crack was observed at δ = 0.38%,
about 0.45 m above the footing. Multiple flexural cracks emerged,
3. Damage evolutions propagated and widened. Flexural cracks stabilized after δ = 1.89%,
see Fig. 9b. Simultaneously, concrete slightly spalled at the extreme
Damage evolutions of the piers with the increasing lateral drift δ are fiber of the shaft. Concrete spalling evolved and its height arrived at
summarized in Table 5. Note that lateral drift δ is adopted, a di- 15 cm at δ = 3.77%, see Fig. 9d. Due to the relative low ρl , no crushing
mensionless variable measured by the ratio of lateral displacement over was observed at the core concrete. Photos of pier UR’s final damage are
effective height.

4
T. Tong, et al. Composite Structures 228 (2019) 111367

Table 3 R400 is shown in Fig. 11 and it shows quite symmetric for push/pull
Tensile testing of longitudinal rebars. cycles. Peak load was witnessed at δ = 2.26%, followed by a small
Longitudinal rebar fsy (MPa) fsp (k3 fsy , MPa) εsy (%) εsp (k2 εsy , %)
degradation of the lateral load-carrying capacity. The maximum lateral
strengths were +126.85 kN and −125.89 kN, respectively. Pier R400
HRB 400 435 574 2.2 20.4 endured 18 complete sets of three push/pull loops before its lateral
436 581 2.2 20.6 load-carrying capacity F dropped to 0.8Fp , at δ = 4.91%. Due to the
426 575 2.1 19.3 limitation of the MTS electrohydraulic servo machine, pier R400 was
not loaded to the fracture of longitudinal rebars. After the cyclic
Note: fsy is yield strength; fsp is peak strength; εsy is yield strain; εsp is strain
loading, monotonic pushover of pier R400 was performed until one
corresponding to peak strength.
longitudinal rebar fractured at the interface between the UHPFRC
jacket and the footing, indicating the flexural-dominated failure mode
and no shift of plastic hinge zone.
Damage evolutions of pier R400 are plotted in Fig. 12 and sum-
marized in Table 5. Photos of pier R400’s final damage are presented in
Fig. 13. When loaded to δ = 0.38%, two flexural cracks formed at 0.25
and 0.55 m above the UHPFRC jacket’s upper surface. At δ = 1.51%,
corner concrete of the shaft slightly spalled above the UHPFRC jacket’s
upper surface, see Figs. 12b and 13a. The test was terminated at
δ = 4.91%.
Around 8 cm length vertical crack was firstly witnessed at
δ = 1.32%, in the vicinity of the upper surface of the UHPFRC jacket,
indicating the apparent passive confinement effect due to the normal
concrete dilation, see Fig. 12b. When loaded to δ = 3.02%, the diagonal
flexural-shear cracks were observed along the direction of external
Fig. 4. A typical strain–stress curve for one sample longitudinal rebar. loading, see Fig. 12c. After the loading was terminated, numerous
vertical expansion cracks and flexural-shear cracks distributed on the
UHPFRC jacket’s surfaces, see Fig. 13.
Table 4
Proportions of the UHPFRC mixture. Different from pier UR exhibiting apparent flexural cracks and
concrete spalling at the bottom shaft (see Fig. 10), the confinement of
UHPFRC Proportion power Steel fiber Superplasticizer Water
the UHPFRC jacket altered the damage distribution of pier R400, and
Mixture (kg/m ) 3
2095 156 22.10 182.4 no flexural cracks or concrete spalling was observed at the shaft en-
closed by the jacket. Due to the relative weak interfacial bond strength
between the UHPFRC and the normal concrete, we witnessed the “de-
presented in Fig. 10. tachment” of the jacket and the footing at δ = 1.51%, and a “gap”
formed at the interface. Before the opening of the gap, flexural cracks
and concrete spalling concentrated at the shaft above the UHPFRC
3.2. Pier R400
jacket. Interestingly, these damages progressed at a much slower rate
after the formation of the gap. On the other aspect, the width of the gap
The hysteretic curve after modification (see Section 2.3) for pier

(a) Dog-bone specimen; (b) Tested tensile displacement-force curves.

Fig. 5. Tensile strength test of UHPFRC.

5
T. Tong, et al. Composite Structures 228 (2019) 111367

Fig. 6. Test setup for pier R850 and loading scheme.

Fig. 8. Hysteretic curve for pier UR.

Fig. 7. Revise of reaction force.

Table 5
Overview of damage evolutions.
Drift (%) Damage description

Pier UR Pier R400 Pier R850

0.75 Flexural cracks formed Flexural cracks formed Flexural cracks formed
0.94 Gap opened at the base shaft
1.32 Vertical expansion crack occurred at the jacket.
1.51 Corner concrete spalled above the jacket; Gap
opened at the base shaft
1.89 Concrete spalled at the base shaft Vertical expansion crack occurred at the jacket
3.02 Diagonal shear cracks occurred at side faces of
UHPFRC jacket
3.77 Concrete spalling height was 15 cm; Strength Strength was greater than 0.8FP; Fracture of longitudinal bar,
degraded to 0.8FP. and the test was terminated.
4.15 Fracture of longitudinal bar, and the test was
terminated.
4.91 Strength degraded to 0.8FP, and the test was
terminated.

