Engineering Structures 98 (2015) 109–117

Contents lists available at ScienceDirect

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

Effects of cast-in-place concrete topping on flexural response of precast

concrete hollow-core slabs
Eray Baran ⇑
Department of Civil Engineering, Atilim University, Ankara 06836, Turkey
Department of Civil Engineering, Middle East Technical University, Ankara 06800, Turkey

a r t i c l e i n f o a b s t r a c t

Article history: Results of a study focusing on the flexural response of precast prestressed concrete hollow-core slabs
Received 25 August 2014 with cast-in-place concrete topping are presented. The experimental part of the study included load test-
Revised 3 April 2015 ing of five precast concrete hollow-core units. The numerically determined flexural response of test spec-
Accepted 11 April 2015
imens was later compared with the experimentally obtained behavior. Results demonstrate that a major
composite action is valid between the hollow-core unit and the topping slab under load levels corre-
sponding to uncracked state of the cross section. Existence of a topping slab resulted in improvements
Keywords:
in the cracking moment and initial stiffness of hollow-core units. The beneficial effect of topping slab
Composite construction
Hollow-core slab
on the ultimate moment capacity was observed to be limited, mainly because of the loss of composite
Prestressed concrete action prior to reaching the ultimate moment capacity. Horizontal shear strength at the interface
Precast concrete between hollow-core unit and topping slab was determined (1) through limited number of pushoff load
Horizontal shear strength tests and (2) through calculations considering the load level corresponding to initiation of significant
relative slip using the basic mechanics of materials approach and the simplified code expression. The
measured and computed interface shear strength values were observed to be significantly lower than
the horizontal shear strength values specified by the ACI and AASHTO Specifications.
Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction transfer mechanism between hollow-core panels and the concrete

topping is through bond at the interface. There is an uncertainty,
Prestressed concrete hollow-core panels have been widely used however, whether the horizontal shear stress capacity at the inter-
throughout the world in concrete and steel structures, including face will be sufficient to allow a fully composite behavior under
buildings, parking structures, and bridges. One of the most com- flexural loading when the concrete topping is placed over machine
mon use of these elements is in floor systems, where precast finished surface of a hollow-core unit with no intentional roughen-
prestressed concrete hollow-core panels are used together with a ing. Current design specifications also addresses the concern for
cast-in-place (CIP) concrete topping to form a load-resisting com- insufficient shear force transfer at interface surfaces and specify a
posite floor system. The composite behavior in these structures is minimum roughening amplitude of 6.3 mm for the surface to be
desired because of the increased strength and stiffness under ver- considered as roughened [1,2].
tical loads and for resisting and transmitting forces resulting from Studies focusing on flexural behavior of composite hollow-core
diaphragm action under lateral loads. units are very scarce. Dowell and Smith tested prestressed concrete
Manufacture of precast concrete hollow-core units mainly hollow-core bridge deck panels with three different levels of
involves machine casting operation. These units are typically man- roughening applied on the surface of hollow-core units prior to
ufactured either by extrusion of really low-slump concrete, or casting of topping concrete [3]. During load testing of specimens,
slip-forming of a relatively higher slump concrete. These manufac- no sign of horizontal slip was observed between the hollow-core
turing techniques do not usually allow the use of any reinforcing units and the CIP topping with any of the surface roughening levels
bars crossing the interface between the hollow-core unit and the studied. Measured moment capacities of the specimens were
CIP concrete topping. Therefore, the primary horizontal shear force reported to exceed the predicted capacities assuming a fully com-
posite behavior. Scott tested a hollow-core slab unit with a CIP
⇑ Address: Department of Civil Engineering, Middle East Technical University,
concrete topping and the agreement between the calculated and
Ankara 06800, Turkey. Tel.: +90 312 210 7475. measured moment capacity and deflection values suggested a full
E-mail address: erayb@metu.edu.tr composite action between the hollow-core unit and the topping

http://dx.doi.org/10.1016/j.engstruct.2015.04.017
0141-0296/Ó 2015 Elsevier Ltd. All rights reserved.
110 E. Baran / Engineering Structures 98 (2015) 109–117

[4]. It was reported that the surface of the hollow-core unit did not Table 1
comply with the surface roughness requirements of the ACI Code, Properties of specimens.

