Journal of Constructional Steel Research 168 (2020) 105868

Contents lists available at ScienceDirect

Journal of Constructional Steel Research

Analytical study and experimental tests on innovative steel-concrete

composite floorings
C.A. de O. Ferrante a, S.A.L. de Andrade a, L.R.O. de Lima b,⁎, P.C.G. da S. Vellasco b
a
Pontifical Catholic University of Rio de Janeiro, PUC-Rio, Brazil
b
Structural Engineering Department, UERJ – State University of Rio de Janeiro, Brazil

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

Article history: Composite floor systems, when compared to reinforced concrete slabs, are a more cost-effective structural solu-
Received 28 June 2019 tion. The overall behaviour of these composite members depends on the shear connection between steel and the
Received in revised form 8 October 2019 concrete encasement. The pre-cast composite flooring system with hollow-core section presents an efficient and
Accepted 11 November 2019
useful solution that helps to reduce the flooring height. This paper reports the results of composite connection
Available online 29 November 2019
push-out and the associated composite beams full-scale tests. The system was composed of a partially encased
Keywords:
asymmetric steel beam, and the shear transfer mechanism was established through an innovative shear connec-
Composite structures tion by chemical bonding, friction, and dowel action (by transversal reinforcing bars passing through the web's
Composite beams holes). The tests results indicated that a rigid connection was produced by the chemical bonding and friction
Castellated steel beams of the embedded flange profile in the concrete slab, as well as the composite beam's ability to develop its plastic
Experimental analysis bending moment capacity. The proposed analytical model provides an efficient way for analysing and designing a
composite beam with encased compression flanges. Based on the results from the ultimate flexural resistance an-
alytically obtained, a sensitivity study was also developed to evaluate the most competitive cases in terms of
structural efficiency.
© 2019 Elsevier Ltd. All rights reserved.

1. Introduction solutions use profiled steel sheeting or shallow cellular beams in com-
posite construction [1,2,9].
Each type of construction requires different building solutions. The Mensinger et al. [10] studied a composite flooring system based
requirements of a construction project vary according to the building on cellular beams partially encased in the concrete slab, where the
purpose, location, environmental conditions, and other demands. How- connection between steel and concrete was obtained through the
ever, there are more subtle matters that must be considered. Larger cit- adhesion between rebars welded along with the web of the section
ies usually have building height limitations, which are influenced by and the concrete. Peltonen and Leskela [11] performed 74 push-out
regulations, economic factors, and aesthetics. The ability to meet these tests using “Deltabeams” to investigate the shear-slip properties of
requirements is an essential factor when selecting a structural system. the concrete dowel connection. Ju et al. [5] published an experimen-
Composite floor systems, when compared to reinforced concrete tal study of composite beams using asymmetrical sections with web
slabs, are a more cost-effective structural solution. As well known in openings (iTECH composite beam). In their analysis, the sections
the literature, the composite members are structural systems that can were coated with concrete on their side and top area, and the longi-
optimise the mechanical properties and material characteristics of tudinal shear resistance was obtained through the adhesion in the
steel and concrete. Additionally, concerning the use of wide flange steel-concrete interface, as well as through the mechanical bracing
shapes and shear studs, various types of steel-concrete composite sys- of the concrete in the openings. Braun et al. [4] developed a compos-
tems have been proposed in the last years. One of the main objectives ite slim-floor beam (CoSFB) characterised by a concrete dowel
of the using composite flooring systems is about the reduction of the placed between the flanges of a hot-rolled section with reinforcing
floor height. Some examples may be pointed out: “slimflor beams” de- bars passing through the openings and a concrete infill. Huo and
veloped by SCI [1], “Thor-beam” and “Delta-beam” conceived in Scandi- D'Mello [7] carried out 24 push-out tests to investigate the
navia [2], Composite Slim Floor Beam (CoSFB) [3,4], iTECH composite concrete-infill-only, tie-bar, ducting and web-welded stud shear
beam [5] and Ultra-Shallow Floor Beams (USFB) [6–8]. Some of these connection in ultra-shallow floor beams (USFB).
As could be observed in the referenced works, the overall behaviour
⁎ Corresponding author. of these composite members depends on the shear connection between
E-mail addresses: lucianolimauerj@gmail.com, luciano@eng.uerj.br (L.R.O. de Lima). steel and the concrete encasement. Many studies have been conducted

https://doi.org/10.1016/j.jcsr.2019.105868
0143-974X/© 2019 Elsevier Ltd. All rights reserved.
2 C.A. de O. Ferrante et al. / Journal of Constructional Steel Research 168 (2020) 105868

Nomenclature

a concrete compressive block height (mm)