6
T. Tong, et al. Composite Structures 228 (2019) 111367

Fig. 11. Hysteretic curve for pier R400.

15th set, with δ = 3.77%.

Damage evolutions of pier R850 are plotted in Fig. 15 and sum-
marized in Table 5. Photos showing pier R850′s final damage are pre-
sented in Fig. 16. The first flexural crack on the shaft appeared 0.25 m
above the UHPFRC jacket, at δ = 0.57%. The second flexural crack
appeared 0.45 m above the jacket, at δ = 0.75%. With the increasing
drift, flexural cracks propagated stably. At δ = 3.77%, the fracture of
one longitudinal rebar occurred at the base shaft, indication the flex-
ural-dominated failure mode and no shift of plastic hinge zone. Mean-
while, the lateral load-carrying capacity was still above 80% of its peak
load Fp .
Fig. 9. Damage evolutions for pier UR. A vertical crack with around 10 cm length was firstly witnessed at
δ = 1.89%, in the vicinity of the upper surface of the UHPFRC jacket,
increased sharply and arrived at around 7–8 mm, which contributed indicating the apparent passive confinement effect, see Fig. 16b.
significantly to the pier R400’s lateral drift. Compared to pier R400, the number of expansion cracks was reduced
significantly and no flexural or flexural-shear cracks were observed on
3.3. Pier R850 the 850 mm-height UHPFRC jacket.
Similar to pier R400, the confinement from the 850 mm-height
The hysteretic curve after modification (see Section 2.3) for pier UHPFRC jacket enabled the integrity of the base shaft, without flexural
R850 is shown in Fig. 14. Peak load was witnessed at δ = 2.64%, fol- cracks or concrete spalling. Concrete cracking propagated apparently of
lowed by a small degradation of the lateral load-carrying capacity. The pier R850 before the opening of the gap (δ =1.13%), located between
maximum lateral strengths were +154.86 kN and −144.84 kN, re- the jacket and the footing. The width of the gap increased to 1 mm,
spectively. Pier R850 endured 14 sets of three push/pull loops, and 4 mm, and even 12 mm at δ = 1.32%, 1.70%, and 3.40%, respectively,
failed with the fracture of one longitudinal rebar at the 2rd loop of the which contributed significantly to the pier R850′s lateral drift.

(a) Perpendicular to the loading direction; (b) Along the loading direction.
Fig. 10. Photos for pier UR’s final damage.

7
T. Tong, et al. Composite Structures 228 (2019) 111367

factor μ , pier R400 also showed the most improvement ( μ =8.80).

However, due to the significant rise in yield drift δ y , pier R850 was the
specimen with the lowest μ = 5.69. Even so, all the piers had agreed
ductility behaviors as μ ≥ 3 [34].

4.2. Strength and stiffness degradations

Strength degradation of a pier is quantatively evaluated by the

strength degradation factor γi , which is defined by the ratio of load-
carrying capacity in the third cycle F3i to that in the first cycle F1i at
each lateral loading level i :
F3i
γi =
F1i (1)
Fig. 19 illustrates strength degradations of piers UR, R400 and
R850. Before lateral drift reached δ = 2.26%, there was a stable period
for three piers, and it showed no apparent difference for the strength
degradations. After that, pier UR showed a dramatic drop stage of
strength degradation and γi reached 0.75 at δ = 3.77%. Pier R850 dis-
played similar strength degradation behavior. Nevertheless, pier R400
exhibited nearly no strength degradation as γi fluctuated around 0.95,
until the termination of the test.
Similarly, lateral stiffness also represented a remarkable degrada-
tion trend under cyclic loading. The stiffness Ki , expressed as the mean
value of that in positive and negative half cycle at each lateral loading
level i , gives as:
1 +
Ki = (K i + K i−)
2 (2)
Fig. 12. Damage evolutions for pier R400. Fig. 20 shows that the stiffness represented a remarkable degrada-
tion phenomenon with the rising drift δ . A number of flexural cracks
4. Impacts of UHPFRC jackets on seismic responses appeared on the piers during the initial stage of the lateral loading,
which resulted in the significant stiffness degradation. With the crack
To further understand the enhancement effects from the UHPFRC propagations, the plastic deformation developed significantly and
jackets, the piers’ hysteretic responses are compared in Table 6. slowed down the degradation. Therefore, the stiffness degradation re-
presented a common characteristic as ‘‘fast followed by slow”. The
4.1. Skeleton curves stiffness degradation trend for tested piers were similar and it seems
that the UHPFRC jackets produced limit effects on the stiffness de-
Skeleton curves are plotted in Fig. 17. Skeleton curves developed gradation.
along a nearly linear line, until reaching the yield point. Before
yielding, little difference of these skeleton curves existed among the 4.3. Self-centering capacity
piers, indicating that the UHPFRC jackets generated ignorable effects
on the pier’s initial stiffness. Residual drift δr is a quantitative index for the post-earthquake
The following performance variables are compared, including the survivability and reparability in the seismic bridge design, and it is
yield point (δ y , Fy ), peak point (δp , Fp ) and ultimate point (δu , Fu ). The generally employed to evaluate the self-centering capacity, as:
yield point (δ y , Fy ) adopts the definition proposed by Sun et al. [37], δr + + δr−
which is the intersection point of the line connecting origin and the δr =
2 (3)
point at the skeleton curve with 0.75Fp , and the horizontal line passing
the point at the skeleton curve with Fp , see Fig. 18. Attention was paid where δr and +
δr−
are the absolute values of positive and negative re-
that the yield point was determined not based on the first yielding of sidual drifts in a certain hysteretic loop, respectively. In this study, the
longitudinal rebars. In addition, ductility factor μ = δu/ δ y is adopted to first hysteretic loop of the three loops in a loading phase was selected
evaluate the piers’ ductility. The ultimate strain δu is charaterized as the for the calculation, since the second and third loops at the identical
drift with the 80% post-peak strength or the fracture of the longitudinal loading phase were quite similar.
bar, whichever occurred first. Fig. 21 shows the column chart for residual drift δr with relationship
These variables in the first and third quadrants are assigned with the to the lateral drift δ . Clearly, at a small lateral drift (δ < 2%), piers
superscripts “+” and “−”, respectively. All the variables summarized R400 and R850 did not have advantages over pier UR in terms of δr . At
in Table 7 are the mean values obtained from skeleton curves in the first a large lateral drift (δ > 2%), we witnessed the significant effects of the
(“+”) and third (“−”) quadrants. 400 mm-height and 850 mm-height UHPFRC jackets in enhancing the
Load-carrying capacity was enhanced significanly with the em- pier’s self-centering capacity. At the last loop of pier UR (δ =3.77%), δr
ployment of a UHPFRC jacket. As shown in Table 7, the 400 mm-height of pier R400 was only 71.4% of pier UR, and δr of pier R850 was only
jacket brought out 15.6% and 16.2% increments in terms of Fy and Fp , 66.5% of pier UR.
in comprsion with pier UR. For the 850 mm-height jacket, the ince-
ments were 36.5% for Fy , and 37.8% for Fp , respectively. 4.4. Energy dissipation
As to the dutility, two criteria were adopted. If ultimate strain δu
was adopted, pier R400 exhibited significant superority (δu =4.91%) Dissipated energy is calculated as the area underneath the experi-
over either pier UR or pier R850 (δu =3.77%). If evaluated by ductility mental lateral force-displacement curve. The dissipated energy per loop

