1. Experimental study on shear behavior of the interface between old and
new deck slabs
Junichiro Niwa a
, Fakhruddin a,⇑
, Koji Matsumoto b
, Yuji Sato c
, Masahiko Yamada c
, Takahiro Yamauchi d
a
Department of Civil Engineering, Tokyo Institute of Technology, Japan
b
Institute of Industrial Science, The University of Tokyo, Japan
c
Engineering Division, Fuji P.S. Corporation, Japan
d
Metropolitan Expressway Company, Japan
a r t i c l e i n f o
Article history:
Received 21 October 2015
Revised 28 July 2016
Accepted 31 July 2016
Keywords:
Widening bridge technique
Deck slab
Interface
Connection
Shear strength
a b s t r a c t
An experimental study of the interface shear transfer between differently aged concrete (old and new
deck slabs) has been performed. The old and new deck slabs parts were crossed by steel bars and sub-
jected to the external prestressing force. The tests were carried out to be representative of a proposed
technique used for widening prestressed concrete (PC) highway decks. The experimental program com-
prised nine specimens tested under double-shear test by taking the initial prestressing levels, connection
methods between steel bars, reinforcement ratio and surface roughness as parameters. The experimental
results indicated that the failure behavior of the interface was greatly affected by the initial prestressing
level, reinforcement ratio and surface roughness of the interface. Finally, a comparison of the experimen-
tal shear strength with those given by JSCE Standard Specification, AASHTO and fib Model Code 2010
showed a conservative result for low and high prestressing levels, low reinforcement ratio and smooth
surface.
Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction
Some highway bridges worldwide become functionally obsolete
due to inadequate width before they become structurally deficient.
The clear width of old bridge has been unable to meet the demand
of increasing traffic flow. If the original old bridge is in service,
widening may be an attractive option [1].
In the common PC highway deck widening technique as shown
in Fig. 1(a), the prestressing tendons in the existing slab should be
connected to the new slab [2]. Consequently, some concrete parts
in the existing slab should be demolished (at least 1000 mm) to
expose the prestressing tendons from the existing slab. The cast-
in-place concrete slabs and beams are required for the widening
structure. Mechanical anchorage is also needed to increase the
integration between the existing and widening structure. There-
fore, some problems still exist, such as the need to connect the pre-
stressing tendons that have great effect on traffic of the existing
bridge, the massive wet work of the cast-in-place concrete, and
high volumes of formwork, which entail great time and financial
cost.
To solve these problems, the new widening PC highway deck
technique has been proposed in this study and schematically illus-
trated in Fig. 1(b). It is to be noted that two PC cables (lower and
upper) are placed inside the precast ribs and prestressed at the dif-
ferent times. First, the precast rib is installed individually to the old
bridge by introducing the first prestressing force to the lower pre-
stressing cables. The distance between two precast ribs is
3000 mm as shown in Fig. 2(a). After that, the precast PC slabs
are placed between adjacent precast ribs. The old steel bars are
then exposed and connected to the new steel bars. Later, the
new RC slab is cast. Finally, the second prestressing force is intro-
duced to the upper PC cable. The advantages of this method are
that the existing prestressing tendons need not be connected to
the widening structures, and the volumes of cast-in-place concrete
and formwork in the construction site can be reduced.
As shown in Fig. 2, the target location of this study was the con-
nection between the old and new deck slabs, exactly at the mid-
span between two precast ribs (spacing of 3000 mm). When the
second prestressing force is introduced to the upper PC cable, some
portion of the compression force will be transferred to the new
deck slab and subsequently to the interface between the old and
new deck slabs through an interface between the precast rib and
http://dx.doi.org/10.1016/j.engstruct.2016.07.063
0141-0296/Ó 2016 Elsevier Ltd. All rights reserved.
⇑ Corresponding author at: M1-17, 2-12-1, O-okayama, Meguro-ku, Tokyo
152-8552, Japan.
E-mail addresses: fakhruddin.m.aa@m.titech.ac.jp, fakhrud.civil05@gmail.com
(Fakhruddin).
Engineering Structures 126 (2016) 278–291
Contents lists available at ScienceDirect
Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

2. precast PC slab and another interface between the precast PC slab
and the new deck slab. The remaining portion of the prestressing
force (introduced by the upper PC cable) is transmitted by the pre-
cast rib to the girder web. These behaviors are illustrated in Figs. 1
(b) and 2(b). To simulate the portion of the compression force
transferred to the interface in the experimental test, the prestress-
ing force was introduced to the interface by using the external pre-
stressing rods. The amount of the compression force transferred to
the interface became the main parameter of this study. Moreover,
various parameters regarding the construction sequences and
details were also carefully selected, such as the connection method
between steel bars, reinforcement ratio, and surface roughness of
the interface. Furthermore, the results of these parameters will
be considered for the direct application of the proposed method.
For this aim, the strip specimen representing the deck slabs sec-
tion subjected to external prestressing force was used. Several tests
have been performed by many researchers to examine the shear
strength of the interface with and without compression forces.
The investigations with the external compression forces have been
conducted by Turmo et al. [3], Zhou et al. [4] and Wakasa et al. [5].
These works, however, were most likely limited to the testing of
shear keyed joints without any reinforcement crossing the inter-
face. The investigations of the interface shear strength without
external compression forces have been performed by many
(a) Common PC box girder bridge widening technique
(b) Proposed PC box girder bridge widening technique
Existing PC
tendon
Sleeves
New PC
tendon
New slab
Demolished
±1000 mm
Mortar
anchorage
New slab
New rebar
Existing bridge Precast ribs PC cables
Old rebar Connection
: Flow of compression force
Fig. 1. Comparison between the common and proposed PC box girder bridge widening technique.
(b) Flow of compression force after introducing the second prestressing to upper PC cable
Lower PC cable
(1st
prestress)
Upper PC cable (2nd
prestress)
2nd
prestress
Target of
this study
: Flow of compression force
Old bridge
New slab
Upper PC cable
(2nd
prestress)
Target of this
study
Lower PC cable
(1st
prestress) @3000 mm
Precast rib
(a) Arrangement of precast ribs and prestressing cables
Fig. 2. Isometric view of the new bridge widening technique.
J. Niwa et al. / Engineering Structures 126 (2016) 278–291 279

3. researchers [6–11]. The results showed that the ultimate shear
strength of the interface is determined by the strength of concrete,
roughness degree and friction coefficient, and the normal compres-
sive stress can increase the ultimate shear strength. It is also
reported that the failure mode is a shear failure of the interface
between the old and new concretes. Nevertheless, literature infor-
mation concerning the shear behavior of the interface crossed by
steel bars and subjected to the external prestressing force is few.
In this regard, this study aims to investigate the shear behavior
of the interface between the old and new deck slabs under double-
shear test by taking the initial prestressing level, the connection
method between steel bars, reinforcement ratio, and surface
roughness of the interface as parameters. The shear capacity, crack
pattern, joint opening, load-displacement response, the stress
increment at the PC rods were examined. Finally, the experimental
shear strength were compared with those given by JSCE Specifica-
tion [12], AASHTO [13] and fib Model Code 2010 [14].