Asb cross-sectional area of bottom tee (mm2)
bc concrete slab effective width (mm)
bf bottom flange width (mm)
bs top flange width (mm)
c steel consumption (kg/m2)
Cr factored compressive resistance of the steel area and the adja-
cent concrete in compression (kN)
C'r factored compressive resistance of the concrete above the top Fig. 1. Composite beams with two precast T-section.
flange (kN)
C''r factored compressive resistance of the concrete below the top
connectors [33]. Another alternative to transfer shear strength from
flange (kN)
the concrete slab to the steel section is the slim-floor steel beam's web
d steel section height (mm)
openings to enable the in-situ concrete to pass and fill the openings,
e internal moment lever arm -Cr (mm)
thus enhancing the interaction between the concrete slab and the web
e' internal moment lever arm - C'r (mm)
post of the steel beam [34].
e ' ' internal moment lever arm - C''r (mm)
However, the pre-cast composite flooring system with
fck concrete compressive strength (MPa)
partially-encased castellated beams presents an efficient, cost-
fy steel yield strength (MPa)
effective solution that helps to reduce building floor heights, as
fu steel ultimate strength (MPa)
can be observed in Fig. 1. This system was conceived as a building
Fb maximum shear force per unit length (N/mm)
alternative to the pre-cast double-T slabs of reinforced concrete
Fn total compressive force transferred at the interface due to an
considering aspects related to precast and ease of transportation
evenly distributed load (kN)
to the construction site. Their main aim was to combine both sys-
hc concrete slab total thickness (mm)
tems, pre-cast concrete and composite structures, allowing for
L composite beam span (m)
faster construction execution while providing a higher quality for
MRd ultimate flexural capacity (kN.m)
the structural elements. This paper focuses on the unconventional
N tensile force in reinforcing bars (kN)
connection between steel cross-section and the concrete slab.
P test loading (kN)
Pu connector ultimate shear resistance (kN)
PRk characteristic load level (kN)
q distributed load (kN/m2)
s longitudinal relative slip (mm)
t concrete slab thickness below the top flange in compression
(mm)
tc concrete slab thickness above the top flange (mm)
tfc top flange thickness in compression (mm)
ts top flange thickness (mm)
tw web thickness (mm)
Tr steel shape factored tensile resistance (kN)
u transverse displacements (mm)
ysb distances from the bottom flange outside edge to the steel
shape centroid (mm)
ϕ steel performance factor
ϕc concrete performance factor Fig. 2. Detail of composite joists with partially-encased castellated steel beams.
τd calculation value for the shear stress in the interface (N/mm2)
∅ transverse reinforcing bars diameter (mm)
δu slip capacity (mm)
σc compressive stresses in the interface (MPa)
τin shear stresses in the interface (MPa)
γ partial safety factor for normal combinations

List of acronyms
P.N.A. Plastic neutral axis
LVDTs Linear Variable Differential Transformers
ASTM American Society for Testing and Materials

in past years investigating various types of shear connectors, such as

stud bolt connectors [12–15], the perfobond connectors and its varia-
tions [16–24]. Moreover, other types of connectors have also been con-
sidered, such as post-installed or retrofitted shear connectors [25,26],
inverted umbrella shear connectors [27], demountable shear connec-
tors [28,29], channel connectors [30], angles [31,32] and V-shaped Fig. 3. The manufacturing process of an asymmetric castellated steel beam.
 C.A. de O. Ferrante et al. / Journal of Constructional Steel Research 168 (2020) 105868 3

Table 1
Bending moment capacity assessment of concrete composite floorings.

PNA location Factored bending moment resistance Plastic stress distribution

Concrete slab MRd = e0 ⋅ ϕ ⋅ Asb ⋅ fy

Top flange MRd = C0 r ⋅ e0 + Cr ⋅ e

Web M Rd ¼ C 0 r  e0 þ C 00 r  e00
þC rs  e

Fig. 4. Shear force transfer mechanism.

4 C.A. de O. Ferrante et al. / Journal of Constructional Steel Research 168 (2020) 105868

Fig. 5. Shear bond used in determining composite properties (uniformly loaded beam).

Instead of the usual shear connections presented before, the 2. Proposed flooring system
unusual connection herein proposed is made of three main ele-
ments: the adhesion between the top flange and the concrete The proposed flooring system is composed of two pre-cast compos-
slab, the dowel action on the transverse reinforcing bars that ite joists joined by a steel wire mesh on top and a concrete solidarization
crosses the web section, and the mechanical interlocking between cast-in-place cover slab. These elements form a double T-section, as
the slab and the section, which is brought about by the web shown in Fig. 1. A castellated steel beam, with an asymmetric I-
openings. section, is partially encased in the concrete slab and constitutes the
There are numerous similar composite floorings systems, but the pe- ribs of each T-section. As illustrated in Fig. 2, the encased part of steel
culiarity of each of them makes it unfeasible to predict the resistance beam provides extra resistance to fire, as well as stabilises the area sub-
values of the steel-concrete connection and to apply them to the system mitted to compressive stresses. The composite beam openings allow the
herein proposed. passage of large pipes and ducts in all directions, thus providing site de-
There are a few standards that address the steel-concrete adhesion signers with more maintainability and flexibility in the use of the space.
topic. The basic hypothesis assumes a perfect interaction between the For ease of transportation of the beams and depending on site free-
two materials, and, for this to be confirmed, there must be no excessive space the fabricator can choose to produce either single or double
relative slip between the materials. It is for this reason that EUROCODE 4 T- sections.
[35], for instance, establishes the maximum shear stress that can occur During the assembly process, the composite beams are located
in the steel-concrete interface – or maximum adhesive stress –, without in parallel to one another. With the composite joists in place a
the need to use shear connections. welded wire mesh, so-called solidarization mesh is laid and a

Fig. 6. Transverse steel reinforcing bars layouts.

 C.A. de O. Ferrante et al. / Journal of Constructional Steel Research 168 (2020) 105868 5

Fig. 7. Curves of load versus span for each effective width (bc).

concrete cover with minimum 40 mm thickness is then cast-in- [36]. The flexural capacity depends on the degree of shear connection
situ. Both cover slab and the joist slab work together as a struc- provided, the compressive resistance of the concrete slab, and the ten-
tural element. The asymmetric steel beams are produced in a sile yield strength of the steel structure. A full interaction was assumed
straight I-section cut in the longitudinal direction. For each T- to occur between the structural steel, the reinforcing bars, and the con-
section, a steel plate is added to the top web, creating regularly- crete. Three calculation models can be used depending on the location
spaced web openings. If necessary, steel plates can be used to of the plastic neutral axis (P.N.A These situations are summarised in
close the web openings at the ends of the beam to increase the Table 1 where the design equations for bending moment capacity are
shear resistance. Fig. 3 show the configuration of these novel sec- introduced associated with the PNA located in the concrete slab, the
tions, with cross-hatched regions representing the steel plates. steel beam top flange or the steel beam web.

