Journal of Constructional Steel Research 194 (2022) 107331

Contents lists available at ScienceDirect

Journal of Constructional Steel Research

journal homepage: www.elsevier.com/locate/jcsr

Experimental study of prestressed steel-concrete composite beams with

profiled steel decking
Marcela Moreira da Rocha Almeida a, *, Alex Sander Clemente de Souza a,
Augusto Teixeira de Albuquerque b
a
Department of Civil Engineering, Federal University of São Carlos, Brazil
b
Department of Civil Engineering, Federal University of Ceará, Brazil

A R T I C L E I N F O A B S T R A C T

Keywords: Steel-concrete composite beams with concrete slab cast on profiled steel decking are widely used in bridges and
Steel-concrete composite beams building construction. Prestressed tendons can be added to the beams to increase their structural performance in
Trapezoidal decking new structures or in retrofitting. However, no experimental data was found in the literature that investigated
External prestressing
prestressing associated with profiled steel decking in composite beams. This paper presents an experimental
Experimental study
study that investigates the behavior of simply supported prestressed steel-concrete composite beams with pro­
Numerical model
filed steel decking under positive bending moment. Tests were carried out both on a beam with straight tendon
profile placed at the beam’s full length and on a conventional non-prestressed composite beam, and deflections
and strains were measured. The beams were mounted in a four-point bending test configuration and then loaded.
The tests terminated due to excessive deformations of the models and safety concerns, and no particular failure
mode was observed. Then, a numerical model to simulate the behavior of the composite beams was developed
and validated with the test data. The numerical models indicated that the failure mechanism in both beams
consisted in yielding of the tensioned steel beam. It was found that adding prestressing can increase the ultimate
moment by up to 19% and significantly reduce deflections under service loads.

Ayyub, Sohn and Saadatmanesh [2] performed experimental tests on

three composite beams under positive moments, two prestressed each
1. Introduction with straight and draped tendons, and one prestressed with straight bars
to verify the influence of prestressing type and tendon profile. They also
Steel-concrete composite floors composed by the association of steel discussed an analytical method of analysis. They concluded that tendons
beams and composite slabs are commonly used in the construction of are more effective than bars in strengthening the beam and that draped
buildings and have proven to be an efficient and economical structural tendons increase the ductility of the beam when compared to straight
solution. On the other hand, prestressing steel-concrete composite tendons.
beams can increase its ultimate capacity, reducing deflections under Chen and Gu [3] carried out experiments in two prestressed steel-
service loads and reducing cracking at concrete slab. Several experi­ concrete composite beams under positive moments to study the ulti­
ments have [1–5] already been performed on prestressed steel-concrete mate moment capacity and incremental tendon stresses. The authors
composite beams with concrete slab, but no previous experimental also proposed simplified expressions to predict the ultimate incremental
research on prestressed steel-concrete composite beams with steel tendon forces. One beam was tested in two loading steps: the first
trapezoidal decking was found. without prestressing, until the yielding of the steel beam and then the
Two steel-concrete composite beams prestressed with bars were beam was unloaded; and the second step was performed with the action
experimentally studied by Saadatmanesh, Albrecht and Ayyub [1], one of prestressing. The other beam was tested in one stage and prestressed
was subjected to positive bending moment and the other to negative before the loading. The authors found out that adding prestressing to the
bending moment. The authors verified that adding prestressed bars composite beam increased the yield and ultimate loads with less
increased the yield and ultimate loads and reduced stresses in the con­ deflection under service loads. Moreover, substantial incremental forces
crete slab in both positive and negative moment regions.

