Engineering Structures 222 (2020) 111145

Contents lists available at ScienceDirect

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

Seismic performance of a novel interior precast concrete beam-column joint T

using ultra-high performance concrete
Zi-Yu Zhanga, Ran Dingb, , Xin Nieb, Jian-Sheng Fanb
⁎

a
Dept. of Civil Engineering, Tsinghua University, Beijing 100084, China
b
Key Lab. of Civil Engineering Safety and Durability of China Education Ministry, Dept. of Civil Engineering, Tsinghua University, Beijing 100084, China

ARTICLE INFO ABSTRACT

Keywords: In traditional precast reinforced concrete (RC) frame structures, the reinforcement details are usually complex in
Ultra high performance concrete (UHPC) the joint zone, which makes the fabrication process difficult and thus affects the construction quality and speed.
Precast concrete beam-column joints Ultra-high performance concrete (UHPC) exhibits substantially higher shear resistance and bond strength than
Seismic performance normal concrete due to its much larger compressive and tensile strengths. As a result, the application of UHPC
Non-steam-cured
can significantly help reducing the amount of stirrups and anchorage length of the longitudinal bar in the joint
Headed bars
Anchorage length
zone of precast RC frames, which can obviously simplify the joint construction. This paper presents the cyclic
Stirrup ratio tests of four interior precast UHPC/RC composite beam-column joints to evaluate quantitatively the effect of
replacing the normal concrete by UHPC in the joint zone. Considering that the condition of steam curing is
difficult to achieve in the actual construction environment, the non-steam-cured UHPC was adopted and proved
to exhibit adequate compressive (> 120 MPa) and tensile strength (> 7 MPa) with the ultimate tensile strain
larger than 1%. The main variables in this study are the anchorage methods, anchorage lengths and stirrup ratios
in the joint zone. The cyclic load tests demonstrate that the cast-in-situ UHPC can work well with the precast
concrete and there is no need of stirrups in the UHPC joint zone. With the use of UHPC instead of normal
concrete, the anchorage length of the beam straight and headed bars can be decreased to 16db and 8.1db,
respectively. Therefore, the application of UHPC in the joint zone is practicable and can greatly ease the
fabrication process, especially with the use of headed bars.

1. Introduction unsatisfied comprehensive economic benefits.

Past researchers have tried to improve the seismic performance and
Precast reinforced concrete (RC) components (slabs [1], columns simplify the reinforcement details of precast RC joint using various
[2–6], beams [6,7], walls [8,9], etc.) and precast RC structures (frames high-performance materials instead of ordinary concrete. Due to the
[10], shear walls [10], etc.) have been widely applied in the world due high toughness and energy dissipation capacity of steel fiber reinforced
to its high efficiency and quality. In particular, China is now vigorously concrete (SFRC), the precast SFRC joints exhibited superior seismic
promoting the research and application of precast RC structures in both behaviors compared to RC joints [13–15]. Besides SFRC, engineered
public and residential buildings to accelerate the industrialization of cementitious composites (ECC), which is characterized by tension
construction [11]. As is well known, the beam-column joints are the key strain-hardening, multi-cracking and ultra-ductility, is also a new
component in precast RC frame structures to connect the prefabricated choice to improve the seismic performance of the joints. Test results
beams and columns [12]. To satisfy the anchoring demands of pro- demonstrated that compared with traditional RC beam-column joints,
truding rebar of the precast beams and the shear demands in the joint both the cast-in-place and precast ECC beam-column joints show better
core area during earthquakes, the joint reinforcement details are seismic performance in terms of damage tolerance and energy dis-
usually complex. Particularly in high earthquake intensity regions, in sipation even with less or none stirrups in the joint core [16–18]. Gou
order to avoid joint shear failure or local bond failure, a large amount of et al. have reported test results of seven large-scale beam-column joints
longitudinal and transverse reinforcement need to be placed in the joint made with ECC or concrete [19]. The main variables of the study are
core, causing extreme difficulty to the in-situ construction. Thus the quantities of stirrups in the joint zone, rebar diameter in the beam,
construction quality and speed are severely affected, leading to anchorage length of the bottom rebar and axial load. Pinching of the

⁎
Corresponding author.
E-mail addresses: dingran@mail.tsinghua.edu.cn (R. Ding), xinnie@tsinghua.edu.cn (X. Nie), fanjsh@mail.tsinghua.edu.cn (J.-S. Fan).

https://doi.org/10.1016/j.engstruct.2020.111145
Received 26 March 2020; Received in revised form 17 July 2020; Accepted 18 July 2020
Available online 01 August 2020
0141-0296/ © 2020 Elsevier Ltd. All rights reserved.
Z.-Y. Zhang, et al. Engineering Structures 222 (2020) 111145

351 0
157 5 360 157 5
12D20

85
D10@80

550
A B
325 550 700 200 4D16
145 0
350

C C
D10@75 D10@150 D10@80 3D18
A B UHPC joint zone
550

250 360 12D20

4D16 4D16 C C
D10@80 D10@80

85
6D10

360
350

3D18

360
A-A B-B C-C

55
250

288
55

351 0

C-C(P-UHPC1)

(a)
360
Heads
55
250

270
55

351 0

C-C(P-UHPC2)

(b)
360
55

Heads
250

162
55

351 0

C-C(P-UHPC3)

(c)
360
Heads
55
250

270
55

351 0

C-C(P-UHPC4)

(d)
Fig. 1. Reinforcement details of (a) P-UHPC1 (b) P-UHPC2 (c) P-UHPC3 and (d) P-UHPC4.

hysteretic curves appeared in the ECC specimens because of the slip of tensile strength of at least 7 MPa [22] and stable strain hardening
top beam reinforcement passing through the joint. properties [23]. According to direct shear test results, the shear strength
Ultra-high performance concrete (UHPC) is a type of new composite of UHPC can exceed 20 MPa [24], much higher than those of normal
material with the compressive strength of at least 120 MPa [20,21], concrete, SFRC and ECC. As a result, the use of UHPC in the beam-

2
Z.-Y. Zhang, et al. Engineering Structures 222 (2020) 111145

Table 1
Basic parameters of the test specimens.
Specimen Fabrication Joint Stirrups in Anchorage method
method material joint ρsv

P-UHPC1 Precast UHPC 0.56% 16db + straight bar

P-UHPC2 Precast UHPC 0.56% 15db + headed bar
P-UHPC3 Precast UHPC 0.56% 9db + headed bar
P-UHPC4 Precast UHPC 0% 15db + headed bar
C-RC [19] Cast-in-place Concrete 1.12% Continuous through the joint
P-RC [19] Precast Concrete 0% 15db + 90-degree hook