3. Analytical investigations 3.2. Shear force transfer mechanism

3.1. Ultimate flexural capacity The overall behaviour of steel-concrete composite members de-
pends on shear connection between steel and concrete encasement,
For this study, plastic analysis of the sections was conducted to eval- which can be accomplished through three main shear-transferring
uate the ultimate flexural capacity of the composite beam, following the mechanisms: chemical bonding (bond between cement paste and sur-
procedure presented in the Canadian design code CAN/CSA-S16–14 face of the steel), friction (assumed to be proportional to the force at
6 C.A. de O. Ferrante et al. / Journal of Constructional Steel Research 168 (2020) 105868

Fig. 10. Steel consumption versus bending moment resistance for each span of the second-
Fig. 8. Curves of steel content versus span.
web sections.

slab uplift (displacement u2) is prevented by the shear stresses τin2 in

the interface) and mechanical interaction (due to embossments, ribs, or the web interface, and by the reinforcing bar. Compressive stresses σc2
shear connectors) [37]. occur on the bottom part of the beam's flange.
In a real structure subjected to shear forces, these mechanisms The calculation value for the shear stress in the interface (τd), caused
can interact with each other producing combined effects in terms by the variety of mechanisms mentioned above, is equivalent to
of slip. After the chemical bond rupture, as the slip increases, the 0.9 N/mm2 [37]. It is assumed that the stress acts on the area of contact
forces are redistributed to the concrete and longitudinal bars between the section and the concrete slab (see Fig. 5). The maximum
while the friction resistance decrease. If there is no reinforcement shear force per unit length is obtained with Eq. (1). The total compres-
in the steel-concrete interface, the shear connection behaviour can sive force transferred from the interface to an evenly-distributed load
become brittle [38]. According to Qureshi et al. [38] the shear con- is given by Eq. (2).
nection strength in a steel-concrete composite beam depends on
two critical factors, namely shear strength of the shear stud F b ¼ ð2  ðbs þ t s þ t 0 Þ−t w Þ  τd ð1Þ
and the resistance of the concrete or composite slab against longi-
tudinal cracking. According to Birkeland and Birkeland [39], the L
Fn ¼ Fb  ð2Þ
friction in the steel-concrete interface can be explained by the 4
well-known interlayer shear slip theory (ISST).
Due to bond transfer, the longitudinal relative slip, s, may simulta- The transverse slab reinforcing bars are needed in steel-concrete
neously occur with transverse displacements u1 e u2, taking place in composite construction to transfer the longitudinal shear in the steel-
the direction perpendicular to the steel-concrete interface (see Fig. 4). concrete interface avoiding the concrete cracking [40]. Zona et al. [40]
The displacement u1 is partially restricted by slab confinement and have mentioned that three design parameters are the most influential
the action of the reinforcing bar. The internal effects associated with on the slip demand in the shear connection: construction sequence
the restrictions of this displacement are the shear stresses in the inter- (propped construction results in slip demand always larger than
face τin1, and tensile stresses in the reinforcing bars, which result in unpropped construction), span length (longer spans are the most criti-
the force N. The equilibrium conditions dictate that stresses (σc1) occur- cal in terms of slip demand), and steel section shape (sections with
ring in the interface equilibrate the τin1 and the force N. Similarly, the equal flanges are less critical). According to Kullman and Hosain [41],
the steel-concrete composite beams with fishbone transverse reinforc-
ing bars, as shown in Fig. 6(a), have higher shear resistance when com-
pared to conventional ones – see Fig. 6(b), retarding the cracking
formation.

3.3. Sensitivity analysis

Based on the results from the ultimate flexural resistance ana-

lytically obtained, a sensitivity study was also developed to evalu-
ate the most competitive cases regarding structural efficiency. This
study used different cross-section dimensions, effective widths of

Table 2
Push-out specimen's main characteristics.

Specimens Transverse Shear bond perimeter Bearing area

reinforcing (mm2) (mm2)

P1 ϕ 10.0 straight bar 176,720 244

P2 ϕ 10.0 straight bar 176,720 1074
P3 ϕ 10.0 “fish-bone” bar 176,720 244
Fig. 9. Study of the relation between steel consumption and bending moment resistance
P4 ϕ 10.0 “fish-bone” bar 50,480 244
for each span of the composite beam.
 C.A. de O. Ferrante et al. / Journal of Constructional Steel Research 168 (2020) 105868 7

Fig. 11. Geometrical characteristics of push-out specimens (dimensions in mm) (a) Specimen P1 and P2, (b) Specimen P3 and P4.
8 C.A. de O. Ferrante et al. / Journal of Constructional Steel Research 168 (2020) 105868

Fig. 12. Angle used to increase the bearing surface in push-out prototype P2.

the slab, and spans. For this issue the following design assump- where: L is the composite beam span; bc is the concrete slab effec-
tions were considered: the slab thickness remained unaltered, i.e. tive width and γ is the partial safety factor for normal combina-
80 mm; the cellular beam steel was ASTM A572 grade 50 with tions equal to 1.4.
yield stress (fy = 345 MPa) and Young's Modulus (E) of 200GPa; Fig. 7 shows the distribution curves of the load versus the slab span
for the concrete slab, the concrete compressive strength was for each investigated sections. Each curve was generated by variating
30 MPa with E = 30.67GPa. For the present analysis, rolled W- the concrete slab effective width: 1000, 1200, 1500, and 2000 mm.
sections (American wide flange beams) were selected; With the The graphs helped to select the composite beam configuration accord-
results of the ultimate bending moment resistance, the distributed ing to the applied load and the required span.
load that causes the beam's collapse according to each model of To use this system in commercial construction projects, the value of
the composite beam can be calculated with Eq. (3). the distributed load (dead load plus imposed load) was set to 10 kN/m2.
This value made possible the calculation of the corresponding span by
isolating the span (L) in Eq. (3). With these values in hand, a study of
M Rd  8
q¼ ð3Þ the relation between the steel consumption (c) and the beam span
L2  bc  γ

