CYCLIC BEHAVIOR OF STEEL BEAM-TO-COLUMN JOINTS:

GOVERNING PARAMETERS OF
WELDED AND BOLTED CONNECTIONS
Luis CALADO1 , Elena MELE2
1 DECivil, Instituto Superior Tecnico, Lisbon, Portugal
2 DAPS, Università degli Studi di Napoli ‘Federico II’, Naples, Italy

Abstract

In this paper the results obtained from the experimental tests on two alternative
connection solutions (fully welded and top-and-seat with web angles) designed for
the same beam-to-column joint are presented. The test program was planned with
the aim of assessing the comparative behaviour of bolted and welded connections,
and for defining the effect of the column size and of the PZ design on the
behaviour of the two types of connection, varying the applied loading history.

1. INTRODUCTION

One of the most important lessons derived form the Northridge and Hyogo Ken-Nanbu
earthquakes has been the particular vulnerability of the beam-to-column connections in SMRF
building structures. Starting from these observations, several research programs have been
world-wide developed in order to enrich the experimental data base for assessing the major
parameters influencing the cyclic behaviour of beam-to-column connections. In addition the
need of accounting for the connection behaviour on the global structure performance has been
underlined and theoretical models for reliably predicting stiffness, strength and deformation
capacity of the connections have been developed.
Since recently bolted connections, in particular top and seat with web angles (TSW)
connections, have not been considered appropriate in seismic applications, due to the partial
strength and semirigidity characteristics. Therefore, TSW, though extensively investigated in the
monotonic range, as reported in Kishi and Chen, (1), received less attention in the cyclic range.
Only recently it has been pointed out (Astaneh, (2); Elnashai et Al., (3)) that the dynamic
behaviour semirigid frames can be particularly favourable due to the period elongation, related
to the connection flexibility and to the damping increase, related to highly dissipative friction
mechanism deriving from a proper "slip capacity design". Both these effects act as a sort of self-
isolation of the frame structure, thus leading to remarkable reduction of the seismic actions. It is
worth to emphasise that also in the context of the SAC Steel Project, started immediately after
the Northridge earthquake to address the specific problem of beam-to-column connections, a
great interest in bolted configurations as alternative to the standard welded connections
(Roeder, (4)), can be found.
In this research framework, a wide experimental program on different types (welded and bolted)
of beam-to-column connections has been carried out at the Material and Structures Test
Laboratory of the Instituto Superior Técnico of Lisbon. The experimental tests have been

194
performed on specimens representative of frame structure beam-to-column joints close to the
ones typical of European design practice. Some preliminary experimental results on the welded
connections have been presented in Mele et Al., (5) and Calado et Al., (6). In this paper a
comparative assessment of the cyclic behaviour of welded and bolted beam-to-column
connections is provided.

2. OBJECT AND AIMS OF THE PAPER

In this paper the results obtained from the experimental tests on two alternative connection
solutions, namely top and seat with web angle (TSW) and fully welded connections (WW),
designed for the same beam-to-column joints are presented. Six experimental tests have been
executed on six (three welded and three bolted) different series of specimens, for a total of 36
tests. The test program was planned with the aim of assessing the comparative behaviour of
bolted and welded connections, and for defining the effect of the column size and of the PZ
design on the behaviour of the two types of connection, varying the applied loading history. In
addition the accuracy of the "component method" of the Eurocode 3 Annex J (CEN, (7)) in
predicting stiffness and strength of the connections is evaluated through the comparison with
experimental monotonic results, and the possibility of extrapolating such theoretical prediction to
the cyclic range is examined.