2. Experimental program
2.1. Specimen design
Four specific test series with a total of nine specimens were per-
formed as summarized in Table 1 and the details of the specimen
are shown in Fig. 3. Generally, each specimen consisted of two
parts (the old and new slabs) and cast in two phases. The both
end parts (old slabs) were cast in the first phase and the middle
part took place (new slab) after seven days. The interface between
the old and new slabs was crossed by steel bars connected by two
different methods and subjected to the external prestressing force
by using four prestressing rods (PC rods).
2.2. Parameters
The main parameters in this study were the initial prestressing
levels (Series-I), the connection method between steel bars (Series-
II), reinforcement ratio (Series-III) and surface roughness of the
interface (Series-IV), as shown in Table 1.
In Series-I, specimens EW-P0.25, EW-P0.5, EW-P1.0 and EW-
P2.0 having the initial prestressing levels of 0.25, 0.5, 1.0 and
2.0 MPa, respectively, were tested. These parameters were deter-
mined based on the construction sequence of the proposed bridge
widening technique. In the bridge case study, the compression
force will be transmitted to the interface between the old and
new slabs after introducing the second prestressing force through
the upper prestressing cables. The amount of initial prestressing
level subjected to the interface was determined from the results
of 3D FEM analysis in the preliminary work. The influence of the
creep properties of precast and new concretes and the construction
sequence to the amount of compression force at the interface has
been considered. The FEM analysis result shows that the prestress-
ing level of 1.0 MPa is sufficient for the serviceability limit state.
Hence, to simulate this condition in the laboratorial test, the
compression forces were then introduced to the interface by using
external prestressing rods and varied from 0.25, 0.5, 1.0 and
2.0 MPa.
In Series-II, the effect of the connection between steel bars on
the behavior of specimens EW-P1.0 (enclosed welding) and SS-
P1.0 (mortar grouted sleeves) were examined and the tested
results were compared with NC-P1.0, which did not have any
splices at steel bars (straight steel bars). Enclosed welding (Fig. 4
(a)) is a process of joining two steel bars. A strong metallurgical
bond between steel bars is created by heating the supplementary
molten metal. Mortar grouted sleeve (Fig. 4(b)) is a mechanical
joint for splicing steel bars and uses a cylindrical shaped steel
sleeve filled with high early-strength grout mortar.
In Series-III, the effect of the reinforcement ratio (q) was inves-
tigated in EW-P1.0, EW-0.79 and EW-1.01 specimens, which had
the reinforcement ratio of 0.51%, 0.79% and 1.01%, respectively,
and the interface cross-sections of those specimens are shown in
Fig. 3(c), (d) and (e), respectively. The reinforcement ratio is calcu-
lated from the total cross-sectional area of steel bars crossing the
interface (As) divided by the area of the interface (Ac = 50,000 mm2
in this study).
Finally, the effect of the surface roughness of the interface was
investigated in Series-IV by comparing the results of EW-P1.0
(rough) and EW-smooth (smooth). It is to be noted that all speci-
mens in this study had a rough surface, except for the EW-
smooth. The rough surface was intentionally roughened by using
retarder on the day before casting and spraying with high-
pressure water after de-molding. The smooth surface was obtained
without roughness treatment.
2.3. Material properties
The material properties, such as compressive strength of con-
crete and yielding strength of steel bars, were determined based
on the general properties in the highway bridge construction.
The design cylinder strength of the old and new slabs at seven days
was 40 and 30 MPa, respectively, with the maximum aggregate
size (Gmax) of 20 mm. Steel bars with a nominal diameter of
13 mm and 16 mm (only in specimen EW-0.79) were used, with
the average yield strength of 345.1 and 361.2 MPa, respectively.
Four prestressing rods with a nominal diameter of 17 mm
(Aps = 226.8 mm2
each) were prepared as the external PC rods.
The yield strength, fpy, the tensile strength, fpu, and the modulus
of elasticity of PC rods, Eps, were 785, 1030 and 200,000 MPa,
respectively.
2.4. Fabrication of specimens
Fig. 5 illustrates the fabrication procedures. The first step is by
connecting the steel bars using enclosed welding or mortar
grouted sleeves. Then, they were assembled into the formwork
together with other steel bars to prevent the flexural failure and
any other local failures, as shown in Fig. 5(a). The day before
Table 1
Experimental cases.
Series Specimens Initial prest. level, ri (MPa) Connection between steel bars Reinforcement ratio, q (%) Surface roughness
I, II, III and IV EW-P1.0 1.0 Enclosed welding 0.51 Rough
I EW-P0.25 0.25 Enclosed welding 0.51 Rough
I EW-P0.5 0.5 Enclosed welding 0.51 Rough
I EW-P2.0 2.0 Enclosed welding 0.51 Rough
II SS-P1.0 1.0 Mortar sleeves 0.51 Rough
II NC-P1.0 1.0 No connection 0.51 Rough
III EW-0.79 1.0 Enclosed welding 0.79 Rough
III EW-1.01 1.0 Enclosed welding 1.01 Rough
IV EW-smooth 1.0 Enclosed welding 0.51 Smooth
280 J. Niwa et al. / Engineering Structures 126 (2016) 278–291

4. casting, the formwork was coated using the retarder to make the
rough surface at the interface. Afterward, the old slab parts were
cast with the target compressive strength of 40 MPa in seven days
(Fig. 5(b)). The next step was casting the new slab part at the mid-
dle of the specimen (f0
c = 30 MPa) as shown in Fig. 5(c). Finally, the
completed specimen is shown in Fig. 5(d).
2.5. Loading method and instrumentation
Before testing, specimens were prestressed using four external
PC rods and anchored at the ends of specimens. The amount of
the initial prestressing levels was similar in all PC rods. After that,
the double-shear test, as one of the shear testing methods was per-
formed on all nine specimens under 2000 kN loading machine as
shown in Fig. 6. The tests were conducted under load control with
the average rate of 9.45 kN/min and all the experimental data were
recorded through the data logger every 3 s.
The locations of transducers and strain gauges are shown in
Fig. 7. Three transducers were attached at each interface to mea-
sure the joint opening. Transducers were also positioned under
the new slab and above the supporting point to measure the dis-
placement at those locations. By subtracting the displacement of
the new slab to the displacement of supporting point, the relative
displacement at the bottom of the interface can be calculated. One
Unit: mm
(b) Top view
1400
135
125250125
PC rods representing
upper PC cable in
widening technique
(a) Front view
Old slab Old slabNew slab
200 Connection
between rebars
250 250900
: Transducer : Strain gauge
(d) EW-0.79 (ρ =0.79%)
125 125250
PC rods
D16
200
150
25 (e) EW-1.01 (ρ =1.01%)
125 125250
PC rods
D13
200
150
25
(c) Series-I, II and IV (ρ =0.51%)
125 125250
PC rods
D13
200
150
25
Fig. 3. Details of the specimen.
(a) Enclosed welding (b) Mortar grouted sleeves
Fig. 4. The connection between steel bars.
(a) Preparing formwork (b) Casting the old slab (c) Casting the new slab (d) Completed specimens
Old
slab
New slab Old
slab
Fig. 5. Fabrication procedures of specimens.