Fig. 13. Push-out prototypes fabrication phases,a) wood casting,b) half of the prototype – first phase,c) final prototype.
 C.A. de O. Ferrante et al. / Journal of Constructional Steel Research 168 (2020) 105868 9

Table 3 steel content than the beam with a ½ W 360 × 32.9 section and
Mechanical properties of steel and concrete. 1500 mm of effective width. Thus, when selecting the most viable solu-
Specimens fy (MPa) Es (MPa) fck (MPa) Ec (MPa) tion, it is necessary to consider the steel consumption and its direct im-
P1 337.5 209,788 42 22,260
pact on the cost of the structure.
P2 337.5 209,788 42 22,260 Fig. 9 presents the relation between steel consumption and the
P3 337.5 209,788 45 35,116 beam's bending moment resistance for each span, showing how
P4 337.5 209,788 45 35,116 the steel mass relates to the resistance values. The graphs reveal
that the lower the relation c/MRd, the better is the system's perfor-
mance, indicating that the steel consumption decreases when the
was performed. In the graphs presented in Fig. 8, the steel consumption strength is equal to or greater than this value. Broadening the
varies depending on the effective width. The curves show that, for the scope of the previous analysis to encompass second-web sections,
same load level and the same span, a composite beam with a ½ W a c/M Rd x L graph was elaborated to include six additional
410 × 38.8 section and an effective width of 2000 mm had a lower W-sections. When comparing Figs. 9 and 10, it is evident that the

Fig. 14. Push-out test setup and instrumentation (a) Setup, (b) LVDT's, (c) Strain gauges (P1 - P2), (d) Strain gauges (P3–P4).
10 C.A. de O. Ferrante et al. / Journal of Constructional Steel Research 168 (2020) 105868

Fig. 15. Geometrical characteristics of composite cross-sections and steel beam (in mm) (a) Specimen CB1, (b) Specimen CB2, (c) Steel beam.

c/MRd relation remained the same for second-web sections. On the investigating a particular feature of the connection. The details of each
other hand, with the same distributed load of 10 kN/m2 , the re- prototype are presented in Table 2.
quired span was larger. Push-out test P1 illustrated in Fig. 11(a), was designed to reproduce
the shear connection setup used in the bending tests that will be

4. Experimental investigation

4.1. Push-out tests

Four push-out tests were conducted aiming to identify the longitudi-

nal shear behaviour of cellular beams partially encased in the concrete
slab. The factors that affected the bond behaviour between the section
and the concrete slab include adhesion in the steel-concrete interface,
the bearing capacity of the concrete, and the transverse reinforcing
bars. The prototypes were designed in such a way as to have the steel-
concrete connection subjected to a direct longitudinal shear force. The
test arrangements were based on Eurocode 4 [35], with each test Fig. 16. Supports and lateral bracing.
 C.A. de O. Ferrante et al. / Journal of Constructional Steel Research 168 (2020) 105868 11

Fig. 17. Test and measurement setup for the composite beams (a) Specimen CB1, (b) Specimen CB2, (c) Steel beam instrumentation, (d) Test overview.

presented in sequence. This specimen was composed of a wire mesh The prototypes were cast in two stages, being each half at a time
and a straight tie-bar with ϕ 10 mm and 244 mm2 of relative bearing using standard concrete and were cured for seven days (see Fig. 13).
surface in the encased web thickness. For specimen P2, the bearing Cylindrical specimens (100 mm × 200 mm) were cast to evaluate the
surface increased to 1074 mm2 due to the addition of a steel plate concrete compressive strength (fck) of each batch. The material proper-
in the section's web (see Fig. 12). As shown in Fig. 11(b), push-out ties of the steel beams and the concrete are shown in Table 3.
tests P3 and P4 used “fish-bone” bars with ϕ 10 mm as transverse re- The tests were conducted in the Laboratory of Structures and Ma-
inforcing bars. For specimen P4, the shear bond area was reduced to terials of the Pontifical Catholic University of Rio de Janeiro, where
50,480 mm2. the concrete slabs were simply supported on the reinforced concrete
12 C.A. de O. Ferrante et al. / Journal of Constructional Steel Research 168 (2020) 105868

Fig. 18. Bending moment versus mid-span deflection curves (a) P1 (b) P2 (c) P3 (d) P4.

slab of the laboratory. Fig. 14 shows the push-out test setup and in- the specimens. The uplift in the steel-concrete interface was also mea-
strumentation in terms of strain-gauges and displacement transduc- sured using 10 mm LVDTs.
ers as well. For prototypes P1 and P2, one strain-gauge was used per
reinforcing steel bar. On the other hand, two strain-gauges were 4.2. Full-scale testing of composite beams
considered for prototypes P3 and P4 where “fish-bone” bars were
used. A hydraulic actuator of 500 kN was used to apply the load to An experimental analysis of two full-scale composite beam speci-
the steel section. The test procedures were divided into two stages. mens (CB1 and CB2) was conducted to investigate their overall behav-
The first stage consisted of the cyclical phase with force control, iour. Each prototype was made with an asymmetric cellular section
wherein 2000 cycles of loading/unloading were applied with a fre- partially encased in the concrete slab. The slab was 1200 mm wide
quency of 0.5 Hz. The load ranged from 5% to 40% of the expected and 80 mm thick, while the encasement of the steel beam measured
failure load. 50 mm. The concrete used for the analysis was processed with an esti-
The second stage consisted of a ramp test conducted up until the mated strength at 30 MPa. The slab was reinforced on the top and bot-
specimen's failure, with the displacement of the actuator with tom with mesh wires of ϕ 4.2 mm, spaced at 150 mm intervals in both
0.01 mm/s as the control parameter. The slip between the concrete directions. Transverse reinforcing bars, with ϕ 10 mm and 600 mm in
and the steel beam was measured using 20 mm LVDTs on the top of length, were inserted through the web openings. Three tie bars equally
 C.A. de O. Ferrante et al. / Journal of Constructional Steel Research 168 (2020) 105868 13