* Corresponding author.
E-mail address: marcelamra@hotmail.com (M.M.R. Almeida).

https://doi.org/10.1016/j.jcsr.2022.107331
Received 28 March 2022; Received in revised form 9 May 2022; Accepted 10 May 2022
Available online 20 May 2022
0143-974X/© 2022 Elsevier Ltd. All rights reserved.
M.M.R. Almeida et al. Journal of Constructional Steel Research 194 (2022) 107331

Notation composite beam

MR- analytical negative moment resistance for conventional
Bc concrete slab width composite beam
bf flange width of the steel beam Mu,EXP ultimate moment from experiments
d height of the steel beam Mu,FE ultimate moment from the finite element analysis
Ecm secant modulus of concrete My yield moment
fck characteristic compressive strength of concrete tf flange thickness of the steel beam
fcm compressive cylinder strength of concrete tw web thickness of the steel beam
ft tensile strength wc crack opening displacement at which stress can no longer
Gf fracture energy be applied
hc slab thickness ΔT increment in prestressing force
he eccentricity of the tendons, measured from the bottom of ΔT,EXP increment in prestressing force from experiment
the steel beam ΔT,FE increment in prestressing force from the finite element
L span length analysis
Li span between applied forces δ deflection
Lt total length of the beam δu,EXP ultimate experimental deflection
M moment ε strain
Mpl plastic moment εc compressive strain in concrete
MR,c analytical positive moment resistance for conventional εc1 compressive strain in concrete at peak stress
composite beam σc compressive stress in concrete
MR,p analytical positive moment resistance for prestressed

are developed in the tendons at the ultimate states of the beams. conventional steel-concrete composite beams with trapezoidal decking
Uy and Craine [4] performed experiments on both conventional and with different levels of shear connection. The authors intended to
prestressed steel-concrete composite beams under positive moment. The investigate the strength and ductility of the shear connectors and pro­
results showed that post-tensioning the composite beam increased the vide benchmark data for the calibration of theoretical models. The re­
ultimate resistance in about 20%. They also developed a numerical sults showed that, regardless the degree of shear connection, the beams
model that was calibrated with the test data and a parametric study to behaved very ductile, and the ultimate moments predicted by the
investigate the influence of several parameters such as effect of slab Eurocode 4 were conservative when compared to the experimental re­
dimensions, initial force and eccentricity in tendons, area of reinforce­ sults. Nie et al. [11] experimentally investigated the behavior of 13 steel-
ment and concrete strength. The parametric study showed that: concrete composite beams, most of them with partial shear connection.
increasing the slab width and depth resulted in an increased beam’s The beams were divided into 3 groups. The first were simply supported
strength; increasing in the initial tendon forces resulted in a higher beams subjected to positive bending moment, the second to negative
carrying capacity but with reduction in ductility; increasing the tendon bending moment and the third consisted in continuous beams. The au­
eccentricity has the effect to increase both the stiffness and strength of thors reported that the experimental results properly agreed with the
the beam; the increase in reinforcement area in the slab slightly in­ theoretical approaches of the designs codes, and that partial shear
creases the stiffness and strength; and increasing the concrete strength connections can be used in the different beam’s configurations that were
slightly influenced the elevation of the yield and ultimate loads. investigated.
Six experimental tests on steel-concrete composite beams under As mentioned, although several experimental studies have been
positive moment were performed by Lorenc and Kubica [5]. One beam developed on prestressed steel-concrete beams with concrete slab, there
was conventional, with no tendons, three were arranged with draped are no experiments with trapezoidal decking, which is a widely practical
tendons and two were arranged with straight tendons. Furthermore, the solution adopted in composite sections. Therefore, the purpose of this
concrete strength varied between the specimens. The conclusions research is to investigate the behavior of simply supported prestressed
pointed out that: external prestressing increases the ultimate load ca­ steel-concrete composite beams with steel trapezoidal decking under
pacity in about 25% when compared to the conventional composite positive moment, by the development of an experimental study. More­
beam; there were no significant differences between draped and straight over, a numerical model is proposed and validated with the tests data.
tendons at the same tendon eccentricity; the failure mechanism con­
sisted in yielding of the steel beam followed by compressive crushing at 2. Experimental program
the concrete slab. Additionally, the authors reported a local failure at the
beam’s ends with no effect of the beams capacity. The experimental program developed is composed by two simply
Almeida et al. [6] discussed an analytical model to predict the ulti­ supported steel-concrete composite beams under positive moment. The
mate strength in steel-concrete composite beams prestressed with first beam was prestressed with two straight tendons placed along the
external tendons and proposed a numerical model to simulate the entire length of the beam on the steel section (PCB) and the second is a
behavior of this configuration of beams. Moreover, the authors devel­ conventional steel-concrete composite beam, without any prestressed
oped a parametric study to investigate the influence of several variables tendons (CCB).
such as position and shape of tendons, initial prestressing force, tendon
length and span length. 2.1. Specimens
Regarding composite prestressed steel-concrete composite beams
with trapezoidal decking, no experimental data were found and conse­ The two beams had identical transversal sections consisting of a
quently no numerical models were validated. Some authors [7–9] pro­ concrete slab with a steel trapezoidal deck, connected to a steel beam by
posed finite element numerical models to predict the behavior of headed studs. The composite slab was 1200 mm wide, and its height
composite steel-concrete composite beams with trapezoidal decking, but varied from 110 mm (from the top of the concrete cover to the bottom of
with no prestressing. Ranzi et al. [10] performed tests on two the trapezoidal deck) to 60 mm (at the crests of the steel trapezoidal