as the unit had a smooth top surface and no intentional roughening Specimen Specimen Topping? Compressive strength of topping
was applied. Different from other more recent experimental stud- width (cm) concrete (MPa)
ies on precast concrete hollow-core units, the specimen in Scott’s W-B 120 No –
study was tested by placing solid concrete blocks on top of the W-T 120 Yes 31.1
composite slab. With this type of loading, it is highly possible that W-T-Ra 120 Yes 30.2

the additional friction force that developed under the effect of the N-B 55 No –
applied weights resulted in relatively higher horizontal shear force N-T 55 Yes 32.3

capacity at the interface between the hollow-core unit and the a

Two ends of concrete topping in this specimen were restrained.
topping concrete, as compared to applying a line load using a
hydraulic cylinder.
Due the fact that precast concrete hollow-core units typically do which has nominal widths of 120 cm and 55 cm, respectively. As
not have any transverse shear reinforcement, the shear strength of shown in Fig. 1, hollow-core slab units were 15 cm deep and rein-
these members has to be carefully looked at during the design forced only with 9.5 mm diameter Grade-270 seven-wire straight
stage. Moreover, with the use of relatively deeper units, the size prestressing strands positioned 4.5 cm from the bottom surface
effect on the shear strength becomes more significant. As a result, of cross section. Hollow-core units in W series specimens had eight
shear strength of hollow-core units has been studied extensively strands, while those in N series specimens had only four strands. In
[5–12]. Through these studies, it is shown that web-shear strength three of the specimens, a 5 cm thick CIP concrete topping was cast
of relatively deep hollow-core units may be smaller than those on top of the hollow-core unit prior to load tests, while the remain-
predicted by current design specifications. Studies focusing on ing two units were tested with no topping. A welded wire mesh
the shear behavior of composite hollow-core units revealed the with 6 mm bar diameter and 15 cm bar spacing was used inside
beneficial effect of CIP concrete topping on improving the shear topping slabs at approximately the mid-depth of slab to control
strength of the units. cracking due to shrinkage and temperature effects.
In order to ensure a proper bond between the CIP concrete top- The hollow-core units used in the test specimens were provided
ping and the precast hollow-core panels, the surface of the panels by a local precast concrete producer in Ankara–Turkey. The units
must be clean, free of laitance, and thoroughly saturated prior to were shipped to the Structural Mechanics Laboratory of Atilim
casting of topping concrete [13]. American Concrete Institute spec- University, where concrete toppings were cast and the load tests
ifies a maximum horizontal shear stress of 0.55 MPa for composite were performed. The design concrete compressive strength for
concrete flexural members with no reinforcement crossing the the hollow-core units was specified as 30 MPa but no information
interface, provided that the contact surface is intentionally rough- was available regarding the measured compressive strength of
ened [1]. A horizontal shear strength of 1.65 MPa is specified in concrete used to cast the units and the level of available stress in
AASHTO LRFD Bridge Design Specifications for concrete placed the strands. The measured compressive strength values for topping
against a surface intentionally roughened to an amplitude of concrete used in composite specimens are given in Table 1.
6.3 mm, while the horizontal shear strength is reduced to The specimens were tested under monotonically increasing dis-
0.52 MPa if the surface is not intentionally roughened [2]. It should placement loading until failure occurred. Loading was applied with
be noted that AASHTO LRFD Bridge Design Specifications uses a an electrically-controlled hydraulic cylinder at the center of a
modified version of the shear friction concept to calculate the hor- 460 cm long clear span as a line load, as shown in Fig. 1. Roller sup-
izontal shear strength and requires a certain amount of reinforce- ports were provided at the two ends of specimens. The load was
ment to be provided at the interface. The horizontal shear strength distributed through the entire width of specimens with the help
at interface between concrete surfaces has been studied exten- of a steel spreader beam attached at the end of the hydraulic cylin-
sively [14–22]. These studies mainly focused on the effect of con- der. In order to achieve a uniform transfer of the load from the
crete strength and the type of surface treatment on the interface spreader beam to the specimens, a 15 cm wide strip of neoprene
shear strength. Because these studies were mostly conducted pad was placed between the spreader beam and the top surface
toward understanding of the shear–friction theory, the specimens of specimens.
usually included dowel reinforcement crossing the interface. The The specimens were instrumented with displacement transduc-
experimental findings obtained from these studies were used to ers (i.e. linear potentiometers) and data from these transducers
evaluate the accuracy of horizontal shear strength equations avail- were continuously collected by a data acquisition system and the
able in several design codes, as well as to propose new design measurements were monitored during the load tests. Location of
equations. the transducers on specimens is indicated in Fig. 1. Two displace-
The current study aims at contributing to the understanding of ment transducers were used to measure the vertical deflection of
the behavior of composite hollow-core units under flexural load- specimens at the midspan section. In composite specimens, the rel-
ing. The main effort focused on identifying the extent to which a ative slip between the hollow-core unit and the topping was mea-
CIP concrete topping placed over the as-cast surface of precast sured with two displacement transducers at each end. Additional
concrete hollow-core units improves the flexural behavior. The slip of prestressing strands during load tests was also monitored
experimental study included load testing of five precast concrete by displacement transducers. For this purpose, four strands were
hollow-core units. The numerically determined flexural response instrumented at each end of hollow-core units.
of test specimens was later compared with the experimentally In an attempt to enhance the horizontal shear force capacity at
obtained behavior. the interface between the hollow-core unit and topping, and to
obtain a fully composite behavior, the restraint mechanism illus-
trated in Fig. 2 was used at ends of Specimen W-T-R. It should be
2. Testing program noted that the use of such restraint mechanism does not represent
a real-life application. It was only used as a method to improve the
Five precast concrete hollow-core units were tested under flex- horizontal shear force transfer between hollow-core unit and top-
ural loading. The naming convention used for the specimens is ping slab so that the case of fully composite behavior could be
given in Table 1. The specimens were divided into W and N series, studied.
 E. Baran / Engineering Structures 98 (2015) 109–117 111