J. Niwa et al. / Engineering Structures 126 (2016) 278–291 281

5. electrical strain gauge was placed at each mid-span of PC rods.
However, strain gauges were not attached to the steel bars crossing
the interface, because the cover tapes to protect them against the
fresh concrete may decrease the bond between steel bars and
concrete.
3. Results and discussions
3.1. Failure behavior
All nine specimens failed in a shear failure of the interface
between the old and new deck slabs. The difference in the failure
behavior of the interface according the level of prestress, reinforce-
ment ratio and surface roughness can be clearly observed through
the examination of the crack pattern.
Generally, two groups of crack patterns were observed in this
study as shown in Fig. 8. The first group was a combination of diag-
onal tension and shear crack at the interface (Fig. 8(a)). This crack
pattern was observed in all specimens, except in EW-1.01 and EW-
Smooth. The diagonal tension crack was extended from the tip of
the joint opening toward the loading point. The second group
was the shear crack of the interface between the old and new deck
slabs (Fig. 8(b)). These crack patterns were observed in EW-1.01
and EW-Smooth, which had the highest reinforcement ratio and
smooth surface, respectively. It is to be noted that the cracks
shown in Fig. 8(b) were the ideal crack pattern for shear strength
tests, where the cracks run vertically from the bottom to the top
interface without the existence of diagonal tension crack.
To investigate the effect of the diagonal tension crack, the com-
parison of the concrete strains at the top fiber of the interface in
the specimens EW-P1.0 (q = 0.51%) and EW-1.01 (q = 1.01%) is
presented in Fig. 9. The positive and negative values indicate the
tension and the compression strains, respectively. It was seen that
both specimens had similarity from the beginning of load, in which
the compression zone was formed at the top fiber of the interface.
However, due to the opening of the joint, the behavior of the inter-
face changed significantly. Since the diagonal tension crack was
formed in EW-P1.0, the top fiber of the interface still behaved as
the compression fiber until failure (Fig. 10(a)). Thus, it can con-
tribute to increase the shear capacity of the interface. On the other
hand, since the diagonal tension crack was not formed EW-1.01,
the measured compression strains gradually reduced and hence,
the tension zone was formed at the top interface until failure
(Fig. 10(b)). In this case, the contribution of the top interface was
not significant on the shear transfer mechanism.
3.2. Joint openings
Table 2 tabulates the measured joint opening (JO) at peak load
and the height of the joint opening at the first crack formation at
the interface (hop). The joint opening was measured at the top
(T1), middle (T2) and bottom (T3) fibers of the interface as shown
in Fig. 7. Increasing the reinforcement ratio from 0.51% (EW-P1.0)
to 1.01% (EW-1.01) significantly decreased the joint opening at the
bottom fiber (T3) from 3.49 mm to 0.19 mm (Table 2). This indi-
cates that the higher the reinforcement ratio, the smaller the joint
opening at failure.
The height of the joint opening at the first crack formation at
the interface (hop) was greatly affected by the initial prestressing
level (Table 2). When the initial prestressing level increased, the
height of joint opening decreased. The specimen EW-P0.25, which
had the lowest initial prestressing level, exhibited the largest hop
(113.3 mm). On the other hand, the specimen EW-P2.0, which
had the highest initial prestressing level, showed the smallest hop
(71.8 mm). In the shear transfer mechanism, the value of hop is
directly proportional to the shear capacity of the interface. The
smaller the height of the joint opening at the first crack formation
at the interface, the higher the shear transfers area at the interface.
3.3. Load-relative displacement response
Fig. 11 shows the response of applied load versus relative dis-
placement. The relative displacement presented in this figure
was the measured displacement at the bottom of the interface
where the failure occurred, which was calculated by subtracting
the displacement of the new slab to the displacement of support-
ing point.
At the beginning, the specimens behaved as the linear elastic
until the opening of the joint between the old and new slabs was
observed and reduced the stiffness of the interface. The joint
Fig. 6. Loading test.
200
: Transducer : Strain gauge
20
80
20
80
55 55
Fig. 7. Measurement items.
282 J. Niwa et al. / Engineering Structures 126 (2016) 278–291

6. opening load (Pc) occurred ranging from 144.6 kN to 740.0 kN
(Table 2). Specimens EW-P2.0 and EW-1.01, which had the highest
initial prestressing level and reinforcement ratio, respectively,
showed the higher first joint opening load among other specimens
in this study. This indicates that the initial prestressing level and
the reinforcement ratio influenced the joint opening load.
After the opening of joint, the displacement increased linearly
with the load until failure. As the load was continued to the ulti-
mate load (Pu), the interface between the old and new slabs
reached the shear capacity, where load decreased drastically until
the range of 34.2–75.8% of the ultimate load. Failure occurred at
the large displacement, ranging from 1.9 mm to 3.7 mm in all
rough surface specimens (except for EW-smooth). In addition, con-
crete crushed at the compression zone of the new slab, near the
loading point. In this stage, the rebars crossing the interface started
to yield.
At the post-peak load (called residual load), the interface
between the old and new slabs can still resist the applied load,
but with a very large displacement. The magnitude of load and dis-
placement were strongly affected by the cross-sectional area of
rebars crossing the interface. The larger the area of steel bars, the
greater the dowel action that can be produced. When the relative
displacement of 20 mm was obtained, the loading test was inter-
rupted due to the limitation of the loading instrumentation.
3.4. Stress in the prestressing rods (PC rods)
Table 3 summarizes the PC rod stress at initial and ultimate
load. Fig. 12 shows the response of applied load versus the average
stress increment at the PC rods (Dfps). The Dfps was calculated by
subtracting the initial prestressing stress, fpe, from the PC rod stress
at ultimate, fps. In this study, the tensile stress in all PC rods never
reached its nominal yield strength at failure.
As presented in Fig. 12, the rate of stress increment was faster at
the load where the interface started to open widely. However, the
rate of stresses increment varied and was affected by each param-
eter in this study. The difference in stress increment among the six
specimens, EW-P0.25, EW-P0.5, EW-P1.0, EW-P2.0, SS-P1.0 and
NC-P1.0, was insignificant as the specimens had similar reinforce-
ment ratio and surface roughness (Fig. 12(a)). However, specimens
having higher reinforcement ratio (EW-0.79 and EW-1.01) and
smooth surface (EW-smooth) showed a significant difference, with
a lower stress increment at the ultimate (Fig. 12(b)).
(a) Failure of EW-P1.0 (ρ = 0.51%)
Compression
arch
(b) Failure mode of EW-1.01 (ρ = 1.01%)
(a) Just before peak load (b) Just after peak load
(a) Just before peak load (b) Just after peak load
: Shear crack at the interface
: Diagonal tension crack
: Flexural crack
: Compression arch
hop
hop
hop: height of the joint opening at the first crack formation at the interface
Fig. 8. Crack pattern of specimens.
Compression Tension
0
100
200
300
400
500
600
-1000 -500 0 500
EW-P1.0 EW-1.01
Shearcapacity,V(kN)
Strain (x10
-6
)
Joint opening
ρ = 0.51% ρ = 1.01%
Fig. 9. Concrete strains at the top fiber of the interface.