Table 4 to a bending test, while the load was applied with two 500 kN actuators
Results of the push-out tests. distributed to four loading points as pointed out in Fig. 17.
Specimens Ultimate load Pu Characteristic load PRk Slip capacity δu The control of the applied load was performed by a 100 kN load cell
(kN) (kN) (mm) located in each actuator. The load was distributed using double steel
P1 172.2 155.0 11.81 sections positioned along the slab width at four different points. The
P2 225.1 202.6 5.92 specimen-measuring setup is detailed in Fig. 17. Five displacement
P3 224.5 202.1 0.74 transducers were used for each specimen, with the T3 displacement
P4 207.2 186.5 0.30
transducer set to measure the mid-span deflections, and the T1 and T5
transducers set to obtain the vertical displacement of the supports.
The T2 and T4 displacement transducers measured the slip in the
spaced at 150 mm intervals were used in each web opening. The com- steel-concrete interface, and strain transducers were set in the mid-
posite beam CB2 was enhanced with a U-section welded to the bottom span section (section A and C) and other sections (sections B, D, and
flange. Fig. 15 shows the longitudinal and transverse sections of the E). Strain gauges 1, 2, 3, and 4 were used at the top flange, and strain
composite beam specimens. gauges 5, 6, 7, and 8 at the bottom flange.
The slab concrete compressive strength was determined by com-
pression tests conducted in 12 specimens. The characteristic compres- 5. Tests results and discussion
sive strength mean value (fck) after 28 days was 32.80 MPa. The
mechanical properties of yield strength (fy) and ultimate strength (fu) 5.1. Push-out tests results
of the steel beams were determined through direct tensile tests. The re-
sults showed a yield strength (fy) of 337.5 MPa and a tensile strength Resistance values and ductility are two essential features when
(fu) of 462.07 MPa. Once again, the tests were conducted in the Labora- evaluating the behaviour of a shear connection. Both can be illus-
tory of Structures and Materials of the Pontifical Catholic University of trated by the relation of shear force and relative slip that is present
Rio de Janeiro. The specimens were considered simply supported, and between the steel and the concrete. The curve that represents this
lateral braced as can be observed in Fig. 16. The beams were subjected relation was obtained through experimental direct shear tests

Fig. 19. Load-slip curves of the specimens (a) P1, (b) P2, (c) P3, (d) P4.
14 C.A. de O. Ferrante et al. / Journal of Constructional Steel Research 168 (2020) 105868

Fig. 20. Failure modes of push-out specimens a) P1 - longitudinal cracking and cracking in the opening's surface, b) P2 – cracking in the opening's surface without reinforcing bar failure,
c) P3 and P4 – longitudinal cracking and transverse reinforcing bar failure.
 C.A. de O. Ferrante et al. / Journal of Constructional Steel Research 168 (2020) 105868 15

Fig. 21. Influence of bearing surface (a) Specimens P1, P3 and P4 (b) Specimens P2.

(push-out) of four models, as shown in Fig. 18. The ultimate shear resistance. When comparing the results of prototypes P1 and P3, chang-
resistance of the connector, Pu, was obtained by dividing the ulti- ing the straight reinforcing bars to the “fish-bone” increased the ulti-
mate load of the test by the number of its openings. The slip capac- mate strength in 30%.
ity, δu, corresponds to the slip measured for the characteristic load In the models with no web plate (P1, P3, and P4), the web of the sec-
level, P Rk . In other words, the slip capacity corresponds to the tion in the opening area was in direct contact with the concrete. As such,
maximum slip in which the connector can still support 90% of its the concrete area under compression was reduced, generating consider-
characteristic shear resistance. able tensile stresses that caused cracking in the opening's surface as also
Most of the longitudinal shear occurred in the steel-concrete inter- can be observed in Fig. 20. In these tests, the failure of the first reinforc-
face by means of adhesion. The adhesive failure occurred close to the ul- ing bar was also noted as pointed out in Fig. 20 (c). On the other hand, in
timate resistance, Pu, initiating the slip at the steel-concrete interface. the prototype P2, where an angle was used to increase the bearing area
The stresses were redistributed through other connection mechanisms, (Fig. 12), only cracking in the opening area was observed without any
such as transverse reinforcing bars. The distribution of adhesive stresses reinforcing bar failure.
is also associated with the cracking state of the concrete slab. The results As commented before, in prototype P2, a welded plate obtained from
of the push-out tests are summarised in Table 4. an angle was added to the web of the section at each end of the opening,
Regarding ductility, the connection presented a rigid behaviour dur- increasing the area under compression allowing the strain to be better
ing the cyclic stage of the test and throughout the loading, eventually distributed in the confined concrete. In this case, the concrete reached
reaching the maximum load. During that stage, no elements of the con- its maximum shear strength in the two cross-sections that coincided
nection suffered deformation. The behaviour of the connection diverged with the sides of the plate, as shown in Fig. 21. As mentioned before,
in the post-failure stage of each test. Push-out test P1 presented ductility
after reaching the maximum load, with a slip capacity of 11.81 mm,
which is higher than the 6 mm prescribed by EUROCODE 4 [35] as the
minimum value for a connector to be considered ductile. Push-out test
P2 showed a slip capacity of 5.92 mm, falling a little below the minimum
requirement. Push-out tests P3 and P4 displayed slip capacity values of
0.74 mm and 0.30 mm, respectively, qualifying the connection as rigid.
The reinforcing bars controlled the opening of the cracks, thus con-
tributing to the shear strength of the concrete in the post-cracking
stage. Fig. 19 shows the curves of strain versus deformation of the trans-
verse reinforcing bars of the four prototypes. The reinforcing bars only
started to work effectively after the adhesive failure between the mate-
rials, with the occurrence of stress redistribution throughout the con-
crete slab and the onset of longitudinal cracking.
For prototypes P1 and P2, the reinforcing bars employed consisted of
straight bars with ϕ 10 mm and 300 mm in length. Although the curves
of strain versus slip showcased a functional slip capacity, the reinforcing
bars were unable to maintain the bearing capacity of the prototypes,
leading to sudden failure and an immediate longitudinal cracking
along the entire length of the concrete slab as can be observed in
Fig. 20 (a). Push-out prototypes P3 and P4 displayed an opposite behav-
iour to the ones obtained in the previous tests. Despite the low slip ca-
pacity, the “fish-bone”-shaped reinforcing bars with ϕ10 mm enabled
a gradual reduction in the load up until its stabilisation. In all tests,
the three steel bars inside each opening were not equally submitted to
the load, with the bars closer to the web contributing with most of the Fig. 22. Load versus mid-span deflection curves for specimens CB1 and CB2.
16 C.A. de O. Ferrante et al. / Journal of Constructional Steel Research 168 (2020) 105868