Hydraulic cylinder

Load spreader
beam
Loadcell

Bare hollow-core unit

specimen

Hydraulic cylinder

Load spreader
beam

Composite
specimen

Fig. 1. Specimen cross section, instrumentation, and loading setup details.

Cast-in-place
moment and the vertical deflection measured at the midspan
Restraint concrete topping
section up to initiation of flexural cracking. Crack initiation was
mechanism
followed by a significant reduction in the stiffness of specimens.
The experimentally determined service stiffness together with
the cracking and ultimate moment capacity of specimens are pre-
sented in Table 2. The stiffness values given in the table are called
the initial stiffness as they were calculated by considering the
slope of the linear portion of the moment versus midspan deflec-
tion plots prior to initiation of cracking. For both the W and N ser-
Precast concrete ies specimens, existence of a topping slab increased the cracking
hollow-core unit
moment and initial stiffness compared to the companion bare hol-
Fig. 2. Restraint mechanism used at the ends of Specimen W-T-R. low-core unit specimens. As indicated in Table 2, the increase in
cracking moment was 22% and 33%, respectively for the W and N
series specimens. Improvement in the initial stiffness of specimens
3. Flexural response of hollow-core specimens with the addition of topping slab is more pronounced than the
cracking moment. For the W and N series specimens the increase
Initiation of flexural cracking on hollow-core units was in stiffness was 93% and 118%, respectively. These observations
observed to occur at midspan section directly below the applied are indication of major composite action between the hollow-core
load. When the cracks were visually detected for the first time, unit and the topping slab under load levels corresponding to
they usually extended from bottom into almost half of the depth uncracked state of the cross section.
of hollow-core unit. As the loading progressed, new cracks were After the initiation of cracking, the difference between the
formed on either side of the loading point. Crushing of concrete moment resisted by the composite and bare hollow-core unit spec-
at the top of hollow-core unit near the midspan section was also imens continued to increase until a relative slip occurred at the
evident. Moment versus midspan deflection response of the W interface between the hollow-core unit and the topping slab in
and N series specimens is shown in Fig. 3(a) and (b), respectively. composite Specimens W-T and N-T. Occurrence of this interface
The plots clearly show the linear relation between the applied slip is evident by a sudden drop in the moment–deflection curves
112 E. Baran / Engineering Structures 98 (2015) 109–117

100 50
(a) (b)
80 40
Moment, kN-m

60 30

40 20
N-B
W-B
N-T
W-T
20 W-T-R
10

0 0
0 20 40 60 80 100 120 140 160 0 20 40 60 80 100 120 140 160
Midspan deflection, mm Midspan deflection, mm

Fig. 3. Moment–deflection response of specimens: (a) W series; (b) N series.

Table 2
Summary of test results.

Specimen Cracking moment Normalized cracking Ultimate moment capacity Normalized moment Initial stiffness Normalized initial
(kN m) momenta (kN m) capacitya (kN m/mm) stiffnessa
W-B 38.6 1.00 68.1 1.00 5.8 1.00
W-T 47.0 1.22 83.5 1.23 11.2 1.93
W-T-R 68.8 1.78 96.1 1.41 11.1 1.91
N-B 20.7 1.00 35.9 1.00 2.8 1.00
N-T 27.6 1.33 40.6 1.13 6.1 2.18
a
Normalized with respect to specimen without topping.

shown in Fig. 3(a) and (b). Following the slip of topping concrete, load testing of composite reinforced concrete slab specimens in
composite specimens started to behave as if there was no compos- another study [23].
ite action between the hollow-core unit and the topping, and the Variation of horizontal slip between the hollow-core unit and
moment–deflection response of these specimens closely followed the topping slab in Specimens W-T and N-T with the applied mid-
the moment–deflection curve of the companion specimens with- span displacement is shown in Fig. 5. The plots indicate that during
out topping. Existence of topping slab increased the ultimate the initial phase of loading, where there was a composite action
moment capacity of wide specimen by 23%, while the increase between the hollow-core unit and the topping slab, no relative slip
in moment capacity remained at 13% for narrow specimen. was measured at the interface. A sudden slip of 1.8 mm and
Photographs showing the extent of concrete damage on some 1.0 mm, respectively in Specimens W-T and N-T resulted in vanish-
specimens during and after the load tests are shown in Fig. 4. As ing of the composite action. After this point the interface slip in
evident in one of the photographs, a separation occurred between both specimens was observed to gradually increase with the
the hollow-core unit and the topping slab following the failure of applied loading. At the end of load testing, the total relative slip
composite specimens. Similar separation was also observed during between the hollow-core unit and the topping slab was measured

