Journal of Constructional Steel Research 63 (2007) 819–832

www.elsevier.com/locate/jcsr

Inelastic response of composite steel and concrete base column connections

L. Di Sarno a,∗ , M.R. Pecce a , G. Fabbrocino b
a Department of Engineering, University of Sannio 82100, Benevento, Italy
b Department S.A.V.A., University of Molise 86100, Campobasso, Italy

Received 24 January 2006; accepted 8 August 2006

Abstract

This paper assesses the experimental results of monotonic (pushover) tests carried out on partially encased composite steel–concrete columns
connected to the foundation block through the traditional bolted steel end plate and an innovative socket type system. These tests show that the
structural response of the traditional connection is significantly influenced by the behaviour of the anchorage bolts. The latter cause large fixed end
rotations and possess limited energy dissipation. Conversely, innovative composite steel–concrete base column connections with socket systems
exhibit adequate overstrength, inelastic deformations and energy absorption capacity. Furthermore, socket-type connections are characterized by
spreading of inelasticity at the base of the composite columns without damage localization on concrete and interface components. It can thus be
argued that the innovative connection type assessed in this study is a viable solution for applications in framed structures fulfilling capacity design
requirements, e.g. structural systems in earthquake prone regions.
c 2006 Elsevier Ltd. All rights reserved.

Keywords: Composite columns; Partially encased columns; Base column joint; Socket-type connection; Overstrength; Ductility; Energy dissipation

1. Introduction
Composite steel and concrete structural systems possess
adequate seismic performance because of their stiffness,
strength and ductility, and have been found to be cost-
effective especially for multi-storey buildings under horizontal
loads ([21,1,18] among many others). Composite systems
for buildings often include steel moment resisting frames,
consisting of steel beams (acting compositely with a metal
deck reinforced slab or solid concrete slab through shear
studs) and encased composite columns, or braced frame
with steel–concrete composite columns [6]. Consequently,
lateral drifts, both interstorey and roof drifts, under horizontal
forces (wind and/or earthquakes) are lowered. Under severe
earthquake loading concrete encasement cracks and reduces
the flexural stiffness of composite beam–columns but the steel
core acts as a back-up system in providing the shear strength
and the ductility to prevent brittle failure. Partial encased
beam–columns with local buckling inhibitors have been found
∗ Corresponding address: Department of Engineering, University of Sannio,
Piazza Roma, 21, Benevento, Italy. Tel.: +39 0824 305566; fax: +39 0824
325246.
E-mail address: disarno@unina.it (L. Di Sarno).
very efficient to prevent local buckling and enhance the global
stability of the frame, thus reducing sensitivity to P–∆ effects
[12,11]. The assessment of structural response of composite
steel and concrete columns is thus of paramount importance,
especially in the earthquake design of framed systems ([13,
25], among many others). The inelastic response of composite
columns is significantly affected by the beam–column, brace-
to-beam, brace-to-column connections and column bases. A
comprehensive review of the experimental tests carried out
on steel–concrete composite beam–columns (both encased
and concrete-filled) can be found, for instance, in [5,23].
It is noteworthy that, to date, analytical and experimental
research focusing on the effects of the base connection
layout on the performance of beam–columns, either partially
or fully encased, is lacking [24]. Few results are available
and were derived chiefly from the steel structures; their
applicability within the capacity-design framework should
be further investigated [17,19]. The composite action, may,
however, affect the failure modes thus endangering the inelastic
performance of the structural member (e.g. [7]).
This paper analyzes the inelastic response of composite
steel–concrete joints at column base. An innovative base
column connection, employing a socket-type system, is
discussed and its response is compared to that of a traditional

c 2006 Elsevier Ltd. All rights reserved.

