See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/321151678

UN MODELLO MECCANICO INNOVATIVO PER

CONNESSIONI COMPOSTE ACCIAIO-
CALCESTRUZZO DI PONTI SOGGETTI A...

Conference Paper November 2017

CITATIONS READS

0 4

4 authors, including:

Manuel Fassin Enrico Cazzador

Universit degli Studi di Trento Universit degli Studi di Trento
6 PUBLICATIONS 12 CITATIONS 11 PUBLICATIONS 5 CITATIONS

SEE PROFILE SEE PROFILE

Fabrizio Paolacci
Universit Degli Studi Roma Tre
120 PUBLICATIONS 470 CITATIONS

SEE PROFILE

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

ASSESSMENT OF THE SEISMIC VULNERABILITY OF AN OLD R.C VIADUCT WITH FRAME PIERS AND
STUDY OF THE EFFECTIVENESS OF DIFFERENT ISOLATION SYSTEMS THROUGH PSEUDODYNAMIC
TEST ON A LARGE SCALE MODEL (RETRO) View project

ASME PVP 2017 View project

All content following this page was uploaded by Fabrizio Paolacci on 18 November 2017.

The user has requested enhancement of the downloaded file.

XXV CONGRESSO C.T.A.

UN MODELLO MECCANICO INNOVATIVO PER

CONNESSIONI COMPOSTE ACCIAIO-CALCESTRUZZO DI
PONTI SOGGETTI A CARICHI SISMICI

AN INNOVATIVE MECHANICAL MODEL FOR STEEL-CON-

CRETE CONNECTIONS OF BRIDGES SUBJECTED TO EARTH-
QUAKE LOADINGS

Manuel Fassin, Fabrizio Paolacci,

Enrico Cazzador, University of Roma Tre
Oreste S. Bursi, Department of Engineering
University of Trento Roma, Italy
Department of Civil, Environmental [email protected]
and Mechanical Engineering
Trento, Italy
[email protected], enrico.cazza-
[email protected], [email protected]

ABSTRACT
Owing to advantages in the optimized use of both concrete and structural steel, composite structures
still remain the best structural solution for buildings and bridges. The composite solution, if com-
bined with Hot Rolled Beams (HRB) means also reduced construction time thanks to the cutback
of in site welding operation. The objective of SEQBRI project was to investigate the possibility to
export the HRB made by High performance fine grain S460M grade for bridges in high seismicity
zone.
This paper deals with an innovative mechanical model based on the components approach, devel-
oped to simulate the behaviour of a composite steel-concrete bridge connection subjected to trans-
versal seismic loadings. An OpenSEES FE non-linear model was implemented. The calibration of
the mechanical model was based on the results provided by the experimental campaign on half-
scaled subassembly specimens of a steel-concrete composite bridge subjected to a transversal load.
An FE stick model of a full-scale steel-concrete composite bridge case study was developed with
OpenSEES, and in order to evaluate the seismic response of the structural system an Incremental
Dynamic Analysis (IDA) was performed.
 SALERNO 1-3 OTTOBRE 2015

SOMMARIO
Grazie ai vantaggi derivanti dalluso di calcestruzzo ed acciaio strutturale, le strutture composte
risultano essere le migliori soluzioni strutturali per edifici e ponti. La struttura composta, se abbinata
allutilizzo di travi in acciaio laminate a caldo (HRB), riduce notevolmente i tempi di costruzione,
in quanto non necessita di saldature in opera. Il progetto SEQBRI si pone come obiettivo la promo-
zione e lutilizzo di travi HRB in acciaio S460M ad alta resistenza (HSS) in zone ad alta sismicit.
Questo articolo intende presentare un modello meccanico innovativo basato sullapproccio per
componenti, sviluppato per simulare il comportamento di una connessione di un ponte composto
acciaio-calcestruzzo soggetto a fenomeni sismici in direzione trasversale. Un modello non lineare
in OpenSEES stato sviluppato per questo scopo. La calibrazione del modello meccanico stata
fatta sulla base dei risultati sperimentali ottenuti su una sottostruttura in scala ridotta del ponte
composto. Al fine di poter valutare la risposta sismica del caso studio scelto tramite lanalisi dina-
mica incrementale (IDA), stato sviluppato in OpenSEES un modello agli EF del ponte intero.