column joint can help substantially reduce the amount of stirrups while columns in the upper floor are connected with longitudinal bars pro-
maintaining equivalent seismic performance compared with the RC truding from the joint by the grout sleeve splicing method or other
beam-column joint. In addition, the bond strength of reinforcement in reliable methods. But for the convenience of specimen fabrication and
the UHPC was found to be higher than 20 MPa and the recommended considering the very little effect on the joint core zone, the two details
anchorage length of straight rebar was no more than 12db (db denotes including the cast-in-situ beam topping and the grouting sleeve con-
the diameter of the rebar) [25–28] under monotonic loading, which is nection of column rebar were not considered in the test. The regions
much less than those embedded in ordinary concrete. Thus it has been where the UHPC was applied are marked in Fig. 1.
demonstrated by all the previous research works that UHPC is a very The section size and the length of the columns were 360 × 360 mm
promising material to solve the rebar congestion problem in precast RC and 1450 mm, respectively, while those of the beams were
beam-column joint. Actually only a few experiments have been carried 250 × 350 mm and 3510 mm, respectively. The reinforcement layouts
out on the cyclic behavior of monolithic UHPC beam-column joint. in the columns were the same for all the four specimens, including
Wang et al. investigated the influence of the axial ratios, stirrup ratios twelve longitudinal bars with a diameter of 20 mm and four-leg stirrups
as well as joint types on the shear capacity of the joints [29]. Other with a diameter of 10 mm and a spacing of 80 mm (ρsv = Ash/
researchers also pointed out that specimens using UHPC have higher bcs = 1.12%, where ρsv was the ratio of transverse reinforcement, Ash
initial cracking strength, bearing capacity and energy dissipation ca- was the total cross-sectional area of transverse reinforcement within
pacity [30,31]. However, very few investigations focused on the ap- spacing s and perpendicular to the column width bc). There were no
plication of UHPC to precast RC beam-column joints. It is still not clear stirrups (ρsv = 0%) in the joint core zone for specimen P-UHPC4 and
that how much stirrups can be eliminated and how many anchorage half of the stirrups (ρsv = 0.56%, two-leg stirrups with a diameter of
lengths can be reduced under cyclic loadings due to the adoption of 10 mm and a spacing of 80 mm) in the joint core zone for specimens P-
UHPC in the joint area. In addition, the conditions of steam curing, UHPC1 ~ 3. The spacings between two adjacent two-leg stirrups with a
which is a common curing method for UHPC to improve strength and diameter of 10 mm in the beams were 80 mm in the plastic hinge and
reduce shrinkage, can hardly be satisfied in the real construction of 150 mm elsewhere. Four longitudinal bars with a diameter of 16 mm
precast buildings. The non-steam-cured UHPC has to be used and the were placed in the top of the beams and extended through the joint.
shrinkage will be large, which might cause potential damage to the Three 18 mm-diameter longitudinal bars with different anchorage
interface between precast concrete and cast-in-situ UHPC. Thus it is also methods and lengths were placed in the bottom of the beams and ter-
an important issue whether UHPC can collaborate well with concrete to minated at different locations in the joint zone.
form a monolithic equivalent beam-column assembly. Important design parameters of the four UHPC/RC composite spe-
In this study, four large-scale interior beam-column joints were cimens as well as two concrete specimens reported by Gou et al. [19]
fabricated and tested cyclically. The UHPC was cast in the joint core are listed in Table 1. Fig. 1 shows the design details including the di-
zone of all the four specimens after the RC beams and columns were mensions and reinforcement layouts of the typical specimens.
prefabricated. In order to simulate the actual construction environment, The ACI 318-14 code [32] suggests a minimum stirrup ratio in the
the non-steam-cured UHPC was used which was just covered with joint zone in order to ensure the axial and shear capacity. Considering
plastic film outdoors and watered several times every day during the that the shear strength of UHPC is much higher than that of normal
curing stage. The effects of the anchorage methods (straight bars or concrete, the amounts of the stirrups in the joint zone of specimens P-
headed bars), anchorage lengths and stirrup ratios in the joint zone on UHPC1 ~ 3 is set to be only about 30% required by the ACI 318-14 [32]
the seismic performance of the precast UHPC/RC composite beam- while no stirrups was applied in the joint zone of specimen P-UHPC4.
column joints were thoroughly investigated. Calculated flexural In order to simplify the reinforcement details, the straight bars were
strength correlates well with the joint strength and demonstrate the used in specimen P-UHPC1 and the headed straight bars were applied in
effects of U-shape bars on the location of plastic hinges. Based on test the other three specimens. According to the ACI 318-14 code [32], for a
results, design recommendations are proposed. straight bar, the development length ld is calculated as Eq. (1):

0.463 × f y db f y db
2. Experimental program ld,318 = N - mm ( lb - in)
fc 26 fc (1)
2.1. Specimen design
where db denotes the nominal diameter of the longitudinal bars, fy and
Considering the much higher cost of UHPC than normal concrete, fc’ denote the yield strength of rebar and cylinder compressive strength
the UHPC was only applied in the joint core zone in this study. The of concrete, respectively. For a headed straight bar, the development
construction processes of the proposed joint in practical projects are as length ldh is calculated as Eq. (2) according to the ACI 318-14 code [32]
follows: (1) precast beams with bottom straight or headed bars pro- and Eq. (3) according to the ACI 352R-02 code [33]:
truding from beam ends are placed on the precast columns in the lower
0.193 × f y db f y db
floor; (2) the beam top longitudinal bars as well as rebar in the slabs ldh,318 = N - mm ( lb - in)
and joint zones are assembled; (3) the joint zones are cast with UHPC fc 62.5 fc (2)
while beam toppings and slabs are cast with concrete; (4) the precast

3
Z.-Y. Zhang, et al. Engineering Structures 222 (2020) 111145

(a) (b)

30 mm

(c) (d)
Fig. 2. Fabrication process of specimen: (a) Reinforcement assembly in the wood mould (b) Size of the anchor head (c) 30 mm grooves on the end of beams and
columns and (4) Placement of U-shape bars.

0.150 × f y db f y db 800με [34] and may cause damage to the interface. Additionally, the
ldh,352 = N - mm ( lb - in)
fc 80 fc interfacial shear sliding failure should be avoided. Thus shear keys and
(3)
additional bars were in need to eliminate these two risks. Hussein et al.
It can be seen that the development length ldh for head bars specified once proposed a new shear key configuration that could enhance the
by the ACI 314-18 code [32] is more conservative than the ACI 352R-02 load capacity under tension or shear loads [35]. However, the shape of
code [33]. For concrete with the cylinder compressive strength as high shear keys was quite complicated and thus leads to construction diffi-
as 120 MPa and steel bar with the yield strength of 500 MPa, the cal- culties. To enhance the integrity of the joint without too much con-
culated ld,318, ldh,318 and ldh,352 are 21.1db, 8.8db and 6.8db. Based on the struction difficulty, the 30 mm grooves were placed on the end of
calculated results above, the straight bars in specimen P-UHPC1 ter- beams as well as columns and the interface was wetted before the cast
minated with the anchorage length of 16db. For specimens P-UHPC2 of UHPC, as suggested by Jang et al. [36]. Fig. 2c shows the grooves in
and P-UHPC4, the anchorage length of the headed straight bars was the column while the grooves in the beams were obscured by the wood
15db and for specimen P-UHPC3, the headed straight bars extended in strip. In addition, three U-shape bars with a diameter of 10 mm were
the joint with 9db. placed at each end of beams to further improve the shear sliding re-
sistance, as shown in Fig. 2d. The total length of U-shape bar was
2.2. Specimen fabrication 300 mm with 200 mm in the beam and 100 mm in the joint zone. The
height of U-shape bar was 170 mm. Four weeks after the cast of beams
First, the reinforcements were assembled in the wood formwork, and columns, the joint zones were cast with the UHPC.
which can be seen in Fig. 2a. After that heads were screwed to the end
of beam bottom longitudinal bars. The external diameter of the head 2.3. Curing regimes of the UHPC
was 43.5 mm, as shown in Fig. 2b. To maintain the integration of
precast joints, the interface between precast concrete and cast-in-situ The curing regimes of the UHPC mainly include the standard curing,
UHPC should be carefully designed. High shrinkage of the UHPC and the steam curing under atmospheric pressure and the autoclave curing
the shear sliding capacity of the interface were two main concerns. [37]. Previous studies have pointed out that steam curing or autoclave
Without steam curing the ultimate shrinkage of the UHPC can exceed curing can further increase the compressive strength of the UHPC

Table 2
Mechanical properties of the UHPC with different curing regimes.
Curing regimes Soaked in Covered with film Steam curing
water and watered

Direct tensile strength 10.5 MPa 10.8 MPa 9.6 MPa

Cylinder compressive strength 131.8 MPa 134.3 MPa 149.0 MPa

4
Z.-Y. Zhang, et al. Engineering Structures 222 (2020) 111145

12 Table 4
Tensile properties of the reinforcing bars.

10 Steel reinforcement Diameter Yield strength Ultimate

(mm) (MPa) strength (MPa)

Longitudinal reinforcement 20 473.6 640.1

8
Tensile stress (MPa)

bars 18 490.8 646.1

16 492.7 676.2

6 Stirrup/U-shape bar 10 469.3 717.1

4
fixed at the self-balancing reaction frame and one actuator each was
connected to the north and south beam. The two actuators moved in the
2 Soaked in water opposite direction to force the joint to rotate clockwise (which is de-
Covered with film and watered fined as the positive direction) or counter-clockwise. When the north
Steam curing beam went down and the south beam went up, the displacement and
0 the force were defined positive. The distance of the two actuators was
0.0 0.5 1.0 1.5 2.0
3150 mm. The translations of both ends of the column were restricted
Tensile strain (%) but the in-plane rotation was allowed. The axial force was applied by
Fig. 3. Typical tensile stress-tensile strain curves of dog-bone UHPC specimens the top hydraulic jack and the axial load ratio N/(bchcfc’) was fixed at
with different curing regimes. 0.10, where N denotes the axial force applied on the top of the column;
bc and hc denote the width and height of the column section, respec-
tively; fc’ denotes the cylinder compressive strength of the concrete used
[38,39]. However, in the real construction of buildings, the conditions
in the columns.
of steam curing or autoclave curing can hardly be satisfied. As for
Fig. 5 shows the loading protocol, as specified by JGJ/T 101-2015
standard curing, the condition of 95% relative humidity is also difficult
[40], a Chinese specification for seismic test of building structures.
to meet. Therefore, to make the application of the UHPC to the building
After applying the axial force, the cyclic force was applied only one
structures practicable, the non-steam-cured UHPC was used in this in-
cycle at each force level. The force increased 0.2Py each cycle until Py,
vestigation. During the curing, four UHPC/RC composite specimens
where Py denotes the predicted yield load. After that, the loading stage
were placed outdoors to simulate the real construction condition. The
turned into the displacement control stage in which each displacement
surfaces of the UHPC were covered with plastic film and watered sev-
level was cycled three times. The applied displacement (Δ) was in-
eral times each day. To reveal the influence of curing regimes on the
creased by Δy (where Δy denotes the predicted yield displacement
mechanical properties of UHPC, material tests with different curing
corresponding to the predicted yield load Py) each time until failure.
regimes were carried out first. The results including the average com-
The lateral and axial loads induced by the actuators were measured
pressive and tensile strengths are listed in Table 2. Fig. 3 shows typical
by the load cell in each actuator. Additionally, photogrammetry method
tensile stress-strain curves of dog-bone specimens with different curing
was used to monitor the deformation of the joint zone. As can be seen in
methods. Test results demonstrate that high humility has little influence
Fig. 4, twenty-five markers were placed uniformly on the surface of the
on the compressive strength, tensile strength and tensile ductility while
joint zone. During the whole process of the experiment, the translations
high temperature can only affect the compressive strength. Even
of every marker could be measured and recorded. Moreover, strain
without high temperature and humility, the UHPC applied in this study
gauges were placed on the different positions of the longitudinal bars
can still achieve high compressive and tensile strength with superior
and stirrups to monitor the rebar strains during the test.
ductility.