0143-974X/$ - see front matter

doi:10.1016/j.jcsr.2006.08.007
820 L. Di Sarno et al. / Journal of Constructional Steel Research 63 (2007) 819–832
steel base plate connection. The latter was designed in
compliance with the rules utilized for the composite frame
tested at JRC Ispra laboratory [3,14]. Several tests under either
monotonic or cyclic lateral loads and different levels of axial
loads were performed. This work focuses on the response
of composite columns under monotonic regime, i.e. pushover
tests. The results of the experiments carried out on specimens
with welded base steel plate (traditional) and socket-type joints
are discussed. It is found that for the traditional connections,
concentrations of inelastic demand occur in the anchorage
bolts and relies chiefly on bond type mechanisms. This type
of structural response is assessed with a simplified fibre-
based numerical model. The latter allows the evaluation of the
different contributions of the lateral deformation at the top of
the column specimens. Such model points out the large inelastic
demand imposed on the anchorage bolts of the steel end
plate in traditional base column connections. Conversely, the
socket-type system, derived from precast reinforced concrete
structures, leads to large energy dissipation. Plastic hinges form
at column base and the strength capacity is not lowered even at
large lateral drifts (greater than 5%–6%). As a consequence,
socket-type foundations may be reliably utilized for steel
and composite steel–concrete framed structures, especially in
regions with moderate-to-high seismicity, to achieve adequate
seismic structural performance.
2. Experimental program and test set-up
Several research programmes were recently launched in
Europe, the US and Japan to investigate the inelastic response
of bare steel and composite steel and concrete composite
structures and/or their subassemblages, both for new and
existing buildings (e.g. [22,7], among others). In Italy, a number
of experimental programmes were funded through national
and European grants to provide technical support for the
implementation and improvement of seismic provisions and
guidelines (e.g. [4,20,11]). These test programmes included
the assessment of the seismic response of composite slabs,
beam–columns and connections, either beam-to-column or
base columns. In particular, the authors of the present work
investigated the inelastic static and dynamic (seismic) response
of base column connections. In so doing, a number of partial
encased column specimens, with different base joints, were
tested in the laboratory of the Department of Structural
Analysis & Design (DAPS, University of Naples, Italy).
The sample specimens included monotonic and cyclic tests;
different levels of axial loads were considered during the
tests to simulate the seismic load combinations implemented
in European standards [11] for composite building structures.
These axial loads were computed with regard to the exterior and
interior composite steel–concrete columns frame tested at JRC
Ispra laboratory [3,14], as shown in Fig. 1. In addition, pull-
out tests were carried out to define the force–slip relationships
of the hooked anchorage bolts; the results of the pull-out
tests were utilized to investigate the deformation capacity of
the base column traditional joint. In this work the results of
the monotonic (pushover) tests are discussed in details. The
experiments were carried out on two types of partially encased
composite columns: HEB260 and HEB280 (see Fig. 1). The
test specimens employed two layouts for the base column
joints as per Fig. 2: traditional (bolted steel base plate) and
innovative (socket-type) joints. The former consist of tapered
steel plates welded onto base plates and anchored to the
foundation block through steel bolted bars (see also [14]). The
latter is an alternative and innovative socket type joint in which
the column is fixed to the foundation block utilizing a special
concrete filler; such joint was developed and designed to benefit
of composite action. Socket type connections are generally
utilized in precast reinforced concrete systems for residential,
office and industrial framed structures.
The first set of specimens (traditional base column
connections) correspond to the columns and base joints
designed in compliance with the guidelines of European
standards [9–11] and used for the full-scale composite frame
tested in ISPRA [3]; the layout of the latter is pictorially
displayed in Fig. 1. The socket-type foundation was designed
in compliance with Eurocode 2 [8] using strut-and-ties design
approach.
In the experimental tests, the lateral loads were applied
at two different locations along the height of the column,
namely at 1.6 m (traditional joint) and 1.7 m (socket type)
above the foundation block to account for the different location
of the restraint. Traditional base plates are generally placed
at floor level, conversely socket type solution enables to use
pavings that cover the foundation block. The horizontal load T ,
simulating the earthquake loading, was applied by means of a
500 kN-hydraulic jack; the test was under displacement control.
As a consequence, the maximum flexural moments M, located
at the base connection, was increased until failure occurred. The
displacement controlled loading regime allowed the softening
branch of the response (capacity) curve to be investigated. The
connection of the jack to the column is ensured by two steel
plates 30 mm thick bolted on two opposite faces of the column.
The reaction wall for the horizontal load is a stiff tapered
cantilever bolted to a steel system; its layout is shown in Fig. 3
along with the test set-up. The cantilever system is connected to
the laboratory floor slab (strong floor) by means of large steel
rebars crossing the slab and the steel elements. These rebars are
loaded in tension to prestress the connection. The vertical load
N is applied by two hydraulic jacks connected with two bars
at the hinges placed at the foundation level. The axial load N
acts along the column centroid axis. The reaction system for
N consists of a steel plate located under the foundation block
and connected to the hinges. This layout ensures that the load
remains along the member axis during the column deformation.
The transversal beam, at the column top, is connected to the
jack by means of large stiffeners. An adequate lateral restraint
along four prestressing bars at the corners were used to prevent
slip and rocking of the foundation block and to guarantee the
transfer of the shear forces to the strong floor level.
The investigation of the interface behaviour between anchor-
age bolts and concrete block was carried out by means of pull-
out tests; the set-up employed to perform the tests is provided
 L. Di Sarno et al. / Journal of Constructional Steel Research 63 (2007) 819–832 821