1 INTRODUCTION
Nowadays, steel-concrete composite beams are successfully used for bridges and buildings, due to
benefits of combining the two construction materials. With reference to bridges, this kind of solu-
tion is very attractive because it offers a wide range of solutions to comply with environmental,
economic, construction, and maintenance requirements. Eurocodes [1], [2] and [3] do not provide
for specific information to design beam-to-beam joints in continuous composite bridges, neither for
gravity loads nor for seismic actions. Thus, an improved methodology to take into account the de-
sign and the construction issues are needed.
In order to fill this gap, both experimental campaign and numerical modelling on representative
substructures must be carried out. To define the level of damage at a certain Engineering Demand
Parameter (EDP) and to capture the local mechanism inside the detail, a mechanical model of the
joint that connects concrete slab I-girders steel beam composite section to the concrete cross-
beam was developed in OpenSEES. In particular, this model was built with non-linear springs
placed in an appropriate configuration that takes into account both the actual geometry of the detail
and the local mechanisms developed during a seismic event.

2 DESIGN AND EXPERIMENTAL CAMPAIGN RESULTS

On the basis of selected case studies, appropriate subassembly specimens were obtained from the
Case study 2.1.1 [4]. This case study is peculiar to the typical straight two-span overpass, with a
total length of 40 m and two spans were each 20 m long; the overpass has a carriageway 6.50 m
wide, harbouring two traffic lanes, and two sidewalks (2.05 m). In total, the width of the cross-
section reaches 10.60 m. The deck is supported, at the intermediate pier as well as at abutments, by
two normal damping rubber bearings (NDRBs). The concrete slab thick 0.25 m is supported by four
row of HEB 600 I-girders, in High Strength Steel (HSS) S460M grade, spaced of 2.65 m, and they
end at reinforced concrete cross-beams (CCBs). At abutments the width of CCBs are of 0.60 m,
whilst at the intermediate pier becomes 0.90 m wide. The pier has a 0.60x7.00 m section, with a
height equal to 7.00 m.
The design of the concrete slab-CCB interface for vertical loads was based on [1], [2], and [3]. Due
to the lack of information about the behaviour of the joint stressed in horizontal direction, three
different detail solutions were designed and experimentally investigated.
XXV CONGRESSO C.T.A.

2.1 Design of subassembly specimens

The aim of this section is to present the design and the investigation of different joint solutions to
connect the longitudinal steel I-girder with the transversal CCB.
As shown in Fig. 1, three solutions were proposed: the first, based on DIN FB 104 Var C. [5], is a
classic solution with a steel head thin plate welded at the end of a steel girder, where compression
forces are transferred to the concrete by the steel plate, and tensile forces are instead transmitted
through shear Nelson stud vertically positioned and welded to the thin steel plate. The second and
third solutions, named DOMI1 and DOMI2, respectively, are a modification of the classic connec-
tion solution, with the prolongation of the web inside the CCB. On this part, the Nelson studs for
the shear are welded perpendicularly to the web. The main difference between these details is the
mechanism for the transferring tensile forces in the case of inversion of the hogging with the sag-
ging bending moment, due to a severe earthquake: in DOMI1 detail, Nelson stud are welded to the
bottom steel thin plate, whilst in DOMI2 the stud are replaced with four prestressed bars. This last
solution is indicated to moderate-high seismic-prone areas, where the inversion of the bending mo-
ment at the intermediate support could be more frequent.
With reference to the setup of tests, two actuators of 1000 kN each provided the load needed to take
the specimen to collapse. Tests were conducted in displacement control with low displacement ve-
locity; in this way, the test became quasi-static, and no effect of forces of inertia were developed.
Two pin load cell hinges, placed at the ends of the specimen, achieved the zero-moment needed to
define the substructure of the specimen, as shown in Fig. 2. Several instruments placed in and on
the specimen, such as strain gauges, linear potentiometer transducers, and wire transductors, pro-
vided the measure of local strains or displacements of more significant areas.

(a) DIN FB 104 Var. C (b) DOMI1 (c) DOMI2

Fig. 1. Steel girder concrete cross-beam connection details

2.2 Experimental tests outcomes

In order to evaluate the response of different joint details proposed under seismic loads, an experi-
mental campaign on test subassembly specimens was carried out. Two type of tests of displacement
history were proposed on them: monotonic and cyclic tests in transversal direction. With reference
to monotonic tests, called SQ1M, SQ2M and SQ3M, the maximum value in term of force at the
failure and the associate displacement varied slightly among tests; in particular, forces vary between
900 and 915 kN, and the maximum displacement at failure is included between 36.8 and 38.3 mm.
The collapse mechanism developed inside the concrete slab at the steel girder CCB interface, as
pointed out by devices installed on the specimen. No detachment of steel head plates from the CCB
could be visually observed, although in some cases Nelson studs underwent considerable strain
states. Cyclic tests on specimens were crucial for the evaluation of the hysteretic behaviour of joints
under transversal loads, as shown in Fig. 3. The displacement history was based on [6]. The SQ1C
and SQ2C tests were characterized by the collapse of the specimen. The strain results, provided by
transducer devices, needed to define the stop of the SQ3C test, according to a limit state of damage
 SALERNO 1-3 OTTOBRE 2015

defined as [7]. This test was carried out with two intentions: i) to replicate the damage limit state at
a certain level of crushing deformation of the concrete, and ii) to understand whether the damaged
specimen was able to bear gravity load without significant losses in term of strength and stiffness.