8
T. Tong, et al. Composite Structures 228 (2019) 111367

(a) Perpendicular to the loading direction;

(b) Along the loading direction; (c) Gap opening.

Fig. 13. Photos for pier R400′s final damage.

EDi and cumulative dissipated energy EDsum were compared. EDi and
EDsum were evaluated until the 20% reduction of the post-peak load-
carrying capacity, or the fracture of longitudinal rebars, whichever
occurred first.
Fig. 22 shows the EDi for the piers. It can be seen that EDi of pier UR
was nearly identical to that of pier R400. Pier R850 consumed a little
higher EDi than that of the other two piers.
As shown in Fig. 23, EDsum was 177.24 kJ for pier UR, was 301.17 kJ
for pier R400, and was 190.21 kJ for pier R850. Obviously, the piers’
EDsum was affected by the UHPFRC jacket height. With the 400 mm-
height jacket, EDsum gained nearly 70.0% enhancement for pier R400,
in comparison with pier UR. Nevertheless, the increment in EDsum
conversely decreased with the 850 mm-height jacket. Pier R850 merely
showed 7.3% increment in EDsum , which was nearly ignorable. The
massive expansion cracks and flexural-shear cracks on the UHPFRC
jacket (see Fig. 13) indicated that pier R400 dissipated the most energy,
Fig. 14. Hysteretic curve for pier R850.
with respect to pier R850 exhibiting merely several expansion cracks

9
T. Tong, et al. Composite Structures 228 (2019) 111367

5.1. Basic equations

Plane strain condition is assumed and it requires Δεz = γzx = γzy = 0 ,

where γzx and γzy are the shear strains in zx and zy planes, respectively.
The coordinate system is illustrated in Fig. 24. By assuming the constant
or zero body force, the stress state is given by the partial derivatives of
the Airy’s function f (x , y ) :
∂ 2f ∂ 2f ∂ 2f
σx = σy = τxy = −
∂y 2 ∂x 2 ∂x ∂y (4)

where σx and σy are concrete axial stresses in x and y directions, re-

spectively; and τxy is the shear stress in xy plane. Compatibility condi-
tion is imposed by the following equation:
∂ 4f ∂ 4f ∂ 4f
+ + =0
∂x 4 ∂x 2∂y 2 ∂y 4 (5)

Furthermore, the boundary condition requires:

σx ax + τxy a y = px τxy ax + σy a y = py (6)

where ax and a y are director cosines of the normal to the boundary; px

and py are boundary forces along x and y directions, respectively.
The problem now converts to find a proper Airy’s function f (x , y )
satisfying Eqs. (5) and (6). Being compatible with the geometry sym-
metry, the second-order and fourth-order Airy’s functions, namely
f2 (x , y ) and f4 (x , y ) , are employed:
Bl 2 2 Bl 2 2 A A A 4
f2 (x , y ) = − x − y f4 (x , y ) = − x 4 + x 2y 2 + y
2 2 12 2 12 (7)
Fig. 15. Damage evolutions for pier R850. By substituting f2 (x , y ) into Eq. (4), we can obtain the corre-
sponding confining stress state, as illustrated in Fig. 24a. Similarly, the
(see Fig. 16). confining stress state of f2 (x , y ) is illustrated in Fig. 24b.
The confining stress state corresponding to
f (x , y ) = f2 (x , y ) + f4 (x , y ) , is expressed in Eq. (8) and plotted in
5. Passive confinement effects Fig. 25.