3. THE EXPERIMENTAL PROGRAM

3.1 The specimens
Two series of full scale specimens have been designed and tested, namely a WW specimen
series (BCC5, BCC6 and BCC8) and a TSW specimen series (BCC9, BCC7 and BCC10). The
specimens of the two series are T-shaped beam-column subassemblages, consisting of a 1000
mm long beam and a 1800 mm long column. The material used for the columns, beams, and
angles is steel S235 JR, with nominal values of yield and ultimate stress respectively equal to
fy=235 MPa and fu=360 MPa. In each series the cross section of the beam is the same
(IPE300), while the column cross section has been varied, being respectively HE160B for the
BCC5 (WW) and BCC9 (TSW) specimens, HE200B for the BCC6 (WW) and BCC7 (TSW)
specimens, and HE240B for the BCC8 (WW) and BCC10 (TSW) specimens. In both the series,
the continuity of the connection through the column has been ensured by horizontal 10 mm thick
plate stiffeners, fillet welded to the column web and flanges.
3.1.1 WW specimens
In the WW specimens, the beam flanges have been connected to the column flange by means
of complete joint penetration (CJP) groove welds, while fillet welds have been applied between
both sides of the beam web and the column flange. The flexural strengths of the beam, column
and panel zone, have been computed on the basis of the nominal yield stress and are reported
in table 1.
Mpb Mpc Mp,PZ
BCC5 147.6 83.2 91.1
BCC6 147.6 151.1 132.4
BCC8 147.6 247.5 182.9

Table 1: moment capacities (in kNm) of the WW specimens

195
From the simple comparison among the nominal plastic moments reported in table 1, it can be
observed that in the three WW specimens the weakest component of the joint configuration is
respectively: the column for the BCC5 specimen, the panel zone for the BCC6 specimen, the
beam for the BCC8 specimen.
3.1.2 TSW specimens
In the BCC9, BCC7 and BCC10 (TSW) specimens, 120x120x10 angles have been adopted.
Two rows of bolts are placed on each leg of the flange angles, while on the legs of the web
angles there is only one row of two bolts. The bolts are M16 grade 8.8 (yield stress fyb=640
MPa, ultimate stress fub=800 MPa, As=157 mm2), preloaded according to the EC3 provisions,
i.e. at FP,CD= 0.7 fub As = 87.9 kN.
It is well known that two major phenomena characterise the behaviour of the TSW connection:
the slippage of bolts and the yielding of the tension angle. For the TSW specimens herein
described the bending moment corresponding to bolt slippage and angle yielding are reported in
table 2, together with the beam and column moment capacities. From the comparison between
the bending moments corresponding to bolt slippage and angle yielding, it derives that the
specimens are "slip critical" connections, since slippage of top and seat angle bolts occurs at a
load level higher than the one corresponding to yielding of the tension angle.
Mslip My,angle Mpb Mpc
BCC9 32 - 47.5 23.3 147.6 83.2
BCC7 32 - 47.5 23.3 147.6 151.1
BCC10 32 - 47.5 23.3 147.6 247.5
Table 2: threshold moment capacities (in kNm) of the TSW specimen

3.2 Loading histories

For each of the six specimen series, the experimental program consisted of six tests, one
monotonic test and five cyclic tests, for a total of 36 tests. The cyclic tests have been carried out
by applying both constant and step-wise increasing amplitude displacement histories. This latter
test type has been carried out according to the basic loading history recommended in ECCS,
(8). In this paper only the increasing rotation tests is reported.

3.3 Experimental set-up and specimen instrumentation

The test set-up, mainly consists in a foundation, a supporting girder, a reaction r.c. wall, a power
jackscrew and a lateral frame. The power jackscrew (capacity 1000 kN, stroke r 400mm) is
attached to a specific frame, pre-stressed against the reaction wall and designed to
accommodate the screw backward movement. The specimen is connected to the supporting
girder through two steel elements. The supporting girder is fastened to the reaction wall and to
the foundation by means of pre-stressed bars. An automatic testing technique was developed to
allow computerised control of the power jackscrew, of the displacement and of all the
transducers used to monitor the specimens during the testing process. Specimens have been
instrumented with electrical displacement transducers (LVDTs), which record the displacement
histories at several points in order to obtain a careful documentation of the various phenomena
occurring during the tests. The same arrangement of LVDTs has been adopted for the three
WW and the three TSW specimen series.

