You are on page 1of 11

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/338312186

Shear Strength of Dry and Epoxy Joints for Ultra-High-Performance Fiber-


Reinforced Concrete

Article  in  Aci Structural Journal · January 2020


DOI: 10.14359/51718078

CITATIONS READS

13 695

4 authors, including:

Farzad Hejazi Milad Hafezolghorani Esfahani


University of the West of England, Bristol Universiti Putra Malaysia
223 PUBLICATIONS   1,467 CITATIONS    13 PUBLICATIONS   265 CITATIONS   

SEE PROFILE SEE PROFILE

Yen Lei Voo


DURA Technology Sdn Bhd
41 PUBLICATIONS   1,683 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

STUDY OF PRECAST FOAMED CONCRETE SANDWICH WALL PANEL BEHAVIOUR UNDER HORIZONTAL DISTRRIBUTION LOAD View project

Development of Earthquake Energy Dissipation Systems View project

All content following this page was uploaded by Yen Lei Voo on 14 August 2021.

The user has requested enhancement of the downloaded file.


ACI STRUCTURAL JOURNAL TECHNICAL PAPER
Title No. 117-S23

Shear Strength of Dry and Epoxy Joints for Ultra-High-


Performance Fiber-Reinforced Concrete
by Balamurugan A. Gopal, Farzad Hejazi, Milad Hafezolghorani, and Yen Lei Voo
Joints between precast segmental bridge girders (PSBGs) represent shear behavior of reinforced concrete multiple-keyed dry
the locations of discontinuity through which compression and shear joints by a finite element (FE) model and proposed a new
forces are transmitted. Therefore, knowledge of the joint behavior design provision model.
to accurately predict the response of PSBGs is essential. The aim Zhou et al.2 experimented using full-scale normal-strength
of this research is to investigate the shear behavior, shear capacity,
concrete shear joints, with different variables (that is, types
and shear-transfer mechanisms of the ultra-high-performance
of joints, numbers of shear keys, and amount of confining
fiber-reinforced concrete (UHPFRC) keyed dry and epoxy joints.
For this purpose, a series of full-scale UHPFRC keyed joint spec- stress). They compared the experimental results with the
imens were manufactured and tested with three variable parame- results from AASHTO and other shear design provision
ters—namely, number of shear keys, confining stress, and the type models and found that the existing shear design provision
of joint (dry or epoxy). The observations from experimental tests model overestimated the results by a value up to 40% for
showed that the capacity of the UHPFRC keyed joints increased single-keyed joints and multiple-keyed epoxy joints. Turmo
with increasing confining pressure applied across the joint, number et al.8 reviewed the different shear design provisions avail-
of shear keys, and the epoxy layers applied on joints. Furthermore, able for PSBGs to evaluate the joint capacity of the exper-
a new design shear model for UHPFRC precast segmental bridges imental tests from the literature. Moreover, he proposed a
was developed and validated with experimental data. new shear design formula for PSBGs with keyed dry joints
Keywords: bridges; dry; epoxy; joint; precast; shear keys; shear strength;
to be included in the Eurocode. Alcalde et al.9 analyzed the
ultra-high-performance fiber-reinforced concrete (UHPFRC). fracture behavior of keyed dry joints under shear loading,
focusing on the influence of the number of keys on the joint
INTRODUCTION capacity and its average shear stress. In this study, they
Ultra-high-performance fiber-reinforced concrete (UHPFRC) concluded that the formulations of different design codes do
is recognized as a solution to precast segmental bridge not conform to the behavior of multiple-keyed joints. In a
girder (PSBG) problems. The overall behavior, including study by Hu and Xie,10 they evaluated the mechanical and
the ultimate strength of segmental bridges, depends very failure characteristics of reinforced concrete multiple-keyed
much on the behavior and integrity of the joints between joints by testing a large-scale model and a three-dimensional
the segments. These joints represent locations of discon- (3-D) FE model was developed for verification purposes.
tinuity through which compression and shear forces are From this research, it was concluded that vertical shear keys
transmitted.1,2 In PSBGs, the shear force at the joint can be mainly contributed to resisting the expansion of segment
transferred through the web and flanges by two different joints in a bending condition. Shamass et al.11 developed an
mechanisms called non-keyed (flat) joints and single/multi- FE analysis using a concrete-damaged plasticity model for
ple-keyed joints. Early segmental bridges typically adopted microcracks in normal-strength concrete material to simu-
a single key in the web section and, often, these were rein- late the shear performance of single and three-keyed epoxy
forced. Contemporary practice, however, is to use multiple joints. In the FE modeling, the epoxy resin was modeled as
unreinforced keys that are distributed through the depth a linear elastic material. Further, an empirical formula was
and width of the joints.3,4 The precast concrete segmental developed for assessing the shear strength of single-keyed
joints can be constructed and erected either using an epoxy epoxy joints of normal-strength concrete. Han et al.12 inves-
layer between the segments or in a dry condition.5 One of tigated the seismic behavior of reinforced concrete exte-
the most used techniques in the precast segmental bridge is rior shear keyed joints through a number of experimental
in their construction by using dry-keyed joints as the speed tests. Also, two analytical models for predicting the force-
of erection, and the lack of reliance on weather conditions displacement backbone curves were determined, which were
make this technique more suitable than using epoxy joints. in good agreement with the experimental results.
However, the capacity of the joints increases by applying Bu and Wu5 investigated the shear behavior of normal-
epoxy layers.6 Buyukozturk et al.1 conducted experimental strength concrete dry non-keyed, and multiple keys precast
tests with variable parameters such as flat and keyed joints, segmental bridge joints under monotonic direct shear
level of confining force, and dry or epoxy joints to assess the loading. The results from shear resistance-vertical displace-
shear strength and deformation behavior of normal-strength ACI Structural Journal, V. 117, No. 1, January 2020.
concrete segmental joints. They determined that the strength MS No. S-2019-033, doi: 10.14359/51718078, received January 27, 2019, and
reviewed under Institute publication policies. Copyright © 2020, American Concrete
of the epoxied joints is consistently higher than the non- Institute. All rights reserved, including the making of copies unless permission is
obtained from the copyright proprietors. Pertinent discussion including author’s
epoxied joints, although the failure of the epoxied joints was closure, if any, will be published ten months from this journal’s date if the discussion
found to be very sudden and brittle. Rombach7 analyzed the is received within four months of the paper’s print publication.