2.4. Material property 3. Test results

The UHPC material used in this study contains 2.5% (by volume) For the specimen P-UHPC1 in which the beam bottom longitudinal
straight steel fibers. The diameter and the length of the straight steel bars terminated straightly without the use of heads, flexural cracks with
fiber were 0.2 mm and 13 mm, respectively. The tensile strength and the maximum width of 0.04 mm were observed both at the bottom of
the elastic modulus were 2700 MPa and 210 GPa, respectively. The the south beam and the top of the north beam near the joint zone at the
water-cement ratio was extreme low to obtain the high compressive load of 15 kN, as shown in Fig. 6a. With the increase of the load, new
strength, thus a large amount of water reducer was needed. The specific flexural cracks formed farther way from the joint zone and the original
proportions of the UHPC material are listed in Table 3. The cylinder cracks became longer and wider. At the load of 45 kN, a hairline shear
compressive strength of the concrete used both in the beams and col- diagonal crack with the width of only 0.01 mm could be seen on the
umns was 45.9 MPa and that of UHPC in the joint zone was 144.0 MPa. surface of the joint zone while the maximum width of the flexural
The tensile strength of UHPC in the joint zone was 9.3 MPa. The yield cracks at the beam plastic hinge zone reached 0.2 mm, which can be
and ultimate strength of the HRB400 steel bars are listed in Table 4. seen in Fig. 6b. Meanwhile, the width of cracks at the interface between
the normal concrete and the UHPC also reached 0.3 mm. At the drift
2.5. Test procedure ratio of 2.14% (30 mm), the flexural crack where the U-shape bars
terminated suddenly opened up and the maximum width reached as
Fig. 4 shows the test setup of the beam-column joint. The joint was large as 4.50 mm, as shown in Fig. 6c. With the increase of the drift

Table 3
Mix proportions for 1 m3 UHPC material.
Component Cement Silica Fume Sand Superfine mineral powder Fiber Water Water reducer

Proportion (kg) 870 180 1000 240 195 199 24

5
Z.-Y. Zhang, et al. Engineering Structures 222 (2020) 111145

Axial actuator

Hinge support
North actuator
South actuator

Marker

Self-balancing reaction frame

Hinge support
South North
Fig. 4. Test setup.

Load control intact, as can be seen in Fig. 7a. Numerous closely spaced diagonal
hairline cracks were found in succession and the maximum width of the
1.0
0.8 diagonal crack was smaller than 0.05 mm. Similarly, the crack width at
0.6 the interface between the precast concrete and cast-in-situ UHPC was
0.4 minor. As can be seen in Fig. 7a, the specimen P-UHPC1 experienced
0.2 typical flexural failure mode near the end of the beams.
Load (P/Py)

0 For the other three UHPC/RC composite beam-column joint speci-

-0.2 mens, as shown in Fig. 7b-7d, the general behaviors and failure modes
-0.4 were nearly the same to those of P-UHPC1, so they are not described in
-0.6 detail for simplicity. For all the four specimens, the interface between
-0.8 precast concrete and cast-in-situ UHPC was in good condition before
-1.0 the test and the damage was minor after the test, demonstrating that the
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
cast-in-situ UHPC can collaborate well with the precast concrete in the
Cycles beam-column joints. In comparisons, the maximum width of diagonal
Displacement control crack in the joint zone of specimen P-UHPC3 was 0.13 mm and was
larger than 0.07 mm of P-UHPC4, 0.05 mm of P-UHPC1 as well as
7
0.04 mm of P-UHPC2. This phenomenon was probably due to the po-
6
5 sition of heads of the beam bottom longitudinal bars. It can be seen that
4 the widest diagonal crack of specimen P-UHPC3 was located in the
3 lower right corner of the joint core zone and passed through the loca-
Displacement (∆/∆y)

2 tion of the heads. Because the anchorage length of headed straight bars
1 was only 9db, the heads had to provide relatively larger anchoring
0 force. Compared with the more uniform distribution of bond stress in
-1 the other three UHPC/RC specimens, the concentrated tensile force
-2 applied to the UHPC by the bar heads in specimen P-UHPC3 was much
-3 larger and further increased the damage of joint zone.
-4 The dimensions of four specimens in this study were the same as
-5 those of specimens C-RC and P-RC reported by Gou et al. [19] except
-6
that the size of columns increased a little from 350 × 350 mm to
-7
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 360 × 360 mm. Compared with C-RC and P-RC [19], the number of
Cycles beam top bars increased from 3 to 4 while the diameter decreased from
18 mm to 16 mm. Thus the total area of beam top bars in UHPC/RC
Fig. 5. Cyclic loading history. specimens was nearly the same as that in two concrete specimens.
These changes were made to meet the ACI limit on the anchorage of the
ratio, the diagonal flexure-shear crack at the end of the beams got wider beam longitudinal bars passing through the joint zone [32]. Besides,
and also became a part of the plastic hinge. It is noteworthy that the each displacement level cycled three times for UHPC/RC specimens
shape of the plastic hinge region was just like an arc due to the ex- while each displacement level only cycled twice for two RC specimens.
istence of additional U-shape bars, which can be seen in Fig. 6d. At the The other differences between two series of specimens include mate-
later stage of the test, spalling of concrete at the plastic hinge region rials used in the joint zone, the anchorage methods and lengths of beam
became more and more severe. However, the joint zone remained quite bottom bars, which can be seen in Table 1. In general, the two series of

6
Z.-Y. Zhang, et al. Engineering Structures 222 (2020) 111145

Fig. 6. Observed phenomena of P-UHPC1: (a) at the load of 15 kN (b) at the load of 45 kN (c) at the displacement of 30 mm and (d) at the displacement of 45 mm.

specimens were very similar, and thus the results can be compared. For specimens. As for the precast RC joint specimen P-RC without stirrups
the cast-in-place monolithic RC joint specimen C-RC with adequate in the joint zone [19], the joint zone underwent the typical shear
stirrups [19], the specimen failed with the formation of typical flexural failure, as shown in Fig. 8b. For the UHPC/RC composite specimen P-
plastic hinges at the beam ends, similar to the four UHPC/RC composite UHPC4 without stirrup in the joint zone, the joint zone was still in good
specimens investigated in this study. However, the beam bottom condition after the test and the maximum width of the diagonal crack
longitudinal bar was pulled out in the final stage and the concrete was only 0.07 mm. Based on the above comparisons, conclusions can be
crushed at the corner of the joint zone in specimen C-RC [19], as shown drawn that the UHPC has much higher shear strength and better an-
in Fig. 8a, which was not found in the four UHPC/RC composite chorage performance than normal concrete.

Fig. 7. Failure modes of (a) P-UHPC1 (b) P-UHPC2 (c) P-UHPC3 and (d) P-UHPC4.

7
Z.-Y. Zhang, et al. Engineering Structures 222 (2020) 111145

Fig. 8. Failure modes of (a) C-RC and (b) P-RC [19].