2
M.M.R. Almeida et al. Journal of Constructional Steel Research 194 (2022) 107331

deck ribs). This width adopted is greater that the effective width of L/8 displacement transducers is shown in Fig. 3a.
on each side of the steel web as recommended by [12] for conventional Strain gauges were mounted in pairs at the midspan section with
steel-concrete composite beams, which according to [13] is also suitable half-bridge configuration. There were eight measurement points of
for prestressed steel-concrete composite beams. The steel deck used in strain: two at the top of the slab (S1 and S2); two on the top flange of the
the specimens was the MF-50, with a 0.8 mm thickness, whose geometry steel beam (S3 and S4); two on the sides of the beam’s web (S5 and S6);
is shown in Fig. 1. The ribs of the decking were placed orthogonally to two on the bottom flange of the steel beam (S7 and S8). For the PCB
the longitudinal axis of the steel beams. The concrete slab was reinforced beam, there were two more measurement points at the midspan section,
with a double welded steel mesh only to prevent the concrete from one in each tendon (S9 and S10). The location of the strain gauges is
cracking. Each steel mesh was 4.2 mm in diameter with 150 mm spacing shown in Fig. 3b.
in the two orthogonal directions.
The steel beam type VS 250 × 24 had its bottom and top flanges with 2.4. Test procedure
120 × 8 mm in dimensions, and the web was 250 × 4.75 mm. In the steel
beam, 12.5 mm thick stiffeners plates were placed at the loading points, Firstly, the PCB was prestressed with post-tensioning technique, after
and at the supports and extremity 22 mm thick plates were placed at the the concrete acquired full strength. The tendons were simultaneously
beam ends to anchor the tendons. The total length of the beams was tensioned with two hydraulic jacks, one placed in each tendon, to avoid
4744 mm, simply supported on a 4500 mm long span. The slab was horizontal displacements. The nominal initial prestressing force in each
connected to the steel beam by headed stud shear connectors. Pairs of tendon was 110 kN. However, a force 5% higher was applied to
19 mm diameter by 90 mm long studs were welded to the top flange of compensate the losses that can occur during the prestressing process.
the steel beam, with 305 mm longitudinal spacing and 76 mm trans­ The elongation of the tendons was measured and checked with the
versal spacing, symmetrical to the centerline of the cross section. The theoretical estimates. One LVDT, on the midspan, and the tendon’s
beams were designed to attain full shear connection. strain gauges (S9 and S10) were recording the amounts during
The only difference between the beams is that PCB has prestressed prestressing.
tendons and CCB had a conventional composite section. Fig. 2 shows the The loading step was the same for both beams. They were tested by
geometry of PCB, which was designed with two seven-wire strands, each applying the loads by a four-point bending test configuration. Initially, a
with a nominal diameter of 12.7 mm. The strands were fixed to anchor load of 20 kN was applied to accommodate the beams, and the beams
plates at the ends of the beam. The stiffeners have holes to allow the were unloaded. Then, the beams were incrementally loaded until the
passage of tendons. A summary of the details of the specimens is pre­ development of excessive deformations of the models and the tests were
sented in Table 1. terminated due to safety concerns, as happened with [10]. After that, the
maximum load was maintained for a moment, and then the beams were