ACI Structural Journal/January 2020 279


ment curves, damage patterns, and failure mechanisms were the complete solution for sustainable constructions. This
investigated in detail. A study by Tawadrous and Morcous13 research will provide an essential contribution to the devel-
used 16 full-scale pushoff normal-strength concrete speci- opment of UHPFRC guidelines for precast segmental bridges
mens with clustered shear connectors on concrete and steel in the future, particularly in the area of the joints. Shear
girders and compared the results with 162 pushoff experi- behavior, shear capacity, and shear-transfer mechanisms of
mental tests from the literature. As mentioned earlier, while the UHPFRC keyed dry and epoxy joints are investigated
considerable research has been undertaken to investigate the through sufficient numbers of experimental tests. New shear
behavior of conventional precast concrete segmental bridges design models are proposed to assess the shear capacity of
with dry and epoxy joints, few researchers4,14 have attempted UHPFRC keys with different types of joints.
to evaluate shear behavior of UHPFRC keyed dry joints.
However, more importantly, no research has been carried out EXPERIMENTAL PROCEDURE
to assess the shear capacity of UHPFRC keyed epoxy joints. High tensile strength and enhanced ductility of UHPFRC
UHPFRC is an ultra-high-strength cementitious material is due to the very-high-strength microsteel fibers and cemen-
that contains a high quantity of cement and silica fume, low titious binders with extremely low water content. These prop-
quantity of water, incorporates large amounts of fibers, and erties make it possible to create very slender structures with
exhibits remarkable characteristics, such as extremely high minimum or no passive steel reinforcement. In the current
mechanical properties, enhanced fracture energy, extremely research, the aim of conducting experimental tests was to
low permeability, limited shrinkage and creep, as well as investigate the shear strength and behavior of typical shear
high corrosion resistance.15-18 Voo et al.4 reported that many joint keys used in PSBGs using UHPFRC concrete through
design models are available in the literature for the calcu- 12 real full-scale joint keys with a geometry closely resem-
lation of the shear strength of unreinforced concrete shear bling the key joint of actual PSBGs. Details of the experi-
keyed joints; no models, however, exist for UHPFRC keyed mental single-keyed, three-keyed, and five-keyed joint spec-
joints. Hence, they proposed a new design model to calcu- imens are depicted in Fig. 1(a), (b), and (c), respectively. It
late shear strength of UHPFRC keyed dry joint specimens can be observed from Fig. 1(d) that all the keyed joints have
using the principal of Mohr’s circle. Jang et al.14 compared a nominal thickness of 200 mm and a height of 1500 mm. As
the shear-off results of 10 dry normal-strength concrete and shown in Fig. 1, each half of the specimens was separated by
UHPFRC multiple keyed joint specimens where the experi- a 50 mm gap at their loading heads so that the applied force
mental results, including shear strength, failure mechanism, is resisted solely by shear through the line of the keys.
and displacement responses, were compared. They reported
that steel fibers had a significant effect on the shear behavior Description of specimens
of the specimens integrated with UHPFRC and size at 20 mm As shown in Table 1, a total of six dry-keyed joints and
aggregate remarkably enhances the shear of keyed joints. In six epoxy-keyed joint samples were fabricated from the
all, investigation on the structures and structural components UHPFRC and tested. In this table, the joint specimens were
constructed with UHPFRC as a bridge building material is designated as SK X-Y (N or E), where X represents the
still in its early stages in many countries compared with number of shear keys, Y represents the average confining
conventional concrete. Over the last decade, experimental pressure across the joint (σn), and N or E are the specimens
studies have been conducted and published on the struc- without and with epoxy at the joint segment interfaces.
tural behavior of UHPFRC members under different loading Furthermore, in Table 1, Pi, Ak, Asm, and Ake are the initially
conditions. However, the focus on the experimental studies applied normal force across the joint, total area for all keys
regarding shear mechanism of the full-scale UHPFRC keyed across the shear plane, area of the smooth section of the
dry and epoxy joints considering variable parameters such joint in shear, and total area for all epoxied keys across the
as number of shear keys and confining stress is almost joint, respectively.
nonexistent. Moreover, the existing provisions3,8,19 tended
to significantly overestimate the shear capacity of the joint Mixing and curing procedures of UHPFRC joint
specimens and are developed for normal-grade concretes, specimens
which cannot be used in UHPFRC joints. To address these The UHPFRC Grade 170 MPa (mean cube compressive
issues, this paper presents new design provision models for strength) was mixed using a fully automated mixer. The
UHPFRC girders with dry and epoxy joints. Validation of UHPFRC consists of a high amount of ordinary portland
the proposed design provision models is accomplished by cement, microsilica, fine sand with maximum diameter of
conducting of 12 full-scale UHPFRC keyed joint specimens 1.5 mm, large quantities of steel fibers, water, and a high-
with three variable parameters—namely, number of shear range water-reducing admixture. The mixture proportion
keys, confining stress, and the type of joint (dry or epoxy). of the UHPFRC is shown in Table 2. The steel fiber has a
tensile strength greater than 2500 MPa. The temperature
RESEARCH SIGNIFICANCE during mixing of UHPFRC was approximately 23°C. After
Joints in precast segmental bridges are weaker than those casting the single-, three-, and five-keyed UHPFRC joint
of adjacent monolithic sections within the segments. Nowa- specimens, initial curing was performed before standard heat
days, researchers are directing increasing attention and initia- curing. Within 30 minutes of casting, the exposed surface of
tive toward exploiting UHPFRC as the future construction the specimens and test control samples were sprayed with a
material, in the belief that UHPFRC technology embraces curing compound. Later, all UHPFRC specimens underwent