Fig. 1. Geometry of the full-scale composite frame tested at Ispra.

in Fig. 4. It consists of a rigid steel frame, designed to prevent 3. Test specimens

the onset of local buckling. The jack is bolted at the horizon-
tal beam of the frame and the anchoring device is gripped with
a tension bar. The anchoring devices were instrumented with a
transducer and strain gauges. It is instructive to note that the use
of strain gauges may, however, modify the mechanism associ-
ated to bond and influence significantly the results due to local
stresses acting on the device.
The experimental programme carried out at the laboratory
of DAPS, in Naples, focused on two sample partially encased
composite steel and concrete columns: HEB260 and HEB280
(see also Fig. 1). The sample specimens tested were cantilever
systems summarised as below (Table 1): partially encased
columns with a steel HEB 260 member and traditional
822 L. Di Sarno et al. / Journal of Constructional Steel Research 63 (2007) 819–832

Fig. 2. Layout of the sample base column connection: traditional (left) and socket-type (right).

Table 1 Table 2
Sample column specimens Mechanical properties of the sample columns

Specimen Axial load (kN) Loading type Connection Property HEB260 HEB280
Web Flange Web Flange
HEB260 330 Monotonic Traditional
HEB260 170 Monotonic Traditional f y (MPa) 406 341 341 300
HEB260 330 Monotonic Socket f u (MPa) 480 449 450 430
HEB280 520 Monotonic Socket fu / f y 1.18 1.32 1.32 1.43
εu (%) 31.8 35.7 34.5 37.1

connection to foundation (stiffening plates and anchoring

devices) and innovative socket-type system; HEB280 with
socket-type foundation. The specimens were instrumented in
a refined fashion to characterize reliably the concentration of
inelastic demand at the base column. Fig. 5 provides a close-up
view of the displacement transducers (LVDTs) located at the
base of the partially encased columns with either traditional or
innovative joints.
The values of axial load N used in the tests were equal to
170 kN, 330 kN and 520 kN. The latter values correspond to
the axial loads of the design load combinations of the full-
scale composite framed building tested in the ELSA laboratory
of JRC in Ispra [3]. The grade of the structural steel of the
specimens was S235; the reinforcing bars grade was B450-
C and the concrete was class C25/30. Actual values of the
mechanical properties of steel and concrete were estimated
from tensile (for steel) and compression (for concrete) tests.
The values computed for the column members are summarized
in Table 2. The width-to-thickness ratios of the sample columns
(b/t f = 20.8 for HEB260 and b/t f = 21.5 for HEB280)
fulfil the limit (b/t f = 44) given in European standards [10]
to prevent local buckling in partially-encased I-sections made
of steel grade S253. Further details on the results of such tests
can be found in [3]. The values of material overstrength ( f u / f y )
relative to the element flange are on average higher than those in
the webs. This result is in agreement with similar experimental
data for steel members produced in Europe (e.g. [2]).
The details of the foundation system are shown in Fig. 2. The
design of such foundation block was based on the ultimate limit
state corresponding to the stress values and distribution at the
column failure. The layout of the composite column specimen
employing HEB 260 steel section and traditional base joint is
provided in Fig. 6, while the specimen HEB260 with socket
type connection is provided in Fig. 7. The thicknesses of the
base of the socket is equal to 300 mm, while the walls of the
socket are 250 mm thick. The total height of the foundation
is 1050 mm. The values of the internal actions, i.e. axial load
N Sd, j , flexural moment M Sd, j and shear VSd, j used for the
design are N Sd, j = 308 kN, M Sd, j = 906 kN m and VSd, j =
444 kN. These values were derived from the ultimate flexural
capacity of the HEB 280; the evaluation of the resistance was
based on conservative assumptions. In fact, for the above cross-
section, the ultimate bending moment is M Rd,col = 755 kN m;
note that the design value M Sd, j accounts for the overstrength
( f u / f y = 1.20) at the base column connection.
For each foundation block, two anchoring devices were
embedded in addition to the ones used for the base column joint
details (Fig. 8). These embedded elements were instrumented
with transducers and were utilized for pull-out tests. Thus,
force–slip relationships were estimated for each anchorage
bolts; these relationships were then utilized to compute the
moment–curvature relationship of the base column section,
as discussed in Section 5. The results of the experimental
tests carried out on composite columns employing innovative
(socket-type) are discussed below.
4. Test results
The test results of the partially encased composite columns
were computed in terms of both local (moment–curvature
M– χ and moment–rotation M–θ) and global (lateral load–top
displacement, F–∆) response parameters. The capacity curves
of the specimens HEB260 with axial load N = 330 kN, both
traditional and innovative base connections, are provided in
Fig. 9. Such curves are expressed in terms of lateral force
(F)-drift (d/H ), where H is the distance of the centreline
of the hydraulic jack from the foundation and d the total
 L. Di Sarno et al. / Journal of Constructional Steel Research 63 (2007) 819–832 823