4. Experimental analysis However, for the north beam, the load still maintained as high as
0.65Pmax in the positive direction and 0.76Pmax in the negative direc-
4.1. Load-displacement curves tion when the displacement reached 6Δy. In general, the hysteretic load-
displacement curves of two beams both were plump. Although the
Table 5 summarizes some important results, including the yield load anchorage length of the straight bars in the joint zone was much smaller
Py and the corresponding drift ratio θy, the maximum load Pmax and the than the requirement specified by the ACI 318-14 code [32], no slip of
corresponding drift ratio θmax, 85% of the maximum load P0.85 and the the longitudinal bars or obvious pinching effect was found in specimen
corresponding drift ratio θ0.85, the ductility factor μ (equals to θ0.85/θy) P-UHPC1, demonstrating that the straight bar with an reduced devel-
and the peak shear stress vj (the maximum value calculated by Eq. (6) opment length can be reliably applied in the UHPC joint zone.
during the test). The yield load Py and the corresponding drift ratio θy For the other three UHPC/RC composite beam-column joint speci-
were decided by using the equivalent energy method [41] for all the six mens of which the beam bottom longitudinal bars were headed, the
specimens. Fig. 9 shows the hysteretic load-displacement curves of the hysteretic load-displacement curves were very similar to that of P-
four precast UHPC/RC composite beam-column joint specimens as well UHPC1, as shown in Fig. 9b-d.
as two concrete specimens reported by Gou et al. [19]. The influence of The hysteretic load-displacement curves of the two RC specimens
the rotation of the columns on the vertical displacement of the beams reported by Gou et al. [19] can be seen in Fig. 9e and f. For the cast-in-
has been eliminated. place monolithic RC joint specimen C-RC, due to the failure mode of the
For the specimen P-UHPC1, before the predicted yield force, the typical flexural plastic hinges at the beam ends, the hysteretic load-
hysteretic load-displacement curve was nearly linear. At the drift ratio displacement curve was also plump. However, typical shear failure
of 2.61% and −4.78%, the load reached its maximum value of 88.52 happened in the precast concrete joint specimen P-RC and thus the
kN and −87.59 kN for the south beam. For the north beam, the max- hysteretic curve was quite pinching. Overall, in terms of failure modes
imum load was 93.51 kN and −88.3 kN with the drift ratio of 2.66% and hysteretic curves, the four precast UHPC/RC composite specimens
and −3.55%. Then during the process that the displacement increased exhibited equivalent seismic performance compared with the cast-in-
from 5Δy to 6Δy, the load decreased by 0.43Pmax in the positive direc- place monolithic RC specimen and much better seismic performance
tion and 0.42Pmax in the negative direction for the south beam. than the precast concrete joint specimen P-RC without stirrups in joint

Table 5
Summary of experimental results.
Specimen Direction Yielding load Peak load Ultimate load Ductility factor μ Peak shear stress vj/MPa

Py (kN) θy Pmax (kN) θmax P0.85 (kN) θ0.85

P-UHPC1 Positive South 80.4 1/52.6 88.5 1/38.3 75.2 1/17.9 2.94 7.05
North 83.8 1/50.0 93.5 1/37.6 79.5 1/19.0 2.63
Negative South 78.8 1/56.1 87.6 1/20.9 74.4 1/17.7 3.17 6.70
North 80.8 1/58.3 88.3 1/28.1 75.0 1/17.0 3.43

P-UHPC2 Positive South 78.6 1/65.1 89.9 1/28.0 76.4 1/19.8 3.29 6.82
North 77.4 1/65.9 89.5 1/37.1 76.1 1/18.6 3.54
Negative South 77.0 1/60.5 88.0 1/45.6 74.8 1/19.3 3.13 6.83
North 77.6 1/51.8 88.1 1/27.7 74.9 1/18.3 2.83

P-UHPC3 Positive South 77.2 1/61.4 82.6 1/28.6 70.2 1/18.6 3.30 6.91
North 85.4 1/57.7 97.0 1/39.3 82.5 1/17.9 3.22
Negative South 80.6 1/70.9 91.0 1/29.2 77.3 1/16.8 4.22 6.64
North 77.8 1/74.0 80.0 1/29.3 68.0 1/18.2 4.07

P-UHPC4 Positive South 76.6 1/67.5 86.0 1/37.1 73.1 1/18.5 3.65 6.78
North 83.9 1/56.0 98.4 1/38.8 83.7 1/18.0 3.11
Negative South 81.1 1/53.3 90.9 1/21.7 77.2 1/18.1 2.94 6.79
North 77.0 1/56.4 84.9 1/28.8 72.2 1/20.5 2.75

C-RC [19] Positive South 80.4 1/76.3 86.6 1/23 73.6 1/17 4.49 7.51
Negative South 74.9 1/71.2 84.4 1/27 71.7 1/17 4.19 7.55

P-RC [19] Positive South 78.1 1/63.2 85.7 1/61 72.9 1/17 3.72 7.53
Negative South 75.1 1/75.9 79.8 1/96 67.8 1/18 4.22 7.61

8
Z.-Y. Zhang, et al. Engineering Structures 222 (2020) 111145

Rotation (%) Rotation (%)

-8.6 -5.7 -2.9 0.0 2.9 5.7 8.6 -8.6 -5.7 -2.9 0.0 2.9 5.7 8.6
120 120

80 80

40 40
Load (kN)

Load (kN)
0 0

-40 -40

-80 -80

-120 -120
-120 -80 -40 0 40 80 120 -120 -80 -40 0 40 80 120
Displacement (mm) Displacement (mm)

(a)
Rotation (%) Rotation (%)
-8.6 -5.7 -2.9 0.0 2.9 5.7 8.6 -8.6 -5.7 -2.9 0.0 2.9 5.7 8.6
120 120

80 80

40 40
Load (kN)

Load (kN)
0 0

-40 -40

-80 -80

-120 -120
-120 -80 -40 0 40 80 120 -120 -80 -40 0 40 80 120
Displacement (mm) Displacement (mm)

(b)

Rotation (%) Rotation (%)

-8.6 -5.7 -2.9 0.0 2.9 5.7 8.6 -8.6 -5.7 -2.9 0.0 2.9 5.7 8.6
120 120

80 80

40 40
Load (kN)

Load (kN)

0 0

-40 -40

-80 -80

-120 -120
-120 -80 -40 0 40 80 120 -120 -80 -40 0 40 80 120
Displacement (mm) Displacement (mm)
(c)
Fig. 9. Hysteresis and skeleton load–displacement curves of (a) P-UHPC1 (b) P-UHPC2 (c) P-UHPC3. (d) P-UHPC4 (e) C-RC [19] and (f) P-RC [19].

9
Z.-Y. Zhang, et al. Engineering Structures 222 (2020) 111145

Rotation (%) Rotation (%)

-8.6 -5.7 -2.9 0.0 2.9 5.7 8.6 -8.6 -5.7 -2.9 0.0 2.9 5.7 8.6
120 120

80 80

40 40
Load (kN)

Load (kN)
0 0

-40 -40

-80 -80

-120 -120
-120 -80 -40 0 40 80 120 -120 -80 -40 0 40 80 120
Displacement (mm) Displacement (mm)
(d)
Rotation (%) Rotation (%)
-8.6 -5.7 -2.9 0.0 2.9 5.7 8.6 -8.6 -5.7 -2.9 0.0 2.9 5.7 8.6
120 120

80 80

40 40
Load (kN)

Load (kN)

0 0

-40 -40

-80 -80

-120 -120
-120 -80 -40 0 40 80 120 -120 -80 -40 0 40 80 120
Displacement (mm) Displacement (mm)

(e) (f)
Fig. 9. (continued)

zone. specimens while the differences among other five specimens were in-
The comparisons of the load-displacement envelopes for the six appreciable. Then at the later phase of the loading process, the dis-
specimens can be seen in Fig. 10. The curves of four UHPC specimens sipated energy of all the specimens decreased, mainly due to the de-
nearly coincide with each other. For these four specimens, the negative crease of the load. In general, the four UHPC/RC composite specimens
peak load should be a little larger than the positive peak load for the exhibited similar performance compared with the cast-in-place mono-
south beams and the result is contrary for the north beams, because the lithic RC specimen in energy dissipation capacity.
total area of the beam top longitudinal bars is slightly larger than the Besides the amount of the dissipated energy, the equivalent
bottom bars. As can be seen in Table 5, most results are in accordance damping coefficient he is also an important index to reflect the effi-
with this conclusion. Ductility coefficient defined as θ0.85/θy are also ciency of the energy dissipation [42]. The equivalent damping coeffi-
listed in Table 5. It can be seen that all the values ranged from 2.5 to cient is calculated by Eq. (4):
4.5. In general, all the specimens exhibited similar initial stiffness,
bearing capacity and ductility. 1 S1
he =
2 S2 (4)

4.2. Energy dissipation capacity where S1 and S2 denote the areas enclosed by different curves which
can be referred in Fig. 12.
The dissipated energy in the first cycle at each displacement level is Fig. 13 shows the relationship between the equivalent damping
plotted with different beam rotation levels in Fig. 11. The value of the coefficient and the drift ratio for the six specimens. It can be seen that
dissipated energy in Fig. 11 equals to the area surrounded by the load- except specimen P-RC [19] with relatively low he values, the he values
displacement curve in each cycle, i.e. the area S1 shown in Fig. 12. Due of the other five specimens were very close to each other and were
to the pinching effect caused by the shear failure of the joint zone, the always increasing with the drift ratio. In the yielding stage, the he va-
dissipated energy of specimen P-RC [19] is the minimum of the six lues for the five specimens P-UHPC1 ~ 4 and C-RC [19] ranged around

10
Z.-Y. Zhang, et al. Engineering Structures 222 (2020) 111145

Rotation (%) Rotation (%)

-8.6 -4.3 0.0 4.3 8.6 -8.6 -4.3 0.0 4.3 8.6
120 120
P-UHPC1 P-UHPC1
P-UHPC2 P-UHPC2
P-UHPC3 P-UHPC3
60 P-UHPC4 60 P-UHPC4
C-RC
P-RC
Load (kN)

Load (kN)
0 0

-60 -60

-120 -120
-120 -60 0 60 120 -120 -60 0 60 120
Displacement (mm) Displacement (mm)
(a) (b)
Fig. 10. Load-displacement envelope curves of (a) the north beams and (b) the south beams.