280 ACI Structural Journal/January 2020


Fig. 1—Details of UHPFRC keyed joint specimens.
Table 1—Experimental specimen details
Type No. Specimen No. of keys σn, MPa Pi, kN Ak, mm2 Asm, mm2 Ake, mm2 Ak/Ajoint, %
1 SK5-10N 5 10 1600 75,000 85,000 0 47
2 SK5-20N 5 20 3200 75,000 85,000 0 47

Keyed dry 3 SK3-10N 3 10 1600 45,000 115,000 0 28


joints 4 SK3-20N 3 20 3200 45,000 115,000 0 28
5 SK1-10N 1 10 1600 15,000 145,000 0 9.4
6 SK1-20N 1 20 3200 15,000 145,000 0 9.4
7 SK5-10E 5 10 1600 75,000 85,000 96,350 47
8 SK5-20E 5 20 3200 75,000 85,000 96,350 47

Keyed epoxy 9 SK3-10E 3 10 1600 45,000 115,000 57,810 28


joints 10 SK3-20E 3 20 3200 45,000 115,000 57,810 28
11 SK1-10E 1 10 1600 15,000 145,000 19,270 9.4
12 SK1-20E 1 20 3200 15,000 145,000 19,270 9.4

heat treatment of 90°C and 100% humidity for 48 hours as Test setup
recommended by the French standard.20 Preparation of the All the UHPFRC joint specimens were tested under vertical
UHPFRC joint specimens after heat treatment is shown in monotonic loading at the centerline of the shear planes with
Fig. 2. The male and female parts of the six epoxy-jointed negligible moment, as shown in Fig. 3. This force was
specimens were joined together using thixotropic epoxy applied directly on a 50 mm thick solid bearing plate at the
adhesive resin and 5 tonne concrete blocks were positioned top of the joint specimens and slip measured using LVDTs.
on top to densely pack them together (refer to Fig. 2(a) The horizontal confining stress was applied to the dry and
through (d)). epoxy-keyed joint surfaces through a stiff steel framework
and loaded using 15.2 mm diameter prestressing strands, as
shown in Fig. 3(a). For the specimens with confining pres-

ACI Structural Journal/January 2020 281


sure levels of 10 and 20 MPa, a total of eight and 16 strands tonic load with a stroke rate of 0.005 mm/s was applied up to
were used, respectively, with each strand stressed to 200 kN the failure of all shear keys. Two LVDTs were used for each
using a hydraulic monojack and then anchored. Each strand keyed joint specimen to measure the slip during loading as
was tensioned at least three times to minimize the loss of shown in Fig. 3(a) and (b).
stress. As shown in this figure, one rigid vertical column
was located on a concrete block and connected through an Mechanical properties of UHPFRC
adequate number of high-strength bolts. After positioning The summary of the mechanical properties for all the
the specimens on the vertical steel frame, a constant mono- keyed joint specimens SK1, SK3, and SK5 are shown in
Table 3, where fcm,cu is the mean cube compressive strength
Table 2—Mixture design of UHPFRC at 1 and 28 days and were measured using 100 x 100 x
Ingredient Amount 100 mm cubes. fctm,el and fctfm are the mean tensile limit of
elasticity and mean post-cracking tensile strength, respec-
Premix (cement, silica fume, and fine sand), kg/m3 2100
tively, according to the French Standard.20 The term fctm,fl is
High-range water-reducing admixture, kg/m3 36 the equivalent elastic flexural strength, which was measured
High-strength steel fiber (straight and hooked), kg/m
3
157 (2% by vol.) using 100 x 100 x 500 mm prism under four-point test.
Free water, kg/m 3
144 To assess the concrete to concrete bond strength, several
test methods exist, and in this research the Slant shear test
3% moisture, kg/m 3
30
was conducted according to ASTM standard,21 which is the
Targeted water-binder ratio (w/b) 0.15 most common method to evaluate the bond strength in the
Total air voids, % 4 shear of epoxy-based resins. For this purpose, 36 UHPFRC

Fig. 2—Preparation of UHPFRC joint specimens.

Fig. 3—Experimental setup of UHPFRC keyed joint specimens under shear test.