Fig. 3. Layout of the test set-up (top) and reaction wall (bottom).

horizontal displacement at the jack height. It is observed that
the traditional connection layout exhibits higher lateral strength
(600 kN m vs. 510 kN m) due to the steel stiffeners used
at the base of the column and to overstrength for seismic
design. Conversely, the ultimate deformation capacity of the
socket type connection is about 75% higher than the counterpart
traditional (about 0.05 rad vs. 0.09 rad); in both cases the
threshold value of 35 m rad given by Eurocode 8 [11] is
exceeded. The latter limit value is more stringent than the 3%
drift, which is assumed as the onset of the ultimate limit state in
steel and composite frames in the US practice [16]. Furthermore
the failure mode of the specimen with steel end plate is related
to anchorage bolt fracture (Fig. 10), while in the case of the
socket type a very ductile mechanism is observed (Fig. 11).
At serviceability, the stiffness of the traditional connection
is slightly higher than that of the socket connection. It can
824 L. Di Sarno et al. / Journal of Constructional Steel Research 63 (2007) 819–832

Fig. 4. Set-up of pull-out tests: layout (left) and actual system (right).

Fig. 5. Close-up view of the electrical displacement transducers (LVDTs): traditional (left) and innovative (right) joint.

thus be argued that the experimental tests carried out both observed for traditional connections. The composite partially
on traditional bolted steel end plate and innovative socket- encased columns with traditional joint yield at about 310 kN,
type connections demonstrate that the former experience brittle which corresponds to a lateral drift of 26 mm (d/ h ∼ 1.65%).
failure modes. Rupture of anchorage bolts as per Fig. 10 was The maximum force is equal to 375 kN for HEB260 with axial
 L. Di Sarno et al. / Journal of Constructional Steel Research 63 (2007) 819–832 825

Fig. 6. Traditional base column joint.

loads N = 330 kN and 340 kN for N = 170 kN. The lower
value found in the second specimen (340 kN) is related to
the premature rupture of the base joint, probably caused by
technological defects of the threaded bars. In both specimens,
i.e. with N = 170 kN and N = 330 kN, under monotonic
regime, the column strength and energy dissipation do not
exhibit significant loss for drift d/ h ∼ 5%–6%. The thick steel
plate and the stiffeners used at the column base ensure that
the end section of the column remains plane (rigid rotation).
Additional tests carried out by the authors under cyclic loads
have demonstrated that the crushed concrete and the inelastic
deformations in the steel components (anchorages), both at
826 L. Di Sarno et al. / Journal of Constructional Steel Research 63 (2007) 819–832

Fig. 7. Socket-type base column joint.