0.2. Then after the drift ratio of 6.5%, the he values increased to 0.25 or are plotted against the beam rotations in Figs. 14 and 15.
larger. Hence conclusions can be drawn that even the load decreased For UHPC/RC composite specimens, the λj,2 values ranged from 0.9
obviously in the later stage, the efficiency of energy dissipation for four to 1 before the drift ratio of 4.3%, which is very similar to those of C-
precast UHPC/RC composite specimens can still maintain a high level RC, indicating that the peak load of the second cycle was nearly the
and he values of them are comparable to those of the monolithic spe- same as that of the first cycle. However, as for specimen P-RC, the λj,2
cimen C-RC. values were smaller than 0.9 at the beginning of the displacement
control stage, demonstrating that the bearing capacity decreased more
4.3. Degradation of strength obviously compared with the other five specimens. Because the loading
on C-RC and P-RC reported by Gou et al. [19] were only cycled twice at
The phenomenon that a structural component loses its bearing ca- each displacement level, Fig. 15 only plots the UHPC/RC composite
pacity with the repeated cycles at the same drift ratio is called the specimens. It can be seen that λj,3 values decreased linearly with the
strength degradation and is described by the strength degradation increase of the rotation level. Compared with the results of λj,2 values, it
coefficient λj. In this study, λj,2 denotes the ratio between the peak load can be seen that the third cycle caused more damage to the specimens.
of the second cycle and that of the first cycle at the same displacement The developing tendency of the four curves were almost identical, il-
level while λj,3 denotes the ratio between the peak load of the third lustrating that the anchorage methods or lengths and the stirrup ratios
cycle and that of the first cycle. The λj,2 and λj,3 of different specimens in the joint zone had no influence on the strength degradation.

Rotation (%) Rotation (%)

Cumulated dissipated Energy(x104kN·mm)

Cumulated dissipated Energy(x104kN·mm)

0.0 2.1 4.3 6.4 8.6 0.0 2.1 4.3 6.4 8.6
12 12
P-UHPC1 P-UHPC1
P-UHPC2 P-UHPC2
10 P-UHPC3 10 P-UHPC3
P-UHPC4 P-UHPC4
8 8 C-RC
P-RC

6 6

4 4

2 2

0 0
0 30 60 90 120 0 30 60 90 120
Displacement (mm) Displacement (mm)

(a) (b)
Fig. 11. Energy dissipation each cycle versus displacement curves of (a) north beams and (b) south beams.

11
Z.-Y. Zhang, et al. Engineering Structures 222 (2020) 111145

considered and bj denotes the effective joint width.

The typical measured shear stress versus shear deformation curves
are shown by Fig. 17. The maximum shear deformation of the UHPC/
RC composite specimens were less than 0.001 rad. However, the max-
imum shear deformation of the RC specimens C-RC and P-RC reached
0.004 rad and 0.05 rad [19], respectively. The maximum diagonal
crack width of UHPC specimens was only 0.13 mm, much smaller than
Load

0.3 mm of the C-RC and 0.6 mm of the P-RC [19]. The comparisons
among the maximum diagonal crack widths also illustrates that the
S1 damage due to the shear deformation of four UHPC/RC composite
specimens is relatively small.
S2 The maximum shear stresses of different specimens are listed in
Table 5. Wang et al. also studied the shear capacity of monolithic UHPC
joints [29]. To obtain the shear strength of the UHPC beam-column
joints, totally nine UHPC beam-column joint specimens were designed
Displacement based on the weak joint–strong component principle so that shear
failure would happen in the joint core [29]. Most specimens reported by
Fig. 12. Typical hysteresis loop and energy dissipation.
Wang et al. exhibited a shear strength larger than 7.05 MPa [29], which
is the maximum shear stress of specimens tested in this study, de-
4.4. Joint shear behavior monstrating the reliability of the precast UHPC joints.
According to the ACI 318-14 code [32], the nominal shear stress of
In order to ensure the enough ductility of the frame structure, the vn = 5/4 fc (MPa) is allowed for RC beam-column joints. The calcu-
shear failure of the joint zone should be avoided. In this study, the lated vn values based on the cylinder compressive of the UHPC and the
displacement of twenty-five markers located at the joint zone were concrete are 15.0 MPa and 8.5 MPa, respectively. Considering the ultra-
monitored and the joint shear deformation can be calculated based on high compressive strength of UHPC, the nominal shear stress of the
the data of four markers at the four corners of the joint zone while the UHPC is difficult to achieve unless excessive beam longitudinal bars are
same calculation method using data of other markers can verify the placed. On the other hand, the maximum shear stresses of UHPC/RC
results. The shear deformation in this study is calculated by Eq. (5), composite specimens were close to the nominal shear stress of the
where a and b denote the lengths of two sides of the joint which are concrete and the joint zones were in good condition after the tests, also
assumed constant during the test; c1 and d1 denote the lengths of ori- demonstrating that the UHPC can be used in the joint zone without any
ginal diagonals while c2 and d2 denote the lengths of deformed diag- stirrups safely.
onals, which can be referred as Fig. 16.

a2 + b2 4.5. Effects of anchorage methods, anchorage lengths and stirrup ratios on

= (|c1 c2| + |d1 d2|)
(5) seismic performance
2·a ·b
The shear stress ν in the joint zone can be calculated by the method The anchorage methods investigated in this study include two ways:
in ACI 352R-02 code [33], as shown in Eq. (6). straight bars and headed bars and their influences can be revealed by
comparing specimen P-UHPC1 and P-UHPC2. The maximum load of
= (C + T Vcol )/ hc / bj (6)
specimen P-UHPC1 with beam bottom straight bars is 4.0% larger than
where C, T and Vcol denote the compressive force from the beam, tensile that of specimen P-UHPC2 with beam bottom headed bars, indicating
force from the beam, and the shear force of the column, respectively; hc the negligible difference in the maximum load. As for energy dissipa-
denotes the depth of the column in the direction of joint shear being tion capacity, these two specimens exhibited similar performance

Rotation (%) Rotation (%)

0.0 2.1 4.3 6.4 8.6 0.0 2.1 4.3 6.4 8.6
0.4 0.4
P-UHPC1 P-UHPC1
Equivalent damping coefficient

Equivalent damping coefficient

P-UHPC2 P-UHPC2
P-UHPC3 P-UHPC3
0.3 P-UHPC4 0.3 P-UHPC4
C-RC
P-RC

0.2 0.2

0.1 0.1

0.0 0.0
0 30 60 90 120 0 30 60 90 120
Displacement (mm) Displacement (mm)
(a) (b)
Fig. 13. Equivalent damping coefficient versus displacement curves of (a) north beams and (b) south beams.