282 ACI Structural Journal/January 2020


Table 3—Summary of mechanical properties of control specimens
Keyed joint specimens
SK1 SK3 SK5
Mechanical properties of UHPFRC Mean S.D. Mean S.D. Mean S.D.
Modulus of elasticity, E0, GPa 46.5 0.57 44.8 0.28 45.8 0.85
Cube compressive strength, fcm,cu,1d*, MPa 91.3 4.7 90.3 4.3 84.8 3.3
Cube compressive strength (fcm,cu,28d†), MPa 172.9 4.81 166.1 3.71 170.4 6.87
Tensile limit of elasticity (fctm,el‡), MPa 9.1 0.28 8.2 0.28 9.4 0.07
Post-cracking tensile strength (fctfm§), MPa 11.9 0.01 12.1 1.27 13.2 0.92
Equivalent flexural strength (fctm,fl‡), MPa 27.7 3.1 32.4 2.1 30.7 2.0
*
For 1-day mean strength, each joint tested with 3 nos. of 100 mm cubes (before heat curing).

For 28-day mean strength, each joint tested with 3 nos. of 100 mm cubes (after heat curing).

For 28-day mean strength, each joint tested with 3 nos. of 100 mm prisms (after heat curing) under 4-point test.
§
For 28-day mean strength, each joint tested with 3 nos. of 100 mm prisms (after heat curing) under 3-point test.

Fig. 4—Example of shear behavior and typical failure of shear keys in SK3-10N.
slant cylinders with a diameter of 75 mm and height of
150  mm were tested and the mean value of bond strength
(C) of 3.7 MPa was calculated, where the standard deviation
and coefficient of variation for the 36 slant shear samples
were 0.23 and 6.2, respectively.

EXPERIMENTAL RESULTS AND OBSERVATION


From the experimental tests, shear behavior, shear capacity,
and shear-transfer mechanisms of the joints were reported.
Three important shear load capacities were monitored and
recorded through shear load-displacement graphs as shown
in Fig. 4(a). As shown in this figure, Vj,cr,exp is the experi-
mental shear load capacity at which the first shear crack Fig. 5—Shear load-displacement comparisons of all joint
occurred at the lower shear key face of the male key compo- specimens.
nent and is taken from the load-displacement curve, where
the assumption of linear elasticity is no longer applicable. figure it can be observed that the slips between the surfaces
After the shear key had cracked, the strength was derived were minimal until the peak load (Vj,u,exp) was achieved.
from the steel fibers crossing the shear plane. The term After the peak load was ached, the keys slid and a sudden
Vj,u,exp is the ultimate shear load capacity of the shear keyed drop in load was observed. All the experimental results are
joints. The term Vj,fric,exp is the residual frictional shear load summarized and compared together in terms of number
capacity on the contact surface after the shear keys sheared of shear keys, confining stress, and the type of joint (dry
off completely, as shown in Fig. 4(b). To find the static or epoxy) in Table 4. From experimental results, it can be
friction coefficients (μ) due to horizontal confining forces concluded that the failure shear loads at first crack (Vj,cr,exp)
between the male and female surfaces, after the failure of and ultimate (Vj,u,exp) are higher with an increasing number
all the shear keys, the load was continually applied. μ is the of keys and at higher confining pressure. For instance, as
ratio of residual frictional shear load (Vj,fric,exp) to the initially reported in Table 4, compared to the UHPFRC keyed dry
applied normal force across the joint (Pi). joint SK1-10N, the failure shear load capacities at first
The experimental shear load-displacement graphs of all crack (Vj,cr,exp) and ultimate (Vj,u,exp) of the UHPFRC keyed
the full-scale joint specimens are shown in Fig. 5. From this dry joint SK5-20N improved by 142% and 152%, respec-

ACI Structural Journal/January 2020 283


Table 4—Summary of experimental results for keyed joint specimens
Specimen No. keys σn, kN Pi, kN Vj,cr,exp, kN Vj,u,exp, kN Vj,fric,exp, kN µ τj,cr,exp, MPa τj,u,exp, MPa
Keyed dry joints
SK1-10N 1 10 1600 930 1278 800 0.50 5.38 7.99
SK1-20N 1 20 3200 1410 1826 1291 0.40 7.69 11.41
SK3-10N 3 10 1600 1270 1953 800 0.50 8.56 12.21
SK3-20N 3 20 3200 1800 2705 1350 0.42 11.88 16.91
SK5-10N 5 10 1600 1700 2914 800 0.50 13.75 18.21
SK5-20N 5 20 3200 2200 3216 1350 0.42 15.00 20.10
Keyed epoxy joints
SK1-10E 1 10 1600 1300 1601 750 0.47 6.88 10.00
SK1-20E 1 20 3200 1800 2288 1250 0.39 10.00 14.30
SK3-10E 3 10 1600 1900 2258 775 0.48 10.00 14.11
SK3-20E 3 20 3200 2000 2864 1270 0.40 12.50 17.90
SK5-10E 5 10 1600 2100 2984 770 0.48 14.38 18.65
SK5-20E 5 20 3200 2600 3495 1310 0.41 16.25 21.84