2.2. Materials gradually unloaded. Fig. 4 presents the arrangement of the beams at the
testing frame. Fig. 5 shows the prestressing of PCB at one of the beam’s
In order to determinate the mechanical properties of the steel beam, ends.
specimens were cut from the same steel sheet lot that the steel profiles
were made. The tensile properties values, determined from the tensile 3. Numerical model
tests performed according to [14], were 312.9 MPa and 313.7 MPa in
yield stress and 480.2 MPa and 474.8 MPa in ultimate stress, and the In order to simulate the behavior of the beams at the experiments, a
average values were adopted. The average Young’s modulus was 204 finite element model was developed on Abaqus Software [15]. The
GPa for the specimens. The nominal tensile properties of the low- models were assembled considering one plane of symmetry at the z axis,
relaxation tendons were 1710 MPa in yield stress and 1900 MPa in ul­ to simplify and reduce the processing cost.
timate strength. The decking was made with type ASTM A 653 Grade 40
galvanized steel, with nominal tensile properties of 280 MPa in yield 3.1. Element types and mesh
stress and 380 MPa in ultimate strength.
The axial compressive strength was measured at 28 days after the The concrete part of the slab was modeled using eight-node linear
concrete casting. Six specimens of 200 mm height and 100 mm diameter hexahedral solid elements with reduced integration (C3D8R). At the
were made during each slab casting. The measured mean compressive steel beam and steel deck, quadrilateral shell element with reduced
strength at 28 days was about 28 MPa for both specimens. A summary of integration (S4R) was adopted. The shear studs were modeled using
the mechanical properties of the specimens is presented in Table 2. two-node beam elements (B31). For the tendons, two-node linear three-
dimensional truss elements (T3D2) were employed. The elements that
2.3. Instrumentation are part of the numerical model are shown in Fig. 6.
A sensitivity analysis was performed to determine the suitable mesh
PCB was assembled with five displacement transducers (LVDT). Two size of the elements. The mesh of the steel beam and shear studs were
LVDT’s were located at the midspan, one in each half of the bottom discretized in 10 mm elements, as used by [6,16]. In the concrete slab
flange (L1 and L2), to measure the beam’s vertical displacement. Two and steel deck meshes, 20 mm side elements were adopted. The tendons
were mounted at one end of the beam, to measure the relative slip be­ were discretized in single elements between each of the stiffeners, to
tween the steel beam and the slab (L2 and L3). The fifth LVDT was assure that each patch behaves continuously. Mesh discretization is
positioned at one of the supports (L5). On the other hand, CCB was displayed in Fig. 7.
assembled with only the LVDT’s L1 and L2. The location of the
3.2. Contact properties and boundary conditions

To simulate the contact between the welded elements such as the

web, flanges, stiffeners, and plates that are part of the steel beam and
also the contact between the shear studs and the top flange, tie constrain
was used. The contact of the bottom surface of the steel deck with the top
flange of the steel beam was defined with a hard property in the normal
direction to the plane, and a penalty property in the tangential direction
Fig. 1. Geometry of the steel deck. with a friction coefficient of 0.01, as adopted by [8]. In the interaction

3
M.M.R. Almeida et al. Journal of Constructional Steel Research 194 (2022) 107331

Fig. 2. Geometry of beam PCB.