12
Z.-Y. Zhang, et al. Engineering Structures 222 (2020) 111145

Rotation (%) Rotation (%)

-8.6 -4.3 0.0 4.3 8.6 -8.6 -4.3 0.0 4.3 8.6
1.1 1.1

1.0 1.0

0.9 0.9
λ j,2

λ j,2
0.8 0.8

P-UHPC1
0.7 0.7
P-UHPC2
P-UHPC1 P-UHPC3
0.6 P-UHPC2 0.6 P-UHPC4
P-UHPC3 C-RC
P-UHPC4 P-RC
0.5 0.5
-120 -60 0 60 120 -120 -60 0 60 120
Displacement (mm) Displacement (mm)

(a) (b)
Fig. 14. Strength degradation coefficient λj.2 versus displacement curves of (a) north beams and (b) south beams.

before the drift ratio of 5.4%, as shown in Fig. 11. After the drift ratio of
5.4%, the load of specimen P-UHPC2 decreased a little more rapidly, as
can be seen in Fig. 10. As a result, the dissipated energy of specimen P-
UHPC1 is larger than that of specimen P-UHPC2 later on. However, it d2
can be seen from Fig. 13 that the equivalent damping coefficients of c1 d1 a c2
these two specimens are nearly the same in all stages of loading, in-
dicating similar efficiency of the energy dissipation. Overall, the spe-
cimens with two anchorage methods exhibit comparable performance
in load capacity, energy dissipation and strength degradation.
The effects of anchorage lengths of headed bars are analyzed in this b
study by comparing specimens P-UHPC2 and P-UHPC3. It can be found
from Figs. 11, 14 and 15 that, the specimen P-UHPC3 with an ancho-
(a) Original (b) Deformed
rage length of 9db exhibited similar and even a little higher bearing
capacity and energy dissipation as well as slower strength degradation Fig. 16. Shear deformation of the joint zone.
than specimen P-UHPC2 with an anchorage length of 15db,

Rotation (%) Rotation (%)

-8.6 -4.3 0.0 4.3 8.6 -8.6 -4.3 0.0 4.3 8.6
1.1 1.1

1.0 1.0

0.9 0.9

0.8 0.8
λ j,3

λ j,3

0.7 0.7
P-UHPC1 P-UHPC1
P-UHPC2 P-UHPC2
0.6 P-UHPC3 0.6 P-UHPC3
P-UHPC4 P-UHPC4

0.5 0.5
-120 -60 0 60 120 -120 -60 0 60 120
Displacement (mm) Displacement (mm)
(a) (b)
Fig. 15. Strength degradation coefficient λj.3 versus displacement curves for four UHPC/RC composite specimens of (a) north beams and (b) south beams.

13
Z.-Y. Zhang, et al. Engineering Structures 222 (2020) 111145

The norminal shear stress of concrete The norminal shear stress of concrete
8 8

Shear stress (MPa)

Shear stress (MPa)

4 4

0 0

-4 -4

-8 -8
The norminal shear stress of concrete The norminal shear stress of concrete
-0.8 -0.4 0.0 0.4 0.8 -0.8 -0.4 0.0 0.4 0.8
-3 -3
Shear deformation (x10 rad) Shear deformation (x10 rad)

(a) (b)
Fig. 17. Shear stress vs shear distortion of (a) P-UHPC1 and (b) P-UHPC3.

Fig. 18. Placement of strain gauges.

100 100

50 50
Load (kN)

Load (kN)

0 0

-50 -50

B1 T1
Yielding strain Yielding strain
-100 -100
-10000 -5000 0 5000 10000 15000 -2000 0 2000 4000 6000 8000
Strain (µε) Strain (µε)

(a) (b)
Fig. 19. Typical load- strain hysteretic curves (a) B1 and (b) T1.

14
Z.-Y. Zhang, et al. Engineering Structures 222 (2020) 111145

shown in Fig. 18. During the test, the maximum strains of the column
South North longitudinal bars and beam stirrups are only 862 με and 1385 με, re-
spectively. As for the joint stirrups, the maximum strain is 1090 με. All
of the bars mentioned above maintained a low level of strain. It can be
seen that the stirrups in the joint zone cannot be made full use of, due to
the limited shear deformation and diagonal crack width, indicating that
the stirrups may be totally removed.
On the other hand, the strains in the beam bottom and top long-
itudinal bars are relatively high, especially in the plastic hinge zone.
The beam longitudinal bars of all the specimens yielded. The typical
measured load-strain hysteretic curves are plotted in Fig. 19. The strain
Buckling of Buckling of
gauges B1 and T1 were attached in the plastic hinge region of the beam
the rebar the rebar
bottom and top longitudinal bar, respectively. For the top rebar, the
Fig. 20. Buckling of the beam bottom longitudinal bars. tensile strain is obviously larger than the compressive strain and when
the load reached the maximum, the tensile strain also achieved its
maximum value which is more than twice of the yielding strain. As for
Concrete UHPC joint zone Concrete the bottom bar, it’s worth noting that the negative strain reached the
beam beam minimum during the process of reverse unloading rather than when the
displacement reached the minimum, which is probably due to the
buckling of the bottom bar. After the formation of flexural hinge, the
6db B3 whole section of the beam was seriously damaged. When the bar was
P-UHPC3 compressed, the cracked concrete could still be compacted, thus pro-
12db B3 viding a certain degree of constraint for the rebar. However, when the
compressive load began to unload, the confinement by the concrete
P-UHPC2
decreased rapidly and the buckling of the rebar might occur. As shown
14db
in Fig. 20, buckling of the bottom longitudinal bars of both the south
db P-UHPC1 and north beam was found while no buckling occurred in the beam top
B1 B2 B3 longitudinal bars, which may be related to the larger total area.
In this study, the anchorage performance of the rebar in the UHPC is
a key issue. The strains of the beam bottom longitudinal bars at dif-
ferent locations were monitored. As shown in Fig. 21, the gauges B1, B2
Fig. 21. Layout of the strain gauges on the beam bottom longitudinal bars.
and B3 were placed at the flexural plastic hinge region, the edge of the
joint zone and the end of the rebar, respectively. Fig. 22 shows the
demonstrating that the development length of 9db is enough in the measured strain-drift ratio skeleton curves of B1 ~ B3 in three speci-
UHPC joint zone for headed bars and can even be smaller. mens with different anchorage methods or lengths. The results of the B1
The reinforcement layout of specimen P-UHPC4 is the same as that indicate that the bottom bar of all the specimens yielded in the flexural
of specimen P-UHPC2, except that there is no stirrup in the joint zone of plastic hinge region. The B1 in P-UHPC2 with the most adequate an-
specimen P-UHPC4. However, the load-displacement curves of these chorage degree yielded prior to the other two specimens. The drift ratio
two specimens were very similar, proving that the two specimens also corresponding to the yielding increase with the decrease of the an-
exhibited comparable performance in bearing capacity, energy dis- chorage degree. As for the B2, the curves of different specimens were
sipation and strength degradation. Besides, the maximum width of di- similar. First the strain increased linearly and then yielded. Compared
agonal cracks in the joint zone of the specimen P-UHPC4 was also very with the results of the B1, the strain values of the B2 were always po-
small, further demonstrating that the effect of joint stirrups is not ob- sitive, indicating no buckling took place at the position of B2. The B3
vious which could be totally removed. strains were relatively small in the P-UHPC1 and P-UHPC2 because the
anchorage lengths were long enough while the B3 strain in P-UHPC3
4.6. Strain analysis was much larger. The largest strain value of the B3 in P-UHPC3 ex-
ceeded 1500 με, proving that the head can provide adequate anchoring
Strain gauges were placed at different positions on the rebar, as

Rotation(%) Rotation (%) Rotation (%)

0.0 2.1 4.3 6.4 8.6 0.0 2.1 4.3 6.4 8.6 0.0 2.1 4.3 6.4 8.6
3000 2500
6000 P-UHPC1
P-UHPC2
2000 P-UHPC3
Yielding strain
3000 2000
1500
Strain (µε)
Strain (µε)

Strain (µε)

0
1000
1000

-3000 P-UHPC1 P-UHPC1 500

P-UHPC2 P-UHPC2
P-UHPC3 P-UHPC3
Yielding strain Yielding strain
-6000 0 0
0 30 60 90 120 0 30 60 90 120 0 30 60 90 120
Displacement (mm) Displacement (mm) Displacement (mm)

(a) (b) (c)

Fig. 22. Strain-displacement envelope curves (a) B1 (b) B2 and (c) B3 for specimens P-UHPC1 ~ 3.

15
Z.-Y. Zhang, et al. Engineering Structures 222 (2020) 111145

B1 B2 B3 B1 B2 12db B3
14db
db
Concrete beam UHPC joint zone Concrete beam UHPC joint zone
Rotation (%) Rotation (%)
0.0 2.1 4.3 6.4 8.6 0.0 2.1 4.3 6.4 8.6
500 500

400 400
Stress (MPa)

Stress (MPa)
300 300

461.6 MPa 392.9 MPa

200 B2 200 B2
B3 B3
Yielding stress Yielding stress
100 100

0 0

0 30 60 90 120 0 30 60 90 120
Displacement (mm) Displacement (mm)
(a) (b)

B1 B2 B3 B1 B2 12db B3
6db

Concrete beam UHPC joint zone Concrete beam UHPC joint zone
Rotation (%) Rotation (%)
0.0 2.1 4.3 6.4 8.6 0.0 2.1 4.3 6.4 8.6
500 500

400 400
177.0 MPa
Stress (MPa)

Stress (MPa)

300 300
412.0 MPa
200 200

100 100

B2 B2
0 B3 0 B3
Yielding stress Yielding stress
0 30 60 90 120 0 30 60 90 120
Displacement (mm) Displacement (mm)
(c) (d)
Fig. 23. Stress-displacement envelope curves calculated using gauges B2 and B3 for (a) P-UHPC1 (b) P-UHCP2 (c) P-UHPC3 and (d) P-UHCP4.