components of the UHPFRC single-keyed joints compared


to the UHPFRC multiple-keyed joints. Moreover, the cohe-
sive effect of the epoxy layer not only increased the shear
resistance of the joint but also allowed the epoxy keyed
joint specimen to displace more compared to the keyed dry
joint specimens before these specimens fully detached. For
instance, as shown in Fig. 5, under the effect of 10 MPa
confining stress, the keyed epoxy joint specimens SK1-10E,
SK3-10E, and SK5-10E displaced more than the keyed
dry joints by 17.2%, 58.7%, and 45%, respectively, before
reaching the ultimate shear loads. However, as shown in
Table 4, the residual frictional shear capacities (Vj,fric,exp)
and the static friction coefficients (μ) of the UHPFRC
keyed dry joint specimens indicated a higher amount in
comparison with the UHPFRC keyed epoxy joint speci-
mens. This would occur due to the presence of the epoxy
powder layer between the UHPFRC surfaces after failure,
having a lubricating effect. Ultimately, by considering the
Fig. 6—Proposed shear model for UHPFRC precast effects of all three variables (that is, the number of keys,
keyed joints. confining stress, and epoxy layer), the failure shear capacity
tively. Also, based on the data presented in Table 4 and the loads (Vj,cr,exp and Vj,u,exp) of the keyed epoxy joint specimens
graph patterns as illustrated in Fig. 5, it provides a vital showed a remarkable increment compared to the keyed dry
point in that specimens with an epoxy layer on key joints joint specimens. In this instance, the Vj,cr,exp and Vj,u,exp of
tend to produce higher failure shear capacity loads (Vj,cr,exp the UHPPFRC keyed joint SK5-20 increased by 184% and
and Vj,u,exp) as compared to specimens with keyed dry joints. 173%, respectively, compared to Specimen SK1-10N. Simi-
For example, under the effect of 20 MPa confining stress, larly, the slip at the ultimate stage in SK5-20E increased by
SK1-20E, SK3-20E, and SK5-20E keyed epoxy joint speci- 87% compared to SK1-10N, which indicates a high energy
mens recorded higher ultimate shear capacity loads (Vj,u,exp) dissipation capacity of the UHPFRC keyed joint SK5-20E.
compared to the keyed dry joint specimens by 25%, 6%, and The shear-strength capacities of the UHPFRC joint speci-
9%, respectively. Similarly, the keyed epoxy joint specimens mens at first crack (τj,cr,exp) and ultimate (τj,u,exp) are presented
produced higher values of Vj,cr,exp compared to the keyed in Table 4.
dry joint specimens (refer to Table 4). Therefore, it can be
concluded that the epoxy layer has an impressive key role DEVELOPMENT OF NEW DESIGN SHEAR MODEL
in achieving higher failure shear capacity loads (Vj,cr,exp and FOR UHPFRC JOINTS
Vj,u,exp) in UHPFRC keyed joints, especially in the single- To develop a new design shear model for UHPFRC keyed
keyed joint specimen compared to the multiple-keyed joints joints, the principle of Mohr’s circle was used in this study,
while, at the same time, the epoxy layer has an impressive as shown in Fig. 6. Moreover, as reported by previous
key role in controlling the slip between the male and female researchers,15,16 UHPFRC can be treated as a perfectly

284 ACI Structural Journal/January 2020


plastic material and plasticity method can be applied for 15,000 mm2 as shown in Fig. 1. The frictional force due to
this concrete. From the principle of Mohr’s circle, the major the cohesion of the epoxy is derived as
principal tensile stress (σ11) can be obtained as follows
Vcoh = C(Asm + Ake) (9)
σx + σ y  σx + σ y 
2

σ11 = +   + τ 2xy (1) where Ake is the area of an epoxy layer on the male compo-
2  2  nent of the key surface (refer to Table 1). C is the cohesion
of the epoxy layer on the UHPFRC surface, whereby this
where σx = σn and σy = 0 are normal stresses in the x- and value is determined through the bond strength of epoxy resin
y-directions, respectively. For limit states design, σ11 = on the concrete surfaces based on the ASTM standard,21 as
ft, where ft is uniaxial tensile strength of the composite. mentioned earlier in this paper.
UHPFRC exhibits strain hardening behavior and ft can be The theoretical shear joint capacity of the keyed dry joints
determined as ft = fct,el at first crack and ft = fctfm at ultimate, at the first crack (Vjd,cr,theo) and at the ultimate state (Vjd,u,theo)
where fct,el and fctfm are tensile limit of elasticity and post- can be calculated from Eq. (10) and (11).
cracking tensile strength respectively, as per French Stan-
dard20 (refer to Table 3). Accordingly, the shear strength (τxy) V jd , cr ,theo = [ (−0.009σ n + 0.59) Asm σ n ]
can be calculated follows
n σn   σn  
2 2

+ ∑ Aki  f ct , el +  −   
 (10)
2 2  i =1  2  2 
 σ  σ   
τ xy =  f t + n  −  n  (2)
 2  2
V jd ,u ,theo = [ (−0.009σ n + 0.59) Asm σ n ]
The joint shear capacity for keyed dry joints (Vjd) shall be n  σ  σ  
2 2

taken as the superposition of the frictional component of the +  ∑ Aki  f ctfm + n  −  n   (11)
 i =1  2  2 
contact surfaces (Vsm) and the shear strength contribution of  
the keys (Vk). Thus, it can be expressed as
The theoretical shear joint capacity of the keyed epoxy
Vjd = Vsm + Vk (3) joints at first crack (Vje,cr,theo) and at the ultimate state (Vje,u,-
theo) can then be determined from Eq. (12) and (13).
While, for keyed epoxy joints, the joint shear capacity (Vje)
is calculated as V je , cr ,theo = C ( Asm + Ake ) + [ (−0.007σ n + 0.54) Asm σ n ]
n σn   σn   (12)
2 2
Vje = Vsm + Vk + Vcoh (4) 
+ ∑ Aki  f ct , el +  −   

 i =1  2  2 
 
where Vcoh is the additional shear strength produced by the
cohesion parameter of epoxy resin. The frictional force (Vsm)
due to the average confining pressure across the joint (σn) V je ,u ,theo = C ( Asm + Ake ) + [ (−0.007σ n + 0.54) Asm σ n ]
can be written as
n σ  σ   (13)
2 2