σx = A (x 2 − y 2 ) − Bl 2
Discussions in Sections 2 and 3 clearly illustrate the enhanced
seismic responses of a RC pier enclosed by a UHPFRC jacket. Based on σy = A (y 2 − x 2) − Bl 2
the observations, the authors summarized the following three τxy = −2Axy (8)
strengthening mechanisms, as:

(1) Cross-section enlargement effect; 5.2. Passive confinement from the transverse reinforcements
(2) Gap opening effect;
(3) Passive confinement effect. The loads P (x ) and P (y ) in Fig. 26, acting orthogonally to the lateral
tie, represent the stress acting on the concrete lateral faces multiplied
Among the three strengthening effects, the enlargement of the cross- by the hoop spacing S :
section due to the 50 mm thick UHPFRC jacket is similar to the con- P (x ) = [A (l 2 − x 2) − Bl 2] S (9)
ventional RC jacket and will not be discussed in detail. Alongside, the
gap between the UHPFRC jacket and the RC footing is already revealed The axial force Nst at the corner, equals the sum of tangential stress
as the Figs. 13c and 16c for piers R400 and R850, respectively. In this and orthogonal pressure:
section, we focus on the passive confinement effect. l l l3 (A + 3B )
For pier UR, transverse reinforcements confined the core concrete Nst = ∫0 P (x ) dx + ∫0 τxy dy = S
3 (10)
and improved the strength and the ductility at the material level.
Moreover, the UHPFRC jacket additionally confined both the cover and Analytical solutions for the coefficients A and B require the de-
core concrete of piers R400 and R850. Different from a circular cross- formation compatibility between the exterior boundary of concrete and
section, how to quantitively evaluate the passive confinement effect on transverse reinforcements. Details can refer to Braga et al. [4]. By using
a rectangular cross-section, either originated from the transverse re- minimum chi-square method to satisfy the deformation compatibility.
inforcements or the UHPFRC jacket, is a difficult problem ever em- The following expression for the A and B are obtained:
pirically being evaluated. In this investigation, considering that the (21SEc2 Es As vl) εz
length (0.50 m) approximated the width (0.45 m) of the shaft, the A =
25S2Ec2 l 4 + 6SEc Es l [315Is (v + 1) + 2l2As (2v + 5)] − 1890Ec2 Is As (v 2 − 1)
rectangular cross-section is simplified as the square cross-section \{ 18Ec Es As v [SEc l3 + 105Es Is (v + 1)]\} εz
B =
(0.475 × 0.475 m) with the equivalent area. Inspired by the work of l2 {25S2Ec2 l 4 + 6SEc Es l [315Is (v + 1) + 2l2As (2v + 5)] − 1890Ec2 Is As (v 2 − 1)} (11)
Braga et al. [4], the passive confinement for the square cross-section is
where Ec is the elastic modulus of concrete; v is Poisson’s ratio; Es is the
obtained.
elastic modulus of transverse reinforcements; As is the cross-section
area of transverse reinforcements; and Is is the moment of inertia of
transverse reinforcements.

10
T. Tong, et al. Composite Structures 228 (2019) 111367

(a) Perpendicular to the loading direction;

(b) Along the loading direction; (c) Gap opening.

Fig. 16. Photos for pier R850′s final damage.

Table 6
Comparison of the piers.
Peak load (Fp ) R850 > R400 > UR
Strength degradation UR > R850 > R400
Stiffness degradation R850 ≈ UR ≈ R400
Ductility-ultimate drift (δu ) R400 > UR ≈ R850
Ductility-ductility factor ( μ ) R400 > UR > R850
Self-centering capacity R400 ≈ R850 > UR
Cumulative energy dissipation R400 > R850 > UR

5.3. Passive confinement from the UHPFRC jacket

Due to the thin thickness of the UHPFRC jacket (50 mm), the au-
thors ignore the moment of inertia for the jacket (Iu = 0 ). Consequently,
A and B in Eq. (11) degrade to: Fig. 17. Skeleton curves of the piers.

11
T. Tong, et al. Composite Structures 228 (2019) 111367

Fig. 21. Residual drifts of the piers.

Fig. 18. Definitions of performance variables.

Table 7
Comparisons of performance variables.
Pier δ y (%) Fy (kN) δp (%) Fp (kN) δu (%) μ

UR 0.502 91.9 2.26 110.6 3.77 7.51

R400 0.558 106.2 2.26 126.4 4.91 8.80
R850 0.662 125.4 2.64 149.9 3.77 5.69

Fig. 22. Dissipated energy per loop EDi of the piers.

Fig. 19. Strength degradations of the piers.

Fig. 23. Cumulative dissipated energy EDsum for the piers.

21Ec Eu tu v
A= ε
l2 [25Ec l + 12Eu tu (2v + 5)] z
18Ec Eu tu v
B= ε
l2 [25Ec l + 12Eu tu (2v + 5)] z (12)

where Eu is the elastic modulus of UHPFRC; tu is the thickness of the

jacket.