Table 6
Summary of bond stresses in UHPC joint zone.
Specimen Δσ MPa s τ/MPa Calculation formula

P-UHPC1 461.6 14db (equals to 252 mm) 8.2 (Δσ × db2 × π/4)/(π × db × 14db)
P-UHPC2 392.9 12db (equals to 216 mm) 8.2 (Δσ × db2 × π/4)/(π × db × 12db)
P-UHPC3 177.0 6db (equals to 108 mm) 7.4 (Δσ × db2 × π/4)/(π × db × 6db)
P-UHPC4 412.0 12db (equals to 216 mm) 8.6 (Δσ × db2 × π/4)/(π × db × 12db)
The minimum anchorage length of the headed bar in the UHPC (6db × 7.4 MPa/8.6 MPa + (9db −6db)) = 8.1db

16
Z.-Y. Zhang, et al. Engineering Structures 222 (2020) 111145

Table 7
Theoretical and test values of flexural capacity at different sections.
Section Calculated maximum bending moment Mc1 and Calculated maximum bending moment Mc2 and Measured maximum bending moment Mt and
corresponding load corresponding load corresponding load

A-A Positive 128.0 kN·m (107.1 kN) 113.0 kN·m (94.6 kN) 108.6 kN·m (92.0 kN)
Negative 122.7 kN·m (102.7 kN) 107.6 kN·m (90.0 kN) 101.5 kN·m (86.0 kN)

B-B Positive 150.1 kN·m (107.6 kN) 141.5 kN·m (101.4 kN) 128.3 kN·m (92.0 kN)
Negative 144.6kN·m (103.7 kN) 136.3kN·m (97.7 kN) 120.0kN·m (86.0 kN)

assumption, as listed in Table 7. Calculated maximum bending moment

Calculated Mc1 considered the ultimate tensile strength of reinforcement while Mc2
12 P-UHPC1 considered the yield tensile strength. Measured maximum bending
P-UHPC2 moment Mt was the average value of four specimens. When the beam
Shear stress (MPa)

P-UHPC3 went down, the direction was defined positive. Section A-A is the po-
P-UHPC4 sition where U-shape bars terminated while section B-B is the end of the
8 beam, as shown in Fig. 1a.
It can be seen from Table 7 that the predicted maximum load cor-
responding to section A-A is smaller than the load corresponding to
section B-B, causing the failure at section A-A where U-shape bars
4 terminated. Besides, the maximum calculated bending moment ne-
glecting rebar strengthening is much closer to the test value since
buckling of beam bottom bars was observed which prevents the rebar
from reaching its ultimate tensile strength.
0
0.0000 0.0002 0.0004 0.0006 0.0008 4.7.2. Calculated shear stiffness of joint zone
The theoretical elastic shear stiffness of UHPC joint zone can be
Shear strain (rad) computed as 0.5Ec/(1 + v). v denotes the Poisson's ratio of UHPC and
Fig. 24. Calculated and measured skeleton shear stress-shear strain curves. equals to 0.213 [43]. Ec denotes the elastic modulus of UHPC. As sug-
gested by Guo et al. [44], Ec is calculated by Eq. (7):

force that can make use of at least 63% (1500 με/yielding strain) of the Ec = 1000/(0.0142 + 1.255/fc ) (7)
yielding stress of the rebar.
where fc denotes the cylinder compressive strength of UHPC.
According to the material tests, the yielding stress and strain-
Fig. 24 shows the comparisons between calculated and measured
hardening modulus of the rebar are 490.8 MPa and 1263.2 MPa, re-
skeleton shear stress versus shear strain curves of four specimens. It can
spectively. Based on the assumption that the material constitutive law
be seen from Fig. 24 that the curves of four specimens were very si-
of the rebar conforms to the bilinear model and kinematic hardening
milar. From the comparison between specimens P-UHPC2 with stirrups
criterion, the stress of the rebar at the positions where B2 and B3 were
in the joint zone and P-UHPC4 without stirrups in the joint zone,
placed can be calculated and is plotted in Fig. 23. The maximum dif-
conclusions that stirrups nearly have no influence on the shear re-
ferences between the stress of B2 and B3 Δσ, the distances between the
sistance of the UHPC joint zone can be drawn. Before the shear strain of
positions of the B2 and B3 s, and the bond stress of the beam bottom
0.0003 rad, measured and calculated curves nearly coincide with each
longitudinal rebar in the UHPC τ for all the four specimens were sum-
other, demonstrating that the UHPC was in the elastic stage. After that,
marized in Table 6. Compared with P-UHPC4 and P-UHPC2, the bond
the shear stiffness dropped due to the development of minor diagonal
stress of the P-UHPC3 is smaller, which means that the bond stress of
cracks under reversed load.
the P-UHPC3 was not made full use of and thus the anchorage length of
the headed bar in the UHPC can at least be reduced to 8.1db, as can be
5. Summary and conclusions
seen in Table 6. Considering that the bond stress in the P-UHPC1 is still
lower than the bond strength since no bond-slip failure was found, the
Based on cyclic tests on four large-scale precast UHPC/concrete
anchorage length of the longitudinal bar in the UHPC could further be
beam-column joint specimens, this paper investigates the application of
reduced, which deserves more investigations. Besides, the bond stress of
UHPC into the joint zone and compares the influence of different an-
the P-UHPC4 without any stirrups in the joint zone is even a little larger
chorage methods as well as different stirrup ratios in the joint zone. The
than that of the P-UHPC2, indicating the negligible effect of stirrups in
conclusions can be drawn as follows:
the joint zone on the bond stress.
(1) Without steam curing, the mechanical properties of the UHPC can
4.7. Theoretical analysis still achieve adequate compressive and tensile strength with stable
strain hardening properties. The curing method in this study is very
4.7.1. Joint strength calculation simple and makes the application of UHPC in the building con-
All the four specimens experienced typical beam flexural failure struction practicable.
mode, thus the joint strength is determined by the beam flexural (2) The cast-in-situ UHPC is able to collaborate well with precast
strength. However, it should be noted that the failure occurred in the concrete with 30 mm grooves and U-shape rebar placed at the in-
position where U-shape bars terminated rather than the end of beams. terface. The influence of shrinkage of the UHPC is acceptable and
To find a precise prediction of joint strength, the flexural capacities of the shear sliding resistance of the interface is adequate.
the two different beam sections were calculated based on plane section (3) The UHPC can be safely used in the joint zone without any stirrups