Vsm = μAsmσn (5) +  ∑ Aki  f ctfm + n  −  n  
 i =1  2  2 
 
As presented in Table 1, Asm is the area of the smooth section
of the keyed joints. μ is the static friction coefficient of Comparison of proposed shear design model with
concrete to the concrete surface and c be estimated by the experimental data
empirical formula as To assess the reliability of the proposed shear design provi-
sion model, a comparison study between the experimental
μ = –0.009σn + 0.59 Keyed dry joints (6) results and the calculated values from proposed model was
conducted in this section. As shown in Fig. 7, comparisons
μ = –0.007σn + 0.54 Keyed epoxy joints (7) were made on the first crack and ultimate stages on the both
keyed dry joints (refer to Tables 5 and 6) and keyed epoxy
The shear force capacity from the shear keys is then given as joints (refer to Tables 7 and 8).
As observed in Fig. 7, the shear capacity loads calculated
n
from the new design provision model show an excellent
Vk = ∑ Aki � τ xy (8)
i =1 match with the experimental results for a single keyed joint.
The estimated first crack loads from the proposed provision
where Aki is the area of the shear key, while i and n are the model and the loads recorded through the experiments on
number of keys. In this study, each shear key had a dimen- the SK1-10N, SK1-20N, SK1-10E, and SK1-20E speci-
sion of 100 x 150 mm and, thus, the area of each key was mens differed by 2.9%, 6.4%, 0.2%, and 0.4%, respectively.
However, a maximum difference of 24.3% is observed

ACI Structural Journal/January 2020 285


between the experimental and proposed model results for ratio of theo/exp for the SK5-10N, SK5-20N, SK5-10E, and
SK1-10N at ultimate. From Tables 5 to 8, the estimated SK5-20E specimens are calculated as 0.79, 0.85, 0.87, and
ratio of theo/exp for the SK3-10N, SK3-20N, SK3-10E, 0.96 at the first crack and 0.69, 0.71, 0.80 and 0.84 at the
and SK3-20E specimens are 0.82, 0.87, 0.92, and 0.98 at ultimate stage, respectively.
the first crack and 0.67, 0.68, 0.85, and 0.85 at the ultimate It can be mentioned that the theo/exp ratios achieved in this
stage, respectively. These ratios indicate a good correlation study, fluctuated at a very minimum range (less than 12%)
between the new provision and the test data, leading to a safe when the confining stress increased from 10 to 20 MPa. It
design as shown in Fig. 8. As presented in Fig. 7 and Tables is quite obvious that the mean value of the theo/exp ratios
5 to 8, comparison of the results indicated a good correlation and the corresponding coefficient of variations (COV) of the
between the proposed provision model and the test data for four groups of specimens—keyed dry joints at first crack
the five-keyed joints, especially for epoxy specimens. The (Table 5), keyed dry joints at ultimate (Table 6), keyed epoxy
joints at first crack (Table 7), and keyed epoxy joints at ulti-
mate (Table 8)—are less than 1.0 and 10%, respectively.
Distribution of the theory against experimental data points
for the keyed dry and epoxy joints are shown in Fig. 8(a)
and (b), respectively. From these figures, it can be seen that
all the first cracks and ultimate shear capacities calculated
through the proposed shear design model are positioned at
the safe zone in comparison to the experimental data. It can
therefore be concluded that the new shear design model will
estimate safe, consistent, and reliable values of failure shear
capacity loads, which can lead to safe design of UHPFRC
precast segmental bridges. However, from the safe zone
comparison graphs, maximum differences of 32% and 20%
are observed between the ultimate shear capacities of theory
and experimental tests of the keyed dry joint SK3-10N
and keyed epoxy joint SK5-10E, respectively, which may
give uneconomic predictions when using in the design of
UHPFRC precast segmental bridges.

CONCLUSIONS
In this research, a new design shear model for UHPFRC
precast segmental bridges with keyed dry and epoxy joints
was developed and validation of the proposed model was
accomplished through comparison with experimental obser-
vations. Further, to assess the shear behavior and shear
capacity of UHPFRC precast segmental bridges, 12 full-scale
shear joint key specimens were cast and tested experimen-
tally up to failure with three variable parameters including
the number of shear joint keys, amount of confining stress,
Fig. 7—Comparison of proposed model and test results at and the type of joint keys (dry and epoxy).
first crack and ultimate.
Table 5—Comparison of proposed model and test results at first crack for keyed dry joints
At first crack (SLS)
Surface friction Shear key Theory versus experiment
Specimen μ Vsm, kN fct.el, MPa τxy, MPa Vk, kN Vj,cr,theo, kN Vj,cr,exp, kN Theo/Exp
SK1-10N 0.50 725.0 9.1 13.2 197.8 922.8 950.0 0.97
SK1-20N 0.40 1160.0 9.1 16.3 244.1 1404.1 1500.0 0.94
SK3-10N 0.50 575.0 8.2 12.2 549.8 1124.7 1370.0 0.82
SK3-20N 0.42 966.0 8.2 15.2 684.3 1650.3 1900.0 0.87
SK5-10N 0.50 425.0 9.4 13.5 1008.8 1433.8 1800.0 0.79
SK5-20N 0.42 714.0 9.4 16.6 1242.4 1956.4 2300.0 0.85
Mean 0.87
STD 0.06
COV 7.71

