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

net/publication/239390903

Experimental Investigation on European Welded Connections

Article in Journal of Structural Engineering · October 2003

DOI: 10.1061/(ASCE)0733-9445(2003)129:10(1301)

CITATIONS READS
27 626

3 authors, including:

Elena Mele Luis Calado

University of Naples Federico II University of Lisbon
65 PUBLICATIONS 1,462 CITATIONS 103 PUBLICATIONS 890 CITATIONS

SEE PROFILE SEE PROFILE

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

Nature inspired structures for tall buildings: patterns, modeling, analysis View project

INNOSEIS - Valorization of innovative anti-seismic devices - RFCS 2015-709434 View project

All content following this page was uploaded by Elena Mele on 07 July 2015.

The user has requested enhancement of the downloaded file.

 Experimental Investigation on European Welded
Connections
Elena Mele1; Luis Calado2; and Antonello De Luca, M.ASCE3

Abstract: In this paper the results of a test program devoted to the assessment of the cyclic behavior of beam-to-column welded
Downloaded from ascelibrary.org by Universita Studi Napoli Fed Ii on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

connections are presented. The specimens are representative of the common European practice, which differs from both U.S. and Japan
practice, due to the size of the connected members and to the welding technique and details. Three specimen groups, characterized by
different values of the relative column-beam-panel zone strengths, are designed and tested under different loading histories. In particular
six tests 共five cyclic and one monotonic兲 have been carried out on each group of specimens, encompassing a total of 18 tests. The tests
have evidenced the effect of column-to-beam strength ratio and of panel zone design on the cyclic behavior and failure modes of the
connections. Furthermore the dependence of the absorbed energy and cumulative plastic rotation of the joint upon the applied loading
history is shown. Finally design implications are suggested on the basis of the test results.
DOI: 10.1061/共ASCE兲0733-9445共2003兲129:10共1301兲
CE Database subject headings: Cyclic loads; Connections, welded; Failure modes; Europe; Experimentation.

Introduction This paper describes a part 共the one devoted to welded con-
nections兲 of a wide experimental program carried out at the Ma-
The 1994 Northridge and 1995 Hyogoken-Nanbu earthquakes can terial and Structures Test Laboratory of the Instituto Superior Téc-
be considered milestones since the research during the subsequent nico of Lisbon on different types 共welded and bolted兲 of beam-
1990s has been strongly inspired by the need of understanding to-column connections. The experimental tests have been
and explaining some ‘‘unexpected’’ aspects of the seismic re- performed on specimens representative of moment resisting frame
sponse exhibited by steel buildings during these events. In par- 共MRF兲 beam-to-column joints typical of European design prac-
ticular the extensive damage and brittle failure modes which oc- tice. In order to provide a reference framework to the experimen-
curred at beam-to-column connections of frame structures were of tal study, some major items of the European current practice in
major concern. designing the frame structural systems and in detailing the welded
Several experimental programs on beam-to-column connec- beam-to-column connections 共Mele 2002兲 are reported in the fol-
tions have been carried out in the United States 共Mahin et al. lowing.
1996; Malley 1998; Kunnath and Malley 2002兲 and in Japan • In Europe the design trend of adopting partial perimeter MRFs
共JCSS 1997; Tanaka et al. 1997; Nakashima et al. 1998; Suita is not so widely established as in the United States; further-
et al. 1998兲 in order to understand the causes of the observed more, building configurations with smaller bay spans 共typi-
damage, to assess the major parameters affecting the cyclic be-
cally 5–7 m兲 than in the United States are usually adopted. For
havior of beam-to-column connections, and to suggest modifica-
the above reasons, beams are usually shallower than the ones
tions in the connection layout for improving the seismic perfor-
adopted in the current U.S. design practice 共Europe beam
mance. The renewed interest in the seismic behavior of
depths, h b ⫽300– 450 mm versus U.S. beam depths, h b
connections has also affected the recent research trends in Europe,
⫽600– 1,000 mm).
where several experimental projects have been undertaken 共Mele
• The beam section is typically welded to the column in the
et al. 1997; Plumier et al. 1998; Taucer et al. 1998; Castiglioni
shop: complete joint penetration welds are applied at the beam
et al. 2000; Calado and Mele 2000兲.
flanges and fillet welds are applied at both sides of the beam
1
web. Continuity plates are commonly utilized for stiffening the
Associate Professor, Structural Analysis and Design Department
column panel zone 共PZ兲, while doubler plates are generally not
共DAPS兲, Univ. of Naples ‘‘Federico II,’’ P. le Tecchio 80, 80125 Naples,
Italy. E-mail: elenmele@unina.it present, thus leading to a substantial contribution of PZ to the
2
Associate Professor, Civil Engineering Dept. 共DECivil兲, Instituto Su- overall deformation of the connection.
perior Tecnico, Av. Rovisco Pais, 1096 Lisboa Codex, Lisbon, Portugal. • The typical welding technique used in Europe is the manual
3
Professor, Structural Analysis and Design Department 共DAPS兲, Univ. metal-arc welding 共MMAW, also known as shielded metal arc
of Naples ‘‘Federico II,’’ P. le Tecchio 80, 80125 Naples, Italy. welding, SMAW兲 although gas shielded welding processes,
Note. Associate Editor: Christopher J. Earls. Discussion open until typically metal inert gas and flux cored arc welding, are gradu-
March 1, 2004. Separate discussions must be submitted for individual ally superseding MMA welding for reasons of productivity
papers. To extend the closing date by one month, a written request must
and cost effectiveness 共The Steel Construction Institute 1992兲.
be filed with the ASCE Managing Editor. The manuscript for this paper
was submitted for review and possible publication on July 17, 2001; A specific procedure for minimizing the porosities in the welds
approved on December 16, 2002. This paper is part of the Journal of is often implemented: After having welded the beam flanges
Structural Engineering, Vol. 129, No. 10, October 1, 2003. ©ASCE, from the top, the root opening is cleaned of impurities from
ISSN 0733-9445/2003/10-1301–1311/$18.00. the bottom, and the opening is filled by means of fillet welds

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / OCTOBER 2003 / 1301

J. Struct. Eng. 2003.129:1301-1311.

 specimens, labeled as BCC5, BCC6, and BCC8 共Fig. 1兲, have
been selected on the basis of a typical European beam section
共IPE300, which could be considered equivalent to W12⫻30 U.S.
wide flange section兲, usually adopted on European spacing 共beam
lengths typically comprised between 5 and 7 m兲, and three differ-
ent column sections:
Fig. 1. Layout of specimens • HE160B 共possibly equivalent to W8⫻24) for the BCC5 se-
ries,
共Taucer et al. 1998兲. This procedure, minimizing the porosities • HE200B 共possibly equivalent to W8⫻40) for the BCC6 se-
that may result at the bottom of the flange, reduces the poten- ries, and
tial of crack formation. • HE240B 共possibly equivalent to W8⫻58) for the BCC8 se-
ries.
Downloaded from ascelibrary.org by Universita Studi Napoli Fed Ii on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

The test program presented in this paper was carried out with
the aim of defining the effects of the column size, which leads to Table 1 provides the main geometrical dimensions of the beam
different PZ properties, and of the loading history on the connec- and columns sections, together with possibly comparable U.S.
tion behavior. In the following section, the design of the speci- wide flange sections.
mens is discussed, the experimental setup, and the loading histo- Some comments on a possible comparison between European
ries which have been utilized in the tests are described. The and U.S. shapes are in order. In this respect it should be under-
experimental results are presented through hysteresis loops and lined that the choice offered to a designer in the United Stated is
global performance parameters and the major factors affecting the wider than in Europe. In the United States, for a given cross
cyclic behavior, the failure modes, and the rotation capacity are section number 共related to height兲 up to 17 different choices are
assessed. provided, with weights, inertia, and plastic moment capacities
consistently variable depending on web and flange thickness,
Experimental Program which leads to different local slenderness. The same is not true in
Europe where, typically, three to five choices are offered for a
Design of Specimens cross-section height. For the above reasons, in Table 1 more than
Three different types of specimens, have been designed and have one comparable U.S. section is provided for each of the European
been subjected to several rotation histories. The three different sections used in this work.

Table 1. Beam and Column Sections of Specimens and Comparison to U.S. Equivalent Profiles
Element section h 共mm兲 b f 共mm兲 t w 共mm兲 t f 共mm兲 h w /t w b f /2t f Mass per length 共kg/m兲 I xx 共cm4兲 Z x 共cm3兲
Beam
IPE 300 300 150 7.1 10.7 39.2 7.0 42.2 8356 628

U.S. section
W12⫻6.5⫻30 313 166 6.6 11.2 44.03 7.41 44.5 9934 708
W12⫻6.5⫻26 310 165 5.8 9.7 50.10 8.51 39.0 8527 612
W10⫻8⫻33 247 202 7.4 11.0 30.41 9.18 49.0 7097 636

Column BCC5
HE 160 B 160 160 8 13 16.75 6.15 42.6 2492 354

U.S. section
W6⫻6⫻25 162 154 8.1 11.6 17.14 6.64 37.1 2220 310
W8⫻6.5⫻24 210 165 6.2 10.2 29.13 8.09 36 3438 379

Column BCC6
HE 200 B 200 200 9 15 18.89 6.67 61.3 5696 643

U.S. section
W8⫻8⫻31 203 203 7.2 11 25.14 9.23 46.1 4545 496
W8⫻8⫻35 206 204 7.9 12.6 22.89 8.10 52.0 5268 569
W8⫻8⫻40 210 205 9.1 14.2 19.96 7.22 59.0 6113 653

Column BCC8
HE 240 B 240 240 10 17 20.6 7.06 83.2 11260 1053

U.S. section
W10⫻10⫻49 253 254 8.6 14.2 26.12 8.94 73.0 11290 986
W8⫻8⫻48 216 206 10.2 17.4 17.76 5.92 71.0 7658 803
W8⫻8⫻58 222 209 13.0 20.6 13.91 5.07 87.0 9467 980

1302 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / OCTOBER 2003

J. Struct. Eng. 2003.129:1301-1311.

 Table 2. Mimimum Specified Values of Mechanical Properties for European and Comparable U.S. Steel Grades
Mimimum specified values of mechanical properties
Steel grade f y 共MPa兲 f u 共MPa兲 ␧ u 共%兲 Notch impact test absorbed energy (at⫹20°) 共J兲
European steel grade: S 235 JR共EN 10025: 1993兲 235 340– 470 26 27
Comparable U.S. steel grade: A36 共ASTM兲 250 400–550 20 —

European steel grade: S 355 JR 共EN 10025: 1993兲 355 490– 630 22 27
Comparable U.S. steel grade: A572 Gr.50 共ASTM兲 345 450 18 —

It is confirmed, from the comparison given in Table 1, that the tested specimens were made of steel S 235 共A36兲, since this
more than one U.S. wide flange section can be considered equiva- material provides better performance in the inelastic range thanks
Downloaded from ascelibrary.org by Universita Studi Napoli Fed Ii on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