17
Z.-Y. Zhang, et al. Engineering Structures 222 (2020) 111145

under the shear stress level of 7 MPa, due to its extreme high shear 2011;137(10):1052–62.
resistance. In this study, no obvious difference was found between [9] Ferdous W, Almutairi AD, Huang Y, Bai Y. Short-term flexural behaviour of concrete
filled pultruded GFRP cellular and tubular sections with pin-eye connections for
specimens P-UHPC2 with the stirrup ratio of 0.56% in the joint zone modular retaining wall construction. Compos Struct 2018;206:1–10.
and P-UHPC4 without stirrups. After the test, the joint zone of P- [10] Kurama YC, Sritharan S, Fleischman RB, et al. Seismic-resistant precast concrete
UHPC4 was still in good condition. Theoretical and test results structures: state of the art. J Struct Eng 2018;144(4):03118001.
[11] Wang J, Zhao JD, Hu ZY. Review and thinking on development of building in-
demonstrate that the UHPC joint zone was in the elastic stage be- dustrialization in China. China Civ Eng J 2016;49(5):1–8. (in Chinese).
fore 0.0003 shear strain. [12] Wu G, Feng DC. Research progress on fundamental performance of precast concrete
(4) The anchorage length of the beam longitudinal straight and headed frame beam-to-column connections. J Build Struct 2018;39(2):1–16. (in Chinese).
[13] Codardi M. Seismic behaviour of steel fiber concrete beam-column. Proceedings of
bar can be reduced to only 16db and 8.1db in the UHPC joint zone the tenth world conference on earthquake engineering. Madrid, Spain; 1992. p.
under cyclic loading. For rebar with a diameter of 18 mm or 169-3174.
smaller, a single head is able to provide the anchorage force as [14] Tuleasca L, Ingham JM, Cuciureanu A. Seismic response of a cast-in-place steel fibre
concrete joint connecting precast beams and columns. Proceedings of Pacific
much as 63% of the yield force in the rebar. It is thus recommended
Conference on Earthquake Engineering; 2003.
to use headed bars in the fabrication of precast beam-column joints, [15] Dora AK, Hamid NA. Seismic performance of SFRC beam-column joint with corbel
so that the congestion of beam bottom longitudinal bar from op- under reversible lateral cyclic loading. Int J Eng Technol 2012;4(1):76–80.
posite sides can be avoided. [16] Parra-Montesinos GJ, Peterfreund SW, Chao SH. Highly damage-tolerant beam-
column joints through use of high-performance fiber-reinforced cement composites.
(5) Based on flexural calculation at different beam sections, theoretical ACI Struct J 2005;102(3):487–95.
analysis can give a precise prediction on the joint strength and [17] Qudah S, Maalej M. Application of engineered cementitious composites (ECC) in
clearly demonstrate the effects of U-shape bars on the location of interior beam–column connections for enhanced seismic resistance. Eng Struct
2014;69:235–45.
plastic hinges. [18] Gou SK, Ding R, Fan JS, Nie X, Zhang J. Experimental study on seismic performance
of precast LSECC/RC composite joints with U-shaped LSECC beam shells. Eng Struct
The precast UHPC/RC composite joint exhibited excellent seismic 2019;189:618–34.
[19] Gou SK, Ding R, Fan JS, Nie X, Zhang J. Seismic performance of a novel precast
performance with significantly simplified anchorage details and totally concrete beam-column connection using low-shrinkage engineered cementitious
eliminated stirrups in the joint zone. In addition, the satisfactory results composites. Constr Build Mater 2018;192:643–56.
also indicate that the anchorage lengths of the beam longitudinal bar in [20] ASTM. Standard practice for fabricating and testing specimens of ultrahigh per-
formance concrete. ASTM C1856/C1586-17; 2017.
the UHPC joint zone can be further reduced. Thus future test of UHPC [21] Yang L, Shi CJ, Wu ZM. Mitigation techniques for autogenous shrinkage of ultra-
joints with less anchorage lengths and higher level shear stress level is high-performance concrete–A review. Compos Part B-Eng 2019::107456.
suggested. [22] Resplendino J, Toulemonde F. Designing and Building with UHPFRC. John Wiley &
Sons; 2013.
[23] Sorelli L, Constantinides G, Ulm FJ, Toutlemonde F. The nano-mechanical signature
CRediT authorship contribution statement of ultra high performance concrete by statistical nanoindentation techniques. Cem
Concr Res 2008;38(12):1447–56.
Zi-Yu Zhang: Data curation, Software, Writing - original draft. Ran [24] Wu PT, Wu CQ, Liu ZX, Hao H. Investigation of shear performance of UHPC by
direct shear tests. Eng Struct 2019;183:780–90.
Ding: Validation, Writing - review & editing, Supervision, Funding [25] Yuan JQ, Graybeal B. Bond of reinforcement in Ultra-high-performance concrete.
acquisition. Xin Nie: Investigation, Methodology. Jian-Sheng Fan: . : ACI Struct J 2015;112(6):851–60.
Conceptualization, Funding acquisition. [26] Marchand P, Baby F, Khadour A, Battesti T, et al. Bond behaviour of reinforcing
bars in UHPFRC. Mater Struct 2016;49(5):1979–95.
[27] Saleem MA, Mirmiran A, Xia J, Mackie K. Development length of high-strength steel
Declaration of Competing Interest rebar in ultrahigh performance concrete. J Mater Civ Eng 2012;25(8):991–8.
[28] Fehling E, Lorenz P, Leutbecher T. Experimental Investigations on Anchorage of
Rebars in UHPC. Proceedings of Hipermat 2012 3rd International Symposium on
The authors declare that they have no known competing financial UHPC and Nanotechnology for High Performance Construction Materials. Kassel,
interests or personal relationships that could have appeared to influ- Germany; 2012, pp. 533-540.
[29] Wang DH, Ju YZ, Zheng WZ, Shen H. Seismic behavior and shear bearing capacity
ence the work reported in this paper. of ultra-high performance fiber-reinforced concrete (UHPFRC) beam-column joints.
Appl Sci 2018;8(5):810.
Acknowledgements [30] Sarmiento PA, Torres B, Ruiz DM, Alvarado YA, Gasch I, Machuca AF. Cyclic be-
havior of ultra-high performance fiber reinforced concrete beam-column joint.
Struct Concrete 2019;20(1):348–60.
The authors gratefully acknowledge the financial support provided [31] Khuzaie HMAA, Rafid S. Comparison study of the structural properties of beam-
by the National Key R&D Program of China (2018YFC0705400) and the column joints using reactive powder concrete (RPC) under reversed cyclic loading.
National Natural Science Foundation of China (51890901, 51978377). J Eng Sci and Technol 2019;14(3):1420–33.
[32] ACI 318-14. Building code requirements for structural concrete (ACI 318-14) and
commentary/reported by ACI Committee 318. Farmington Hills, Mich.: American
References Concrete Institute; 2014.
[33] Joint ACI-ASCE Committee 352. Recommendations for Design of Beam‐Column
Connections in Monolithic Reinforced Concrete Structures (ACI 352R-02).
[1] Al-Rubaye M, Manalo A, Alajarmeh O, Ferdous W, Lokuge W, Benmokrane B, et al. Farmington Hills, Mich.: American Concrete Institute; 2002.
Flexural behaviour of concrete slabs reinforced with GFRP bars and hollow com- [34] Chen BC, Li C, Huang W, An MZ, Han S, Ding QJ. Review of ultra-high performance
posite reinforcing systems. Compos Struct 2020;236:111836. concrete shrinkage. J Traffic Transp Eng 2018;18(1):13–28. (in Chinese).
[2] AlAjarmeh OS, Manalo AC, Benmokrane B, Karunasena K, Ferdous W, Mendis P. [35] Hussein HH, Sargand SM, Steinberg EP. Shape optimization of UHPC shear keys for
Hollow concrete columns: review of structural behavior and new designs using precast, prestressed, adjacent box-girder bridges. J Bridge Eng
GFRP reinforcement. Eng Struct 2020;203:109829. 2018;23(4):04018009.
[3] Haber ZB, Saiidi MS, Sanders DH. Seismic performance of precast columns with [36] Jang HO, Lee HS, Cho K, Kim J. Experimental study on shear performance of plain
mechanically spliced column-footing connections. ACI Struct J construction joints integrated with ultra-high performance concrete (UHPC). Constr
2014;111(3):639–50. Build Mater 2017;152:16–23.
[4] Mohammed AA, Manalo AC, Ferdous W, Zhuge Y, Vijay PV, Pettigrew J. [37] Shi CJ, Wu ZM, Xiao JF, Wang DH, Huang ZY, Fang Z. A review on ultra high
Experimental and numerical evaluations on the behaviour of structures repaired performance concrete: Part I. Raw materials and mixture design. Constr Build Mater
using prefabricated FRP composites jacket. Eng Struct 2020;210:110358. 2015;101:741–51.
[5] Billington SL, Yoon JK. Cyclic response of unbonded posttensioned precast columns [38] Zhang YS, Sun W, Liu SF, Jiao CJ, Lai JZ. Preparation of C200 green reactive
with ductile fiber-reinforced concrete. J Bridge Eng 2004;9(4):353–63. powder concrete and its static–dynamic behaviors. Cem Concr Compos
[6] Siddika A, Al Mamun MA, Ferdous W, Alyousef R. Performances, challenges and 2008;30(9):831–8.
opportunities in strengthening reinforced concrete structures by using FRPs-A state- [39] Yazici H. The effect of curing conditions on compressive strength of ultra high
of-the-art review. Eng Fail Anal 2020;104480. strength concrete with high volume mineral admixtures. Build Environ
[7] Cuenca E, Serna P. Shear behavior of prestressed precast beams made of self-com- 2007;42(5):2083–9.
pacting fiber reinforced concrete. Constr Build Mater 2013;45:145–56. [40] JGJ/T 101-2015. Specification for Seismic Test of Buildings. Beijing, China: China
[8] Smith BJ, Kurama YC, Mcginnis MJ. Design and measured behavior of a hybrid Architecture & Building Press; 2015 (in Chinese).
precast concrete wall specimen for seismic regions. J Struct Eng [41] Feng P, Qiang HL, Ye LP. Discussion and definition on yield points of materials,

18
Z.-Y. Zhang, et al. Engineering Structures 222 (2020) 111145

members and structures. ENG MECH 2017;34(3):36–46. (in Chinese). reinforced concrete (uhpfrc): a feasibility study for using ultrasonic and resonant
[42] Yuan F, Pan JL, Xu Z, Leung CKY. A comparison of engineered cementitious com- frequency testing techniques. Constr Build Mater 2012;35:361–7.
posites versus normal concrete in beam-column joints under reversed cyclic [44] Guo XY, Kang JF, Zhu JS. Constitutive relationship of ultrahigh performance con-
loading. Mater Struct 2013;46(1–2):145–59. crete under uni-axial compression. J SOUTHEAST U: Nat Sci Ed
[43] Hassan AMT, Jones SW. Non-destructive testing of ultra high performance fibre 2018;47(2):369–76. (in Chinese).

19