196
4. CYCLIC TESTS
4.1 Premise
In the following the experimental results obtained in the test program are provided. In particular
the cyclic behaviour and the failure modes observed for the six sets of specimens are
described, and the moment rotation hysteresis loops obtained in the increasing amplitude tests
are provided. In the moment rotation hysteresis loops hereafter presented, reference is made to
three different values of rotation, namely: (1) the “unprocessed” total rotation given by the
applied interstory drift angle d/H; (2) the beam rotation )b and (3) the panel zone rotation )PZ,
both obtained through the measured LVTDs displacements of the specimens. Correspondingly,
in the M-d/H and M-)PZ experimental curves the moment is evaluated at the column centreline,
while in the Mb-)b curves the moment is evaluated at the column face.
4.2 WW specimens
In figure 1 (a) the moment - total rotation (M-d/H) experimental curves resulting from the
BCC5C, BCC6C and BCC8D tests (cyclic increasing stepwise amplitude) are plotted, while in
figure 1 (b) both the corresponding moment – beam plastic rotation )b,pl and the moment - panel
zone rotation )PZ curves are plotted. The beam plastic rotation has been obtained through the
measured displacements at the beam instrumented section by subtracting the contributions of
the beam and column elastic rotations as well as of the panel zone distortion.
4.2.1 BCC5
As can be derived from the curves reported in figure 1 (a) and (b), and as demonstrated also
throughout the experimental program, the cyclic behaviour of the specimen BCC5 is
characterised by a great regularity and stability of the hysteresis loops up to failure, with no
deterioration of stiffness and strength properties. In the very last cycle the specimen has
collapsed with a sudden and sharp reduction of strength, due to fracture initiated in the beam
flange and propagated also in the web. During the tests, significant distortion of the joint panel
zone has been observed, while not remarkable plastic deformation in the beam occurred.
4.2.2 BCC6
Throughout the test program, two different kinds of cyclic behaviour have been observed for the
BCC6 specimens. In some cases the behaviour of the specimens is close to the behaviour
observed for the BCC5 type, with almost no deterioration of the mechanical properties up to the
last cycle, during which the collapse occurred. For the other tests a gradual reduction of the
peak moment at increasing number of cycles is evident. In these cases, starting from the very
first plastic cycles, local buckling of the beam flanges occurred, and a well defined plastic hinge
has formed in the beam. The contribution of the panel zone deformation has not been as
significant as in the BCC5 specimen type.

197
 400 400

M [kNm]

M [kNm]
BCC5C ) b,pl
200 200 ) PZ

0 0
10% -5% 0% 5% 10% 10% -5% 0% 5% [rad%] 10%
-200 -200
d/H [rad%]
-400 -400

400 400
M [kNm]

M [kNm]
) PZ
BCC6C
200 200

0 0
10% -5% 0% 5% 10% 10% -5% 0% 5% [rad%] 10%
-200 -200
d/H [rad%] ) b,p
l
-400 -400
400 400
) PZ
M [kNm]

M [kNm]
BCC8D
200 200

0 0
10% -5% 0% 5% 10% 10% -5% 0% 5% [rad%] 10%
-200 -200
) b,pl
d/H [rad%]
-400 -400

Figure 1: WW specimens: (a) moment-global rotation curves; (b) moment-beam plastic

rotation and moment-panel rotation curves

4.2.3 BCC8
The hysteresis loops obtained from the tests on the BCC8 specimens show a gradual reduction
of the peak moment starting from the second cycle, where the maximum value of the applied
moment has been usually registered. This deterioration of the flexural strength of the connection
is related to occurrence and spreading of local buckling in the beam flanges and web. A well
defined plastic hinge in the beam has formed in all the tested specimens. In the specimens
BCC8 the panel zone deformation has not been remarkable, and the plastic deformation mainly
took place in the beam.

4.3 TSW specimens

In figure 2 (a) the moment - total rotation (M-d/H) experimental curves resulting from the
BCC9D, BCC7C and BCC10C tests (cyclic increasing stepwise amplitude) are plotted, while in
figure 2 (b) both the corresponding moment – beam rotation )b and the moment - panel zone
rotation )PZ curves are plotted. As can be derived from the curves reported in figure 2, the
shape of hysteresis loops of the three TSW specimens is very similar. The cyclic behaviour, the
phenomena observed during the tests and the collapse modes are the same for the three
specimen series, thus the following unique paragraph is devoted to describe the above issues
for the three specimens.

198
4.3.1 BCC9 / BCC7 / BCC10
The cyclic behaviour of the TSW connections is characterised by bolt slippage and yielding and
spreading of plastic deformation in the top and bottom angles, cyclically subjected to tension.
Plastic ovalization of the bolt holes have also been observed mainly in the leg of the angle
adjacent to column flange. The experimental curves, typical of this type of connection, shows
pinched hysteresis loops, with a large slip plateau (very low slope of the experimental curve)
and subsequent sudden stiffening. In fact when the specimen position is at d = 0, due to the
concomitant effects of bolt slippage, hole ovalization and the plastic deformation of the angle
legs adjacent to the column flange, the beam is completely separated from the column (gap
open).

150 150

Mb [kNm]
M [kNm]

) PZ )b
BCC9D 100 100

50 50

0
d/H 0
-10% -5% 0% 5% 10% -10% -5% 0% 5% 10%
-50 -50

-100 -100

-150 -150

150 150
) PZ
Mb [kNm]
M [kNm]

BCC7C 100 100 )b

)
50 50
d/H
0 0
-10% -5% 0% 5% 10% -10% -5% 0% 5% 10%
-50 -50

-100 -100

-150 -150

150 150
) PZ
Mb [kNm]
M [kNm]

BCC10C 100 100 )b

50 50
d/H
0 0
-10% -5% 0% 5% 10% -10% -5% 0% 5% 10%
-50 -50

-100 -100

-150 -150

Figure 2 TSW specimens: (a) moment-global rotation curves, (b) moment-beam net
rotation and moment - panel rotation curves

At large applied displacements, which impose large rotations to the connection, the contact of
the compression angle and the beam web to the column flange (gap closure) give rise to
sudden stiffening of the connection, which is evident in the experimental curves. No significant
rotation of the column and distortion of the panel zone has been observed throughout the
experimental tests carried out on the three specimens. At each step on the test, slight
deterioration of the joint resistance in the three applied cycles can be observed in the
experimental curves, mainly due to yielding and spreading of plastic deformation in the top and
bottom angles, cyclically subjected to tension. In all the test carried out on the three specimen

199
series, the collapse of the connection occurred due to fracture in the leg angle located on the
beam flange, immediately after the fillet. Negligible scatters can be observed in the moment
capacity of the three connection series, as it is expected, since the inelastic behaviour of the
connection is governed by the angle. Also the maximum values of global rotation experienced
by the specimens is the same for the BCC9 and BCC10 series, and slightly larger for the BCC7
one.

5. MONOTONIC TESTS
The moment rotation curves obtained from the monotonic tests carried out on the six specimens
are presented in figures 3 (a) and (b) which respectively report the results of the WW and TSW
specimens. In these curves the moment is evaluated at column centreline and the rotation is
given by the total interstory drift angle (d/H). In each figure also the moment panel zone rotation
)PZ are reported. From the experimental results on the two series of specimen the effect both of
the connection typology (TSW and WW) and of the column cross section (HE160B, HE200B,
HE240B) can be derived.
By comparing the two series of experimental curves it must be noticed that the three WW
specimens show significant differences in the initial stiffness, maximum strength and
deformation capacity, thus confirming the strong effect of the column cross section size already
observed in the cyclic tests. On the contrary, the three TSW specimens present quite close
experimental responses. This difference between the behaviour of WW and TSW specimens is
mainly related to the design of the specimens, since in the TSW connections the weakest
component is the same in the three specimens (the angle in tension), thus the beam, column
and panel zone strength ratios does not affect the response of the specimens. Slight scatters
can be observed in the initial stiffness, due to the different column and panel zone deformability,
but the nonlinear portion of the curve and the maximum bending moment are very similar.

300 300 M [kNm]

BCC8
M [kNm]

250 250

200 BCC5 200 )PZ BCC10

BCC9
150 150
BCC7
100 )PZ 100
BCC6 50
50
d/H [rad%]
0 0 d/H [rad%]
0% 5% 10% 15% 20% 25% 30% 0% 5% 10% 15% 20% 25% 30%

Figure 3 Monotonic experimental curves: (a) WW specimens; (b) TSW specimens

As already evidenced, the behaviour of the WW connections is affected by the column

dimensions since the three combinations of beam and column framing in the joint give rise to
panel zone strength values respectively: smaller than, approximately equal to and larger than
the plastic moment of the beam, for the BCC5, BCC6 and BCC8 specimens. These
observations are confirmed by analysing the main test data provided in table 3.
In particular it can be noticed that while in the WW specimens the panel zone distortion )PZ
significantly contributes to the specimen global rotation d/H (at the maximum value of the

200
bending moment registered in the relevant test), though at different extent in the three
specimens, a completely different order of magnitude of this contribution is registered in the
TSW connections. For the TSW specimens, instead, the rotations due to bolt slippage )slip
computed on the basis of the LVDTs measured displacements constitutes a major contribution
to the total rotation d/H.

M [kNm] Mb [kNm] d/H [%] )PZ [%] )b,pl - )PZ [%]

BCC5 214 196 21.4 16.5 0.66
WW

BCC6 231 207 10.5 4.5 4.9

BCC8 292 251 8.3 3.3 5.3
M [kNm] Mb [kNm] d/H [%] )PZ [%] )slip [%]
BCC9 135 122.5 10.8 2.4 5.2
TSW

BCC7 144 127 10.5 1.15 6.8

BCC10 146 126 9.4 0.19 4.8

Table 3: main experimental data from the monotonic tests on the WW and TSW specimens

6. PREDICTION OF THE MONOTONIC BEHAVIOUR

Comparing the monotonic and the cyclic curves both for the BCC5 WW specimen, and for the
BCC9 TSW specimen, it can be observed that the monotonic curve perfectly envelope the cyclic
one. Thus it seems of primary importance assessing the prediction capacity of the monotonic
behaviour offered by the available numerical models.
In figures 4 and 5 the monotonic curves, respectively of the BCC5 and BCC9 specimens, are
compared to theoretical moment-rotation curves. In figure 4 (a) and (b) the reference
experimental curves report the moment at the column centreline, M, vs the different rotation
values, i.e.: the global rotation d/H, the beam and the PZ rotations, )b and )PZ, obtained from
the LVDT measured displacements, and depurated of: rigid rotation of the specimen supports,
elastic rotations of the beam and column. The theoretical curve is the one obtained through the
application of the "component method" as outlined in the EC3 Annex J. In the figure 5 (a) and
(b) the reference experimental curve reports the moment at the column centreline, M, vs the
global rotation d/H. The theoretical curves obtained through: the application of the EC3 Annex J
procedure, the model proposed by Kishi & Chen, (1) and the simplified approach proposed by
De Luca et Al., (9) are also shown in the figure. Both the figure 4 (b) and 5 (b) provide the detail
of the curves up to 1% rotation.
6.1 WW BCC5 specimen
In the case of welded connections stiffened by means of continuity plates as the ones described
in this paper, the Annex J procedure suggests to account for only one component in the
evaluation of the rotational stiffness of the joint, i.e. the column web panel in shear, while the
flexural strength of the joint is given by the resistance, multiplied by the lever arm, of the
weakest among five basic components, namely the panel in shear, the column web in tension
and in compression, the column flange in bending and the beam flange and web in
compression. For the BCC5 specimen the panel shear resistance resulted the minimum one,
thus confirming the governing role of the PZ in the deformation and failure modes of the BCC5

201
specimen observed in the tests. The moment rotation curve obtained through the Annex J
procedure, improved as suggested in Faella et Al., (10), shows a reasonable accuracy both in
terms of strength (figure 4 (a)) and in terms of initial stiffness (figure 4 (b)). By observing that the
monotonic curve of the BCC5 specimen envelopes the experimental curves obtained under
cyclic loading, it seems that a reliable prediction of the joint monotonic behaviour is of major
importance also for assessing the cyclic performance.

300 BCC5 150 BCC5

M [kNm]

M [kNm]
EC3 EC3
250 ) PZ 125
)b exp
)b exp
200 100

150 75
) PZ
100 50
d/H d/H
50 25

0 rotation [rad%] 0 rotation [rad%]

0% 3% 6% 9% 12% 15% 0.0% 0.2% 0.4% 0.6% 0.8% 1.0%
Fi
Figure 4 experimental vs theoretical curves: WW BCC5 specimen
M [kNm]

M [kNm]
150 BCC9 75 BCC9
De Luca, Eh/E=1/50
exp
exp
100 50
EC3 EC3

Chen
50 25 De Luca, Eh/E=1/50
Chen
De Luca, Eh/E=0 De Luca, Eh/E=0

0 d/H [rad%] 0 d/H [rad%]

0% 2% 4% 6% 8% 10% 0.0% 0.2% 0.4% 0.6% 0.8% 1.0%

Figure 5 experimental vs theoretical curves: TSW BCC9 specimen

6.2 TSW BCC9 specimen

In the experimental curve of figure 5 (a) four major branches can be identified: an initial linear
branch; a second, remarkably nonlinear branch, going approximately from to 35 to 55 kNm,
where the both the slip of bolts and yielding of the angle occur; a third sensibly linear branch
which arrives at 85 kN, characterised by a reduced slope, where hardening of the angle occurs
up to the failure of the first row of bolts; and a final, always approximately linear branch but with
a reduced slope, where the angle, also due to geometry variation related to the loss of a bolt
row, is further stretched and carry the load in simple tension. It is worth to notice that the ratio of
the slope of the third to the first branch of the curve is close to the ratio Eh/E derived from the
tension tests on the angle coupon, while the maximum bending moment registered in the test at
the column face is close to the moment corresponding to the pure tensile bearing capacity of the
angle. Similar considerations still hold by observing the other TSW experimental curves.
The prediction models do not represent all these phenomena, since the ultimate bearing
capacity developed by the connection is due to a deformation state which is not compatible with
the global stability of the structure. Therefore the purpose of the models is the prediction of the
initial stiffness, the plastic threshold and the consequent nonlinear branch up to a given

202
deformation value, which should be compatible with the structure and checked against the
deformation capacity of the connection. In this perspective, the EC3 model, applied with the
improvements suggested by Faella et Al., (11), matches quite closely the initial stiffness, slightly
overstimates the plastic threshold given by 2/3 of the resistance MjRd, and provides a reasonable
value of MjRd. In addition, by considering a hardening slope equal to 1/55 and accounting for a
strength ratio fu/fy=1.48, as derived from the angle coupon tests, also a portion of the fourth
branch can be match. The model of Kishi and Chen (1) also leads to a good estimation of
strength, but does not match the initial stiffness, since it does not allow for bolt preloading. The
simplified approach of De Luca et Al., (9) provides a range which includes the experimental
curve. A deeper assessment of the prediction capacity of the different models is not possible
here due to space limitations.

7. SUMMARY AND CONCLUSIVE REMARKS

The experimental results obtained in this research allow to define the collapse modes, the
rotation capacity and the ultimate bending strength of bolted and welded beam-to-column
connections. In this paper the major aspects governing the cyclic and the monotonic behaviour
of bolted (TSW) and welded (WW) connections have been evidenced against experimental
results. It has been shown that the panel zone does not affect the behaviour of the TSW
connections, which instead is mainly related to the tension angle geometry and strength
properties. On the contrary the panel zone has demonstrated to affect at large extent all the
response parameters (stiffness, strength and deformation capacity) of welded connections.
Finally the application of the EC3 Annex J model (with some modifications suggested by other
Authors) to two single examples of each connection typology, has shown to provide reasonable
results.

ACKNOWLEDGMENTS
The financial support of MURST (Italy) and FCT (Portugal) is gratefully acknowledged.

NOTATION
Abbreviations
TSW = top and seat with web angles connection
WW = fully welded connection
PZ = panel zone
CJP = complete joint penetration

Symbols
fy = yield stress
fu = ultimate stress
fyb = yield stress of bolts
fub = ultimate stress of bolts
FP;CD = bolt preloading force
M = bending moment at column centerline (experimental value)

203
Mb = bending moment at column face (experimental value)
Mslip = bending moment corresponding to bolt slippage (theoretical value)
My,angle = bending moment corresponding to angle yielding (theoretical value)
Mpb = plastic moment capacity of the beam section (theoretical value)
Mpc = plastic moment capacity of the column section (theoretical value)
MpPZ = plastic moment capacity of the panel zone (theoretical value)
d = applied displacement
H = distance between the column centreline and the beam tip
d/H = applied interstory drift angle
)b = beam rotation
)b,pl = beam plastic rotation
)PZ = panel zone rotation
)slip = rotation due to bolt slippage
E = Young modulus
E / Eh = hardening ratio

REFERENCES
1. Kishi, N., Chen, W.F.(1990) Moment rotation relations of semirigid connections with angles. J. of
Struct. Eng. ASCE, Vol..116.
2. Astaneh-Asl, A. (1995). “Seismic design of bolted steel moment-resisting frames.” Steel Tips,
Structural Steel Educational Council, Moraga, California, July.
3. Elnashai, A.S., Elghazouli, A.Y., Denesh-Ashtiani, F.A. (1998) Response of semirigid steel frames to
cyclic and earthquake loads. J. of Struct. Eng. ASCE, Vol.124, No.8.
4. Roeder, C.W. (1998). Design models for moment resisting steel constructions. Proc. SEWC 1998.
Paper T158-4
5. Mele, E., Calado, L., De Luca, A. (1999) Experimental behaviour of beam-to-column welded
connections: effect of the panel zone design. Proc. XVII C.T.A., Napoli, Oct.
6. Calado, L., Mele, E., and De Luca, A. (1999). “Experimental investigation on the cyclic behaviour of
welded beam-to-column connections”. Proc. 2nd European Conf. on Steel Structures, PRAHA,
Czech Republic, Paper No.215.
7. CEN TC250/SC3-PT9. (1997). Eurocode 3, Part 1.1. Rev. Annex J. “Joints in building frames”. Ed.
Approved Draft: Jan.
8. ECCS (1986). “Seismic Design. Recommended testing procedure for assessing the behaviour of
structural steel elements under cyclic loads”. Tech. Comm. 1 - Structural Safety and Loadings,
TWG1.3 – Rep. No.45.
9. De Luca, A., De Martino, A., Pucinotti, R., Puma, R. (1995) (Semirigid ?) Top & Seat angle
connections: Critical review of Eurocode 3 approach.. Proc. XV C.T.A., Riva del Garda, Oct.
10. Faella C., Piluso, V. and Rizzano, G. (1995). Modelling of the moment-rotation curve of welded
connections: proposals to improve Eurocode 3 Annex J. Proc. XV C.T.A., Riva del Garda, Oct., 405-
418.
11. Faella C., Piluso, V., Rizzano, G. (1996). Prediction of the flexural resistance of bolted connections
with angles. Proc. IABSE Coll. on Semirigid Structural Connections, Istanbul, Sept.

204