Fig. 20. Stiffness degradations of the piers. 5.4. Strength and ductility enhancements of concrete

After obtaining values for the coefficients A and B , the equivalent

confining pressure from fl′ is obtained for a square section as [4]:

12
T. Tong, et al. Composite Structures 228 (2019) 111367

(a) from f 2 ( x, y ) ;

(b) from f 4 ( x, y ) .
Fig. 24. Distributions of confining stresses.

fl′ = −Bl 2 (13) fl′ = fl, i + fl′, e for core concrete

fl′ = fl′, e for cover concrete (14)
For piers R400 and R850, the core concrete of the shaft, enclosed by
the stirrups, is also confined by the UHPFRC jacket. Meanwhile, their
where fl, i is the confining pressure from transverse reinforcements; fl′, e is
cover concrete is merely confined by the UHPFRC jacket. The confining
the confining pressure from the UHPFRC jacket.
pressure for core and cover concrete of UHPFRC jacket-retrofitted piers
The confining pressure fl, i is plotted as the curved marked with
is evaluated as: “Stirrups” in Fig. 27, varying with the axial strain εz . Two specific points
are mentioned on the curve. The first point corresponds to the limit

Fig. 25. Distributions of lateral stresses from f (x , y ) .

13
T. Tong, et al. Composite Structures 228 (2019) 111367

Fig. 26. Stirrup schematization.

where εcu is the strain when stress drops to 20% of peak stress fcc′ ,
and is related to the volume fraction of transverse reinforcements ρsv
and yield strength of transverse reinforcements f yh , as:
f yh
εcu = 0.004 + 0.9ρsv
300 (18)

The Z slop in Eq. (17) is the strain softening slope of the segment
joining the peak point ( fcc′ , εcc ) to the ultimate point ( fcu
′ , εcu ), of which
in essence corresponds to a linearization of the descending branch, as:
fcc′ − fcu
′ 1
Z =
εcu − εcc fcc′ (19)

The stress–strain relationships for concrete are illustrated in Fig. 28,

including unconfined concrete, concrete confined by stirrups, concrete
Fig. 27. Lateral confining stress. confined by the UHPFRC jacket, and concrete confined by combined
stirrups and UHPFRC jacket.
value of the Poisson’s ratio v = 0.5, see Eq. (20). Beyond this point, the It should be noted that concrete behaves nonlinearly and the values
confining pressure fl, i vary linearly until the yielding of transverse re- of Ec and v are functions of the strain state. Therefore, the secant
inforcements. modulus of concrete is adopted for Ec . The Poisson’s ratio v mentioned
For simplicity, a elastic-perfectly plastic stress-strain relationship is above is related to the axial strain εz [16], as:
assumed for UHPFRC, see Fig. 5b. The confining pressure fl′, e varying εz εz εz
with the axial strain εz is plotted as the curved marked with “UHPFRC v (εz ) = v0 [1 + 0.2( ) − ( )2 + 1.55( )3] ≤ 0.5
εc 0 εc 0 εc 0 (20)
jacket” in Fig. 27. The marked point on the curve corresponds to the
peak tensile stress of UHPFRC. where v0 = 0.2 and εc0 is the strain at stress peak of the unconfined
The combined confining pressure fl′ for core concrete is also shown concrete. The upper limit value of v is assumed to be 0.5, which cor-
as the curved marked with “Combined” in Fig. 27, including con- responds to diffuse cracking of the unconfined concrete.
tributions from the stirrups and the UHPFRC jacket.
With reference to the equation proposed by Moghaddam et al. [23], 6. Numerical modeling strategies
the peak strength fcc′ is obtained as:
fcc′ fl′ fl′ Searching an accurate simulation tool to represent seismic responses
=1+8 − 4( )1.2 of the UHPFRC jacket-retrofitted piers is indispensable to perform the
fc′0 fc′0 fc′0 (15)
The corresponding peak strain εcc is given as:
fl′
εcc = εc 0 ( )1.1
fc′0 (16)
′ , εc0 ) is the peak point at the stress-strain
In Eq. (15) and Eq. (16), ( fc0
curve of the unconfined concrete, whereas ( fcc′ , εcc ) is the peak point of
confined concrete.
The Kent-Scott-Park model [27,33] is adopted to characterize the
stress-strain relationship of confined concrete. This model consists a
parabolic ascending part, a linear descending part, and a horizontal
part, as:
c ε c 2 ε
⎧ fcc′ [2( εcc ) − ( εcc ) ] εc ≤ εcc
⎪
σc = fcc′ [1 − Z (εc − εcc )] εcc ≤ εc ≤ εcu
⎨
⎪ 0.2fcc′ εc > εcu (17)
⎩ Fig. 28. Compressive stress–strain relationships for concrete.