Specimen W-B Specimen W-B

Specimen N-T

Fig. 4. Extent of concrete damage on some specimens during and after load testing.
 E. Baran / Engineering Structures 98 (2015) 109–117 113

100 16 50 16
W-T (a) N-T (b)
14 14
80 40
12 12

Interface slip, mm

Interface slip, mm
Moment, kN-m

Moment, kN-m.
10 10
60 30

8 8
40 20
6 6

4 4
20 10
Moment Moment
2 2
Interface slip Interface slip
0 0 0 0
0 20 40 60 80 100 120 140 160 0 20 40 60 80 100 120
Midspan deflection, mm Midspan deflection, mm

Fig. 5. Variation of interface slip: (a) Specimen W-T; (b) Specimen N-T.

to be 13.8 mm and 7.1 mm, respectively in Specimens W-T and are shown in separate plots. Superimposed on the same plots is the
N-T. It should be mentioned that the interface slip values shown overall moment–deflection response of the specimens.
in the plots in Fig. 5(a) and (b) are measured only on one shear Loss of the composite behavior in Specimen N-T, which was evi-
span of these specimens. No relative slip was detected between dent by a sudden reduction in the load resisted by the specimen
the topping slab and the hollow-core unit at the other shear span and a sudden relative slip between the hollow-core unit and the
of both specimens. topping slab, also corresponded to a sudden slip of some of the pre-
As mentioned earlier, a restraint mechanism was used at both stressing strands inside the hollow-core unit, as seen in
ends of Specimen W-T-R, in order to improve the horizontal shear Fig. 6(c) and (d). It is also worth to note that the prestressing
force capacity at the interface between the hollow-core unit and strands that had slip at one end did not show any appreciable slip
topping slab. The relation between the overall response of this at the other end. The strand slip behavior in the noncomposite
specimen and the other W-series composite specimen (i.e., specimen (Specimen N-B) was slightly difference. In this case,
Specimen W-T) is shown in Fig. 3(a). As opposed to Specimen W- one of the strands started to slip at the south end starting from a
T, Specimen W-T-R did not experience a sudden drop in the load midspan deflection of approximately 80 mm and continued to slip
capacity. This was due to the fact that the restraint mechanisms with a relatively fast rate until the end of load testing. The same
provided at the ends of the specimen controlled the relative slip strand also showed a sudden slip at the north end near the end
at the interface and prevented the occurrence of a sudden initiation of load testing. This strand had an initial slip of 9.7 mm and
of slip similar to the ones shown in Fig. 5(a) and (b). As a result, a 5.1 mm, respectively at the south and north ends. Such large values
composite behavior was valid throughout the load testing for of initial strand slip might be an indication of poor bond between
Specimen W-T-R. this strand and the surrounding concrete.
As evident in Fig. 3(a) Specimen W-B was unloaded after reach-
ing the load level corresponding to 90% of the ultimate moment
capacity. This level of load was well beyond the initiation of flexu- 4. Comparison between measured and predicted response
ral cracking on the hollow-core unit. Upon removal of the applied
load on the specimen, 83% of the deflection at the midspan section Sectional analyses based on strain compatibility between pre-
was recovered. stressing strands and the surrounding concrete were conducted
Another measure that can be used to make comparison among using Response-2000, a sectional analysis program by Bentz and
deformation characteristics of specimens is the maximum deflec- Collins [25]. These analyses incorporated realistic nonlinear
tion. As evident in Fig. 3(a) and (b), there is no clear trend between stress–strain models for concrete and prestressing strands.
the maximum midspan deflections attained during load tests and Concrete compressive and tensile strength values were assumed
the existence of a topping slab or width of the specimen. to be 30 MPa and 1.75 MPa, respectively and the effective strand
prestress after all losses was assumed to be 60% of strand ultimate
strength, fpu. Nonlinear moment–curvature relation of slab cross
3.1. Additional slip in prestressing strands during load testing sections obtained this way was then converted into load versus
midspan deflection curves presented in Fig. 7(a)–(d). The load–de-
Slip of prestressing strands at the ends of hollow-core units was flection behavior obtained through nonlinear strain compatibility
apparent prior to load tests. This strand slip that is available with analyses is observed to agree remarkably well with the measured
no service load acting on the member is very common to saw- response for the non-composite specimens (Specimens W-B and
cut precast prestressed concrete members, and it is termed as ‘‘ini- N-B). For specimens tested with concrete topping, on the other
tial slip’’. Previous studies suggest that excessive initial strand slip hand, the measured and numerically determined load–deflection
is a sign of poor concrete quality and insufficient compaction, and responses differ significantly, mainly because the sectional analysis
may result in reduced flexural and shear capacity of hollow-core assumes a perfectly composite behavior between the hollow-core
units [7,24]. The initial strand slip at both ends of hollow-core unit and the topping slab.
units used in the test specimens was measured. Additional slip of Also indicated on the plots in Fig. 7 is the moment capacity pre-
the strands during load testing of the hollow-core specimens was dicted based on the ACI-318 procedure [1]. The procedure utilizes
monitored and the data for Specimens N-B and N-T are presented an equivalent rectangular compressive stress block for concrete
in Fig. 6. For each of these specimens, the additional strand slip val- and an expression to calculate the strand stress at ultimate capac-
ues measured at the south and north ends of the hollow-core units ity. Moment capacity calculated with this procedure provides an
114 E. Baran / Engineering Structures 98 (2015) 109–117