Fig. 23. Load versus slip curves (a) CB1, (b) CB2.

the plate model (P2) had a 30% increase in the ultimate resistance when state, followed by a nonlinear behaviour and the failure of the beam.
compared to the corresponding prototype with no plate (P1). The bending moment at the end of the elastic domain was equivalent
to 90kN.m in both specimens, representing 74% and 60% of the ultimate
5.2. Experimental results of composite beam tests resistance for composite beams CB1 and CB2, respectively. Beams 1 and
2 had ultimate bending moments of 122kN.m and 150kN.m, and maxi-
The curves referring to the bending moment versus the mid-span mum deflections in the mid-span of 50 mm and 70 mm, respectively.
deflections of the tested specimens are depicted in Fig. 22. The speci- According to the test results, the composite beam CB2 showed a higher
mens showed a linear elastic behaviour during the initial loading ductility and strength.

Fig. 24. Load versus strain curves, (a) CB1, (b) CB2.
 C.A. de O. Ferrante et al. / Journal of Constructional Steel Research 168 (2020) 105868 17

Fig. 25. Progression of the cracks (a) concrete slab bottom (b) concrete slab top (c) crushing of the concrete in the web opening.

Fig. 26. Deformed shape of composite beam CB2.

Fig. 23 depicts the bending moment versus slip curves, which

were measured at the ends of each composite beam. The compos-
ite action behaviour described experimentally consisted of two
stages. The first stage was characterised by the elastic phase,
where the composite elements exhibited full composite action
produced by the adhesion between steel and concrete. A plastic
behaviour, defined by a plateau, was present in the second stage.
The adhesion between the two materials was responsible for
most of the longitudinal shear resistance. The adhesion failed
upon reaching a bending moment around 100kN.m for CB1, and
around 120kN.m for CB2. The presence of transverse reinforcing
bars improved the connection's slip capacity in the plastic phase,
with composite beams CB1 and CB2 showing maximum slip values
of 1.1 mm and 5.8 mm, respectively.
Fig. 24 shows the bending moment versus strain curves of the two
performed tests. In composite beam CB1, the strains were measured in
two central parts of the section, on the top and bottom flanges. The av-
erage values obtained by each strain gauge pair were plotted on the
graph. The yield stress (fy) of ASTM A572 grade-50 steel (web and bot-
tom flange) and ASTM A36 steel (top flange and U-section of composite
beam CB2) is also shown in the graphs. The average 7–8 for composite
beam CB1 reached the yield stress that was expected in the mid-span,
while composite beam CB2 did not reach the corresponding yield stress
specified for each steel type. In both tests, the top flange showed small
compressive strains.
The concrete slabs cracks were monitored in the two tests. During
the testing of composite beam CB1, no significant cracks were observed. Fig. 27. Curves of bending moment versus mid-span deflection.
18 C.A. de O. Ferrante et al. / Journal of Constructional Steel Research 168 (2020) 105868