the column base, endanger the global lateral stiffness of
the composite column [15]. Bond-related phenomena give
rise to degrading effects, especially at large drifts, thus
reducing significantly the energy dissipation capacity of the
member. These findings point out that traditional connections
are not fully satisfactory, especially when the contribution
of reinforced concrete component increases, i.e. due to
cross section dimensions. Conversely, innovative socket-type
connections possess adequate ductility. Under monotonic load
conditions, the test results show strain hardening of the base
column equal to 1.32. This is due chiefly to the material
overstrength of structural steel (compare, for instance, values of
 L. Di Sarno et al. / Journal of Constructional Steel Research 63 (2007) 819–832 827

Fig. 8. Steel reinforcement used for the traditional base column joint and layout of the additional anchorage bolts for pull-out tests.

Table 3
Material mechanical properties used to computed the interaction curves

Property Design Yield Collapse

f cd (MPa) 14.2 21.3 38.0
f sd (MPa) 391.3 450.0 450.0
f ad (MPa) 213.6 235.0 450.0
γc 1.50 1.00 1.00
γs 1.15 1.00 1.00
γa 1.10 1.00 1.00
Key: f cd = concrete compression strength at yield ( f c); f sd = compres-
sion/tensile strength of reinforcement steel (bars); f ad = compression/tensile
strength of reinforcement steel (profiles); γc = partial safety factor for con-
crete; γs = partial safety factor for steel (bars); γa = partial safety factor for
steel (profiles).
Fig. 9. Capacity curves for the traditional and socket-type connections for
HEB260 (N = 330 kN).
member-to-foundation joint, the bending–axial load M–N
f u / f y of the member flange in Table 2). The contribution of the interaction curves (Fig. 12) were computed for the specimens
hardening of the longitudinal reinforcement bars is in fact very HEB 260 and HEB 280. These M–N curves were derived
small (1.13 vs. 1.32). The tests carried out on the specimens in compliance with the simplified method of design for
employing the socket joint do not exhibit strength deterioration doubly symmetric and uniform cross-section over the member
even at large lateral drifts, e.g. d/ h > 0.04–0.05 rad. The length with rolled steel sections implemented in the European
formation of the plastic hinge occurs at the base column, as standards for composite steel and concrete structures [10]. The
observed during the tests. Fig. 11 shows the occurrence of resistance of the cross-section to combined compression and
inelastic deformations at the base column during the test on the bending is calculated assuming stress blocks (plastic stress
HEB 260 with N = 330 kN; the spreading of such inelasticity is distribution); the tensile strength of the concrete is neglected
also evident. At very large drifts, the flange plate of the column and the interaction curve is replaced by a polygonal diagram
tends to bend outwards and the bond between the inner concrete (the dashed line in Fig. 12). The material mechanical properties
and the exterior steel plate is broken as demonstrated by the used to derive the simplified interaction curves are provided in
close-up view of the Fig. 11, without any relevant effect on Table 3. The values of the simplified M– N interaction curves
the global response. The tensile resistance of the concrete is for the test with HEB260 (N = 170 kN and N = 330 kN) and
exceeded and inclined (flexural) cracks initiate and propagate HEB280 (N = 330 kN and N = 520 kN) estimated at yield
above the foundation block. and collapse are summarized in Tables 4 and 5, respectively.
To shed light on the structural performance of the sample The computed values of the M–N curves demonstrate that
columns with either traditional or innovative socket-type the failure mechanism is controlled in the case of traditional
828 L. Di Sarno et al. / Journal of Constructional Steel Research 63 (2007) 819–832

Fig. 10. Close-up view of the deformations of the specimen HEB260 with N = 330 kN (traditional connection).

Table 4
Values of interaction curves for the test with HEB260 (N = 170 kN and N = 330 kN) and traditional joint at different performance state

Yield Collapse Test

M (kN m) N (kN) M (kN m) N (kN) M (kN m) N (kN)
Section above stiffening plate 343 330 633 330 488 330
339 170 628 170 433 170
Column base connection 440 330 550 330 596 330
395 170 510 170 530 170