286 ACI Structural Journal/January 2020


Table 6—Comparison of proposed model and test results at ultimate for keyed dry joints
At ultimate stage (ULS)
Surface friction Shear key Theory versus experiment
Specimen μ Vsm, kN fctfm (MPa) τxy, MPa Vk, kN Vj,cr,theo, kN Vj,cr,exp, kN Theo/Exp
SK1-10N 0.50 725.0 11.9 16.1 242.2 967.2 1278.0 0.76
SK1-20N 0.40 1160.0 11.9 19.5 292.3 1452.3 1826.0 0.80
SK3-10N 0.50 575.0 12.1 16.5 735.9 1310.9 1953.0 0.67
SK3-20N 0.42 966.0 12.1 19.7 886.9 1852.9 2705.0 0.68
SK5-10N 0.50 425.0 13.2 17.5 1308.6 1733.6 2514.0 0.69
SK5-20N 0.42 714.0 13.2 20.9 1565.9 2279.9 3216.0 0.71
Mean 0.72
STD 0.05
COV 6.73

Table 7—Comparison of proposed model and test results at first crack for keyed epoxy joints
At first cracking stage (SLS)
Epoxy cohesion Surface friction Shear key Theory versus experiment
Specimen C, MPa Vcoh, kN μ Vsm, kN fct,el (MPa) τxy, MPa Vk, kN Vj,cr,theo, kN Vj,cr,exp, kN Theo/Exp
SK1-10E 3.7 607.8 0.47 681.5 9.1 13.2 197.8 1487.1 1490.0 0.99
SK1-20E 3.7 607.8 0.39 1131.0 9.1 16.3 244.1 1982.9 1990.0 0.99
SK3-10E 3.7 639.4 0.48 552.0 8.2 12.2 549.7 1741.1 1900.0 0.92
SK3-20E 3.7 639.4 0.40 920.0 8.2 15.2 684.3 2243.7 2300.0 0.98
SK5-10E 3.7 671.0 0.48 408.0 9.4 13.5 1008.8 2087.8 2400.0 0.87
SK5-20E 3.7 671.0 0.41 697.0 9.4 16.6 1242.4 2610.4 2700.0 0.96
Mean 0.95
STD 0.05
COV 5.31

Table 8—Comparison of proposed model and test results at ultimate for keyed epoxy joints
At ultimate stage (ULS)
Epoxy cohesion Surface friction Shear key Theory versus experiment
Specimen C, MPa Vcoh, kN μ Vsm, kN fctfm, MPa τxy, MPa Vk, kN Vj,cr,theo, kN Vj,cr,exp, kN Theo/Exp
SK1-10E 3.7 607.80 0.47 681.5 11.9 16.1 242.2 1531.5 1601.0 0.96
SK1-20E 3.7 607.80 0.39 1131.0 11.9 19.5 292.3 2031.1 2288.0 0.89
SK3-10E 3.7 639.40 0.48 552.0 12.1 16.4 735.9 1927.3 2258.0 0.85
SK3-20E 3.7 639.40 0.40 920.0 12.1 19.7 886.9 2446.3 2864.0 0.85
SK5-10E 3.7 671.0 0.48 408.0 13.2 17.5 1308.6 2387.6 2984.0 0.80
SK5-20E 3.7 671.0 0.41 697.0 13.2 20.9 1565.9 2933.9 3495.0 0.84
Mean 0.87
STD 0.05
COV 6.12

The following conclusions are subsequently drawn from joints indicated higher values compared to the keyed epoxy
these experimental results: joints. This can be due to the presence of epoxy powder layer
1. From the experimental tests, the shear capacities of joint between the UHPFRC surfaces after failure, having a lubri-
specimens improved up to 150% with an increasing number cating effect. However, further investigations are needed to
of keys and at higher confining stress. clearly address this difference.
2. By considering the effect of the epoxy layer, the keyed Ultimately, this research will provide an essential contri-
epoxy joints tended to produce higher shear failure capacity bution to the development of UHPFRC guidelines for
up to 25% compared to the specimens with dry joints. precast segmental bridges in the future, particularly in the
3. The residual frictional shear capacities (Vj,fric,exp.) and area of the joints.
the static friction coefficients (μ) of the UHPFRC keyed dry

ACI Structural Journal/January 2020 287


design of reinforced concrete and ultra-high-performance fiber-reinforced
concrete structures.

Yen Lei Voo is an Adjunct Associate Professor in the School of Civil and
Environmental Engineering at the University of New South Wales (UNSW),
Sydney, NSW, Australia, and the Managing Director and Founder of Dura
Technology Sdn. Bhd in Malaysia. In 1997, he was offered a place at UNSW,
where he received his bachelor’s in engineering (civil) degree with First
Class Honors in 2000. Upon graduating, he was being immediately offered
a full scholarship by the same university for his noteworthy achievements to
complete his PhD in civil engineering. His research interests include design
of precast ultra-high-performance fiber-reinforced concrete bridges.