J. Niwa et al. / Engineering Structures 126 (2016) 278–291 283

7. (a) EW-P1.0 (ρ = 0.51%) (b) EW-1.01 (ρ = 1.01%)
Tensioned
CCL
Compressed
CL
Fig. 10. Behavior of the top fiber of the interface.
Table 2
Experimental results.
Specimens f0
c (MPa) du (mm) hop (mm) JO at Pu (mm) Load (kN)
Old slab New slab T1 T2 T3 Pcr Pc Pu Pdrop
EW-P1.0 54.6 41.7 3.6 81.2 À0.07 1.61 3.49 249.0 401.7 1011.4 525.2
EW-P0.25 52.1 41.5 2.3 113.3 À0.04 0.55 2.04 87.0 270.0 905.0 568.0
EW-P0.5 52.2 41.6 2.4 91.4 À0.02 À0.02 2.21 189.0 300.0 879.1 643.6
EW-P2.0 54.7 43.5 2.3 71.8 À0.06 À0.09 2.02 285.0 470.0 1105.7 626.0
SS-P1.0 55.6 43.9 3.7 61.9 À0.03 0.07 2.73 229.0 400.0 1112.8 630.8
NC-P1.0 55.2 44.5 2.3 92.9 À0.05 À0.20 4.16 266.0 399.0 1139.0 651.0
EW-0.79 55.3 43.8 2.0 76.9 0.08 0.02 0.52 268.0 392.0 979.0 761.2
EW-1.01 54.5 43.3 1.9 60.7 0.05 0.13 0.19 225.0 740.0 866.2 805.0
EW-smooth 53.6 41.7 20.0 101.9 0.12 0.19 0.66 – 144.6 364.8 –
f0
c: compressive strength of concrete, du: relative displacement at the bottom interface where failure occurred, hop: height of joint opening at the first crack formation at the
interface, T1, T2 and T3 are joint opening (JO) at the top, middle and bottom of the interface, respectively, (À): compression, (+): crack opened, Pcr: first flexural cracking load
at the new slab, Pc: first joint opening load, Pu: ultimate load, Pdrop: sudden drop of load.
0
200
400
600
800
1000
1200
0 5 10 15 20 25
EW-P1.0
EW-P0.25
EW-P0.5
EW-P2.0
Load(kN)
Relative displacement (mm)
0
200
400
600
800
1000
1200
0 5 10 15 20 25
EW-P1.0
SS-P1.0
NC-P1.0
Load(kN)
Relative displacement (mm)
0
200
400
600
800
1000
1200
0 5 10 15 20 25
EW-P1.0
EW-0.79
EW-1.01
Load(kN)
Relative displacement (mm)
0
200
400
600
800
1000
1200
0 5 10 15 20 25
EW-P1.0 (rough)
EW-smooth
Load(kN)
Relative displacement (mm)
(a) Effect of prestressing levels (Series-I) (b) Effect of connection method (Series-II)
(c) Effect of reinforcement ratio (d) Effect of surface roughness
(Series-IV)(Series-III)
Fig. 11. Relation between load and relative displacement at the bottom interface.
284 J. Niwa et al. / Engineering Structures 126 (2016) 278–291

8. It is interesting to note that the relation between the loads ver-
sus the displacement in Fig. 11 had similarity with the loads versus
PC rod stress in Fig. 12. It indicates that the displacement and
stress increment at PC rods influenced each other. A similar test
result was also obtained for testing the shear strength of segmental
beams with the external prestressing tendon, which was con-
ducted by previous researchers [15–19].
Table 3 tabulates the ratio between the average stress incre-
ment and initial stress (Dfps/fpe) at the upper and lower PC rods.
The Dfps/fpe ratio at the upper PC rod was about À0.5 to 2.8. Mean-
while, the Dfps/fpe ratio at the lower PC rods was about 1.4–19.7.
The negative value means the initial PC rod stress decreased during
the loading test. These results indicate that: (1) the prestressing
stress at the PC rods, especially at the lower PC rods, increased sig-
nificantly during the loading test, and (2) the stress increment at
the lower PC rods was much greater than at the upper PC rods
and thus, resulted in the different value of the bending moment
applied to the interface. Furthermore, the value of the bending
moment influenced the joint behavior in each specimen. This is
further discussed in Section 3.6.
3.5. Shear capacity
3.5.1. Effect of initial prestressing levels
When comparing the relationship between the initial prestress-
ing level and the ultimate shear capacity in Fig. 13(a), it can be seen
that the larger the initial prestressing level across the interface
facilitated greater the ultimate shear capacity. This is expected,
since greater initial prestressing level across the interface
promotes greater transfer capacity and delays the failure [20].
Moreover, the increase in the initial prestressing level from 0.5 to
1.0 MPa enhanced the shear capacity (13.1%) more effectively than
the increase in the initial prestressing level from 1.0 to 2.0 MPa
(9.3%). This is in agreement with Tassios and Vintzeleou [21], that
the shear capacity was not increased proportionally with the initial
prestressing level, in which the positive influences of the prestress-
ing level tended to be reduced with a greater amount of the initial
prestressing level.
3.5.2. Effect of connection between steel bars
The influences of connection methods between steel bars were
discussed based on the results of EW-P1.0 (enclosed welding), SS-
P1.0 (mortar grouted sleeves) and NC-P1.0 (control specimen with-
out connection or splices) as shown in Fig. 13(b). It is noted that
the location of splice of steel bars was 135 mm from the interface
(Fig. 3(b)). This indicates that the cross-sectional area of reinforce-
ment at the interface was similar in all specimens. However, the
shear capacity of the interface decreased by using the enclosed
welding and mortar grouted sleeves. The reduction rate of enclosed
welding (11.2%) was higher than that of mortar grouted sleeves
(2.3%). This could be because a local stress is concentrated at the
welded connection, or the welded connection is the weakest point
compared to the steel bar itself. Based on this result, however, it
cannot be concluded that the mortar grouted sleeves are better
than the enclosed welding because the result in this study was
obtained from the single test specimen. Further research is neces-
sary to clarify this phenomenon.
Table 3
Increase in the PC rods stress at ultimate load.
Specimens du (mm) Joint opening T3 (mm) fpe (MPa) fps (MPa) Dfps (MPa) Dfps/fpi Vu (kN)
Upper Lower Upper Lower Upper Lower
EW-P1.0 3.6 3.49 55.1 88.3 388.3 33.2 333.2 0.6 6.0 505.7
EW-P0.25 2.3 1.04 13.8 53.1 285.6 39.3 271.8 2.8 19.7 452.5
EW-P0.5 2.4 2.21 27.5 53.7 302.1 26.2 274.6 1.0 10.0 439.5
EW-P2.0 2.3 2.02 110.1 134.1 451.4 24.0 341.3 0.2 3.1 552.8
SS-P1.0 3.7 2.73 55.1 85.3 462.1 30.2 407.0 0.5 7.4 556.4
NC-P1.0 2.3 4.16 55.1 85.4 409.7 30.3 354.6 0.5 6.4 569.4
EW-0.79 2.0 0.52 55.1 62.4 195.3 7.3 140.2 0.1 2.5 489.5
EW-1.01 1.9 0.19 55.1 48.8 139.2 À6.3 84.1 À0.1 1.5 433.1
EW-smooth 20.0 0.66 55.1 25.0 134.7 À30.1 79.6 À0.5 1.4 182.4
du: displacement at the interface where failure occurred, T3: transducer at the bottom fiber of the interface, fpe: average PC rod stress at initial load, fps: average PC rods stress
at ultimate load, Dfp: average stress increment at PC rods, Vu: ultimate shear capacity (Pu/2).