#1 #2 #3 #4
Strand numbering scheme
50 4,0 50 0,5
Specimen N-B Specimen N-B
(a) (b)

Strand slip at south end, mm

Strand slip at north end, mm

40 40 0,4
Moment 3,0 Moment

Moment, kN-m .
Moment, kN-m

30 30 0,3
Initial slip
Strand
2,0 number
at North
end(mm)
20 20 1 5.1 0,2
Strand #1 2 3.5
Initial slip
Strand
at South 3 2.3
number
end(mm) 4 1.5
1 9.7 1,0 Strand #1
10 2 1.8 10 0,1
3 2.1
Strands #2-4 Strands #2-4
4 1.8

0 0,0 0 0
0 30 60 90 120 150 0 30 60 90 120 150
Midspan deflection, mm Midspan deflection, mm

50 0.5 50 0.5
Specimen N-T Specimen N-T
(c) (d)

Strand slip at north end, mm

Strand slip at south end, mm

40 Strand #4 0.4 40 0.4

Moment, kN-m.

Moment, kN-m .

Moment Moment
30 0.3 30 0.3
Initial slip
Strand
at North
number
end(mm)
1 1.7
20 0.2 20 2 1.2 0.2
Initial slip 3 2.1
Strand #1
Strand
number
at South Strand #3 4 1.3
end(mm)
1 1.5
10 2 1.7 0.1 10 Strand #2 Strands #3&4 0.1
Strands #1&2
3 2.2
4 3.4

0 0 0 0
0 30 60 90 120 150 0 30 60 90 120 150
Midspan deflection, mm Midspan deflection, mm

Fig. 6. Additional slip in prestressing strands: (a) and (b) Specimen N-B; (c) and (d) Specimen N-T.

accurate prediction of the capacity of hollow-core slab units tested 4.1. Analysis of horizontal shear stress at interface
without concrete topping. For the composite specimens tested in
this study, on the other hand, the ACI-318 procedure with full com- Because the amount of relative slip between topping slab and
posite behavior assumption leads to a gross over-prediction of hollow-core unit has a vital importance on the behavior of a com-
moment capacity. posite system, additional effort was devoted to investigate the
Bilinear load–deflection behavior utilizing the two points cor- interface shear strength available in composite specimens. For this
responding to the initiation of flexural cracking and the ultimate purpose, pushoff tests were conducted on 40  40 cm pieces of
capacity for each specimen is also plotted in Fig. 7. The deflec- topping slab on the south shear span of Specimen W-T following
tion at the point of crack initiation was calculated assuming an the flexural load testing. No appreciable relative slip was measured
uncracked section, while cracked section properties were used between the topping slab and the hollow-core unit on the south
in calculating the deflection corresponding to the ultimate shear span during the flexural load testing of this specimen.
capacity. Cross-sectional properties corresponding to cracked Therefore, results of these pushoff tests are believed to be repre-
and uncracked states and used to calculate the sectional sentative of the actual bond strength between the CIP topping slab
response are given in Table 3. Cracked section properties were and the hollow-core units used in this study. Three pushoff tests
determined following the basic mechanics of materials were conducted and the load at which relative slip initiated
approach. The bilinear load–deflection model determined this between the topping slab piece and the hollow-core unit was
way accurately represents the experimentally determined divided by the interface surface area to determine the interface
load–deflection response of hollow-core units tested with and shear strength. Based on these three pushoff tests, the average
without concrete topping. As evident, the bilinear model is able interface shear strength was determined to be 0.19 MPa.
to represent the stiffness for both uncracked and cracked states Horizontal shear stress at the interface between topping slab
realistically. It is worth to mention that more complex simula- and hollow-core unit in the composite specimens can be calculated
tion methods have also been used in the past by other research- using Eq. (1), which is a fundamental equation derived considering
ers for analysis of precast concrete hollow-core slabs under equilibrium of horizontal forces resulting from bending of the
different loading scenarios [9,26]. member. Even though the derivation of this expression is based
 E. Baran / Engineering Structures 98 (2015) 109–117 115