Table 1
Details of the specimens.
Model L (mm) Lt (mm) Li (mm) Tendon profile Bc (mm) he (mm) hc (mm) d (mm) bf (mm) tf (mm) tw (mm)

PCB 4500 4744 1500 Straight 1200 50 110 250 120 8 4.75
CCB 4500 4744 1500 – 1200 – 110 250 120 8 4.75

between the concrete slab and the top surface of the steel deck a similar
Table 2
procedure was adopted, considering a hard property in the normal di­
Summary of mechanical properties of the specimens.
rection to the plane, and a penalty property in the tangential direction
Steel beam Steel decking Concrete High strength with a friction coefficient of 0.5. The headed shear studs were defined as
slab strands
embedded within the concrete slab, where the nodes of the studs are
fy fu fy fu fcm (MPa) fy fu constrained in the concrete slab, which is the host element. This
(MPa) (MPa) (MPa) (MPa) (MPa) (MPa)
approach was also adopted by [6].
313.3 477.5 280 380 28 1710 1900 To simulate the rollers that supports the beam, a boundary condition
restricting the vertical translation of the supports was used. A boundary
condition representing the symmetry plane on the Z axis was also
adopted.

3.3. Loading steps

The prestressing and loading steps that occurred in the experiment

were also simulated in the modeling. Firstly, the initial step imple­
mented all the boundary conditions and the tendon stresses. The com­
mand predefined field was used to implement the tendon loads, in which
can be defined an initial value of the mechanical variable stress applied
to the tendons. In the first effective step, the stresses were gradually
transferred from the tendons to the composite beam by the contact at the
anchorages placed at the beam extremity, simulating the state of stresses
Fig. 3. Location of LVDT’s and strain gauges.
at the model after the prestressing application. On the beginning of the
second step, the beam already behaves as a prestressed composite beam,
and then the external load is applied. This approach was also adopted by

4
M.M.R. Almeida et al. Journal of Constructional Steel Research 194 (2022) 107331

(a) PCB (b) CCB

Fig. 4. Arrangement of beams in tests.

curve, as presented by Han et al. [17]. For the headed shear connectors
an elastoplastic model with positive linear hardening was adopted.
Finally, the model used for the steel trapezoidal deck was the perfect
elastoplastic.

3.4.2. Concrete
To simulate the behavior of the concrete slab, the concrete damage
plasticity (CDP) presented by Eurocode 2 [18] was adopted. The stress-
strain relationship of concrete under compression is given by Eqs. (1–2).

σc k (εc /εc1 ) − (εc /εc1 )2

= (1)
f cm 1 + (k − 2) (εc /εc1 )

Ecm |εc1 |
k = 1.05 (2)
f cm
To model the tensile behavior of concrete, a stress-displacement
relationship was adopted instead of a stress-strain relationship. This
relationship is based on the fracture energy (Gf), derived from Hillerborg
[19], as given in Eqs. (3–4). The fracture energy is a material property,
with typical values in the range from 40 N/m, for typical construction
concrete, to 120 N/m, for high strength concrete. The value of 40 N/m
was adopted in the analysis. Tensile strength is determined as presented
in Eurocode 2 [18] (Eq. 5). This approach to model the tensile behavior
of concrete was also used by [6].
( )
Gf ft
wc = 0.8 for 0 < f t ≤ (3)
ft 3
( )
Gf ft
wc = 3.6 for f t > (4)
ft 3

Fig. 5. Prestressing of PCB. f t = 0.3 f ck (2/3) (5)

[6]. On the other hand, the conventional beam, without prestressing, 4. Results and discussion
had only one step in which the loading was applied.
For PCB, the nominal initial prestressing force was 110 kN in each
tendon. After prestressing, with the hydraulic jacks still in place, the
3.4. Material modeling maximum measured force was about 115 kN. When the jacks were
removed and the tendons were anchored at the extremity plates, there
3.4.1. Steel was an accommodation of the tendons, and the prestressing force on
For the steel beam, an idealized trilinear stress-strain relationship each of the tendons dropped to about 95 kN. The stress loss during
was adopted, in which the hardening path is constituted by a trilinear prestressing was approximately 13% when compared to the design force

5
M.M.R. Almeida et al. Journal of Constructional Steel Research 194 (2022) 107331

Concrete slab
(C3D8R)
Headed
stud (B31)
Steel deck
(S4R)