0
200
400
600
800
1000
1200
-100 0 100 200 300 400 500
EW-P1.0
EW-P0.25
EW-P0.5
EW-P2.0
SS-P1.0
NC-P1.0
Load(kN)
Average stress increment,Δ f
ps
(MPa)
0
200
400
600
800
1000
1200
-100 0 100 200 300 400 500
EW-0.79
EW-1.01
EW-Smooth
Load(kN)
Average stress increment,Δf
ps
(MPa)
(a) Effect of prestressing and
connection method
(b) Effect of reinforcement ratio and
surface roughness
Rough,
ρ = 0.79%
Smooth,
ρ = 0.51%
Rough,
ρ = 1.01%
All: Rough,
ρ = 0.51%
Joint
opening
Joint
opening
Fig. 12. Relation between load and average stress increment at the lower PC rods.
J. Niwa et al. / Engineering Structures 126 (2016) 278–291 285

9. 3.5.3. Effect of reinforcement ratio at the interface
The experimental results in Fig. 13(c) show that the reinforce-
ment ratio greatly affected the shear capacity of the interface.
The specimen EW-1.01 (Vu = 433.1 kN), which had the largest rein-
forcement ratio, exhibited the smallest shear capacity than that of
EW-0.79 (Vu = 489.5 kN) and EW-P1.0 (Vu = 505.7 kN). This indi-
cates that the greater the reinforcement ratio, the smaller the shear
capacity.
Above result was contradicted by the general agreement by pre-
vious researchers [7,9–11,21,22], who concluded that as the rein-
forcement ratio increases, the ultimate shear capacity increases
significantly. The different results in this study can be explained
by investigating the crack pattern and the stress increment at the
PC rods from the specimens EW-P1.0 (q = 0.51%) and EW-1.01
(q = 1.01%). According to Section 3.1, the crack pattern in specimen
EW-P1.0 (Fig. 8(a)) reveals a combination of the diagonal tension
and shear cracks at the concrete joint. Meanwhile, the cracking
observed in specimen EW-1.01 (Fig. 8(b)) can be attributed to the
shear crack at the concrete joint. As the diagonal tension crack
occurred at the interface, the top fiber of the interface tended to
be compressed and contributed to increase the shear capacity as
illustrated in Fig. 10(a). Moreover, according to Section 3.4, the
stress increment at the PC rods in EW-1.01 reduced compared to
that EW-P1.0 due to the decrease in the opening of joint. Further-
more, the smaller stress increment, the smaller shear capacity can
be achieved at ultimate.
In this study, even though the reinforcement ratio of 1.01%
(EW-1.01) gives a negative effect on the ultimate shear capacity,
the interface can provide a good contribution to the serviceability
limit state; whereas the first joint opening load (Pc) and the stiff-
ness of the interface still increased.
3.5.4. Effect of surface roughness of the interface
The shear capacity of the interface was greatly affected by the
surface roughness of the interface. It is seen from Fig. 13(d) that
the shear capacity of the EW-Smooth (smooth) extremely
decreased by 63.9% as compared to that of EW-P1.0 (rough). This
is because the surface roughness acts as a mechanical interlocking
to grip the newly added concrete from being separated. The higher
the degree of roughness of a particular surface, the higher the fric-
tion and cohesion coefficients to enhance the shear capacity of the
interface [23,24]. Moreover, according to the saw-tooth model
[25,26], when a surface is rough, shear stressing causes not only
parallel displacement but also the opening of the joint, which sets
up tensile stresses in any steel bars crossing the interface and
external PC rods. These, in turn, create equalizing compressive
stress in the interface, permitting frictional forces to become
established.
3.6. Bending moment effect
The value of the bending moment applied to the interface
(Minter) affected the interface behavior between the old and new
slabs as a consequence of the opening of the joint. It is seen from
Fig. 14 that the value of bending moment consisted of two compo-
nents, one due to the applied force (Mext) and the other due to pre-
stressing force (Mps). The Mext and Mps gave the positive and
negative bending moment, respectively. Thus, the Minter is given as
+9.3%
0 100 200 300 400 500 600 0 100 200 300 400 500 600
Shear capacity, Vu (kN) Shear capacity, Vu (kN)
0 100 200 300 400 500 600
Shear capacity, Vu (kN)
0 100 200 300 400 500 600
Shear capacity, Vu (kN)
EW-P1.0
EW-P2.0
EW-P0.5
EW-P0.25
+13.1%
Prestressinglevel
(a) Effect of initial prestressing level (Series-I)
EW-P1.0 (Enclosed welding)
SS-P1.0 (Mortar sleeves)
NC-P1.0 (No splices)
-11.2%
-2.3%
(b) Effect of connection method (Series-II)
Reinforcementratio,ρ
EW-P1.0 (rough)
SL-Smooth
EW-P1.0 (ρ = 0.51%)
EW-1.01 (ρ = 1.01%)
EW-0.79 (ρ = 0.79%)
(c) Effect of reinforcement ratio (Series-III) (d) Effect of surface roughness (Series-IV)
-3.2%
-14.4%
-63.9%
Fig. 13. Ultimate shear capacity.
286 J. Niwa et al. / Engineering Structures 126 (2016) 278–291

10. Minter ¼ Mext À ðMps-lower þ Mps-upperÞ ð1Þ
where
Mext ¼ Pb=2 ð2Þ
Mps-lower ¼ Apsfpselower ð3Þ
Mps-upper ¼ ÀA0
psf
0
pseupper ð4Þ
In which P is the applied forces; b: distance from the center of the
supporting plate to the interface (b = 55 mm in this study); A0
ps
and Aps are the total area of the upper and lower PC rods, respec-
tively; f0
ps and fps are stress at the upper and lower PC rods, respec-
tively; and e is the eccentricity of the external PC rod. Under the
loading, the initial eccentricity of the external the PC rod (em) was
changed by the amount equal to the relative displacement at the
bottom interface (d) as shown in Fig. 15. That is, the second
order-effects in externally prestressed beams can be accounted by
taking the value of e [17] in Eqs. (5) and (6); those are the eccentric-
ity of the upper (eupper) and lower (elower) PC rods, respectively.
eupper ¼ em þ d ð5Þ
elower ¼ em À d ð6Þ
In which em is the initial tendon eccentricity (50 mm in this study)
and d is the relative displacement at the bottom interface, which
was calculated by subtracting the displacement of the new slab to
the displacement of supporting point.