On the other hand, composite beam CB2 presented cracks in several also responsible for the post-failure behaviour, maintaining a grad-
points when tested. One of these points was the centre of the span, ual slip between the section and the concrete slab;
where the concrete slab bottom presented the cracks illustrated in f) push-out tests P1 and P2 failed abruptly and were unable to sustain
Fig. 25(a). These cracks occurred due to the tensile forces produced by the load. Push-out models P3 and P4, in turn, proved to be able to
the bending moment. Fig. 25(b) shows a longitudinal cracking on the sustain the load, which later underwent a gradual reduction.
top side of the concrete slab. This crack occurred in tandem with the
concrete failure, located in the opening area near the section's web, as The experimental results provided comprehensive information
shown in Fig. 25(c). Both crack and failure in the concrete were caused on the flexural behaviour of the composite beams and the behaviour
by the longitudinal shear, which led to a relative slip between the sec- of the shear connection. The following conclusions could be drawn
tion and the concrete slab. In Fig. 26 is shown the deformed shape of from the composite beams study:
composite beam test CB2 where can be observed a 70 mm deflection
in the mid-span. a) the results of the parametric analysis show that among the sections
with lower masses, the ½ W 410 × 38.8 section proved to be advan-
5.3. Comparison between theoretical and experimental results tageous in most cases (for spans of 7, 8, and 9 m), by only changing
the slab effective width;
Fig. 27 shows the analytical and experimental evolution curves of b) the experimentally-measured behaviour of the composite beams
the bending moment in the composite beams. Both models show con- was very close to the one predicted in theory. However, a global col-
sistent results in the linear phase. The experimental model's behaviour lapse mode was not observed;
changed at a bending moment of 100 kN.m, due to the materials nonlin- c) according to the experimental results, composite beam CB2 had a
earity and the breaking of bond in the steel-concrete interface. The end 36% increase in the ultimate resistance and was 40% more ductile
of the linear domain represents the onset of slippage between the two when compared to composite beam CB1;
materials, with the emergence of cracks along with the slab and a con- d) the values obtained with the theoretical model, considering the
siderable reduction in the beam stiffness. In the final phase, the CB1 ex- section's plastic resistance, were adequate for a preliminary analysis
perimental curve showed a tendency to be closer to the analytical of the composite system's global behaviour. The CB1 and CB2 tests
model, reaching 97% of the expected value of the ultimate flexural respectively reached 97%, and 84% of the ultimate resistance pre-
capacity. dicted in theory;
The CB2 experimental model was not as consistent with the analyt- e) the testing of the composite beams displayed adequate composite
ical model as the first CB1 beam. This experimental model reached 86% behaviour and longitudinal shear strength.
of the stiffness value predicted by the analytical model in the elastic
phase. The welding of the U-section to the bottom flange may have Acknowledgements
been the cause of the inconsistent stiffness in the models, unlike what
occurred in the testing of composite beam CB1. The stiffness degrada- The authors would like to thank CAPES (Finance code 001), CNPq
tion in the CB2 experimental model increased with the start of the rela- (305143/2015-8) and FAPERJ (E-26/203.186/2015) for the financial
tive slip between steel/concrete and the yielding of the materials. The support given to this research program.
experimental test reached 84% of the ultimate flexural capacity pre-
dicted by the analytical model. References

6. Final considerations [1] R.M. Lawson, P. Beguin, R. Obiala, M. Braun, Slim-floor construction using hollow-
core and composite decking systems, Steel Constr. 8 (2015) 85–89.
[2] X. Lu, P. Mäkeläinen, Slim floor developments in Sweden and Finland, Struct. Eng.
This paper presents an experimental and analytical investigation of Int. 6 (1996) 127–129.
the structural behaviour of two composite joists with partially encased [3] M. Braun, O. Hechler, R. Obiala, Untersuchungen zur verbundwirkung von
betondübeln, Stahlbau 83 (2014) 302–308.
castellated steel beams. Four push-out tests were also conducted to in- [4] M. Braun, R. Obiala, C. Odenbreit, Analyses of the loadbearing behaviour of
vestigate the shear connection. The composite beams had asymmetric deepembedded concrete dowels, cosfb, Steel Constr. 8 (2015) 167–173.
steel I-sections with regularly-spaced rectangular web openings. The [5] Y.K. Ju, S. Chun, S.-D. Kim, Flexural test of a composite beam using asymmetric steel
section with web openings, J. Struct. Eng. 135 (2009) 448–458.
top flange and part of the web of the steel section were encased in the
[6] B.Y. Huo, C. D'mello, Shear transferring mechanisms in a composite shallow cellular
concrete slab. The push-out tests consisted of the adhesion between floor beam with web openings, Struct 9 (2017) 134–146.
the encased section and the concrete slab, and tie-bars with the infill [7] B.Y. Huo, C. D'mello, Push-out tests and analytical study of shear transfer mecha-
concrete traversing the web openings. nisms in composite shallow cellular floor beams, J. Constr. Steel Res. 88 (2013)
191–205.
Concerning the push-out tests, the following conclusions can be [8] S. Chen, T. Limazie, J. Tan, Flexural behavior of shallow cellular composite floor
pointed out: beams with innovative shear connections, J. Constr. Steel Res. 106 (2015) 329–346.
[9] D. Lam, X. Dai, U. Kuhlmann, J. Raichle, M. Braun, Slim-floor construction–design for
ultimate limit state, Steel Constr. 8 (2015) 79–84.
a) the connection between concrete and steel showed a rigid behav- [10] M. Mensinger, M. Fontana, A. Frangi, Entwicklung eines multifunktionalen
iour. Prototypes P1, P2, P3, and P4 had slip capacities of 11.81 mm, deckensystems mit erh¨ohter ressourceneffizienz, Stahlbau 79 (2010) 282–297.
5.92 mm, 0.74 mm, and 0.30 mm, respectively; [11] S. Peltonen, M.V. Leskelä, Connection behaviour of a concrete dowel in a circular
web hole of a steel beam, International conference on composite construction in
b) after reaching the maximum load, the adhesive failure between the steel and concrete V – ASCE. South Africa 2006, pp. 544–552.
materials caused the slip to rapidly increase, submitting other mech- [12] J. Qureshi, D. Lam, J. Ye, The influence of profiled sheeting thickness and shear
anisms, such as the transverse reinforcing bars, to the direct effect of connector's position on strength and ductility of headed shear connector, Eng.
Struct. 33 (2011) 1643–1656.
the load; [13] D.L. Araújo, Sales MWR, S.M. Paulo, Debs ALH, Headed steel stud connectors for
c) push-out models P1, P2, P3, and P4 had ultimate load values of 172.2 composite steel beams with precast hollow-core slabs with structural topping,
kN, 225.1 kN, 224.5 kN, and 207.2 kN, respectively; Eng. Struct. 107 (2016) 135–150.
[14] J. Bonilla, L.M. Bezerra, E. Mirambell, Resistance of stud shear connectors in compos-
d) the resistance of push-out model P2 was 30% higher than P1model,
ite beams using profiled steel sheeting, Eng. Struct. 187 (2019) 478–489.
as a result of the web plate increasing the triaxial compression area [15] M.C. Rodrigues, Andrade S.A.L. de, Lima L.R.O. de, S. Vellasco PCG da, F.B. Ramires,
in the concrete; Experimental assessment of the composite joints shear connector component, J.
e) the “fish-bone”-shaped reinforcing bars used in models P3 and P4 Constr. Steel Res. 132 (2017) 203–216.
[16] C. Vianna J da, L.F. Costa-Neves, S. Vellasco PCG da, Andrade S.A.L. de, Structural be-
contributed to a 30% increase in strength when compared to the haviour of T-Perfobond shear connectors in composite girders: an experimental ap-
models with straight reinforcing bars. The reinforcing bars were proach, Eng. Struct. 30 (2008) 2381–2391.
 C.A. de O. Ferrante et al. / Journal of Constructional Steel Research 168 (2020) 105868 19