Steel beam Tendon

(S4R) (T3D2)

Fig. 6. Numerical model elements.

development of the bending moment vs the deflection at midspan for

PCB measured by the LVDT’s L1 and L2. Fig. 9 presents the development
of the moment vs the incremental prestressing force in each tendon
determined through the strains measured by the gauges S9 and S10. The
numerical results are also shown in Fig. 8 and Fig. 9.
At the conventional non-prestressed composite beam, the beam was
loaded until the midspan moment reached about 209 kN.m. Fig. 10
shows the development of the moment vs the deflection at midspan for
CCB measured by the LVDT’s L1 and L2, and the numerical results are
also presented.
In both beams, no particular failure mechanism was observed in the
development of the experiments. The ultimate moments determined
from the numerical models are compared with the experimental data in
Table 3, which also compare the ultimate deflections and increments in
the tendon forces. The numerical results are presented for the deflection
limit of L/100, to satisfy a reasonable serviceability limit state. In
Table 4 are presented the analytical moment resistance in positive mo­
Fig. 7. Mesh discretization. ments regions for the conventional and prestressed composite beams;
the moment resistance in negative moments regions for the composite
and 21% when compared to the actual applied force. An upward beam; and the plastic and yield moments for the steel section. The
deflection of 3.6 mm was registered at the end of the prestressing step. determination of the resistance of prestressed composite beams was
At the loading step, the prestressed composite beam was loaded until based in [6]. The theoretical estimates of positive moment resistance for
the midspan moment was about 249 kN.m. Fig. 8 presents the the conventional and prestressed composite beams are consistent with

300
300
L1
L2 250
250 Numerical

200
200
M (kN.m)

150
M (kN.m)

150

100
100
S9
50 S10
50 Numerical

0
0 0 10 20 30 40 50
-10 0 10 20 30 40 50 60 ΔT (kN)
δ (mm)
Fig. 9. Moment vs incremental tendon forces for numerical and test results
Fig. 8. Moment vs deflection curves for numerical and test results for PCB. for PCB.

6
M.M.R. Almeida et al. Journal of Constructional Steel Research 194 (2022) 107331

300 Fig. 13 shows the deformed beams after the unloading process. It can
L1
be seen the permanent deflection that occurred in both PCB and CCB.
L2 The horizontal displacement in the concrete slab and steel beam,
250
Numerical measured by the transducers L3 and L4, respectively, are shown in
Fig. 14, where a detail of the maximum horizontal displacements
200 measured is illustrated. The results indicate that the behavior of the
beam was highly ductile with large values of slip recorded at the beam
M (kN.m)