70 100
(a) (b)
60
80
50
Load, kN

Load, kN
60
40

30 Specimen W-B measured behavior 40

Specimen W-T measured behavior
20 Strain compatibility analysis
Strain compatibility analysis
ACI-318 predicted capacity 20 ACI-318 predicted capacity
10
Predicted bilinear behavior Predicted bilinear behavior
0 0
0 20 40 60 80 100 120 140 160 0 20 40 60 80 100 120 140 160
Midspan deflection, mm Midspan deflection, mm

35 50
(c)
30 (d)
40
25
Load, kN

Load, kN
20 30

15 Specimen N-B measured behavior Specimen N-T measured behavior

20
Strain compatibility analysis Strain compatibility analysis
10
ACI-318 predicted capacity 10 ACI-318 predicted capacity
5
Predicted bilinear behavior Predicted bilinear behavior
0 0
0 20 40 60 80 100 120 140 160 0 20 40 60 80 100 120 140 160
Midspan deflection, mm Midspan deflection, mm

Fig. 7. Relation between measured and predicted behavior: (a) Specimen W-B; (b) Specimen W-T; (c) Specimen N-B; (d) Specimen N-T.

Table 3 where d is the distance from extreme compression fiber to centroid

Cross-sectional properties used to calculate specimen response.
of the prestressing strands in composite cross section.
Specimen Uncracked section Cracked section The horizontal shear stress at the interface between topping
ytra (cm) Atr (cm2) Itr (cm4) ycr (cm) Icr (cm4) Qcr (cm3) slab and hollow-core unit for the composite specimens
(Specimens W-T and N-T) were calculated using Eqs. (1) and (2)
W-B 7.4 1223 30,140 2.0 2337 n/a
W-T 9.2 1798 69,150 2.5 5347 359 considering cracked transformed cross section properties given in
Table 3. The calculated interface shear stress values are shown in
N-B 7.4 578 13,980 2.1 1158 n/a
N-T 9.3 843 32,360 2.6 2656 179 Table 4. These interfacial shear stress values correspond to the
a
occurrence of a sudden relative slip between topping slab and hol-
Distance from centroid to top of cross section.
low-core unit. Therefore, these values represent the horizontal
shear stress that caused slip of topping slip, which can also be
called horizontal shear stress capacity or horizontal shear strength.
on a linear elastic behavior, it can also be used to evaluate horizon- Accordingly, these computed values agree acceptably well with the
tal shear stress in cracked sections as long as Q and I are computed interface shear strength value of 0.19 MPa determined from the
based on the cracked section [18]. push-off tests.
V Q The computed interface shear strength values shown in Table 4
vh ¼ ð1Þ can be compared against the horizontal shear strength values spec-
I  bv
ified in the design codes in an attempt to demonstrate how accu-
In Eq. (1), vh is the horizontal shear stress at the interface, V is rately they represent the members involving concrete topping
the vertical shear force at each shear span of the hollow-core unit, slab cast against the machine finished surface of a hollow-core unit
Q is the first moment of the cross-sectional area of topping slab with no intentional roughening. American Concrete Institute spec-
with respect to the neutral axis of the composite section, I is the ifies maximum horizontal shear strength of 0.55 MPa for compos-
moment of inertia of the composite section with respect to the ite concrete flexural members with no reinforcement crossing the
neutral axis, and bv is the width of the interface. interface, provided that the contact surface is intentionally
Recent versions of ACI-318 implicitly specify Eq. (2) for calcula-
tion of horizontal shear stress. Even though the expression in Eq.
(2) appears to be significantly different from the one in Eq. (1), Table 4
its derivation is also based on the consideration of horizontal forces Interface shear stress in composite specimens.
develop in the topping slab as a result of flexural effects and Specimen Interface shear stress from Eq. Interface shear stress from Eq.
equilibrium of these forces. (1) (MPa) (2) (MPa)

V W-T 0.21 0.20

vh ¼ ð2Þ N-T 0.23 0.22
d  bv
116 E. Baran / Engineering Structures 98 (2015) 109–117

80 40
(a) (b)
60 30
Load, kN

Load, kN
40 20

20 Specimen W-T measured behavior 10 Specimen N-T measured behavior

Predicted behavior Predicted behavior

0 0
0 20 40 60 80 100 120 140 160 0 20 40 60 80 100 120
Midspan deflection, mm Midspan deflection, mm

Fig. 8. Predicted load–deflection behavior of composite specimens considering interface slip: (a) Specimen W-T; (b) Specimen N-T.