REFERENCES
1. Buyukozturk, O.; Bakhoum, M. M.; and Beattie, S. M., “Shear
Behavior of Joints in Precast Concrete Segmental Bridges,” Journal of
Structural Engineering, ASCE, V. 116, No. 12, 1990, pp. 3380-3401.
2. Zhou, X.; Mickleborough, N.; and Li, Z., “Shear Strength of Joints
in Precast Concrete Segmental Bridges,” ACI Structural Journal, V. 102,
No. 1, Jan.-Feb. 2005, p. 3.
3. Rombach, G. A., and Specker, A., “Segmentbrücken,” Beton-Kal-
ender, Teil 1, Verlag Ernst und Sohn, Berlin, Germany, 2004, pp. 177-211.
4. Voo, Y. L.; Foster, S. J.; and Voo, C. C., “Ultrahigh-Performance Concrete
Segmental Bridge Technology: Toward Sustainable Bridge Construction,”
Journal of Bridge Engineering, V. 20, No. 8, 2014, p. B5014001.
5. Bu, Z. Y., and Wu, W. Y., “Inter Shear Transfer of Unbonded
Prestressing Precast Segmental Bridge Column Dry Joints,” Engineering
Structures, V. 154, June 2018, pp. 52-65.
6. Issa, M. A., and Abdalla, H. A., “Structural Behavior of Single Key
Joints in Precast Concrete Segmental Bridges,” Journal of Bridge Engi-
neering, V. 12, No. 3, 2007, pp. 315-324.
7. Rombach, G., “Dry Joint Behavior of Hollow Box Girder Segmental
Bridges,” FIP Symposium, Segmental Construction in Concrete, New
Delhi, India, 2004.
8. Turmo, J.; Ramos, G.; and Aparicio, A. C., “Shear Strength of Dry Joints
of Concrete Panels with and Without Steel Fibers: Application to Precast
Segmental Bridges,” Engineering Structures, V. 28, No. 1, 2006, pp. 23-33.
9. Alcalde, M.; Cifuentes, H.; and Medina, F., “Influence of the Number
of Keys on the Shear Strength of Post-Tensioned Dry Joints,” Materiales de
Construcción, V. 63, No. 310, 2013, pp. 297-307.
10. Hu, Z. N., and Xie, Y. L., “Mechanical and Failure Characteristics of
Shear Keys on Immersed Tunnel Segment Joints under Differential Settle-
ments,” Procedia Engineering, V. 166, 2016, pp. 373-378.
11. Shamass, R.; Zhou, X.; and Wu, Z., “Numerical Analysis of Shear-Off
Failure of Keyed Epoxied Joints in Precast Concrete Segmental Bridges,”
Journal of Bridge Engineering, V. 22, No. 1, 2016, p. 4016108.
12. Han, Q.; Zhou, Y.; and Ou, Y., “Seismic Behavior of Reinforced
Fig. 8—Safe zone comparison of proposed shear model with Concrete Sacrificial Exterior Shear Keys of Highway Bridges,” Engi-
experimental results. neering Structures, V. 139, 2017, pp. 59-70.
13. Tawadrous, R., and Morcous, G., “Interface Shear Resistance of
A detailed parametric study using FE software is suggested Clustered Shear Connectors for Precast Concrete Bridge Deck Systems,”
Engineering Structures, V. 160, No. 1, 2018, pp. 195-211.
to assess the overall behavior of dry and epoxy UHPFRC 14. Jang, H. O.; Lee, H. S.; and Cho, K., “Experimental Study on Shear
precast segmental bridges. Performance of Plain Construction Joints Integrated with Ultra-High
Performance Concrete (UHPC),” Construction and Building Materials,
V. 152, 2017, pp. 16-23.
AUTHOR BIOS 15. Voo, Y. L.; Foster, S. J.; and Gilbert, R. I., “Shear Strength of Fiber
Balamurugan A. Gopal is a PhD Candidate in the Department of Civil Reinforced Reactive Powder Concrete Prestressed Girders without Stirrups,”
and Structural Engineering at the University of Putra Malaysia, Seri Journal of Advanced Concrete Technology, V. 4, No. 1, 2006, pp. 123-132.
Kembangan, Malaysia. He received his BE and MS from the University of 16. Voo, Y. L., and Foster, S. J., “Characteristics of Ultra-High Performance
Science, Malaysia, George Town, Malaysia, in 1998 and 2002, respectively. ‘Ductile’ Concrete and Its Impact on Sustainable Construction,” IES Journal
His research interests include shear strength of precast segmental bridges Part A: Civil and Structural Engineering, V. 3, No. 3, 2010, pp. 168-187.
with ultra-high-performance fiber-reinforced concrete. 17. Buttignol, T. E. T.; Sousa, J. L. A. O.; and Bittencourt, T. N., “Ultra
High-Performance Fiber-Reinforced Concrete (UHPFRC): A Review of
Farzad Hejazi is an Associate Professor, Innovation and Research Coor- Material Properties and Design Procedures,” Revista IBRACON de Estru-
dinator in the Civil Engineering Department at the University of Putra turas e Materiais, V. 10, No. 4, 2017, pp. 957-971.
Malaysia. He received his BS, MS, and PhD in civil and structural engi- 18. El-Tawil, S.; Tai, Y.; and Belcher, J. A., “Field Application of Nonpro-
neering from Kerman University and University of Putra Malaysia, respec- prietary Ultra-High-Performance Concrete,” Concrete International, V. 40,
tively. His research interests include seismic design, finite element simula- No. 1, Jan. 2018, pp. 36-42.
tion, reinforced concrete and steel structures, and optimization of structural 19. AASHTO, “Guide Specifications for the Design and Construction
seismic control systems. of Segmental Concrete Bridges,” second edition, Washington, DC, 1999,
pp. 3-118.
Milad Hafezolghorani is the R&D Manager of Dura Technology Sdn. 20. NF P 18-470, “Concrete-Ultra-High Performance Fiber-Reinforced
Bhd in Malaysia. He received his BS, MS, and PhD in civil and structural Concrete-Specifications, Performance, Production and Conformity,” 2016.
engineering from Isfahan University, Isfahan, Iran; National University 21. ASTM C882/C882M-05, “Bond Strength of Epoxy-Resin Systems
of Malaysia, Bangi, Malaysia; and the University of Putra Malaysia in Used With Concrete By Slant Shear,” ASTM International, West Consho-
2010, 2013, and 2017, respectively. His research interests include seismic hocken, PA, 2005.

288 ACI Structural Journal/January 2020


View publication stats

You might also like