150 ends, as reported by [10]. The vertical displacement measured by the

transducer L5 at one of the supports was very small, which indicates that
the concrete supports did not present a significant shortening that could
100
influence the measuring of the beam’s deflection.
In Fig. 15 the extremities of PCB after the end of the test are shown.
50 End 1 showed no damage at the concrete slab, but the concrete slab at
End 2 was crushed locally. The crushed extremity was the one where
0 there was a rib at the trapezoidal sheet, and the one that remained intact
0 10 20 30 40 50 60 was the one with reduced thickness of the slab. This local failure did not
δ (mm) influence the ultimate resistance of the composite beam. There were no
relevant cracks at the rest of the concrete slab, especially at midspan,
Fig. 10. Moment vs deflection curves for numerical and test results for CCB.
indicating that the concrete did not present crushing failure. Even with
no particular failure mechanism observed in the experiments, when
analyzing the numerical model at the same load level in which the ex­
Table 3
Comparison between experimental and numerical results. periments were ended, considerable higher strains than the ones re­
ported in Fig. 11 and Fig. 12 were identified locally on the bottom web
Model Experimental Numerical Mu,EXP/
near to the loading points (Fig. 16), where no instrumentation was
Mu,FE
Mu,EXP δu,EXP ΔT,EXP Mu,FE ΔT,FE installed, indicating that the failure may have started due to excessive
(kN.m) (mm) (kN) (kN.m) (kN)
yielding at the steel beam. The failure due to yielding of the steel section
PCB 249.0 51.3 38.4 251.0 43.5 0.99 without concrete crushing indicates a more ductile behavior than the
CCB 209.0 45.5 – 204.2 – 1.02 beams with concrete slab presented by [1–3,5], who reported failure due
to yielding of the steel section followed by concrete crushing at the slab.
The de-bonding between the concrete slab and trapezoidal decking was
Table 4 not very significant in both beams (Fig. 17).
Analytical moments.
MR,c (kN.m) MR,p (kN.m) MR- (kN.m) Mpl (kN.m) My (kN.m) 5. Conclusion
183.4 233.3 94.9 93.0 83.3
The present work investigated the behavior of prestressed steel-
concrete composite beams with steel trapezoidal decking under posi­
the experimental and numerical results. The comparison between nu­ tive moment by the development of an experimental and numerical
merical and test results presented in Fig. 8, Fig. 9 and Fig. 10 shows a study. Based on the presented results, the following conclusions were
satisfactory correlation between the curves, indicating that the numer­ drawn:
ical models were able to represent the behavior that occurred
experimentally. • The use of prestressing in composite beams can significantly improve
Fig. 11 and Fig. 12 show the development of strains measured at the their behavior, increasing the yielding and ultimate loads and
midspan by the gauges along the height of PCB and CCB beams, reducing deflections under service loads. The experimental data
respectively, at different moment values. The presented moments are showed an increase of 19% in ultimate capacity on the prestressed
equivalent to increments of 50 kN in loading until the ultimate load. beam when compared to the control beam without prestressing;

Fig. 11. Development of strains along PCB height.

7
M.M.R. Almeida et al. Journal of Constructional Steel Research 194 (2022) 107331

Fig. 12. Development of strains along CCB height.

300

250

200
M (kN.m)

150

100

L3
50
L4

0
-14 -12 -10 -8 -6 -4 -2 0 2
δ (mm)

(a) PCB Fig. 14. Horizontal displacements measured by L3 and L4.

(b) CCB
Fig. 13. Post-test deformed beams.

• Prestressing losses represented 21% of the initial prestressing force (a) End 1 (b) End 2
that was applied, which was already 5% higher than the designed
Fig. 15. Damage to PCB ends.
initial prestressing force. To achieve the desired increase in the
strength of the composite beams, higher losses should be considered
when applying prestressing; no rib with only the slab cover, or suitable reinforcement should be
• The slab failed locally at the beam’s end where there was a located placed to avoid localized damage;
rib, but no significant effects on the ultimate resistance of the beams • During the experiments, no particular failure mechanism was iden­
were identified. The slab should end with the portion where there is tified in both beams, but the numerical models indicates that the
failure mechanism consisted in yielding of the tensioned steel beam.

8
M.M.R. Almeida et al. Journal of Constructional Steel Research 194 (2022) 107331

the work reported in this paper.