Table 5 electrical displacement transducers located at the column base

Values of interaction curves for the test with HEB280 (N = 330 kN and as displayed in Fig. 5. These LVDTs record the lateral
N = 520 kN) and socket-type joint at different performance state (at column displacement of several points of the column flanges in
base connection)
order to define reliably the various inelastic phenomena,
Yield Collapse Test e.g. yielding, local buckling, fracture initiation, occurring
M (kN m) N (kN) M (kN m) N (kN) M (kN m) N (kN)
at those locations during the tests. Fig. 13 provides force-
390 330 740 330 608 330 rotation curves computed at the base column of the socket-
420 520 750 520 762 520
type connection under monitonic loads. All monitored sections
shows significant nonlinear response as the deformations
increase; this behaviour characterizes also the cyclic tests [15].
joint by the base column component. By contrast, for socket- Comparison between total cord rotation and the base column
type connections, the failure mode is due to the formation of one derived from LVDT #6 measures enable to recognise that
the plastic hinges in the columns (see also Fig. 11, where the the higher is the drift, the higher is the contribution to the total
spreading of inelasticity due to flexure is evident). drift given by the deformation of the socket type connection.
The inelastic response of the critical region of the specimens Furthermore, due to the location of LVDT #1, 45.0 cm far
with socket-type joints was monitored carefully through six from the foundation, the socket type shows a yielding spreading
 L. Di Sarno et al. / Journal of Constructional Steel Research 63 (2007) 819–832 829

Fig. 11. Close-up view of the deformations of the specimen HEB260 with N = 330 kN (socket-type connection).

Fig. 12. Simplified interaction curve for combined axial load–uniaxial bending
moment (N –M).

that is nearly twice the width of the section (45.0 cm vs.
52.0 cm). As far as seismic design is concerned, the longer
the spreading of inelasticity the higher the energy dissipation.
By contrast, traditional connection systems, employing bolted
steel end-plate generate high concentrated inelastic demand on
the anchorage bolts, which fail prematurely in a brittle manner.
Fig. 13. Force–rotation of the base column of the socket-type connection:
LVDTs #1 to LVDT #6.
5. Numerical simulations
Numerical simulations of the pushover tests were carried out
through a fibre-based model, which may accommodate both
830 L. Di Sarno et al. / Journal of Constructional Steel Research 63 (2007) 819–832

Fig. 14. Evaluation of fixed end rotation: system response (top-left), cross-section response (top-right) and computational model (bottom).

Fig. 15. Moment–curvature of the columns and the base joints: HEB260 with N = 330 kN (left) and N = 170 kN (right).

mechanical (material nonlinearity and residual stresses) and
geometrical (P–∆ effects) nonlinearities. The effects of the
fixed end rotations are also accounted for; the computational
model is pictorially shown in Fig. 14. The constitutive
relationships employed in the numerical simulations were
based on experimental tests carried out on structural steel
and reinforcement bars (see Table 2); concrete stress strain
including size effect was used to simulate behaviour of
concrete. The collapse of the sample columns was caused
by base column joint failure in the case of the traditional
connection as shown in Fig. 10. The ultimate resistance of
the columns is thus controlled by this failure mechanism.
Consequently, the ultimate bending moment of the base
joint was computed for purpose of comparisons. The
moment–curvature relationship of the base column was derived
through the constitutive relationship of anchorage bolts and the
Mander’s model for unconfined concrete for the high strength
material which supports the rigid steel end-plate. The peak
of resistance for concrete ( f c = 70 MPa) is assumed at
εc = 0.002, which corresponds to the upper bound of the
compressive strength estimated via experimental tests. The
actual yield and ultimate strengths of the bolts are equal to
400 MPa and 570 MPa, respectively; the strain hardening
starts at εsh = 0.01. Fig. 15 provides the moment–curvature
curves of the columns and the base joints for the sample
specimens HEB260 with both N = 170 kN and N = 330 kN.
 L. Di Sarno et al. / Journal of Constructional Steel Research 63 (2007) 819–832 831

Fig. 16. Pull-out tests for anchorage bolts used for HEB260 with N = 330 kN (left) and N = 170 kN (right).