14
T. Tong, et al. Composite Structures 228 (2019) 111367

Fig. 29. The fiber-based FE model for pier UR.

vulnerability assessment of a bridge. 6.2. Piers R400 and R850

Conventional 3D solid FE model built in general purpose FE soft-
ware (ABAQUS, ANSYS, etc.) requires high-level modeling strategies In the fiber-based FE models for piers R400 and R850. The three
and demanding computational costs [39], therefore is not suitable for strengthening mechanisms presented in Section 5 are incorporated
the seismic bridge analysis. In this study, the fiber-based FE model through the following approaches:
(within the framework of OpenSees), convenient for design-oriented or
entire bridge simulations, is adopted for the seismic simulation of the (a) Cross-section enlargement effect is automatically captured by including
UHPFRC jacket-retrofitted piers. the UHPFRC jacket in the “fiberSec” section, as illustrated in Fig. 31.
(b) Gap opening effect is embraced through modifying the constitutive law of
longitudinal rebars, see Eq. (22).
6.1. Pier UR
(c) Passive confinement effect is taken into account by assigning cover
concrete and core concrete with proper stress–strain relationships.
The OpenSees framework is mature in simulating the seismic re-
Specifically, the stress–strain relationship labelled with “UHPFRC
sponses of a typical CIP RC pier. The fiber-based FE model for pier UR is
jacket” in Fig. 28 is assigned for cover concrete, and the one labelled
presented in Fig. 29.
with “Combined” is assigned for core concrete.
The “fiberSec” section consists a bundle of fibers. Fig. 29 illustrates
a typical section for the shaft, formed by 320 fibers for core concrete, 80
The fiber-based FE model for pier R850 shown in Fig. 31. Conrete02
fibers for cover concrete and 8 fibers for longitudinal rebars. Pier UR is
embedded in OpenSees is selected for UHPFRC, which can consider the
discretized into 10 elements and 2 Gaussian (integration) points are set
tensile strength. Cross-section discretization scheme of pier R850 is
for each element.
identical to pier UR for the shaft out of the UHPFRC jacket. The jacket is
Uniaxial material model Conrete01 embedded in OpenSees is selected
discretized into 100 fibers, see Fig. 31. The whole shaft is discretized
for core concrete and cover concrete. Stress–strain relationships for
into 10 elements and 2 Gaussian (integration) points are set for each
cover concrete and core concrete are obtained in Section 5.2, see the
element.
curves labelled with “Unconfined” and “Stirrups” in Fig. 28, respec-
For pier UR, concrete cracks smeared inside the plastic hinge zone
tively.
(see Fig. 10), and the bond-slip behavior between the longitudinal re-
Uniaxial material model ReinforcingSteel in OpenSees is adopted for
bars and the shaft did not play a significant role. On the other aspect,
longitudinal rebars [5,9]. ReinforcingSteel can consider damage accu-
the two UHPFRC jacket-retrofitted piers displayed the concentrated
mulation, stiffness degradation, and strength reduction of the re-
cracks (gaps) between the UHPFRC jacket and the RC footing. As a
inforcement caused by cyclic loading. The stress–strain curve of long-
consequence, the bond-slip behavior illustrated in Fig. 32 should be
itudinal rebars is expressed as [9]:
incorporated in the FE model.
With respect to the work done by Pan et al. [26], the constitutive
⎧ Es εs for 0 < εs ≤ εsy
⎪ law of the longitudinal rebars is modified to account for the bond-slip
σs = fsy for εsy < εs < k1 εsy behavior. The modified skeleton curves for longitudinal rebars under
⎨ Es (1 − k3) 2
⎪ k3 fsy + εsy (k2 − k1)2 (εs − k2 εsy ) for k1 εsy < εs < k2 εsy tension can be expressed as:
⎩ (21)

where Es , fsy , εsy are summarized in Table 3; k1, k2 , and k3 are applied to ⎧ Es εs / k 0′ for 0 < εs ≤ k 0′ εsy
⎪ ′ ′
define the shape of the stress–strain curve and their definitions are il- σs = fsy for k 0 εsy < εs < k1 εsy
lustrated in Fig. 4. ⎨ Es (1 − k3) 2
⎪ k3 fsy + εsy (k2′ − k1′)2 (εs − k 2′ εsy ) for k1′ εsy < εs < k 2′ εsy
Experimental and simulation are compared in Fig. 30, in terms of ⎩ (22)
hysteretic loop, skeleton curve, residual drift, and energy dissipation
per loop. Experimental curves are plotted in solid lines and fiber-based where k 0′, k1′ and k 2′ are the three factors defining the shape of the
FE predictions are plotted in dash lines. modified uniaxial stress–strain curve, and they are:

15
T. Tong, et al. Composite Structures 228 (2019) 111367

(a) Hysteretic curve; (b) Skeleton curve;

(c) Residual drift; (d) Energy dissipation per loop.

Fig. 30. Experimental and simulation results for pier UR.

fsy db distribution proposed by Ou et al. [25], see Fig. 32 and Eq. (24).
k 0′ = 1 + 8¯τ b1 Le
fsy db τ¯b1 = 1.0 fc′ for εs ≤ εsy
k1′ = k1 + 8¯τ b1 Le
fsy db fsy db 2 1
τ¯b2 = 0.5 fc′ for εs > εsy (24)
k 2′ = k2 + + (k3 − 1)( 3 k1 + 3 k2 )
8¯τ b1 Le 4¯τ b2 Le (23)
Experimental and simulation results for pier R400 and R850 are
where Le is the length of the footing fiber element; db is the diameter respectively compared in Figs. 33 and 34, in terms of hysteretic loop,
of longitudinal rebars; τ̄b1 and τ̄b2 are the simplified bond shear stress skeleton curve, residual drift, and energy dissipation per loop. Good

Fig. 31. The fiber-based FE model for pier R850.