Table 4 tabulates the values of bending moment due to
the applied forces (Mext), due to the prestressing forces (Mps =
Mps-lower + Mps-upper), and applied to the interface (Minter) at the same
load level (P = 866 kN, the peak load of EW-1.01). It was seen that
the value of Mps increased with the increase in the initial prestress-
ing level (ri), and on the other hand, it decreased with the increase
in the reinforcement ratio (q). Consequently, it affected the value
of bending moment applied to the interface (Minter). With the
higher Mps, the value of the bending moment applied to the inter-
face (Minter) at the same load level became smaller. This is caused
by the significant differences between the various prestressing
forces in the external PC rods. In the real structure, this factor is
subjected to the important size effect.
The size effect on the RC interface stress transfer behavior is an
essential consideration since laboratory tests are limited to small-
scale experiments. In the proposed widening technique, the varia-
tion of the prestressing force in the external PC rod due to the size
effect depended on the length of the new deck slab, the area of the
interface and diameter of PC cables. For example, increasing the
new slab length resulted in the decrease in the prestressing force
at the PC rods due to the reduction in the entire deformation of
the member and the tendon eccentricity (second-order effects).
In this study, however, the influence of size effect on the variation
of the prestressing force in the external PC rod has not been consid-
ered. This effect needs to be considered in the future work.
3.7. Resisting mechanism of the interface between the old and new
slabs
Fig. 16 shows the resistance mechanism of the interface crossed
by steel bars and subjected to prestressing force. According to the lit-
erature review on design expression for shear-friction conducted by
Santos and Júlio [26], the load transfer mechanisms at the concrete-
to-concrete interface is due to: (1) cohesion; (2) friction; and (3)
dowel action. The cohesion is mainly dominated by the degree of
roughness, area of the interface and concrete strength. The friction
depends on the prestressing levels, reinforcement ratio, roughness
degree, area of the interface and yield strength of rebars crossing
the interface. The dowel action depends on the reinforcement ratio,
roughness degree, steel bars diameter, concrete cover to the rein-
forcement, concrete strength and yield strength of reinforcement.
Compared to other mechanisms contributing to interface shear
transfer, it is found that the frictional force, which resulted from
the external compression forces and reinforcement crossing the
interface, is the most dominant factor that contributes the inter-
face shear resistance in the interface with rough surface.
4. Simplified mechanical model of the interface
The mechanical model of composite joints has been proposed
by Nie et al. [9] as shown in Fig. 17(a). However, the effect of the
stress increment has not been considered in that research. In this
study, the new simplified model of RC interface considering the
PC rod stress increment was proposed based on the behavior of
the interface; including the crack patterns, joint opening and rela-
tive displacement at the bottom of the interface, and the stress
increment at the PC rods as shown in Fig. 17(b). This new simpli-
fied model is classified into four stages as follows:
4.1. Stage I
The flexural crack (Pcr) was observed in the new slab (constant
moment region), but the interface between the old and new slabs
fps Apselower+ f'ps A'pseupper
Mps-lower+Mps-upper
Bending moment diagram
due to prestressing
Bending moment diagram
due to applied forces
(-)
(+)
Mext=Pa/2
a
c.g.c
2/P2/P
b
elower=em-δ
eupper=em+δ
Mext =Pb/2
P/2 P/2
Fig. 14. Bending moment diagram.
c.g.c
em+δ
em-δ
A'ps f'ps
Aps fps
PC rods
Fig. 15. Stresses and forces applied to the interface due to prestressing.
J. Niwa et al. / Engineering Structures 126 (2016) 278–291 287

11. was still linear elastic. A small displacement at the interface was
only produced by the shear deformation of the concrete.
4.2. Stage II
The joint opening occurred and subsequently, the relative dis-
placement increased. Consequently, the PC rod stress increased
in response to the joint opening and displacement. If the tensile
strength of concrete was smaller than the simultaneous action of
normal stress and shear stress, the diagonal tension crack will be
propagated from the tip of the first joint opening to the location
of loading point. From this stage, the steel bars crossing the inter-
face and prestressing forces contributed very significantly to the
shear resistance mechanism of the interface.
4.3. Stage III
The interface between the old and new slabs reached the shear
capacity. The bond of the interface was lost and a sudden drop
occurred indicating a large displacement at the interface. In addi-
tion, concrete crushed at the compression zone of the new slab
near the loading point. In this stage, the reinforcement crossing
the interface completely yielded.
4.4. Stage IV
This stage is called the post-peak load, where the relative dis-
placement increased disproportionally with the load. The magni-
tude of load depended on the cross-sectional area of steel bars
Table 4
The values of bending moment at similar load level (P = 866 kN).
Specimens ri (MPa) q (%) d (mm) Total stress in
PC rod (MPa)
Aps fps (kN) A0
ps f0
ps (kN) Mps-lower (kN m) Mps-upper (kN m) Mext (kN m) Mps (kN m) Mint (kN m)
fps f0
ps
EW-P1.0 1.0 0.51 2.68 324.5 74.7 147.2 33.9 7.0 À1.8 23.8 5.18 18.6
EW-P0.25 0.25 0.51 2.04 268.6 47.5 121.9 21.6 5.8 À1.1 23.8 4.72 19.1
EW-P0.5 0.5 0.51 2.30 292.6 52.4 132.8 23.8 6.3 À1.2 23.8 5.09 18.7
EW-P2.0 2.0 0.51 1.30 377.5 119.8 171.3 54.4 8.3 À2.8 23.8 5.55 18.3
SS-P1.0 1.0 0.51 2.78 338.2 75.4 153.5 34.2 7.2 À1.8 23.8 5.44 18.4
NC-P1.0 1.0 0.51 1.89 297.6 71.8 135.0 32.6 6.5 À1.7 23.8 4.81 19.0
EW-0.79 1.0 0.79 1.73 140.3 46.2 63.7 20.9 3.1 À1.1 23.8 1.99 21.8
EW-1.01 1.0 1.01 1.94 137.2 46.9 62.3 21.3 3.0 À1.1 23.8 1.89 21.9
ri is the initial prestressing level; q is the reinforcement at the interface; d: relative displacement at the bottom interface; f0
ps and fps are stress at the upper and lower PC rods;
A0
ps and Aps are the total area of the upper and lower PC rods, respectively; Mext is the value of bending moment due to the applied forces; Mps is the valued of bending moment
due to the prestressing force at the upper and lower PC rods; Minter: the value of bending moment applied to the interface.
(a) Deformational characteristics (b) Contribution of rebars crossing the interface,
area and roughness of the interface
σn : external prestressing level
Αs : cross-sectional area of steel bars
σs: tensile strength of steel bars
σn
Joint opening
σn
Old slab
Asσs
Area of the interface
Surface
roughness
Asσs
Fig. 16. Shear resistance mechanism at the interface crossed by steel bars and subjected to prestressing.
(b) Four mechanical stages of the interface
Load (kN)
Relative displacement (mm)
Pc
Pu
I
II
IV
δu
III Unloading
Interface completely cracking
Pcr
New concrete
cracking
First crack at the interface
I : Flexural crack in new concrete
II: Interface cracking and diagonal
tension crack occurred
III: Interface completely cracked
and new concrete crushed
IV: Embedded bars completely
yielding
(a) Three mechanical stages of the
interface proposed by Nie et al. [9]
Load (kN)
Relative displacement (mm)
Pc
Pu
I
III
δr
II
Pr
New concrete cracking
Interface cracking
I : No crack in concrete, embedded
bars not yielding
II : Interface cracking, embedded
bars not yielding
III: Interface cracking, embedded
bars completely yielding
Fig. 17. Comparison between mechanical stages of the interface proposed by Nie et al. [9] and this study.