Reference lines relative to the onset of the yield in the base

joint and the column are also plotted in Fig. 15. For both
specimens HEB 260, the former component yields at a value of
about 300 kN m, while the column elastic threshold is higher
(452 kN m and 475 kN m for HEB260 with N = 330 kN
and N = 170 kN, respectively). The total deformation of the
columns were computed by integrating the moment–curvature
along the member. This relationship was derived by assuming
that member cross-sections remain plane and slip between
steel and concrete does not take place (ideal bond); the
computational model employed for the numerical simulation
is outlined in Fig. 14. The fixed end rotation was estimated
through the moment–curvature of the base joint (position of
neutral axis) and a fitting curve of experimental bond-slip
constitutive relationships provided in Fig. 16 (slip effects).
Comparisons between numerical simulations and experi-
mental test are provided in Fig. 17 for the sample column HEB
260 with N = 330 kN; results for traditional and socket-type
joints are displayed. For the former, it may be observed that the
value of the yield of the base joint is close to the elastic thresh-
old measured during the experiment, thus demonstrating that
the failure of the base column controls the global response of
the composite steel and concrete column. This behaviour was
Fig. 17. Comparisons between numerical simulations and experimental test for
also experienced by the sample HEB 260 with N = 170 kN. HEB260 with N = 330 kN for traditional (H = 1310 mm) (top) and socket
Force–deformation curves of the tested columns are also shown type (H = 1700 mm) (bottom) joints.
in the same figure. The total lateral deformations and the con-
tributions caused by the fixed-end rotation and the column flex- moment computed through the numerical simulation of the can-
ibility are included. The deformation of the column was esti- tilever column with height of 1700 mm is smaller than the value
mated with respect to the length of 1310 mm, i.e. the distance estimated experimentally; the difference between these bending
between the column tip and the steel stiffening end plate at moments is about 25%. Similarly, Fig. 17 shows that the yield
the base. The comparison between experimental and numerical of the sample column evaluated numerically is higher than that
simulation of the socket-type joint provided in Fig. 17 demon- measured experimentally. The latter may be caused by the inter-
strates that also for this type of connection layout the fixed action between the composite column and the surrounding con-
end rotation can be significant. However, the contribution of crete. It is observed that for the socket type connection the fixity
the end rotation to the overall lateral displacement of the spec- of the column is not located at the interface between the end-
imen with socket-type joint tends to be lower than in the case plate and the foundation block, but it penetrates in the socket,
of the traditional bolted base plate connection. The end rotation as proved by the profile of the lateral displacement determined
in the case of socket joints is caused by concrete–steel interface by means of the LVDTs shown in Fig. 5.
mechanisms that occur within the socket, i.e. slip in the embed- 6. Conclusions
ded length. This slip is a function of the concrete strength, the
number and configuration of studs, if any, welded on the col- Experimental tests carried out on composite steel and
umn web. Moreover, for socket-type connections the bending concrete columns were presented in this paper. Two layouts
832 L. Di Sarno et al. / Journal of Constructional Steel Research 63 (2007) 819–832
for the base column connections were assessed: the traditional
system employing the bolted steel end plate and the innovative
socket type. The experimental results demonstrate that the
socket system is beneficial for the spreading of inelasticity
at the base of the composite columns. To assess the
inelastic structural performance, the composite specimens
were subjected to monotonic loads at increasing lateral drifts
(pushover experimental tests). It was found that the maximum
drift of the socket-type connection is nearly 75% higher than the
traditional bolted steel end plate. Traditional base connections
fail in a less ductile fashion because of the fracture of the
anchorage bolts. Conversely, socket connections exhibit a
ductile response due the formation of the plastic hinge at the
base of the column, which extends over a length much higher
than the cross section depth. As a result, socket-type joints can
be reliably used for design of structures which may experience