roughened [1]. A horizontal shear strength of 1.65 MPa is specified 5. Conclusions

in AASHTO LRFD Bridge Design Specifications for concrete placed
against a surface intentionally roughened to an amplitude of Flexural load tests were performed on precast prestressed con-
6.3 mm, while the horizontal shear strength is reduced to crete hollow-core panels in order to demonstrate the extent to
0.52 MPa if the surface is not intentionally roughened [2]. The which concrete topping placed over the as-cast surface of precast
interface shear strength values both determined through pushoff concrete hollow-core units improves the flexural response.
load tests and calculated using the measured load capacity of the Following conclusions can be drawn based on the results:
composite specimens remained significantly lower than the hori-
zontal shear strength values specified by the ACI and AASHTO  A 5 cm thick topping slab placed over 15 cm deep hollow-core
Specifications. The reason for such low interfacial shear strength units were observed to increase the cracking moment and ini-
values can be attributed to relatively smooth surface of hollow- tial stiffness of 120 cm wide bare hollow-core units by 22%
core units used in the testing program as a result of the machine and 93%, respectively. This improvement in the behavior is an
finishing operation. It should be mentioned here again that the indication of major composite action between the hollow-core
interface surface in the test specimens was not intentionally unit and the topping slab under load levels corresponding to
roughened. uncracked state of the cross section.
 Initiation of significant relative slip between hollow-core unit
and topping slab resulted in a sudden reduction in moment
4.2. Predicted load–deflection response considering incomplete resisting ability of composite specimens. Following the slip of
composite action topping slab, composite specimens started to behave as if there
was no composite action between the hollow-core unit and top-
As mentioned previously, following the loss of composite ping slab, and the moment–deflection response of these speci-
behavior as a result of a sudden slip at the interface between top- mens closely followed the moment–deflection curve of the
ping slab and hollow-core unit, the load–deflection behavior of companion specimens without topping slab. Existence of
composite specimens closely followed that of bare hollow-core topping slab resulted in a 23% increase in ultimate moment
specimens. Based on this observed behavior, the predicted bilinear capacity of 120 cm wide bare hollow-core unit.
load–deflection representation for the composite and bare hollow-  Effect of topping slab in terms of increase in cracking moment,
core units were combined to produce a complete load–deflection initial stiffness, and ultimate moment capacity is observed to be
response that covers the behavior of test specimens both before smaller on specimens tested with 55 cm width, as opposed to
and after the loss of composite behavior. The criteria that defines 120 cm.
the transition from composite behavior to non-composite behavior  At the end of load testing, total relative slip of 13.8 mm and
was taken as the horizontal shear stress at the interface between 7.1 mm were measured between the hollow-core unit the top-
hollow-core unit and topping slab. The limiting value of the inter- ping slab on one shear span of composite specimens. No relative
face shear stress was taken as 0.20 MPa. Considering that the aver- slip was detected between the topping slab and the hollow-core
age interface shear strength obtained from the pushoff tests was unit at the other shear span of these specimens.
0.19 MPa and that the interface shear strength calculated for the  Slip of prestressing strands at the ends of hollow-core units
two composite specimens using Eqs. (1) and (2) were between were apparent prior to load tests. In addition to this initial
0.20 and 0.23 MPa, using an interface shear stress value of 0.20 strand slip, additional slip of strands was observed during load
as the transition between the composite and non-composite testing. Loss of the composite action, which was evident by a
behavior was deemed reasonable. sudden reduction in the load resisted by the specimen and a
As shown in Fig. 8 for both the wide and narrow specimens, the sudden relative slip between the hollow-core unit and the top-
initially uncracked and cracked parts of the response is dictated by ping slab, also corresponded to a sudden slip of some of the pre-
the behavior predicted assuming a complete composite action. This stressing strands inside the hollow-core units. It was observed
response is valid until the initiation of significant interface slip, that the prestressing strands that had slip at one end did not
which was assumed to occur when the horizontal shear stress at show any appreciable slip at the other end.
the interface between hollow-core unit and topping slab reached  Load–deflection response obtained from sectional analysis
a value of 0.20 MPa. At this point, the load drops to the level corre- based on strain compatibility between prestressing strands
sponding to the response of the bare hollow-core unit and the and the surrounding concrete, and utilizing realistic nonlinear
behavior starts to be dictated by that of the bare hollow-core unit. stress–strain models for concrete and prestressing strands is
 E. Baran / Engineering Structures 98 (2015) 109–117 117