Fig. 2. Test setup Fig. 3. SQ2C cyclic test Force- Fig. 4. SQ2C damaged
Displacement relationship specimen

3 THE INNOVATIVE MECHANICAL MODEL

A novel simplified 3D non-linear mechanical models of the DOMI2 solution (see Fig. 5) has been
implemented in OpenSEES environment, as illustrated in Fig. 6. The model has been developed
considering all the components acting on the CCB using nonlinear spring elements [8]. More spe-
cifically, non-linearities have been explicitly considered in the steel girders, concrete slab, pier, and
connections between the CCB and girder-to-slab. Steel girder and concrete slab have been modelled
using nonlinear fiber section beam elements. Relevant to elements outside the CCB, the mechanical
behaviour of steel girders and rebars have been modelled according to experimental characteriza-
tion tests. Thanks to numerous tests on rebar and girder steel samples, steel components have been
accurately characterized.
Due to space reasons the characterization results have not been reported, nevertheless these results
allowed to consider Steel girders flange and steel web separately inside the model by using the
Steel02 Giuffre-Menegotto-Pinto material type. Input parameters are the yield strength fy, the initial
elastic tangent E, the strain-hardening ratio b and parameters to control the transition from elastic
to plastic branches R0, cR1 and cR2.
The fiber-based elements and springs, i.e. TC1 and TC2 components, of the concrete slab have been
modelled according to the experimental response based on displacement transducer sensors for the
transversal FE model. The non-linearity has been considered within Concrete01 material type, ac-
cording to which the stress strain relationship is based on the Kent-Scott-Park concrete model,
with no tensile strength. Youngs modulus and tensile strength of C35/45 concrete were calculated
according to [1]. In detail, we assumed: the maximum concrete strength fpc= 51.9 MPa with a cor-
responding deformation of p=0.28%; the asymptotic concrete strength fpcu= 41.52 MPa with a rel-
evant deformation of u=0.67%.
The strength of the single shear stud of components TC3 and TC4 has been calculated according to
[2], while the related nonlinear behaviour has been evaluated by means of the shear load-slip curves
defined in [9]. The vertical head plates welded on the steel girders and directly in contact with the
transverse concrete beam have been modelled by rigid links. The links standing for shear headed
studs within the joint are connected to these rigid links according to the DOMI2 concrete cross-
beam configuration. In order to reproduce the constraint in compression due to the presence of the
concrete crossbeam, two elements have been modelled at both sides of the CCB joints. A simplified
evaluation of the gaps stiffness has been performed, assuming that the compressive force coming
from the girder would uniformly act on a surface equal to the area of the vertical head plate. The
XXV CONGRESSO C.T.A.

two levels of the pretension bars, component TC6 in the bottom area of the CCB, have been mod-
elled by two elastic truss elements. The shear contribute of the inner plate has been considered with
the component TC5, whilst the shear in z-direction of the concrete slab TC2 has been implemented
as proposed in [10] and [11]. The main component for the transversal loading is the component
TC1, for this reason its description is reported in a dedicated chapter.

Fig. 5. View of the DOMI2 solution; Fig. 6. View of the 3D mechanical model

Table 1. Main components of the mechanical 3D model

Main Components

TC1 - Concrete and re-bars axial response in the TC4 - Shear on the studs of the inner plate (y direction)
slab (x direction)

TC2 - Shear of the concrete slab (z direction) TC5 - Shear on the inner plate (z direction)

TC3 - Shear on the studs of the inner plate (x direc- TC6 - Prestressing bars (x direction)
tion)

Table 2. Hypothesis considered to fit the concrete slab response

Hypothesis Y
X
The tensile strength of concrete has been halved and
I is equal to fctm = 0.5fctm* = 2.09 MPa
Z