significant inelastic excursions, such as those in earthquake
prone regions.
Acknowledgements
The present work was funded by the Italian Ministry
of Research, through the grant PRIN-COFIN04 (Composite
steel and concrete earthquake-resistant frames: advanced
dissipative joint systems, methods for damage assessment and
seismic design guidelines); the financial support is gratefully
acknowledged.
References
[1] Broderick BM, Elnashai AS. Seismic response of composite frames-I.
Response criteria and input motion. Engineering Structures 1996;18(9):
696–706.
[2] Byfield MP, Nethercot DA. Material and geometric properties of
structural steel for use in design. The Structural Engineer, Journal of the
Institution of Structural Engineers 1997;75(21):363–7.
[3] Bursi OS, Caramelli S, Fabbrocino G, Molina J, Salvatore W, Taucer F.
3D Full-scale seismic testing of a steel–concrete composite building at
ELSA. contr. no. HPR-CT-1999-00059. European Community; 2004.
[4] CNR 10016. Composite steel and concrete structures. Guidelines for
applications in construction. CNR bulletin n. 192. Rome (Italy); 1999 [in
Italian].
[5] Cosenza E, Pecce MR. Le colonne composte acciaio-calcestruzzo: Analisi
sperimentali, modelli di calcolo, indicazioni normative. Costruzioni
Metalliche 2001;10(2):49–60 [in Italian].
[6] Cosenza E, Zandonini R. Composite construction. In: Chen WF, editor.
Handbook of structural engineering. Boca Raton (USA): CRC; 1997.
[7] Di Sarno L, Elnashai AS. Seismic retrofitting of steel and composite
building structures. Mid-America earthquake center report, CD Release
02-01. IL (USA): University of Illinois at Urbana-Champaign; 2002.
[8] Eurocode 2. Design of concrete structures. Part 1.1: General rules and
rules for buildings. Brussels (Belgium): Eur. Comm. for Stand.; 2004.
[9] Eurocode 3. Design of steel structures. Part 1.1: General rules and rules
for buildings. Brussels (Belgium): Eur. Comm. for Stand.; 2004.
[10] Eurocode 4. Design of composite steel and concrete structures. Part 1.1:
General rules and rules for buildings. Brussels (Belgium): Eur. Comm. for
Stand.; 2004.
[11] Eurocode 8. Design provisions for earthquake resistance of structures.
Part 1.3: General rules. Specific rules for various materials and elements.
Brussels (Belgium): Eur. Comm. for Stand; 2004.
[12] Elnashai AS, Elghazouli Y. Performance of composite steel–concrete
members under earthquake loading, part 2: Parametric studies and design
considerations. Earthquake Engineering and Structural Dynamics 1993;
22(4):347–68.
[13] Fabbrocino G, Manfredi G, Cosenza E, Pecce MR. Some remarks on
deformation capacity of composite frames in seismic areas. In: Proc. of
the 1st international conference on steel and composite structures. 2001.
[14] Fabbrocino G, Pecce MR, Di Sarno L. Inelastic response of steel and
concrete columns. In: Proc. of the fourth international conference on steel
and composite structures. 2004. CD-ROM.
[15] Fabbrocino G, Pecce MR, Di Sarno L. Inelastic response of steel and
concrete columns. In: Proc. of the fourth international conference on
advances in steel sructures. vol. I. 2005. p. 793–8.
[16] Federal Emergency Management Agency. Prestandard and commentary
for the seismic rehabilitation of buildings. FEMA report no. 356.
Washington DC (USA); 2000.
[17] Hajjar JF. Composite steel and concrete structural systems for seismic
engineering. Journal of Constructional Steel Research 2002;58(5–8):
702–23.
[18] Lee TKL, Pan ADE. Analysis of composite beam–columns under lateral
cyclic loading. Journal of Structural Engineering, ASCE 2001;127(2):
186–93.
[19] Mazzolani FM. Steel and composite structures in European seismic areas:
Research, codification, design and applications. Earthquake Spectra 2003;
19(2):415–52.
[20] OPCM. General criteria for new seismic zonation in Italy and seismic
standards. Rome (Italy): Department of National Emergency Agency;
2003 [in Italian].
[21] Ricles JM, Paboojian SD. Seismic performance of steel-encased
composite columns. Journal of Structural Engineering, ASCE 1994;
120(8):2474–94.
[22] Roeder CW. Overview of hybrid and composite systems for seismic
design in the United States. Engineering Structures 1998;20(4–6):355–63.
[23] Shanmugam NE, Lakshmi B. State of art report on steel–concrete
composite columns. Journal of Constructional Steel Research 2001;
57(10):1041–80.
[24] Spacone E, El-Tawil S. Nonlinear analysis of steel–concrete composite
structures: State of the art. Journal of Structural Engineering, ASCE 2004;
130(2):159–68.
[25] Thermou GE, Elnashai AS, Plumier A, Doneux C. Seismic design
and performance of composite frames. Journal of Constructional Steel
Research 2004;60(1):31–57.