observed to agree remarkably well with the measured response [2] AASHTO. LRFD bridge design specifications. 5th ed. Washington
(DC): American Association of State Highway and Transportation Officials;
for the non-composite specimens. For specimens tested with
2010.
concrete topping, on the other hand, the measured and numer- [3] Dowell RK, Smith JW. Structural tests of precast, prestressed concrete deck
ically determined load–deflection responses differ significantly, panels for California freeway bridges. PCI J 2006;51(2):76–87.
mainly because the sectional analysis assumes a perfectly com- [4] Scott NL. Performance of precast prestressed hollow core slab with composite
concrete topping. PCI J 1973;18(2):64–77.
posite behavior between the hollow-core unit and the topping [5] Girhammar UA, Pajari M. Tests and analysis on shear strength of composite
slab. slabs of hollow core units and concrete topping. Constr Build Mater
 Pushoff tests conducted on 40  40 cm pieces of topping slab in 2008;22:1708–22.
[6] Ueda T, Stitmannaithum B. Shear strength of precast prestressed hollow slabs
one of the specimens indicated the average interface shear with concrete topping. ACI Struct J 1991;88(4):402–10.
strength to be 0.19 MPa. The interface shear strength values [7] Palmer KD, Schultz AE. Experimental investigation of the web-shear strength
were also calculated for the two composite specimens consider- of deep hollow-core units. PCI J 2011;56(4):83–104.
[8] Palmer KD, Schultz AE. Factors affecting web-shear capacity of deep hollow-
ing the load level corresponding to initiation of significant rela- core units. PCI J 2011;55(2):123–46.
tive slip and using (1) the basic mechanics of materials [9] Brunesi E, Bolognini D, Nascimbene R. Evaluation of the shear capacity of
approach and (2) the simplified code expression. These com- precast–prestressed hollow core slabs: numerical and experimental
comparisons. Mater Struct. DOI:http://dx.doi.org/10.1617/s11527-014-0250-
puted interface shear strength values were between 0.20 and 6.
0.23 MPa. The reason for the measured and computed interface [10] Hawkins NM, Ghosh SK. Shear strength of hollow-core slabs. PCI J
shear strength values being significantly lower than the hori- 2006;51(1):110–4.
[11] Yang L. Design of prestressed hollo core slabs with reference to web shear
zontal shear strength values specified by the ACI and AASHTO
failure. J Struct Eng 1994;120(9):2675–96.
Specifications is attributed to relatively smooth surface of hol- [12] Hegger J, Roggendorf T, Kerkeni N. Shear capacity of prestressed hollow core
low-core units used in the testing program as a result of the slabs in slim floor constructions. Eng Struct 2009;31:551–9.
[13] PCI. PCI Manual for the design of hollow core slabs. 2nd ed. Chicago
machine finishing operation.
(IL): Precast/Prestressed Concrete Institute; 1998.
 Predicted bilinear load–deflection representation for the com- [14] Mones RM, Brena SF. Hollow-core slabs with cast-in-place concrete toppings:
posite and bare hollow-core units can be combined to produce a study of interfacial shear strength. PCI J 2013;58(3):124–41.
a complete load–deflection response that covers the behavior of [15] Djazmati B, Pincheira JA. Shear stiffness and strength of horizontal
construction joints. ACI Struct J 2004;101(4):484–93.
test specimens both before and after the loss of composite [16] Kahn LF, Slapkus A. Interface shear in high strength composite T-beams. PCI J
behavior. For cases including cast-in-place concrete topping 2004;49(4):102–10.
placed against the machine finished surface of hollow-core [17] Julio ENBS, Branco FAB, Silva VD. Concrete-to-concrete bond strength,
influence of the roughness of the substrate surface. Constr Build Mater
units with no intentional roughening, an interface shear stress 2004;18:675–81.
of 0.20 MPa can be used to define the transition from composite [18] Loov RE, Patnaik AK. Horizontal shear strength of composite concrete beams
to non-composite behavior. The predicted load–deflection with a rough interface. PCI J 1994;39(1):48–69.
[19] Bass RA, Carrasquillo RL, Jirsa JO. Shear transfer across new and existing
response produced this way represents the actual response of concrete interfaces. ACI Struct J 1989;86(4):383–93.
composite specimens with an acceptable accuracy. [20] Hofbeck JA, Ibrahim IO, Mattock AE. Shear transfer in reinforced concrete. ACI J
1969;66(1):119–28.
[21] Julio ENBS, Costa DD, Branco FAB, Alfaiate JMV. Accuracy of design code
expressions for estimating longitudinal shear strength of strengthening
concrete overlays. Eng Struct 2010;32:2387–93.
Acknowledgments [22] Santos PMD, Julio ENBS. A state-of-the-art review on shear-friction. Eng Struct
2012;45:435–48.
[23] Mosallam A, Taha MMR, Kim JJ, Nasr A. Strength and ductility of RC slabs
Technical assistance provided by Murat Guney and laboratory
strengthened with hybrid high-performance composite retrofit system. Eng
assistance provided by a former graduate student Sadi Ibrahim Struct 2012;36:70–80.
Haruna are acknowledged. [24] Brooks MD, Gerstle KH, Logan DR. Effect of initial strand slip on the strength of
hollow-core slabs. PCI J 1988;33(1):90–111.
[25] Bentz E, Collins M. User manual for Response-2000 reinforced concrete
sectional analysis using the modified compression field theory. Toronto,
References Ontario, Canada): University of Toronto; 2001.
[26] Hegger J, Roggendorf T, Teworte F. FE analysis of shear-loaded hollow-core
[1] ACI. Building code requirements for structural concrete: ACI318-05. Detroit slabs on different supports. Mag Concr Res 2010;62(8):531–41.
(MI): American Concrete Institute; 2005.