16
T. Tong, et al. Composite Structures 228 (2019) 111367

load-carrying capacity. Specifically, the 400 mm- and 850 mm-

height UHPFRC jackets leaded to 16.2%, and 37.8% increments in
peak strength, respectively.
(2) The 400 mm-height UHPFRC jacket apparently enhanced a pier’s
ductility. With the jacket height increasing to 850 mm, the ductility
of pier R850 was impaired to some extent. In addition, the UHPFRC
jacket improved a RC pier’s self-centering capacity and significantly
mitigated the residual lateral drift.
(3) With proper selection of jacket height, pier R400 displayed the
highest cumulative energy dissipation, with apparent expansion
cracks and shear-dominated cracks on the jacket.
(4) Three strengthening mechanisms originated from the UHPFRC
jacket are summarized. Among them, the passive confinement ef-
fect is analytically evaluated. The three strengthening mechanisms
are incorporated into the fiber-based FE models, which can accu-
rately capture seismic responses of piers R400 and R850.
Fig. 32. Bond stress distribution of longitudinal rebars at the gap.
In this paper, the authors investigate a novel seismic retrofitting
technique with the UHPFRC jacket, and also provide and the appro-
agreements validate the effectiveness of the proposed fiber-based FE priate fiber-based FE model. It should be noted that the “gap opening
model in capturing seismic behaviors of the two UHPFRC jacket-ret- effect” was obtained based on the experimental observations and fur-
rofitted piers. ther in-depth investigations are indispensable to verify the postulation.
Alongside, more experiments are necessary to enrich the seismic ret-
7. Conclusions rofitting effect of a RC pier’s with the UHPFRC jacket, in terms of both
flexural and shear capacities.
Pseudo-static cyclic loading test was performed to illustrate the
potential of a UHPFRC jacket in seismic retrofitting of a RC bridge pier Acknowledgements
with insufficient longitudinal rebars. Based on the experimental and
simulation results, the following conclusions can be drawn: This research is supported by the Fundamental Research Funds for
the Central Universities (China), the Priority Academic Program
(1) The UPHFRC jacket was effective in increasing a RC pier’s lateral Development (PAPD) of Jiangsu Higher Education Institutions (China,

(a) Hysteretic curve; (b) Skeleton curve;

(c) Residual drift; (d) Energy dissipation per loop.

Fig. 33. Experimental and simulation results for pier R400.

17
T. Tong, et al. Composite Structures 228 (2019) 111367

(a) Hysteretic curve; (b) Skeleton curve;

(c) Residual drift; (d) Energy dissipation per loop.

Fig. 34. Experimental and simulation results for pier R850.