References

[1] H. Saadatmanesh, P. Albrecht, B.M. Ayyub, Experimental study of Prestressed

composite beams, J. Struct. Eng. 115 (1989) 2348–2363, https://doi.org/10.1061/
(ASCE)0733-9445(1989)115:9(2348).
[2] B.M. Ayyub, Y.G. Sohn, H. Saadatmanesh, Prestressed composite girders under
positive moment, J. Struct. Eng. 116 (1990) 2931–2951, https://doi.org/10.1061/
(ASCE)0733-9445(1990)116:11(2931).
[3] S. Chen, P. Gu, Load carrying capacity of composite beams prestressed with
external tendons under positive moment, J. Constr. Steel Res. 61 (2005) 515–530,
Fig. 16. Strains at the steel beam in the numerical model for PCB. https://doi.org/10.1016/j.jcsr.2004.09.004.
[4] B. Uy, S. Craine, Static flexural behaviour of externally post-tensioned steel-
concrete composite beams, Adv. Struct. Eng. 7 (2004) 1–20, https://doi.org/
10.1260/136943304322985729.
[5] W. Lorenc, E. Kubica, Behavior of composite beams prestressed with external
tendons: experimental study, J. Constr. Steel Res. 62 (2006) 1353–1366, https://
doi.org/10.1016/j.jcsr.2006.01.007.
[6] M.M.R. Almeida, A.S.C. Souza, A.T. Albuquerque, A. Rossi, Parametric analysis of
steel-concrete composite beams prestressed with external tendons, J. Constr. Steel
Res. 189 (2022), https://doi.org/10.1016/j.jcsr.2021.107087.
[7] F. Tahmasebinia, G. Ranzi, A. Zona, Beam tests of composite steel-concrete
members: a three-dimensional finite element model, International Journal of Steel
Structures 12 (2012) 37–45, https://doi.org/10.1007/s13296-012-1004-3.
[8] F. Tahmasebinia, G. Ranzi, A. Zona, Probabilistic three-dimensional finite element
study on composite beams with steel trapezoidal decking, J. Constr. Steel Res. 80
(2013) 394–411, https://doi.org/10.1016/j.jcsr.2012.10.003.
[9] U. Katwal, Z. Tao, M.K. Hassan, Finite element modelling of steel-concrete
composite beams with profiled steel sheeting, J. Constr. Steel Res. 146 (2018)
1–15, https://doi.org/10.1016/j.jcsr.2018.03.011.
[10] G. Ranzi, M.A. Bradford, P. Ansourian, A. Filonov, K.J.R. Rasmussen, T.J. Hogan,
B. Uy, Full-scale tests on composite steel-concrete beams with steel trapezoidal
decking, J. Constr. Steel Res. 65 (2009) 1490–1506, https://doi.org/10.1016/j.
jcsr.2009.03.006.
[11] J. Nie, J. Fan, C.S. Cai, Experimental study of partially shear-connected composite
beams with profiled sheeting, Eng. Struct. 30 (2008) 1–12, https://doi.org/
10.1016/j.engstruct.2007.02.016.
[12] Eurocode 4. EN 1994-1-1, 2004 - Design of composite steel and concrete structures
- Part 1–1: General rules and rules for buildings. Comite Europeen de
Normalisation, Brussels, Belgium, 1994.
[13] S. Chen, Z. Zhang, Z, effective width of a concrete slab in steel-concrete composite
beams prestressed with external tendons, J. Constr. Steel Res. 62 (5) (2006)
493–500, https://doi.org/10.1016/j.jcsr.2005.08.009.
[14] American Society of Testing and Materials, E8/E8M − 13a Standard Test Methods
for Tension Testing of Metallic Materials. https://doi.org/10.1520/E0008_E
0008M-13A.
[15] DASSAULT SYSTÈMES, ABAQUS v6.12, 2012.
Fig. 17. De-bonding between concrete and trapezoidal decking.
[16] A. Rossi, C.H. Martins, R.S. Nicoletti, A.S.C. Souza, Reassesment of lateral torsional
buckling in hot-holled I-beams, Structures (2020) ISTRUC801, https://doi.org/
There were no significant cracks on the concrete slab, which means 10.1016/j.istruc.2020.04.041.
[17] L.-H. Han, X.-L. Zhao, Z. Tao, Tests and mechanics model for concrete-filled SHS
that the concrete did not present crushing failure; stub columns, columns and beam-columns, Steel Compos. Struct. 1 (2001) 51–74,
• The proposed numerical model developed with Abaqus software can https://doi.org/10.1296/SCS2001.01.01.04.
properly simulate the behavior of the composite beams. [18] Eurocode 2. EN 1992-1-1, 2004 - Design of concrete structures - Part 1–1: General
rules and rules for buildings. Comite Europeen de Normalisation, Brussels,
Belgium, 1992.
Declaration of Competing Interest [19] A. Hillerborg, The theoretical basis of a method to determine the fracture energy GF
of concrete, Mater. Struct. 18 (1985) 291–296, https://doi.org/10.1007/
BF02472919.
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence