Computers and Structures 176 (2016) 50–69

Contents lists available at ScienceDirect

Computers and Structures

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

An enhanced component based model for steel connection in a hybrid

coupled shear wall structure: Development, calibration and
experimental validation
F. Morelli a,⇑, M. Manfredi b, W. Salvatore a
a
Department of Civil and Industrial Engineering, University of Pisa, Italy
b
General Infrastructure Services Department, Site Engineering Group, CERN, Switzerland

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

Article history: In the present paper the development, calibration and experimental validation of two component-based
Received 14 December 2015 models of dissipative steel links connecting a reinforced concrete wall to a steel gravity frame is pre-
Accepted 5 August 2016 sented.
Available online 1 September 2016
The dissipative capacity of such structures is greatly influenced by the effective hysteretic behavior of
the wall-to-column coupling system. Within the present work, experimental tests results on two differ-
Keywords: ent configurations of wall-to-column coupling system are presented. The development and calibration of
Steel-concrete hybrid structures
two non-linear cyclic component-based models, one for each configuration, are described, allowing a bet-
Coupled shear wall
Dissipative systems
ter understanding of the force transmission mechanisms and their influence on the global structural
Experimental behavior behavior.
Component-based model Ó 2016 Elsevier Ltd. All rights reserved.
Angle joint model

1. Introduction particular attention to the ones capable to guarantee good ductile

During the last decades, numerous researches were carried out
aiming at developing structural solutions capable to withstand the
seismic lateral loads, assuring the life safety in case of strong inten-
sity earthquakes and minimizing the damages to structural and
non structural elements in case of low-to-mid earthquakes. The
scientific community, in fact, understood that, besides the need
to assure the safety of the building and of its occupants, related
mostly to the structural behavior under strong earthquakes, the
seismic risk, and so the economic loss, is deeply related to the
behavior of the building under moderate earthquake. A recent
proof was given from the Emilia (Italy) May 2012 earthquakes that,
further to all the damages and losses of human life caused by the
collapse of buildings located close to the epicenters, were charac-
terized by huge economic indirect losses. The damages on the
industrial buildings located on one of the most industrialized Ital-
ian region and the need of speed up their recovery pushed the Ital-
ian Government to issue special laws [1,2] and the scientific
community to develop suitable tools [3,4] for the applications of
these new-generation laws.
On the basis of previous considerations, several structural
typologies, materials and innovative devices were studied, with
⇑ Corresponding author.
E-mail address: francesco.morelli@dic.unipi.it (F. Morelli).
behavior allowing the design of lighter and more competitive
structures [5–9].
Among all the possible solutions, concrete shear wall structure
are characterized by high horizontal strength and stiffness allow-
ing the control of horizontal displacements and storey drifts. The
ductility of such structures is strongly dependent on rotational
capacity of plastic hinges usually located at the base of the shear
walls. In order to assure a ductile behavior, brittle mechanisms
(for example concrete crushing caused by shear in boundary ele-
ments or in diagonal strut and tie elements, instability of thin
walled sections or buckling of reinforcing bars under compression)
should be avoided. Moreover, due to the predominant stiffness of
the concrete wall compared to the one of the other structural ele-
ments (usually concrete frames), a very high ratio of the seismic
force is absorbed by the shear wall itself so transmitting high local-
ized force to the foundation.
In order to overcome aforementioned problems Coupled Shear
Walls systems (CSW) were deeply studied by several authors in
the past [10–12]. The functioning principle of CSW is schematized
in Fig. 1: the presence of the coupling beams allows the formation
of vertical forces, acting at foundation level, lowering the intensity
of the bending moments on the walls’ base.
The entity of vertical forces and, by consequence, of the bending
moment acting at foundation level of each wall depends on the

http://dx.doi.org/10.1016/j.compstruc.2016.08.002
0045-7949/Ó 2016 Elsevier Ltd. All rights reserved.
 F. Morelli et al. / Computers and Structures 176 (2016) 50–69 51

Fig. 1. Comparison between two isolated walls (left) and coupled walls (right).

walls coupling ratio, CR, function of the stiffness of the beams with steel coupling beam
respect to the one of the walls and generally evaluated as: side column
T S
CR ¼ ð1Þ
M 2s þ M2d þ ðT  SÞ
r.c. wall
with reference to symbols introduced in Fig. 1.
The major drawbacks in using reinforced concrete beams for the
coupling of shear walls stem from the brittle behavior of the con-
crete. In fact, high level of detailing is needed in order to assure a
stable hysteretic behavior of the concrete beams under cyclic
reverse loading, including confinement of the concrete and suffi-
cient anchorage of steel reinforcement in the connected walls
[13]. In order to solve such problems, steel connecting beams can
be used and the resulting structure can be addressed as a Hybrid
Coupled Shear Wall (HCSW) system. In HCSW, the steel coupling
beams can dissipate energy by yielding under either bending or
Seismic resistant Gravity structure
shear forces, similarly to the eccentrically braced systems. Provid-
structure
ing a sufficiently long embedment length to steel connecting
beams, if enough vertical reinforcing steel is assured in the wall
Fig. 2. Structural scheme of the innovative hybrid coupled shear wall system.
boundary element to prevent excessive cracking, special reinforc-
ing details, such as bars welded to the beam flanges or passing
through them, are not more required [14]. Embedment length To these purposes, in the present study, the real hysteretic
can be reduced and the strength and stiffness of the connection behavior of the elements connecting the r.c. wall to the side steel
can be increased if the embedded element is provided with shear columns were investigated, developing an accurate mechanical
studs and some additional reinforcing details [15,16]. HCSW solu- model.
tion, however, requires necessarily the presence of at least two The general methodology schematically shown in Fig. 3 was fol-
concrete walls, reducing greatly its field of application due to pos- lowed. Referring to the HCSW case study designed within the
sible architectural limitations or requirements. INNO-HYCO project, an experimental campaign was carried out
A recent research, carried out in the framework of the European on sub-systems representing the wall-to-column dissipative
project INNO-HYCO (INNOvative HYbrid and Composite steel- connection and executing both monotonic and cyclic tests, with
concrete structural solutions for building in seismic area) [17], constant and increasing amplitude. To fully understand and repre-
funded by the Research Fund for Coal and Steel (RFCS), proposed sent the cyclic behavior of the coupling system, a component based
an innovative typology of HCSW extending basic principles of model was then fully developed and calibrated on experimental
CSW and HCSW. The new solution was obtained by coupling a r. results, as already successfully done in past researches such as
c. shear wall with two steel side column by means of steel links [18–21] among others.
where the energy dissipation can take place, while the gravity
loads are carried by a steel pendular structure, see Fig. 2. Several
advantages can be so obtained, such as the possibility of foreseeing 2. Innovative hybrid coupled shear wall systems
the presence of just one r.c. shear wall and an easier repair in case
of damages to the dissipative elements. The innovative HCSW system, extensively described in [22,23]
It is straightforward that, in the case of design of a dissipative and shown in Fig. 2, is composed by an earthquake resistant struc-
structure, the dissipative capacity of the steel coupling beams is ture realized by a reinforced concrete wall linked in correspon-
a fundamental design parameter and, consequently, great atten- dence of each side to steel columns by steel coupling beams. The
tion should be given to the geometrical and mechanical character- HCSW system is connected to a steel hinged-frame in such a way
istics of the embedded elements, of the dissipative ones and of that the HCSW system absorbs horizontal actions while gravity
their connection with the steel side columns. frame only withstands gravity loads. The overturning moment
52 F. Morelli et al. / Computers and Structures 176 (2016) 50–69

INNO-HYCO results: Preliminary calibration of Using data available

− Development of HCSW
− Design procedure
− Design of case study
Experimental tests on
sub-systems
Selection of the main
mechanical components
the model on the state of the art
Comparison with the
experimental results
Final calibration of the Using experimental
results on tests with
model constant amplitude
Comparison with the
experimental results

Fig. 3. Methodology.

acting on the HCSW system is subdivided between the bending
moment acting on the concrete wall and the couple of axial forces
acting on the two steel side columns, as schematically shown in
Fig. 4. The coupling beams, designed by a proper capacity design
approach, can dissipate, through hysteretic cycles, a portion of
the seismic energy, so limiting the value of forces transmitted to
the other structural elements; the concrete wall essentially
remains in the elastic field or experiences limited damages in the
case of strong earthquakes.
Analyzing Fig. 4, it is evident that, neglecting the contribution of
gravity loads on the HCSW system, the axial force NEd,C acting on
the steel column can be obtained as the sum of the shear forces
transmitted by the dissipative links, VEd,link,i:
X
nlink
NEd;C ¼ V Ed;link;i
i¼1
consequently the overturning moment acting on the steel side col-
umns, MEd,C, and the associated coupling ratio, CR, can be obtained
as follows:
Starting from the definition of a CR value, the design procedure,
proposed and detailed described in [24], can be resumed in the fol-
lowing steps:
– Step 1: Reinforced concrete wall for the flexural action, MWall. The
overall dimensions of the wall are fixed, considering a wall
thickness, twall, sufficient to accommodate the dissipative link.
Specific local enlargements of the wall thickness can be fore-
seen for the accommodation of large dissipative links. The lon-
gitudinal reinforcement is dimensioned in order to resist to a
bending moment equal to MWall. Indications supplied by [25]
can be usefully adopted for the design of specific details of
the dissipative link framing into the wall.
– Step 2: Dissipative links. The shear action VEd,link,i in each link is
ð2Þ evaluated considering a uniform or non-uniform distribution
of shear among the link themselves as follows:
NC 1 CR M Ed;Wall
V Ed;link;i ¼ wi ¼ wi : ð5Þ
nlink nlink 1  CR LTOT

M Ed;C ¼ LTOT  NEd;C ð3Þ In the case of uniform distribution, the coefficients wi are equal
to 1 and the shear action Vlink,i can be simply evaluated as the
M Ed;C
CR ¼ : ð4Þ value of the axial force on the steel side column, NC, divided
M Ed;C þ M Ed;Wall by the number of links. In the case of non-uniform distribution,
It is straightforward that the coupling ratio, CR, is one of the suitable relationships for the evaluation of the coefficients wi
most important design parameter: lower values of CR imply higher are proposed in [24]. Once the shear force in each link is evalu-
overturning moment on the concrete wall, MWall and high damages ated, it is possible to define its cross section and, consequently,
at its base, while higher values of CR determine lower values of the connections to the r.c. wall and to the side steel columns.
MWall and a more diffused action on the wall. – Step 3: Steel side columns. Side columns are not-dissipative ele-
ments and are so designed following the capacity design rules
amplifying the design axial force, NEd,C, by a safety factor. The
eccentricity between the column axis and the application point
of the link shear force shall be also taken into account.
Vbase – Step 4: Reinforced concrete wall, transverse reinforcement. The
wall design shear force, VEd,Wall, is evaluated hypothesizing the
formation of a plastic hinge at the wall base and the complete
yielding of all the dissipative links considering suitable safety
H1 factors so that it results:

cw MEd;Wall þ 1:1cov MEd;C

V Rd;Wall > V Ed;Wall ¼ ð6Þ
H1
where H1 is the resultant height of the fundamental mode iner-
Vwall tial force distribution or a fundamental-mode based equivalent
lateral force distribution, see Fig. 4.
NC Mwall NC
Following the aforementioned design procedure and referring
LTOT
to the requirements of Eurocode 8 [25], within the INNO-HYCO
Fig. 4. Innovative hybrid coupled shear wall system: schematic distribution of project [1], an 8-storeys HCSW buildings was designed. The plan
horizontal forces among the seismic resistant structural elements. and front views are schematically reported in Fig. 5. The
 F. Morelli et al. / Computers and Structures 176 (2016) 50–69 53

=
800 cm

=
=
800 cm

27.20 m

=
=
800 cm

= 340 cm =
40 m

800 cm

800 cm 800 cm 800 cm 800 cm 800 cm

65cm 600cm 65cm

800 cm

800 cm 800 cm 800 cm 800 cm 800 cm

40 m

Fig. 5. 8 storeys case study. Plan view and front view.

gravity-resisting structure is composed by pinned beam-to- configuration 2 the moment is balanced by a couple of forces dis-
column joints (non-moment-resisting frame) with storey height tribution orthogonal to the embedded element axis line, see Fig. 7.
equal to 3.40 m. Columns are continuous and pinned at the base; Two different solutions for the link splice were considered. In
beams are realized with IPE500 profiles. The design of the innova- the first case it was placed at a distance from the concrete wall suf-
tive seismic-resisting system was executed using a spectrum rep- ficient to allow an easy bolting of the replaceable part (solution
resentative of a medium intensity seismic event (ag = 0.25 g, type associated to configuration 1), in the second case, the splice con-
1 spectra, ground type C). nection was placed in correspondence of the wall face and
Two different hypotheses for the link distribution were made threaded bushings were used to allow fastening of the link (solu-
(uniform and non-uniform), obtaining cross-sections of links tion associated to configuration 2). In both cases, the link is con-
reported in Tables 1 and 2 for different coupling ratios CR. nected to the steel column by an angle joint, assumed unable to
During the design of the 8-storey case study, two different cou- transfer significant bending moments.
pling system configurations were considered, see Fig. 6, in the fol- It is evident that a reliable knowledge concerning the effective
lowing referred as ‘‘configuration 1” and ‘‘configuration 2”. behavior of the elements connecting the concrete wall to the side
It is evident that, in order to assure a good dissipative capacity steel columns is a basic point for an efficient design of the HCSW
to the steel links and to allow their eventual repair after mid-to- system.
strong earthquakes, a primary role is played by the connections
to the r.c. wall and to the steel side columns. They should be char- 3. Experimental campaign
acterized by a suitable stiffness, in order to avoid excessive elastic
deformation that would limit the dissipative capacity of the cou- An experimental campaign on 8 specimens, 4 for each coupling
pling beam, and by a sufficient strength in order to avoid damages system configurations, aiming at studying their real hysteretic
not easy to be repaired. behavior, was carried out.
An embedded element, a dissipative link, a beam splice (able to
transfer shear and flexural actions between the link and the
3.1. Design of specimens and test setup
embedded element) and a link-to-column joint make up both con-
nection typologies. In configuration 1, the bending moment is
A sub-structure representing the HCSW structural portion
transferred by the link to the wall primarily by the shear action
involved in the seismic energy dissipation, which is specifically
of shear studs welded to the embedded element flanges, while in
made up of the dissipative link and the connected structural ele-
ments, was properly designed. In order to facilitate the application
Table 1
of the external forces during tests, the sub-structure was ‘‘rotated”
Designed links for the 8-storey case study (uniform distribution). by 90 deg, as schematically shown in Fig. 8.
The specimens were designed starting from the steel link pro-
Storey CR
files obtained for the 8-storeys case study and considering the
0.60 0.70 0.80 0.85 0.90 loading capacity of the testing facility. Considering a safety factor
1 to 8 IPE330 IPE400 IPE500 IPE550 IPE750 equal to 1.5 with respect to the jack capacity, the designed inter-
mediate link of length 660 mm with section IPE330 steel S355,
determined a down-scaled link with section IPE200, length equal
Table 2 to 360 mm and steel grade S275. The technical drawings of each
Designed links for the 8-storey case study (non-uniform distribution). coupling system configurations are shown in Fig. 9. In order to
Storey CR study the influence of fastening torque of the angle joint bolts on
0.60 0.70 0.80 0.85 0.90
the overall behavior, for connection 1 a fastening torque of about
50 N m, estimated as the mean value applicable using by simple
1 to 4 IPE300 IPE300 IPE330 IPE360 IPE450
wrench, was used, while for connection 2 the fastening torque
5 to 8 IPE360 IPE450 IPE600 IPE750 IPE750
was fixed equal to 250 N m, as indicated by Eurocode 3 [26].
54 F. Morelli et al. / Computers and Structures 176 (2016) 50–69

Replaceable
R.c. wall dissipative link
Side column

Beam splice

Embedded element
Link-to-column angle
connection

Replaceable
R.c. wall dissipative link
Side column

Beam splice
Link-to-column angle
Embedded element
connection

Fig. 6. Wall-to-column coupling systems: configuration 1 (top) and configuration 2 (bottom).

Fig. 7. Schematic behavior of the embedded elements of the two coupling system configurations.

The experimental tests were organized connecting the r.c. wall
to the lab strong floor in order to minimize its vertical and horizon-
tal displacement, while the cyclic shear force was transmitted to
the link by a hydraulic actuator connected to the element repre-
senting the steel side column, as shown in Fig. 10.
The force transmitted by the actuator was recorded by a load
cell, while the displacements and the strains of the steel and con-
crete elements by displacement transducers and strain gauges, dis-
posed and numbered as shown in Fig. 11.
For each coupling system configurations, 4 cyclic tests were
executed. Two specimens, hereafter referred as ‘‘first group”, were
subjected to an increasing displacement cyclic history, equal to 0.1
d, 0.2 d, 0.3 d, 0.5 d, 0.7 d and 1.0 d, being d the displacement
expected at failure, evaluated equal to 32 mm considering the
limitation given by Eurocode 8 [25] for the dissipative links of
eccentrically braced structures. Five cycles were repeated for each
intermediate step and ten cycles for the last step (1.0 d). Two other
specimens, hereafter referred as ‘‘second group”, were subjected to
a constant amplitude cyclic history, corresponding to the design
displacement d up to failure, with a maximum number of imposed
cycles equal to 20.
3.2. Configuration 1: test results
In Fig. 12 the results of the tests for configuration 1, in terms of
force-displacement curves, are reported. In particular, the blue line
represents the link top displacement (evaluated at a distance of
50 mm from the link top by the LVDT sensor n.11 of Fig. 11), while
 F. Morelli et al. / Computers and Structures 176 (2016) 50–69 55

Sub system extrapolated from the structure Test specimen

Side column
Replaceable Seismic
link element action
R.c. wall Side column
Replaceable
link element
90°

R.c. wall

Seismic action

Fig. 8. Schematic extrapolation of the subsystem from the global structure.

29.5 129 129 129 148.5 25

29.5 70.5 70.5

4 M24 4 M24
cl. 10.9 cl. 10.9
HEB 200 IPE200 HEB 220

35
HEB 200 S355 S275 S355

135
65
35
40 60 50 L 150 x 90 x 10
150

200 2525
490 100 350 70 10 220
590 420 10 220
(a)

8 threaded 8 M24
bushes M24 cl. 10.9
cl.10.9 4 M24 4 M24
cl. 10.9 cl. 10.9
HEB 200 HEB 220
35

S355 S355
135
65

IPE200
35

S275

40 60 50 L 150 x 90 x 10
150
8 8 10
150 134 175 2525 125 120 70 80
350
500

(b) 420 220

Fig. 9. Coupling system: (a) configuration 1 and (b) configuration 2.

the red1 one represents column element axis displacement
(recorded by the LVDT sensor n.12 of Fig. 11).
On the basis of test results, it can be seen that the system is
characterized by important relative displacements between the
link top and the column axis, mainly due to the deformability of
the angle joints. This deformability causes evident pinching
phenomena in the hysteretic behavior of the wall-to-column con-
1
For interpretation of color in Fig. 11, the reader is referred to the web version of
this article.
nection (red line), while the sub-system wall-dissipative link (blue
line, without so taking into account the influence of angle joints)
shows a relatively ‘‘fat” hysteretic behavior. Moreover, analyzing
the external force applied versus the relative displacement
between the column axis and the top link curve, see Fig. 13, it is
evident that:
(i) the relative displacement increases as the maximum
imposed displacement increases and its maximum value is
greater than the bolt-hole allowance (equal to about
10 mm);
56 F. Morelli et al. / Computers and Structures 176 (2016) 50–69

Fig. 10. Global test setup.

Fig. 11. Sensors position.

(ii) a linear variable friction between the angle profile and the
link or column element causes energy dissipation;
(iii) there is a sudden stiffness increase of the angle joint behav-
ior due to the contact between the bolts and the link web
and the consequent plastic ovalization of the dissipative link
web holes, as evident from the specimens residual deforma-
tion after the tests, see Fig. 14.
No evident damages on the concrete wall were detected at the
end of the tests.
3.3. Configuration 2: test results
Similarly to configuration 1, in Fig. 15 the results of the tests
obtained for configuration 2, in terms of force-displacement
curves, are shown.
It is clear that this configuration is able to dissipate an higher
quantity of seismic energy with respect to the first one. The
increase of the dissipated energy is very likely due to the presence
of higher friction forces, if compared with the ones of the configu-
ration 1, between the angle elements and the dissipative link. In
 F. Morelli et al. / Computers and Structures 176 (2016) 50–69 57

(a) (b)

(c) (d)
Fig. 12. Experimental results on configuration 1. Force – displacement curves of the first (a, b) and second group (c, d).

(a) (b)
Fig. 13. Force versus link-to-column relative displacement for the test (a) n° II of the first group and (b) n°IV of the second group of configuration 1.

fact, as shown in Fig. 16, the area within the force-relative dis-
placement curve, is greater than the one relative to configuration
1, see Fig. 13. A linear variable friction between the angle profile
and the link or column element can be still individuated, but the
stiffness increase due to the contact between the bolts and the link
web is less evident and smoother.
All the configuration 2 specimens showed the same collapse
mechanism: the link flanges experienced local buckling phenom-
ena in the zone of the plastic hinge and a cracking on the buckled
zone, see Fig. 17, probably because of flexural low cycle fatigue.
The link web zone connected to the angle element was deformed
also because of the plastic ovalization of the link holes and the
angle elements were damaged, showing cracks triggered from
the bolt holes and propagated toward the edges of the angle pro-
file. The main cause of these cracks can be attributed to the cyclic
flexural plasticization of the angle elements that, during the tests,
acted as series of ‘‘T-stub” elements, see Fig. 17e. No evident dam-
age of the r.c. wall was observed.
4. Component-based models
In order to extend the obtained experimental results to joints
characterized by different mechanical and geometrical characteris-
tics and so to develop a tool able to predict the link capacity
to dissipate seismic energy under cyclic action, suitable
58 F. Morelli et al. / Computers and Structures 176 (2016) 50–69

Fig. 14. Test n° IV of the second group: (a) position of the seat angle during the test and (b) plasticization of the link web at the end of test.

(a) (b)

(c) (d)
Fig. 15. Force – displacement graphs of the first (a, b) and second group (c, d) of connection typology 2.

(a) (b)
Fig. 16. Force versus link-to-column relative displacement for the test (a) n° II of the first group and (b) n°IV of the second group of configuration 2.
 F. Morelli et al. / Computers and Structures 176 (2016) 50–69 59

Fig. 17. Damages to the elements at the end of test n°II of the first group: (a) local instability of the link flange; (b) cracking of the link flange due to the cyclic plasticization;
(c) plasticization of the link web; (d) damage to the angle element due to the cyclic plasticization; and (e) deformation of the angle elements during the tests.

component-based models were developed and implemented in
OpenSEES (‘‘The Open System for Earthquake Engineering Simula-
tion” [27]) for both the two coupling system configurations.
To these purposes, four main functional zones involved in the
force transmission between the steel column and the r.c. wall
were individuated: the steel element embedded in the concrete
wall, the beam splice, the steel dissipative element and the angle
connection between the dissipative element and the steel column,
see Fig. 18.
Each functional zone was then decomposed in functional ele-
ments/components. In particular, in Table 3 the functional ele-
ments/components considered for each zone are reported.
Each component was modeled by suitable linear on nonlinear
laws and schematized by spring or beam elements. The constitu-
tive non linear laws were preliminarily calibrated, when possible,
using the experimental data available in the technical literature.
The unsatisfactory behavior evidenced by the link-to-column
angle connection during the experimental results obtained in con-
figuration 1, highlights the necessity to apply a convenient fasten-
ing torque also in the case of angle connections behaving as pins.
For this reason, given the low scientific and technical significance
of such solution, the link-to-column angle connection has been
modeled only for configuration 2.
4.1. Component model of coupling system configuration 1
In Fig. 19, the coupling system configuration 1 and the associ-
ated mechanical scheme are shown.
With reference to Table 3, in the following, the linear and non-
linear models adopted for each components and the eventual cal-
ibration procedure are reported.
4.1.1. Steel embedded element
The steel embedded profile, designed to remain elastic even if
the dissipative element undergoes plastic deformations, was
schematized as an elastic beam with an HEB200 cross section
whose axis was assumed coincident with the embedded profile
axis (with reference to Fig. 19, element I-C).
4.1.2. Steel shear studs
Each row composed by three shear studs was schematized by a
nonlinear spring connected to the embedded element through
60 F. Morelli et al. / Computers and Structures 176 (2016) 50–69
Fig. 18. Main zones involved in the mechanism of force transmission for config-
uration 1 (top) and 2 (bottom).
rigid links describing the eccentricity between the embedded pro-
file axis and the base of the shear studs.
The cyclic force-slip relation (F-c) of each spring was modeled
in OpenSEES using a Pinching4 material [27], taking so into account
the stiffness and strength cyclic degradation.
The monotonic tri-linear model proposed by Eurocode 4 [28]
was used as backbone curve for the cyclic model. The monotonic
model was calibrated using the Aribert et al. [29] monotonic exper-
imental results, see Fig. 20a, obtained on specimens characterized
by rows containing a single /19 Nelson shear stud. The embedded
element was characterized by three /19 Nelson shear studs for
each row so that the parameters for the monotonic behavior of 3
studs in a row were obtained adopting the hypothesis that each
stud works in parallel with the others without interaction, scaling
the forces obtained in [29] by a factor equal to 3, see Fig. 20b. The
values of parameters characterizing the monotonic calibrated
curve are reported in Table 4.
Using the tri-linear monotonic model as ‘‘backbone” curve, the
cyclic model was firstly obtained for the single shear stud in a row
considering a symmetric pinched force-slip model (Pinching4 of the
OpenSEES material library [27]) introducing strength and loading
Table 3
Functional elements/components considered in the connections models.
Embedded zone:
– steel embedded element;
– shear studs (only for connection 1);
– concrete in compression;
Dissipative steel link:
– dissipative steel element;
Fig. 19. Mechanical model for the coupling system configuration 1.
stiffness degradation effects as a function of the maximum dis-
placement and total energy dissipated.
The form of the damage rules are a more general version of the
damage index proposed by Park and Ang [30].
  a3  a4 
dmax Ei
di ¼ a1  þ a2  ð7Þ
dcap;max Emon
where i refers to the current displacement increment, di is the dam-
aging index (di = 0 means no damage, di = 1 means maximum dam-
age), dmax and dcap,max are, respectively, the maximum displacement
experienced by the system and the displacement that define the
failure of the system, Ei is the energy dissipated up the i cycle, Emon
is the energy required for the failure of the system under monotonic
loading. a coefficients are to be calibrated from the experimental
results.
The loading stiffness ki and the strength Fmax,i at each cycle are
then computed in the following way:
ki ¼ k0  ð1  dki Þ ð8Þ
F max;i ¼ F max  ð1  dF i Þ ð9Þ
where k0 is the initial stiffness, Fmax is the undamaged envelope
strength, dki and dFi are, respectively, the stiffness and strength
damage indexes.
The a coefficients governing the cyclic behavior of the Pinching4
model were calibrated using the results related to the cyclic tests
described in Aribert et al. [29] by a trial-and-error procedure aim-
ing at minimizing errors in terms of strength and of dissipated
energy, see Fig. 21a. In the case of 3 shear studs per row, the
parameters of the related Pinching4 model were obtained
Beam splice:
– embedded steel profile: flange and web in tension or compression;
– dissipative link: flange and web in tension or compression;
– endplates in bending;
– bolts in tension;
Link-to-column angle connection (only for configuration 2):
– angle elements modeled as ‘‘T-stubs”;
– bolts in shear;
 F. Morelli et al. / Computers and Structures 176 (2016) 50–69 61

120 350
(PMAX, sMAX)
100 300
(PRk, spl)
250
80
3 shear studs
200
F [kN]

F [kN]
60 Experimental 1 shear stud
(Py, sy) Numerical Model 150
40
100
20 50

0 0
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0
s [mm] s [mm]
(a) (b)
Fig. 20. Calibration of the monotonic tri-linear model on the base of the experimental results of Aribert et al. [29].

Table 4
Parameter for the definition of the tri-linear monotonic model for a single stud in a
row and for 3 studs in a row.
Py sy PRk spl
Single stud
61.5 kN 0.22 mm 87.9 kN 0.848 mm
3 studs per row
184.5 kN 0.22 mm 263.7 kN 0.848 mm
modifying the parameters related to the backbone curve definition,
see Fig. 21b.
4.1.3. Concrete in compression
The component is constituted by the contact between the con-
crete surface and the embedded profile flanges and the mechanical
behavior of the concrete in compression. It was schematized by a
unilateral element, constituted by a no-tension element schema-
tizing the unilateral contact, in series with a nonlinear extensional
spring, schematizing the concrete in compression. Unilateral ele-
ments were disposed all along the embedded zone, on both sides
of the embedded element axis, with inter-axis equal to 10 mm.
Two different concrete models were used to distinguish the por-
tion of concrete confined by the steel stirrups and the unconfined
portion: the Mander nonlinear model [31] was assumed for the
concrete in compression, in the confined zone, comprised between
the E-E0 section and the I point, see Fig. 19; in the remaining zone
(unconfined concrete) the Popovic model [32] was used.
Both the concrete models were calibrated on the basis of the
mean value of the results of compression tests executed on six
specimens realized with the same concrete material used to build
the specimens tested in the laboratory. Results of compression
tests are reported in Table 5.
PMAX sMAX
In Tables 6 and 7 the constitutive laws, the values of related
parameters and the resulting stress-strain curves used in the com-
ponent model, respectively, for the unconfined and confined con-
103.4 kN 6.00 mm
crete, are reported.
310.2 kN 6.00 mm
4.1.4. Beam splice
The beam splice, designed as a full strength joint, was schema-
tised as a linear rotational spring with a constant stiffness evalu-
ated using the component-based method described in Eurocode 3
part 1–8 [25]. With reference to the indications and symbols used
by this standard, the considered components and the mechanical
scheme adopted for the evaluation of the beam splice initial stiff-
ness, Sj,ini, are shown in Fig. 22. The stiffness associated to each
component is resumed in Table 8 while the equivalent values
and the beam splice initial stiffness, Sj,ini, are reported in Table 9.
4.1.5. Dissipative link
Starting from the experimental evidence, the dissipative link,
made up by an IPE 200 profile, was modeled using an element with
lumped plasticity: the position of the plastic hinge was assumed
coincident with the center of the stiffened zone of the dissipative
element, point B of Fig. 19; the remaining part of the dissipative
link was modeled as an elastic beam, element AB of the same
figure.
Several methods exist for the calibration of the moment-chord
rotation (M-h) curve of the plastic hinge, some of them also
described in international standards [33,34]. However, in order to
obtain a numerical behavior as close as possible to the experimental

100 250
F [kN]

F [kN]

80 200
60 150
40 100
20 50
0 0
-4.0 -3.0 -2.0 -1.0 -20 0.0 1.0 2.0 3.0 4.0 -4.0 -3.0 -2.0 -1.0 -50 0.0 1.0 2.0 3.0 4.0
-40 -100
Experimental
-60 Numerical model -150
-80 -200
-100 -250
s [mm] s [mm]
(a) (b)
Fig. 21. Force-slip relationship [29] for the (a) single shear stud and (b) three shear stud per row.
62 F. Morelli et al. / Computers and Structures 176 (2016) 50–69

Table 5
Compression strength of the r.c. wall elements for connection 1 and connection 2.

Compression strength of r.c. wall of configuration 1 [MPa] Compression strength of r.c. wall of configuration 2 [MPa] Total mean value
(conn. 1 and conn. 2) [MPa]
Test 1 Test 2 Test 3 Mean value Test 1 Test 2 Test 3 Mean value
40.8 28.6 27.7 32.3 27.8 30.1 37.2 31.7 32.0

Table 6
Unconfined concrete model parameter [32].

Unconfined concrete
0
Constitutive law f c eec0 r
f c ¼ r1þc ec r
e0c
0
Unconfined concrete strength f c = 32.0 MPa
Deformation e0c associated to f 0c e0c = 0.002

0
Shape parameter r ¼ 0:4  103  f c ðpsiÞ þ 1:0 ¼ 2:86

Table 7
Confined concrete model parameters [31].

Confined concrete
0
Constitutive law f cc ee0c r
f c ¼ r1þccec r
e0cc
0
Unconfined concrete strength f c = 32.0 MPa
0
Deformation associated to f c e0c = 0.002
0
Confined concrete strength f cc = 42.0 MPa
0
Deformation associated to f cc e0cc = 0.005

Beams web
in tension

Endplates in
bending K5,HEB,1 K10 K5,IPE,1
Row 1

IPE200 K5,HEB,2 K10 K5,IPE,2

HEB 200 HEB 200 IPE200
236 mm

S355 Row 2 S275 S355 S275

K7,HEB K7,IPE
141 mm

K5,HEB,3 K10 K5,IPE,3

Row 3
51 mm

K8,HEB K8,IPE

Bolts in
tension
(b)
Centre of
compression Beams web and flange
in compression

(a)
Fig. 22. Evaluation of the beam splice initial stiffness: (a) components considered and (b) mechanical scheme adopted.

one, in the present study the aforementioned M-h relationship was behavior. In particular, the flexural overstrength, S, and the rota-
obtained processing experimental data available in technical tion capacity, R, were evaluated through two different empirical
literature. expressions, hereafter reported, functions of the geometrical and
Landolfo et al. [35] carried out an experimental program on a mechanical characteristics of the cross section:
wide range of cross section typologies, comprising I sections, under
monotonic and cyclic loading, also developing a methodology for ðsÞ ðsÞ ðsÞ ðsÞ bf ðsÞ E ðsÞ eh
1=S ¼ C 1 þ C 2  k2f þ C 3  k2w þ C 4  þ C5  þ C6  ð10Þ
the estimation of the main parameter affecting the non-linear Lv Eh ey
 F. Morelli et al. / Computers and Structures 176 (2016) 50–69 63

Table 8
Single components stiffness.

Component Stiffness Component Stiffness

* *
Endplate of the HEB200 profile in bending k5;HEB;1 ¼
0:9leff ;min t 3p
¼ 71:27 mm Endplate of the IPE200 profile in bending k5;IPE;1 ¼
0:9leff ;min t 3p
¼ 46:85 mm
m3x m3x
0:9leff ;min t 3p 0:9leff ;min t 3p
k5;HEB;2 ¼ m3x
¼ 28:75 mm k5;IPE;2 ¼ m3x
¼ 24:63 mm
0:9leff ;min t 3p 0:9leff ;min t 3p
k5;HEB;3 ¼ mx3
¼ 23:50 mm k5;IPE;3 ¼ m3x
¼ 21:97 mm
Beams web and flange in compression k7,HEB200 = 1 Beams web in tension k8,HEB200 = 1
k7,IPE200 = 1 k8,IPE200 = 1
Bolts in tension k10 ¼ 1:6A
LB ¼ 7:65 mm
S

*
A different stiffness was evaluated for each bolt row in tension level.

(3) Evaluation of SIPE200 and RIPE200 for the IPE200 profile using
Table 9
Equivalent stiffness of bolts in tension and beam splice initial stiffness. (10) and (11).
(4) Calibration of the monotonic multi-linear behavior of the
Equivalent stiffness Lever arm
IPE200 profile. The IPE300 multi-linear curve was scaled,
Equiva lent stiffness and keff ;1 ¼ 1
¼ 6:02 mm h1 ¼ 51 mm see Fig. 23b, using, for the x axis as scale factor the ratio
1
þ 1 þ 1
k5;HEB;1 k10 k5;IPE;1
lever arm for each bolt
keff ;2 ¼ 1
¼ 4:85 mm h2 ¼ 141 mm SIPE200/SIPE300 while for the y axis the ratio RIPE200/RIPE300,
row in tension level 1
þ 1 þ 1
k5;HEB;2 k10 k5;IPE;2
obtained respectively in steps 1 and 3.
keff ;3 ¼ 1
¼ 4:57 mm h3 ¼ 236 mm
1
þ 1 þ 1
k5;HEB;3 k10 k5;IPE;3 (5) Calibration of the cyclic model of the IPE300 profile using
P the cyclic experimental results data available in Landolfo
Equivalent lever arm keff ;r h
2

zeq ¼ Pr k hr ¼ 159:6 mm

r eff ;r r et al. [35], see Fig. 24a. The OpenSEES Hysteretic material
P
Equivalent stiffness of k h
r eff ;r r
was adopted, using the monotonic multi-linear curve
keq ¼ zeq ¼ 12:3 mm
bolts in tension obtained in step 4 as backbone curve and calibrating the
Initial beam splice Ez2eq strength and stiffness degradation parameters by a trial-
Sj;ini ¼ 1 ¼ 920 956 kNm=rad
stiffness keq
and-error procedure with the goal of equalizing the dissi-
pated energy cycle by cycle.
(6) Calibration of the cyclic model of the IPE200 profile adopting
ðRÞ ðRÞ 1 ðRÞ 1 ðRÞ bf  h ðRÞ bf  t f ðRÞ Af the same scale factor used in step 3, see Fig. 24b.
R ¼ C1 þ C2  þ C3 þ C4 þ C5 þ C6
k2f k2w L2v h  Lv ATOT
Lm 4.1.6. Numerical results
ðRÞ ðRÞ
þ C7 þ C8 S ð11Þ On the base of the mechanical models developed for the cou-
Lv
pling system configuration 1, see Fig. 19, and for all its single com-
The values of the coefficients C(R) and C(s) as well as the meaning ponents, an overall component-based model was developed and
of the terms in (10) and (11) are reported and explained in [35]. implemented in OpenSEES. Comparing the results of the compo-
Lacking tests on IPE200 profiles, monotonic cyclic behavior of nent model and the experimental ones related to the first group
dissipative link were calibrated analyzing firstly an IPE300 profile of tests (increasing amplitude), see Fig. 25, a very satisfactory
and then scaling results to obtain an estimation of the IPE200 agreement, especially in terms of strength and loading stiffness,
behavior. The following procedure was so followed: was noticed. On the contrary, some differences existed for the
unloading stiffness so that, the energy dissipated by the compo-
(1) Evaluation of the parameters SIPE300 and RIPE300 for the IPE300 nent model resulted higher. These differences were probably due
profile using (10) and (11). to the indirect process of calibration used that did not consider
(2) Calibration of the numerical multi-linear model of the the actual behavior of the dissipative element adopted during the
IPE300 using the experimental data available in Landolfo tests.
et al. [35]. It was assumed a multi-linear model defined A better agreement was obtained calibrating the unloading
using 6 points and assuring the same dissipated energy of stiffness parameter (leaving all the other parameters unchanged)
the experimental results, as shown in Fig. 23a. on the base of the experimental results of the first group of

250 250

200 200
Moment [kNm]
Moment [kNm]

150 Numerical Model 150

Experimental results Numerical model
100 100 IPE300
Numerical model
50 50 IPE200

0 0
0 0.02 0.04 0.06 0.08 0.1 0.12 0 0.02 0.04 0.06 0.08 0.1 0.12
Rotaon Rotaon
(a) (b)
Fig. 23. (a) Moment-rotation curve for the IPE300 profile: experimental curve (dashed line) extrapolated from [35] and multi-linear model (solid line) and (b) Moment-
rotation curve for the IPE300 and IPE200 profiles.
64 F. Morelli et al. / Computers and Structures 176 (2016) 50–69

experimental tests (increasing amplitude). The second group of numerical results and the experimental results are shown respec-
tests (constant amplitude) was then used for the validation of tively in Figs. 26 and 27, respectively for group 1 and group 2 of
the newly calibrated model. The comparison between the updated tests.

300 300

250 250

200 200

150 150

100 100
Moment [kNm]

Moment [kNm]
50 50

0 0
-0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04 -0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04
-50 -50

-100 -100

-150 -150

-200 -200
Numerical model IPE200
-250 Experimental IPE300 -250
Numerical model IPE300
Numerical model IPE300
-300 -300
Rotaon Rotaon

(a) (b)
Fig. 24. (a) Calibration of the cyclic numerical model for the IPE300 profile and (b) IPE200 profile cyclic numerical model obtained downscaling the IPE300 profile one.

300 300

200 200

100 100
Force [kN]
Force [kN]

0 0
-20 -15 -10 -5 0 5 10 15 20 -20 -15 -10 -5 0 5 10 15 20
-100 -100
Experimental Experimental
Numerical Numerical
-200 -200

-300 -300
Displacement [mm] Displacement [mm]
(a) (b)
Fig. 25. Comparison between the link force-displacement curve of the theoretical model (purple line) and the experimental results (orange line) of the first test group. (For
interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

300 300

200 200

100 100
Force [kN]
Force [kN]

0 0
-20 -15 -10 -5 0 5 10 15 20 -20 -15 -10 -5 0 5 10 15 20
-100 -100
Experimental Experimental
Numerical Numerical
-200 -200

-300 -300
Displacement [mm] Displacement [mm]

(a) (b)
Fig. 26. Comparison between the link force-displacement curve of the updated theoretical model (solid green line) and the experimental results (dashed orange line) of the
first test group. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
 F. Morelli et al. / Computers and Structures 176 (2016) 50–69 65

The improved calibration of the proposed model is also evident
from the comparison of the total dissipated energy, as shown in
Fig. 28. The results obtained with the calibrated model are practi-
cally coincident with the experimental ones even for the second
tests group (tests III° and IV°).
4.2. Component model of coupling system configuration 2
The experimental behavior of the coupling system configura-
tion 2 highlighted that, even if a sufficient fastening torque is
applied to the bolts of link-to-column angle joint, the presence of
the angle connection can induce relevant relative displacements
between the link and the column. Such displacements can lower
the effective stiffness and dissipation capacity of the wall-to-
column connection.
A more complex model was so developed, including the func-
tional elements/components related to the angle elements and to
the bolts in shear. The coupling system configuration 2 and the
mechanical scheme of its component model are shown in Fig. 29,
where the dissipative link and the angle element-column systems
are represented separately for a clearer representation.
The embedded elements and the link were modeled using the
same approach adopted for the connection 1. The L, M, N, O points
of Fig. 29, representing the location of the holes in the link, are
respectively connected to the L’, M’, N’, O’ points, representing
the holes in the angle elements, by the functional element/
component representing the behavior of bolts in shear.

300 300

200 200

100 100

Force [kN]
Force [kN]

0 0
-20 -15 -10 -5 0 5 10 15 20 -20 -15 -10 -5 0 5 10 15 20
-100 -100
Experimental Experimental
Numerical Numerical
-200 -200

-300 -300
Displacement [mm] Displacement [mm]

(a) (b)
Fig. 27. Comparison between the link force-displacement curve of the updated theoretical model (solid green line) and the experimental results (dashed orange line) of the
second test group. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 28. Comparison between the energy dissipated during the four tests and the numerical results obtained by the theoretical model and the calibrated one.
66 F. Morelli et al. / Computers and Structures 176 (2016) 50–69

unilateral constraint offered by the contact with the column flange,

was applied to each non-linear spring.

4.2.2. Bolts in shear

Fig. 29. Mechanical model for connection 2.
4.2.1. Angle elements
The angle elements were modeled following the approach pro-
posed by [36]. They were so subdivided in 4 ‘‘strips”, see Fig. 30,
and each of them modeled as a non-linear beam element using a
fiber element discretization. An elasto-plastic stress-strain law,
evaluated adopting the nominal mechanical characteristics of the
angle elements, was applied to each fiber. The general model then
includes the axial/flexural behavior of each strip, considering all
the geometrical and mechanical non-linearities. Following the
approach proposed by [36], a gap element, representing the
The bolted connection between the angle elements and the dis-
sipative one is characterized by two important phenomena: (1) the
friction due to the sliding between the angle profile and the link
causing energy dissipation; (2) the pinching due to the contact
between the bolts and the link web and the consequent plastic
ovalization of the web holes of dissipative link.
The friction was modeled with a rigid-plastic behavior, see
Fig. 31a, where the friction force, FFr, was evaluated as the product
between the friction coefficient, lF, and the bolt pre-tensioning
force, Fp,C.
F Fr ¼ F p;C  lF ð12Þ
The friction coefficient lF was assumed equal to 0.30 (the sur-
faces of the steel elements were cleaned before the fastening of
the bolts but not specifically treated), while for the estimation of
the pre-tensioning force of the bolts the relation proposed by Euro-
code 3 [26] was used:
F p;C ¼ 0:7  f ub  AS ð13Þ
where fub is ultimate tensile stress of the bolt, while AS is the effec-
tive bolt cross section. The pinching due to the contact between
the link web and the bolt was modeled using the multilinear
force-displacement relationship showed in Fig. 31b. The initial free
sliding, dGap, before the contact between the bolt and the link
web was assumed equal to 1 mm, corresponding to the design

Fig. 30. Model of the angle elements (adapted from [36]).

Force
Force
FBe
KBe

FFr

Slip Gap Slip

(a) (b)
Fig. 31. Force-slip relation for the (a) friction and (b) bearing phenomena.
 F. Morelli et al. / Computers and Structures 176 (2016) 50–69 67

Fig. 32. Modeling of the angle elements-link connection.

400 400
F [kN] F[kN]
300 300

200 200

100 100
d [mm] d[mm]
0 0
-40 -30 -20 -10 0 10 20 30 40 -40 -30 -20 -10 0 10 20 30 40
-100 -100

-200 -200

-300 -300

-400 -400
Force - Column Disp. (Numerical) Force - Link Disp. (Numerical)
Force - Column Disp. (Experimental) Force - Link Displ. (Experimental)

(a) (b)
Fig. 33. Comparison between the experimental and the numerical results for test I (first test group) on coupling system configuration 2: (a) Force-Column axis displacement
and (b) Force-Top link displacement.

400 400
F [kN] F [kN]
300 300
200 200
100 100
d [mm] d [mm]
0 0
-40 -30 -20 -10 0 10 20 30 40 -40 -30 -20 -10 0 10 20 30 40
-100 -100
-200 -200
-300 -300
-400 -400
Force - Column Disp. (Numerical) Force - Link Disp. (Numerical)
Force - Column Disp. (Experimental) Force - Link Displ. (Experimental)
(a) (b)
Fig. 34. Comparison between the experimental and the numerical results for the test III (second test group) on coupling system configuration 2: (a) Force-Column axis
displacement and (b) Force-Top link displacement.

bolt-hole clearance. The bearing resistance, FBe, was evaluated as springs, one for the horizontal and one for the vertical direction,
indicated in Eurocode 3 [26], assuming a safety factor cM2 equal to 1. as schematically shown in Fig. 32. The force-slip relation of each
The connection between the angle element and the link pro- non linear spring is obtained as resultant of the friction and bearing
vided by each bolt was then schematized by two non-linear phenomena working in parallel.
68 F. Morelli et al. / Computers and Structures 176 (2016) 50–69
Fig. 35. Comparison between the energy dissipated during the I and III and the numerical results obtained by the numerical model.
4.2.3. Numerical results
Following the same procedure shown for the coupling system
configuration 1 and introducing the components relative to the
angle elements and bolts in shear, a component-based model for
the configuration 2 was developed and implemented in OpenSEES.
The comparison between the numerical model and the experimen-
tal results are reported in Figs. 33 and 34. The comparison was exe-
cuted with reference to both the system total displacement
(recorded by the LVDT sensor n.12 of Fig. 11) and the top link dis-
placement (evaluated at a distance of 50 mm from the link top by
the LVDT sensor n.11 of Fig. 11). The proposed model is capable of
representing with very good precision the behavior of both of them
catching the influence of the main physical phenomena, even if it
lightly underestimates the total dissipated energy, see Fig. 35. This
difference can be mainly due to the uncertainty in the estimation
of the friction coefficient.
5. Conclusions
In the present paper, the behavior of the dissipative coupling
systems of an innovative hybrid coupled shear wall systems
(HCSW), obtained coupling an r.c. wall with two side steel col-
umns, is presented and analyzed. In the framework of the INNO-
HYCO project [17], two different coupling system configurations
were proposed, as shown in Fig. 9. In order to investigate the influ-
ence of the friction between the angle elements and the link web,
two different values of the fastening torque were also used for the
two configurations. The experimental campaign, carried out on a
total of 8 specimens, 4 for each configuration, highlighted a good
dissipative capacity of the system without evident damages to
the concrete wall. The experimental tests showed also that the
angle connection has a great influence in the overall behavior; sev-
eral non-linear phenomena were highlighted, such as linear vari-
able friction and pinching due to the contact between the bolts
and the profiles edges (especially in the configuration 1, where
the pretension of the bolts was lower).
Two component-based mechanical models, one for each config-
uration, were realized. Each component was calibrated using
experimental results available in the technical literature. The resul-
tant global mechanical models were able to approximate in a very
satisfactory way the actual behavior of both the configurations in
terms of strength and loading stiffness. A suitable calibration using
the experimental results of first test group allowed to obtain, for
connection typology 1, a very good agreement in terms of energy
dissipation, too.
The model of the configuration 2 took accurately into account
the influence of the link-to-column connection realized by angle
elements and it was effectively able to numerically reproduce the
behavior of the whole system and the of its components. In fact,
both the force-column displacement and force-link displacement
relationships well described the experimental results. Some
improvements of the second model can be still perhaps executed,
measuring in a more accurate way the friction coefficient; however
this task was not inside of the scopes of the present study.
Both numerical and experimental results showed the impor-
tance of the friction force developed between the web link and
the angle element in the link-to-column connection. In fact, the
hysteretic behavior of configuration 2 is characterized by hys-
teretic cycles assuring a greater energy dissipation to the whole
HSCW system.
Acknowledgments
The research leading to these results has received funding from
the European Union’s Research Fund for Coal and Steel (RFCS)
research programme under grant agreement n° [RFSR-CT-2010-
00025] and from the Italian Department of Civil Protection within
the Italian Research Project RELUIS-DPC 2014-2018.
References
[1] Governo Italiano. Decreto Legge 6 giugno 2012, n. 74, Interventi urgenti in
favore delle popolazioni colpite dagli eventi sismici che hanno interessato il
territorio delle province di Bologna, Modena, Ferrara, Mantova, Reggio Emilia e
Rovigo, il 20 e il 29 maggio 2012, Gazzetta Ufficiale della Repubblica Italiana
n.131 of June 7 2012 (in Italian).
[2] Parlamento Italiano. Legge 1 agosto 2012, n. 122, Conversione in legge, con
modificazioni, del decreto legge 6 giugno 2012, n. 74, recante interventi
urgenti in favore delle popolazioni colpite dagli eventi sismici che hanno
interessato il territorio delle province di Bologna, Modena, Ferrara, Mantova,
Reggio Emilia e Rovigo, il 20 e il 29 maggio 2012, Gazzetta Ufficiale Ufficiale
della Repubblica Italiana n. 180, August 3, 2012 (in Italian).
[3] Braga F, Gigliotti R, Monti G, Morelli F, Nuti C, Vanzi I, et al. Post-seismic
assessment of existing constructions: evaluation of the shakemaps for
identifying exclusion zones in Emilia. Earthq Struct 2014;8:37–56. http://dx.
doi.org/10.12989/eas.2015.8.1.895.
[4] Braga F, Gigliotti R, Monti G, Morelli F, Nuti C, Vanzi I, et al. Speedup of post
earthquake community recovery. The case of precast industrial buildings after
the Emilia 2012 earthquake. Bull Earthq Eng 2014;12(5):2405–18. http://dx.
doi.org/10.1007/s10518-014-9583-.
 F. Morelli et al. / Computers and Structures 176 (2016) 50–69
[5] Braconi A, Bursi OS, Fabbrocino G, Salvatore W, Tremblay R. Seismic
performance of a 3D full-scale high-ductility steel-concrete composite
moment-resisting structure – Part I: Design and testing procedure. Earthq
Eng Struct Dyn 2008;37:1609–34. http://dx.doi.org/10.1002/eqe.829.
[6] Braconi A, Bursi OS, Fabbrocino G, Salvatore W, Taucer F, Tremblay R.
Seismic performance of a 3D full-scale high-ductility steel-concrete
composite moment-resisting structure – Part II: Test results and analytical
validation. Earthq Eng Struct Dyn 2008;37:1635–55. http://dx.doi.org/
10.1002/eqe.843.
[7] Braconi A, Salvatore W, Tremblay R, Bursi OS. Behaviour and modelling of
partial-strength beam-to-column composite joints for seismic applications.
Earthq Eng Struct Dyn 2007;36:142–61. http://dx.doi.org/10.1002/eqe.629.
[8] Badalassi M, Braconi A, Caprili S, Salvatore W. Influence of steel mechanical
properties on EBF seismic behaviour. Bull Earthq Eng 2013;11(6):2249–85.
http://dx.doi.org/10.1007/s10518-013-9498-4.
[9] Braconi A, Morelli F, Salvatore W. Development, design and experimental
validation of a steel self-centering device (SSCD) for seismic protection of
buildings. Bull Earthq Eng 2012;10:1915–41. http://dx.doi.org/10.1007/
s10518-012-9380-9.
[10] Pauley T, Santhakumarm AR. Ductile behavior of coupled shear walls. J Struct
Div-ASCE 1976;102(1):93–108.
[11] Aristizabal-Ochoa JD. Dynamic response of coupled wall systems. J Struct Div-
ASCE 1982;108(ST8):1846–57.
[12] Aristizabal-Ochoa JD. Seismic analysis of slender coupled walls. J Struct Div-
ASCE 1983;109(7):1538–52.
[13] Shahrooz BM, Remetter ME, Qin F. Seismic design and performance of
composite coupled walls. J Struct Eng 1993;119(11):3291–309. http://dx.doi.
org/10.1061/(ASCE)0733-9445(1993) 119:11(3291).
[14] El-Tawil S, Kuenzli CM, Hassan M. Pushover of hybrid coupled walls. II:
analysis and behavior. J Struct Eng 2002;128(10):1282–9. http://dx.doi.org/
10.1061/(ASCE)0733-9445(2002) 128:10(1282).
[15] Park WS, Yun HD. Seismic behaviour of coupling beams in a hybrid coupled
shear walls. J Constr Steel Res 2005;61:1492–524. http://dx.doi.org/10.1016/j.
jcsr.2005.04.006.
[16] Park WS, Yun HD. Seismic behaviour and design of steel coupling beams in a
hybrid coupled shear wall systems. Nucl Eng Des 2006;236:2474–84. http://
dx.doi.org/10.1016/j.nucengdes.2006.03.008.
[17] Dall’Asta A, Leoni G, Zona A, Hoffmeister B, Bigelow H, Degée H, et al.
INNOvative HYbrid and composite steel-concrete structural solutions for
building in seismic area – RFSR-CT-2010-00025 project – final
report. Brussels: European Commission; 2014.
[18] Kim JH, Ghaboussi J, Elnashai AS. Mechanical and informational modeling of
steel beam-to-column connections. Eng Struct 2010;32:449–58. http://dx.doi.
org/10.1016/j.engstruct.2009.10.007.
[19] Elnashai AS, Elghazouli AY. Seismic behaviour of semi-rigid steel frames. J
Constr Steel Res 1994;29:149–74. http://dx.doi.org/10.1016/0143-974X(94)
90060-4.
[20] Díaz C, Martí P, Victoria M, Querinb OM. Review on the modelling of joint
behaviour in steel frames. J Constr Steel Res 2011;67:741–58. http://dx.doi.
org/10.1016/j.jcsr.2010.12.014.
69
[21] Del Savio AA, Nethercot DA, Vellasco PCGS, Andrade SAL, Martha LF.
Generalised component-based model for beam-to-column connections
including axial versus moment interaction. J Constr Steel Res 2009;65(8–
9):1876–95. http://dx.doi.org/10.1016/j.jcsr.2009.02.011.
[22] Dall’Asta A, Leoni G, Tassotti L, Zona A. Seismic performance of innovative steel
and concrete hybrid coupled shear walls. In: COMPDYN 2015, the 5th
ECCOMAS thematic conference on computational methods in structural
dynamics and earthquake engineering, Crete Island, Greece, 25–27 May 2015..
[23] Manfredi M, Morelli F, Salvatore W. An enhanced component based model for
steel links in hybrid structures. In: Proceedings of COMPDYN 2015, 5th
ECCOMAS thematic conference on computational methods in structural
dynamics and earthquake engineering, Crete Island, Greece, 25–27 May 2015.
[24] Zona A, Degée H, Leoni G, Dall’Asta A. Ductile design of innovative steel and
concrete hybrid coupled walls. J Constr Steel Res 2016;117:204–13. http://dx.
doi.org/10.1016/j.jcsr.2015.10.017.
[25] EN 1998-1. Eurocode 8: design of structures for earthquake resistance, part 1:
General rules, seismic actions and rules for buildings. Brussels: European
Committee for Standardization (CEN); 2004.
[26] EN 1993-1-8. Eurocode 3: design of steel structures, part 1–8: Design of joints.
Brussels, Belgium: European Committee for Standardization (CEN); 2005.
[27] Mazzoni S, McKenna F, Scott MH, Fenves GL, et al. Opensees command
language manual; 2006.
[28] EN 1994-1-1 Eurocode 4: design of composite steel and concrete structures –
part 1: General rules and rules for buildings. Brussels, Belgium: European
Committee for Standardization (CEN); 2005.
[29] Aribert JM, Dinga ON, Lachal A. Modélisation et étude expérimentale
d’assemblages de type poutre mixte-poteau en acier. Construction
Metallique; 2000, No. 1.
[30] Park YJ, Ang AHS. Mechanistic seismic damage model for reinforced concrete. J
Struct Eng 1985;111(4):722–39. http://dx.doi.org/10.1061/(ASCE)0733-9445
(1985) 111:4(722).
[31] Mander JB, Priestley MJN, Park R. Theoretical stress-strain model for confined
concrete. J Struct Eng 1988;114(8):1804–25. http://dx.doi.org/10.1061/(ASCE)
0733-9445(1988) 114:8(1804).
[32] Popovic S. A numerical approach to the complete stress-strain curves for
concrete. Cem Concr Res 1973;3(5):583–99. http://dx.doi.org/10.1016/0008-
8846(73)90096-3.
[33] Federal Emergency Management Agency, FEMA. Prestandard and commentary
for the seismic rehabilitation of buildings. Washington, D.C.: FEMA 356. p.
2000.
[34] EN 1998-3. Eurocode 8: Design of structures for earthquake resistance – Part
3: Assessment and retrofitting of buildings. Brussels: European Committee for
Standardization (CEN); 2005.
[35] D’Aniello M, Landolfo R, Piluso V, Rizzano G. Ultimate behavior of steel beams
under non-uniform bending. J Constr Steel Res 2012;78:144–58. http://dx.doi.
org/10.1016/j.jcsr.2012.07.003.
[36] De Stefano M, De Luca A, Astaneh-Asl A. Modeling of moment-rotation cyclic
response of double-angle connections. J Struct Eng 1994;120(1):212–29.
http://dx.doi.org/10.1061/(ASCE)0733-9445(1994) 120:1(212).