288 J. Niwa et al. / Engineering Structures 126 (2016) 278–291

12. crossing the interface. The larger the area of steel bars, the greater
load increment that could be produced.
5. Comparison of test results and design codes
In this section, the test results of nine specimens were com-
pared with code predictions from JSCE Standard Specifications
[12], AASHTO [13] and fib Model Code 2010 [14]. The shear capac-
ity predicted by these guidelines is based on the assumptions that:
(a) the partial safety factors for the material properties are not
applied; (b) reinforcement at the interface yields at ultimate load;
(c) the original cross-section of the interface, Ac is used to calculate
the ultimate shear strength of the interface (Ac = 50,000 mm2
in
this study).
5.1. JSCE (Japan Society of Civil Engineers)
According to JSCE Standard Specifications [12], the design
capacity for shear transfer is computed using the following equa-
tion when the reinforcement is provided in a shear plane, Vcwd
under axial force:
Vcwd ¼ ððsc þ pss sin
2
h À apfy sin h cos hÞAc þ VkÞ=c ð7Þ
where:
sc ¼ lf
0b
c ðapfy À rnÞ
1Àb
ð8Þ
ss ¼ 0:08fy=a ð9Þ
a ¼ 0:75f1 À 10ðp À 1:7rn=fyÞg ð10Þ
where Vcwd is the shear capacity of joints; p is reinforcement ratio
along the shear plane; h is angle between interface and reinforce-
ment provided in the interface (h = 90° in this study); fy is yield
strength of reinforcement provided in the interface; Ac is area of
the interface; Vk is shear capacity of shear key (Vk = 0 in this study);
rn is average normal compressive stress acting on the interface; c is
the member factor (c = 1.0 in this study); l is coefficient of friction
with value of 0.45; and b is coefficient representing the effect of sur-
face roughness of interface. If construction joint surface is not prop-
erly treated, 0.5 is used for the coefficient b.
5.2. AASHTO-LRFD 2007
The AASHTO provision [13] gives the following design formula
to estimate the nominal shear resistance of interface, Vni:
Vni ¼ cAc þ lðAvf fy þ PcÞ ð11Þ
but not greater than the lesser of Eqs. (12) and (13).
Vni 6 K1f
0
cAc ð12Þ
Vni 6 K2Ac ð13Þ
where Pc is permanent net compressive force normal to the inter-
face; Avf is area of reinforcement crossing the interface; c is cohesion
coefficient; K1 is friction of the concrete strength available to resist
the interface shear; and K2 is limiting the interface shear resistance.
The values of c and l coefficients are presented in Table 5.
5.3. fib Model Code 2010 (MC2010)
In fib MC2010 [14], the main contributions to the overall shear
resistance result from the mechanical interlocking and adhesive
bonding (sR,interlocking), friction due to external compression forces
perpendicular to the interfaces and clamping forces due to rein-
forcement and connector (sR,friction); and the dowel action of rein-
forcement/connectors crossing the interface (sR,dowel action). The
ultimate shear resistance at the interface with the dowel action
effect can be approached with the following equation:
su ¼ sc
|{z}
Interlock
þ l rn þ j1qfy
À Á
|fflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflffl}
Friction
þ j2q
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
fc;cubefy
q
|fflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflffl}
Þ
Dowel action
6 bvf
0
c ð14Þ
where sc is the adhesive bonding or interlocking, i.e., adhesive
forces due to the chemical and physical bonding at the interface;
j1 is the interaction factor of reinforcement due to simultaneous
bending and or reduced anchorage of the bars; j2 is the interaction
factor for the dowel action effects; fc,cube is the cube compressive
strength; b is a coefficient allowing for angle of diagonal concrete
strut (0.5 for rough surface and 0.4 for smooth surface) and v is
reduction factor for strength of diagonal concrete strut, which cal-
culated by Eq. (15). Table 5 tabulates the values of sc, l, c, j1, j2
and the maximum shear strength assumed by AASHTO and fib
MC2010 guidelines.
v ¼ 0:55ð30=f
0
cÞ
1=3
6 0:55 ð15Þ
Following the previous researchers [7,10,11], the comparison of
the shear strength (su = Pu/2Ac) between the test results and design
codes with the effect of prestressing level, reinforcement ratio and
surface roughness of the interface is presented in Figs. 18, 19 and
20, respectively. The effects of each parameter on the accuracy of
guidelines are discussed below.
5.4. Effect of prestressing level
In these guidelines, the prestressing level which acts on the
interface is assumed to be the value of the initial prestressing level.
In fact, the prestressing level increased significantly as a conse-
quent of the joint opening. For this objective, Fig. 18(a) and (b)
are presented to compare the relationship between the shear
strength and the prestressing level according to each design guide-
lines if it was calculated from the initial prestressing level and the
ultimate prestressing level, respectively. It can be seen that if the
shear strength was calculated from the initial prestressing level,
all design guidelines gave too conservative results. However, if it
was calculated from the ultimate prestressing level, the accuracy
in all design guidelines improved.
5.5. Effect of reinforcement ratio
Fig. 19 shows the relationship between the shear strength (su)
and the reinforcement ratio at the interface (q). The prestressing
level at ultimate was used in calculation. For low reinforcement
Table 5
The values of sc, l, c, j1, j2 and maximum shear strength.
Codes Surface roughness sc (MPa) l c (MPa) j1 j2 Maximum shear strength (MPa)
AASHTO Smooth – 0.6 0.075 0.2 0.8 5.52
Rough – 1.0 0.28 0.3 1.8 12.41
fib MC2010 Smooth – 0.5–0.7 – 0.5 1.1 6.19
Rough 1.5–2.5 0.7–1.0 – 0.5 0.9 10.46
J. Niwa et al. / Engineering Structures 126 (2016) 278–291 289

13. ratio (q = 0.51%), it can be observed that all the selected guidelines
gave too conservative solution. For high reinforcement ratio
(q = 1.01%), it can be seen that AASHTO provision was not safe
and fib MC2010 gave the best solution.
5.6. Effect of surface roughness
Fig. 20 shows the comparison between the design guidelines
and the test result with the effect of the smooth surface. The pre-
stressing level at ultimate was used in calculation. Due to a single
test specimen, only the specimen EW-Smooth was plotted in
Fig. 20. It can be seen that all the selected guidelines predicted
the shear strength conservatively.
6. Conclusions
Tests were conducted on nine double-shear test specimens to
study the shear behavior of the interface between the old and
new deck slabs crossed by steel bars and subjected to prestressing
force. The following are the main conclusions of the study:
(1) Increasing the initial prestressing level increases the first
joint opening load and the shear capacity of the interface.
However, the increment of the shear capacity was not pro-
portional to the initial prestressing level. More positive
influences of the prestressing level are observed when the
initial prestressing level enhances from 0.5 to 1.0 MPa.
(2) The connection methods between steel bars do not affect the
shear capacity and the crack pattern of the interface signifi-
cantly, as the location of the connection between steel bars
is sufficiently far from the interface.
(3) Increasing the reinforcement ratio from 0.51% to 1.01%
decreased the shear capacity of the interface. This is caused
by the decrease in the opening of joint, which further reduces
the external prestressing force acting perpendicularly to the
interface. However, the enormity of reinforcement ratio has a
notable influence on the serviceability limit state of the rein-
forced concrete interface; whereas the first cracking load and
the stiffness of the interface increase.
(4) The surface roughness of the interface has important influ-
ences on the behavior of the interface; including the crack
pattern, joint opening, displacement, stress increment at
the prestressing rod and the shear capacity of the interface.
(5) Due to the opening of joint and displacement, a remarkable
increment of the tensile stress at the prestressing rods is
observed. Furthermore, it affects the value of bending
moment applied to the interface between the old and new
deck slabs.
(6) Comparing the experimental shear strength with design
guidelines, it can be stated that all the selected guidelines
are conservative for low and high prestressing levels acting
on the rough interface. For high reinforcement ratio
(q = 1.01%), AASHTO provision was not safe and fib
MC2010 gave the best solution. Moreover, for the smooth
interface, all the selected guidelines predict the shear
strength conservatively.
Acknowledgement
The authors are would like to grateful to Fuji P.S. Corporation
and Splice Sleeves Japan, Ltd. for their kind support to this research
project.
References
[1] American Concrete Institute. Guide for widening highway bridges, 345.2R;
1998.
0
2
4
6
8
10
12
14
0 2 4 6 8 10
Prestressing level (MPa)
EXP
JSCE
AASHTO
MC2010
0
2
4
6
8
10
12
14
0 2 4 6 8 10
Prestressing level (MPa)
EXP
JSCE
AASHTO
MC2010
(a) Initial prestressing (b) Ultimate prestressing
Shearstrength(MPa)
Shearstrength(MPa)
Fig. 18. Accuracy of design guidelines with the effect of prestressing level.
0
2
4
6
8
10
12
14
0 0.5 1 1.5 2
Reinforcement ratio (%)
EXP
JSCE
AASHTO
MC2010
Shearstrength(MPa)
Fig. 19. Accuracy of design guidelines with the effect of reinforcement ratio at the
interface.
0
2
4
6
8
10
12
14
0 2 4 6 8
Prestressing level (MPa)
EXP
JSCE
AASHTO
MC2010
Shearstrength(MPa)
Smooth surface
Fig. 20. Accuracy of design guidelines with the effect of the smooth surface.
290 J. Niwa et al. / Engineering Structures 126 (2016) 278–291

14. [2] Cavanillas JAT, Munoz JM, Gayo AH, Manso AN. Widening of San Pedro de la
Ribera Viaduct. Hormigón Acero 2011;62(260):11–45.
[3] Turmo J, Ramos G, Aparicio AC. Shear strength of dry joints of concrete panel
with and without steel fibers: application to precast segmental bridges. Eng
Struct 2006;28:23–33.
[4] Zhou X, Mickleborough N, Li Z. Shear strength of joints in precast concrete
segmental bridges. ACI Struct J 2005:3–11.
[5] Wakasa T, Otsuka H, Yabuki W. Experimental study of the shear strength of
precast segmental beams with external prestressing. Struct Concr 2005;6
(2):63–80.
[6] Santos PMD, Júlio ENBS, Silva VD. Correlation between concrete-to-concrete
bond strength and the roughness of the substrate surface. Constr Build Mater
2007;21(8):1688–95.
[7] Júlio ENBS, Dias-da-Costa D, Branco FAB, Alfaiate JMV. Accuracy of design code
expressions for estimating longitudinal shear strength of strengthening
concrete overlays. Eng Struct 2010;32:2387–93.
[8] Santos PMD, Júlio ENBS. Factors affecting bond between new and old concrete.
ACI Mater J 2011;108(4):449–56.
[9] Nie JG, Wang YH, Zhang XG, Fan JS, Cai CA. Mechanical behavior of composite
joints for connecting existing concrete bridges and steel-concrete composite
beams. J Constr Steel Res 2012;75:11–20.
[10] Randl N. Design recommendations for interface shear transfer in fib Model
Code 2010. Struct Concr 2013;14(3):230–41.
[11] Rahal KN, Khaleefi AL, Al-Sanee. An experimental investigation of shear-
transfer strength of normal and high strength self compacting concrete. Eng
Struct 2016;109:16–25.
[12] Japan Society of Civil Engineers (JSCE). Standard specification for design of
concrete structures, structure performance verification; 2007.
[13] AASHTO LRFD bridge design specifications. American Association of State
Highway and Transportation Officials; 2007.
[14] International Federation for Structural Concrete (fib). Model code 2010-final
draft. vol. 1; March 2012.
[15] Sivaleepunth C, Niwa J, Diep BK, Tamura S, Hamada Y. Prediction of tendon
stress and flexural strength of externally prestressed concrete beams. J Mater
Concr Struct Pavem JSCE 2006;62(1):260–73.
[16] Aparicio AC, Ramos G, Casas JR. Testing of externally prestressed concrete
beams. Eng Struct 2002;24:73–84.
[17] Ng CK, Tanh KH. Flexural behaviour of externally prestressed beams. Part I:
analytical model. Eng Struct 2006;28:609–21.
[18] Ng CK, Tanh KH. Flexural behaviour of externally prestressed beams. Part II:
experimental investigation. Eng Struct 2006;28:622–33.
[19] Ghallab A, Beeby AW. Factors affecting the external prestressing stress in
externally strengthened prestressed concrete beams. Cement Concr Compos
2005;27:945–57.
[20] Haskett M, Oehlers DJ, Mohamed Ali MS, Sharma SK. Evaluating the shear-
friction resistance across sliding planes in concrete. Eng Struct
2011;33:1357–64.
[21] Tassios TP, Vintzeleou EN. Concrete-to-concrete friction. J Struct Eng
1987;113:832–49.
[22] Maekawa K, Qureshi J. Stress transfer across interfaces in reinforced concrete
due to aggregate interlock and dowel action. J Mater Concr Struct Pavem JSCE
1997;34(34):159–72.
[23] Mohamad ME, Ibrahim IS, Abdullah AB, Rahman AB, Kueh ABH, Usman J.
Friction and cohesion coefficients of composite concrete-to-concrete. Cement
Concr Compos 2015;56:1–14.
[24] Papastergiou D, Lebet J-P. Experimental investigation and modelling of the
structural behaviour of confined grouted interfaces for a new steel-concrete
connection. Eng Struct 2014;74:180–92.
[25] Birkeland PW, Birkeland HW. Connection in precast concrete constructions. J
Am Concr Inst 1996;63(3):345–68.
[26] Santos PMD, Júlio ENBS. A state-of-the-art review on shear-friction. Eng Struct
2012;45:435–48.
J. Niwa et al. / Engineering Structures 126 (2016) 278–291 291