1105007002), the National Natural Science Foundation for Young [15] Kawashima K, MacRae GA, Hoshikuma JI, Nagaya K. Residual displacement re-
Scientists of China (51808113), and the National Natural Science sponse spectrum. J Struct Eng 1998;124(5):523–30.
[16] Kupfer H, Hilsdorf HK, Rusch H. Behavior of concrete under biaxial stresses. J Proc
Foundation for Young Scientists of Jiangsu Province (BK20180389). 1969;66(8):656–66.
[17] Kurama YC, Sritharan S, Fleischman RB, et al. Seismic-resistant precast concrete
References structures: state of the art. J Struct Eng 2018;144(4).
[18] Lavorato D, Nuti C, Santini S, Briseghella B, Xue J. A repair and retrofitting inter-
vention to improve plastic dissipation and shear strength of Chinese RC bridges.
[1] Albanesi T, Lavorato D, Nuti C, Santini S. Experimental tests on repaired and ret- IABSE Symposium Report. Vol. 105. No. 9. International Association for Bridge and
rofitted bridge piers. Proceedings of the international FIB symposium. 2008. Structural Engineering, 2015.
[2] Billah AHMM, Muntasir M, Alam MS, Bhuiyan MAR. Fragility analysis of retrofitted [19] Lavorato D, Azeredo J, Bergami AV, Salvador Filho JAA, Nuti C, Santini S, et al.
multicolumn bridge bent subjected to near-fault and far-field ground motion. J Column Repaired with HPFRCC and Confined with CFRP: Numerical Analyses to
Bridge Eng 2012;18(10):992–1004. Evaluate the Column Section Capacity. Global Civil Engineering Conference.
[3] Billah AHMM, Alam MS. Seismic performance evaluation of multi-column bridge Singapore: Springer; 2017.
bents retrofitted with different alternatives using incremental dynamic analysis. Eng [20] Rodriguez M, Park R. Seismic load tests on reinforced concrete columns strength-
Struct 2014;62:105–17. ened by jacketing. Struct J 1994;91(2):150–9.
[4] Braga F, Gigliotti R, Laterza M. Analytical stress–strain relationship for concrete [21] Ministry of Communication (MOC) of China. General code for design of highway
confined by steel stirrups and/or FRP jackets. J Struct Eng 2006;132(9):1402–16. bridges and culverts. JTG D60-2004 2004.
[5] Chang GA, Mander JB. Seismic energy-based fatigue damage analysis of bridge [22] Ministry of Housing and Urban-Rural Development (MOHURD) of China. Standard
columns: part 1 - evaluation of seismic capacity. Technical Report NCEER 1994. for test method of mechanical properties of normal concrete. GB/T50152-2012
[6] Choi Y, Yuan RL. Experimental relationship between splitting tensile strength and 2012.
compressive strength of GFRC and PFRC. Cem Concr Res 2005;35(8):1587–91. [23] Moghaddam H, Samadi M, Pilakoutas K, Mohebbi S. Axial compressive behavior of
[7] Dagenais MA, Massicotte B, Boucher-Proulx G. Seismic retrofitting of rectangular concrete actively confined by metal strips; part A: experimental study. Mater Struct
bridge piers with deficient lap splices using ultra high-performance fiber-reinforced 2010;43(10).
concrete. J Bridge Eng 2017;23(2). [24] Nagayama M, Miyashita T. Tensile stress-strain relationship of high-performance
[8] Del Zoppo M, Di Ludovico M, Balsamo A, Prota A. Comparative analysis of existing fiber reinforced cement composites. The Research Report of the Chinese branch of
RC columns jacketed with CFRP or FRCC. Polymers 2018;10(4):361. the Japanese Architecture Society 2008; 31.
[9] Esmaeily A, Xiao Y. Behavior of reinforced concrete columns under variable axial [25] Ou YC, Tsai MS, Chang KC, Lee GC. Cyclic behavior of precast segmental concrete
loads: analysis. ACI Struct J 2005;102(5):736–44. bridge columns with high performance or conventional steel reinforcing bars as
[10] Frangopol D, Sause R, Kusko C. Bridge maintenance, safety, management and life- energy dissipation bars. Earthq Eng Struct Dyn 2010;39(11):1181–98.
cycle optimization. Proceedings of the Fifth International IABMAS Conference, [26] Pan WH, Tao MX, Nie JG. Fiber beam–column element model considering re-
Philadelphia, USA, 11–15 July 2010. CRC Press; 2010. inforcement anchorage slip in the footing. Bull Earthq Eng 2017;15(3):991–1018.
[11] Ghasemi H, Otsuka H, Cooper JD, Nakajima H. Aftermath of the Kobe earthquake. [27] Park R, Kent DC, Sampson RA. Reinforced concrete members with cyclic loading. J
Public Roads 1996;60(2). Struct Division 1972;98(7).
[12] He R, Yang Y, Sneed LH. Seismic repair of reinforced concrete bridge columns: [28] Park SH, Kim DJ, Ryu GS, et al. Tensile behavior of ultrahigh performance hybrid
review of research findings. J Bridge Eng 2015;20(12). fiber reinforced concrete. Cem Concr Compos 2012;34(2):172–84.
[13] Itani R, Liao X. Effects of retrofitting applications on reinforced concrete bridges. [29] Priestley MN, Seible F, Xiao Y, Verma R. Steel jacket retrofitting of reinforced
Columns 2003. concrete bridge columns for enhanced shear strength-part 1: theoretical con-
[14] Japan Road Association (2002) Design Specifications of Highway Bridges. Part V: siderations and test design. Struct J 1994;91(4):394–405.
Seismic Design. Japan. [30] Priestley MJN, Verma R, Xiao Y. Seismic shear strength of reinforced concrete

18
T. Tong, et al. Composite Structures 228 (2019) 111367

columns. J Struct Eng 1994;120(8):2310–29. Eng 2016;14(2):501–27.

[31] Priestley MJN, Seible F, Calvi GM. Seismic design and retrofit of bridges 1996. [38] Tastani SP, Pantazopoulou SJ, Zdoumba D, Plakantaras V, Akritidis E. Limitations
[32] Saadatmanesh H, Ehsani MR, Jin L. Seismic strengthening of circular bridge pier of FRP jacketing in confining old-type reinforced concrete members in axial com-
models with fiber composites. ACI Struct J 1996;93(6):639–738. pression. J Compos Constr 2006;10(1):13–25.
[33] Scott BD, Park R, Priestley MJN. Stress-strain behavior of concrete confined by [39] Teng JG, Fernando D, Yu T. Finite element modelling of debonding failures in steel
overlapping hoops at low and high strain rates. J Am Concr Inst 1982;79(1):13–27. beams flexurally strengthened with CFRP laminates. Eng Struct 2015;86:213–24.
[34] SDC (The Seismic Design Criteria). Caltrans seismic design criteria version 1.7. [40] Xu S, Wu C, Liu Z, et al. Experimental investigation of seismic behavior of ultra-high
California Department of Transportation. Sacramento, CA, USA. 2013. performance steel fiber reinforced concrete columns. Eng Struct 2017;152:129–48.
[35] Seible F, Priestley MN, Hegemier GA, Innamorato D. Seismic retrofit of RC columns [41] Yang C, Okumus P. Ultrahigh-performance concrete for posttensioned precast
with continuous carbon fiber jackets. J Compos Constr 1997;1(2):52–62. bridge piers for seismic resilience. J Struct Eng 2017;143(12):04017161.
[36] Sun Z, Bingjun SI, Wang D, et al. Review on the repair techniques for earthquake [42] Yazıcı H. The effect of curing conditions on compressive strength of ultra high
damaged RC bridge piers. J Earthq Eng Eng Vibr 2009;29(5):128–32. strength concrete with high volume mineral admixtures. Build Environ
[37] Sun Z, Wang D, Bi K, Si B. Experimental and numerical investigations on the seismic 2007;42(5):2083–9.
behavior of bridge piers with vertical unbonded prestressing strands. Bull Earthq

19