[17] J.P.S. Cândido-Martins, L.F. Costa-Neves, P.C.G. da S. Vellasco, Experimental evalua- [28] R.M. Sencu, Y.C. Wang, J. Yang, D. Lam, Performance evaluation of demountable
tion of the structural response of Perfobond shear connectors, Eng. Struct. 32 shear connectors with collar step at ambient and elevated temperatures, Eng. Struct.
(2010) 1976–1985. 194 (2019) 94–105.
[18] L.F. Costa-Neves, J.P. Figueiredo, P.C.G. da S. Vellasco, J. da C. Vianna, Perforated shear [29] N. Rehman, D. Lam, X. Dai, A.F. Ashour, Experimental study on demountable shear
connectors on composite girders under monotonic loading: an experimental ap- connectors in composite slabs with profiled decking, J. Constr. Steel Res. 122
proach, Eng. Struct. 56 (2013) 721–737. (2016) 178–189.
[19] H. Allahyaria, I.M. Nikbin, S.R. Rahimi, A. Heidarpour, A new approach to determine [30] E. Baran, C. Topkaya, An experimental study on channel type shear connectors, J.
strength of Perfobond rib shear connector in steel-concrete composite structures by Constr. Steel Res. 74 (2012) 108–117.
employing neural network, Eng. Struct. 157 (2018) 235–249. [31] P. Keo, C. Lepourry, H. Somja, F. Palas, Behavior of a new shear connector for U-
[20] C. Zhao, Z. Li, K. Deng, W. Wang, Experimental investigation on the bearing mecha- shaped steel-concrete hybrid beams, J. Constr. Steel Res. 145 (2018) 153–166.
nism of Perfobond rib shear connectors, Eng. Struct. 159 (2018) 172–184. [32] Y. Liu, L. Guo, B. Qu, S. Zhang, Experimental investigation on the flexural behavior of
[21] C. Vianna J da, Andrade S.A.L. de, S. Vellasco PCG da, L.F. Costa-Neves, Experimental steel-concrete composite beams with U-shaped steel girders and angle connectors,
study of Perfobond shear connectors in composite construction, J. Constr. Steel Res. Eng. Struct. 131 (2017) 492–502.
81 (2013) 62–75. [33] L.M. Bezerra, O.O. Cavalcante, L. Chater, J. Bonilla, V-shaped shear connector for com-
[22] S.-H. Kim, K.-S. Kim, S. Park, J.-H. Ahn, M.-K. Lee, Y-type perfobond rib shear connec- posite steel-concrete beam, J. Constr. Steel Res. 150 (2018) 162–174.
tors subjected to fatigue loading on highway bridges, J. Constr. Steel Res. 122 (2016) [34] E. Hosseinpour, S. Baharom, W.H.W. Badaruzzaman, A.W.A. Zand, Push-out test on
445–454. the web opening shear connector for a slim-floor steel beam: experimental and an-
[23] S.-H. Kim, O. Han, K.-S. Kim, J.-S. Park, Experimental behavior of double-row Y-type alytical study, Eng. Struct. 163 (2018) 137–152.
perfobond rib shear connectors, J. Constr. Steel Res. 150 (2018) 221–229. [35] EN, 1994-1-1: Eurocode 4 - Design of composite steel and concrete structures, part
[24] J.-C. Gu, D. Liu, W.-Q. Deng, J.-D. Zhang, Experimental study on the shear resistance 1–1: General rules and rules for buildings, Brussels, 2004.
of a comb-type perfobond rib shear connector, J. Constr. Steel Res. 158 (2019) [36] CAN/CSA-S16-14, Design of Steel Structures, Ontario, 2014.
279–289. [37] fib Bulletin 55, Model Code 2010, Vol. 1, Lausanne, 2010.
[25] I.E.J. Henderson, X.Q. Zhu, B. Uy, O. Mirza, Dynamic behaviour of steel–concrete [38] J. Qureshi, D. Lam, J. Ye, Effect of shear connector spacing and layout on the shear
composite beams with different types of shear connectors. Part I: experimental connector capacity in composite beams, J. Constr. Steel Res. 67 (2011) 706–719.
study, Eng. Struct. 103 (2015) 298–307. [39] P.W. Birkeland, H.W. Birkeland, Connections in precast concrete construction, J. Am.
[26] I.E.J. Henderson, X.Q. Zhu, B. Uy, O. Mirza, Dynamic behaviour of steel-concrete com- Concr. Inst. 63 (1966) 345–367.
posite beams retrofitted with various bolted shear connectors, Eng. Struct. 131 [40] A. Zona, G. Ranzi, Shear connection slip demand in composite steel–concrete beams
(2017) 115–135. with solid slabs, J. Constr. Steel Res. 102 (2014) 266–281.
[27] J.-L. Zhang, B.-D. Liu, P.-Y. Zhang, Z.-H. Wang, Small-scale test and analysis of [41] R.B. Kullman, M.U. Hosain, Shear capacity of stub-girders: full scale tests, J. Struct.
corrugated-steel-plate–concrete composite member adopting novel shear connec- Eng. 111 (1985) 56–75.
tors, Eng. Struct. 184 (2019) 369–383.