The yield strength of 8 and 10 re-bars obtained GEFRAN 1 GEFRAN 3

from tensile tests were reduced by a factor of 0.9. SP 1 (right) SP 1 (left)
II Nonetheless, values are higher than those required
SP 2 (right) SP 2 (left)
in [13].
The ultimate steel strength was reduced to 50% of
SP 3 (right) SP 3 (left)
the original value. Instead of a hardening branch, a
softening one was imposed. Deterioration processes SP 4 (right) SP 4 (left)
that occurred inside the slab before failure (e.g. slip SP 5 (right) SP 5 (left)
between concrete and re-bars) were not measured GEFRAN 5 GEFRAN 7
III with dedicated instruments. To take this type of phe-
nomena into account, this assumption is made,
CCB outline
which exclusively affects the model response for
high transversal displacement values that might not
be achieved in case of realistic seismic events. Fig. 7. View of the slab springs
 SALERNO 1-3 OTTOBRE 2015

3.1 TC1 component

With reference to the top and bottom layer springs representing re-bars and concrete, a Hysteretic
material has been defined to take into account the behaviour of concrete in compression and of steel
in tension. As shown in Fig. 7, each spring has been implemented considering equivalent properties,
i.e. the single spring includes the properties of a part of concrete slab with relevant re-bars, in this
case each spring represents 1/5 of the slab depth. Moreover, the tension-stiffening effect is imple-
mented by following the specifications of [12] and the force displacement relationship considered
as reference for this component is based on the Kent-Park Model. The springs response in com-
pression has been evaluated in order to best fit the local response of the specimen provided by the
experimental outcomes of LVDT displacement transducers. The aim was to accurately reproduce
experimental tests results of the part of the specimen which collapsed, therefore additional hypoth-
esis have been considered as shown in Table 2.

3.2 Validation of Mechanical model

To validate the mechanical model, a comparison between SQ3M test data and finite element re-
sponse in terms of curvature was made. The error between finite element analysis output and ex-
perimental results has been evaluated from an energetic point of view in order to investigate it on a
local scale. The most sophisticated component is represented by the concrete slab CCB interface
and a related concept of curvature is introduced to validate the FE model, as shown in equation (1).
The reference length for displacement measurement of Gefran devices is Lref and is set to 200 mm,
which corresponded to the instruments initial length. L is the distance between Gefran in the com-
pressed slab region and the devices located where the concrete is under tension and it is equal to
1225 mm. To compare test and model results, the curvature was calculated for the slab springs of
the mechanical model (with reference length Lref). The curvature, equation (1), has been related to
the bending moment at the concrete slab - CCB connection so that test and model responses could
be compared, adopting the energetic approach defined in equation (2), in which E stands for energy.
( ) ( )
+ +
(1) = (2)
= =

Numerical analysis results are displayed as dashed lines in Fig. 8. The symmetric response of the
mechanical model, for moment curvature plots are overlapped, is of particular interest.
As far as the left interface is concerned, the error in terms of the previously defined energy is 1%,
while for the right interface, it rises to almost 8%. Fig. 9 shows a comparison in terms of spring
model elongations with respect to the average Gefran elongation (represented as a dashed line) over
the entire concrete slab width, for the maximum transversal force. As reported by experimental
results, elements of the DOMI2 connection component suffered from negligible stresses and defor-
mations, hence these quantities are not further investigated herein.

4 PROBABILISTIC SEISMIC DEMAND EVALUATION OF THE CASE

STUDY
A 3D non-linear stick model of the overall bridge has been implemented in OpenSEES environment
by considering the local mechanical model of the CCB validated on the basis of experimental
results. A group of 14 accelerograms was been selected from the PEER ground motion database
and scaled so that the mean spectrum best approximated the target spectrum, according to the
prescriptions of [3]. The scaling factor modifies the original record only in amplitude. Incremental
dynamic analyses (IDAs) [14] have been performed. On the basis of the following EDPs reported
hereinafter, nonlinear time history simulations and multi-record IDA curves have been built for: i)
XXV CONGRESSO C.T.A.

global: Pier drift; ii) intermediate: Bending moment-curvature, Bending moment-rotation of the
CCB, Rotation of one side of the joint; and iii) local: Maximum force and relative displacements
of shear studs, Maximum deformation of the concrete slab.

Fig. 8. Comparison between numerical and exper- Fig. 9. Comparison between numerical and
imental curvature experimental section elongation

4.1 Results of numerical simulations

The selected scan of PGA is in the range (0.1 g - 2.0 g). The multi-record IDA response of 20 and
16 steel re-bars are shown in Fig. 10. In line with the experimental tests, the transversal earthquake
loading is not significant for the case study investigated. Only in case of very strong earthquake,
i.e. 1.0 g PGA, the limit state of concrete cracking is exceeded in terms of mean maximum response.

a) IDA results for 20 re-bars b) P[EDP>EDPlimit|IM] 20 re-bars

of concrete slab (tension) of concrete slab (tension)
Fig. 10. IDA-Drift ratio of the pier

5 CONCLUSIONS
The calibration of an innovative mechanical model of a steel-concrete composite section of a bridge,
based on the experimental campaign results from the SEQBRI project has been presented. Subas-
sembly specimens of different joint details were designed according to Eurocodes. In order to un-
derstand the behaviour of the connection under seismic events, several monotonic and cyclic trans-
versal tests were carried out in the laboratory. All the tests clearly showed that the collapse of spec-
imens was driven by the concrete slab. Therefore the different detail connections did not influence
the response of the bridge subassembly. The comparison between the numerical results and the
experimental tests demonstrates that the local mechanical model replicates the actual behaviour of
 SALERNO 1-3 OTTOBRE 2015

the connection, with an acceptable error of less than 8% in terms of energy. The proposed numerical
approach is completely general and can be applied in modular fashion to different case studies with
DOMI2 detail solution, i.e. longer and/or multiple spans bridge. In addition, a probabilistic demand
analysis considering the collapse mechanisms and structural damage effects has been performed by
the implementation of the local mechanical model in the complete bridge. The numerical results
have shown a favourable seismic response of the proposed details, even in case of for severe earth-
quakes. The applications of the PBEE approach in terms of losses deserves further studies.

ACKNOWLEDGMENT
This work was carried out with a financial grant from the Research Fund for Coal and Steel of the
European Community, within the SEQBRI project: Performance-Based Earthquake Engineering
Analysis of Short-Medium Span Steel-Concrete Composite Bridges, G.A. RFSR-CT-2012-00032.

REFERENCES
[1] CEN Eurocode 2: Design of concrete structures. Part 1-1: General rules and rules for buildings.
Bruxells: CEN 2011
[2] CEN Eurocode 4: Design of composite steel and concrete structures. Part 1-1: General rules
and rules for buildings, and Part 2: General rules and rules for bridges. Bruxelles: CEN; 2012
[3] CEN. Eurocode 8: Design of structures for earthquake resistance. Part 1: General rules, seismic
actions and rules for buildings, and Part 2: Bridges. Bruxelles: CEN; 2012
[4] D1.1 Definition of representative case studies. Selection of I-girder subassemblies to be
tested, SEQBRI project, 2013
[5] DIN FB 104, Composite steel and concrete bridges, Deutsches Institut fr Normung e.V., Ger-
many, Fachbericht, 2009
[6] ECCS, Recommended Testing Procedure for Assessing the Behaviour of Structural Elements
under Cyclic Loads, Technical Committee 1 Structural Safety and Loading, 1986
[7] Mackie, K.E., Wong, J.-M., and Stojadinovic, B., Integrated Probabilistic Performance-Based
Evaluation of Benchmark Reinforced Concrete Bridges, PEER 2007/09, January 2008
[8] Rassati, G.A., Leon, R.T., and No, S., Component Modelling of Partially Restrained Compo-
site Joints under Cyclic and Dynamic Loading, Journal of Structural Engineering, 2004.
130:343-351
[9] Gattesco, N., and Giuriani, E., Experimental study on shear stud connectors subjected to cyclic
loadings, Journal of Constructional Steel Research, Vol.38, pg. 1-22, 1996
[10] Kelly, T., Nonlinear Analysis of Reinforced concrete Shear Wall Structures, Bulletin of
the New Zealand Society for Earthquake Engineering, Vol.37, No.4, December, 2004
[11] Bentz, E.C., Vecchio, F.J., and Collins, M.P., Simplified Modified Compression Field
Theory for Calculating Shear Strength of Reinforced Concrete Elements, ACI Structural Jour-
nal, Vol.103, December, 2007
[12] CEB-FIP Model Code 90, Fib-International, 1993
[13] NTC08 - Decreto Ministeriale 14.01.2008 Norme tecniche per le costruzioni, 2008 (in
Italian)
[14] Vamvatsikos, D., and Cornell, C.A., Incremental Dynamic Analysis, Earthquake Engr.
& Struct. Dynamics, J. Wiley & Sons, v. 31, pp. 291-514, 2001

KEYWORDS
Steel-Concrete Composite Bridge, Concrete Cross Beam, Hot Rolled Beam, PBEE, OpenSEES,
Mechanical Model, Incremental Dynamic Analysis, Nonlinear Time History Simulation.

View publication stats