lent to each European section, depending on the parameter which to the smaller nondimensional slenderness which results from the
is considered as the most influential for the comparison. In fact, same section. In any case the S 235 共A36兲 is still used in Europe,
from an economical point of view, the most important parameter especially in seismic regions, even though the attempt to replace
should be the weight per unit length, while from a structural point it with the grade S 355 共A572 Gr.50兲 does exist. From the com-
of view, in the elastic range the main parameter is the moment of parison of properties given in the Table 2, it is evident that the
inertia, in the inelastic range the plastic moment capacity, together European and U.S. materials are quite similar. Some differences
with local slenderness of web and flanges (h/t w and b f /2t f , re- worth being mentioned are that for European steel, not only a
spectively兲, are the parameters which govern the behavior. The minimum but also a maximum strength value is specified, and a
comparison in terms of local slenderness indicates that a possibly minimum toughness value is required according to the Charpy
equivalent European section can be locally stockier than U.S. notch test.
counterparts. In order to obtain the actual values of the mechanical proper-
The choice of the three specimens provides cases which can be ties of beam and column elements, several tension tests on
considered representative of typical beam-column assemblage oc- samples taken from the specimen elements have been carried out.
curring at different story levels in a low rise ‘‘European’’ steel The average values of yield and ultimate stress have been com-
framed structure. puted for both flanges and web of column and beam elements,
Furthermore, having taken the beam section constant in the and are reported in Table 3.
three specimens, and having adopted three column sections which For the three specimens, the plastic moment capacities of the
are characterized by different plastic moment capacities, it has beam, column, and PZ have been computed on the basis both of
been possible to investigate the influence of the beam-to-column the nominal values and of the actual measured values of yield
strength ratios on the inelastic behavior. It can be underlined that stress. The PZ flexural strength 共moment capacity兲 has been
the BCC5 specimens are characterized by the weakest column evaluated according to the EC3 Annex J 共CEN 1997兲 provisions,
since the nominal plastic moment of the column is slightly larger by means of the following formula:
than half of the nominal plastic moment of the beam. In the BCC8
specimen this situation is inverted since the nominal plastic mo- 共 V PZ⫹⌬V PZ兲 •z
M p,PZ⫽ (1)
ment of column is almost twice the one of the beam. In the BCC6 ␤
specimen the situation is intermediate. In the following para- where VPZ⫽column web shear strength; ⌬VPZ⫽additional shear
graph, the values of bending strength capacities 共plastic moments兲 strength due to the contribution of transverse stiffeners 共continuity
of beam, column, and PZ computed both with nominal and actual plates兲; z⫽lever arm, which are, respectively, computed accord-
measured values of yield stress are provided in detail. ing to the following relationships:

Specimen Properties 0.9• f y,wc•A vc

V PZ⫽ (2)
)
The steel utilized for both beam and column elements of the
specimens is S 235 JR 共EN 10025: 1993兲, which is quite similar 4M pl,fc,Rd 2M p,fc,Rd⫹2M pl,st,Rd
to the U.S. steel grade A36. In Table 2 the minimum specified ⌬V PZ⫽ ⭐ (3)
hs hs
values for the main mechanical properties of European and U.S.
equivalent steel grades are provided. In this table both the steel z⫽h b ⫺t f b (4)
grades S 235 JR 共U.S. equivalent: ASTM A36兲 and S 355 JR
共U.S. equivalent: ASTM A572 Gr. 50兲 are considered, since the h c /2 h b
␤⫽1⫺ ⫺ (5)
latter seems to be used quite extensively at this time. However, Lb Lc

Table 3. Average Values of Actual Measured Material Properties

BCC5 BCC6 BCC8
Beam IPE300 Column HE160B Beam IPE300 Column HE200B Beam IPE300 Column HE240B
Variable Flange Web Flange Web Flange Web Flange Web Flange Web Flange Web
f y 共MPa兲 274.8 305.5 323.1 395.6 278.6 304.9 312.6 401.6 292 300 300 309
f u 共MPa兲 404.6 412.6 460.2 490.1 398.8 411.4 434.9 489.8 445 450 457 469
YR 1.47 1.35 1.42 1.24 1.43 1.35 1.39 1.22 1.53 1.50 1.52 1.52

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / OCTOBER 2003 / 1303

J. Struct. Eng. 2003.129:1301-1311.

 Table 4. Flexural Strengths 共in kNm兲 of Beam, Column, and Panel
Zone 共Nominal and Actual Values兲
Specimen M pb M p,PZ M pc 2 M pc
series Stress value 共kNm兲 共kNm兲 共kNm兲 共kNm兲
BCC5 Nominal values 147.6 91.1 83.2 166.4
Actual values 172.6 149.9 114.4 228.8
BCC6 Nominal values 147.6 132.3 151.1 302.2
Actual values 175 220.6 201.0 402.0
Fig. 2. Weld details of specimen
BCC8 Nominal values 147.6 182.8 247.5 495.0
Actual values 183.4 239.6 315.9 631.8
mens, can be considered particularly valuable for experimental
Downloaded from ascelibrary.org by Universita Studi Napoli Fed Ii on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

investigations on the inelastic behavior and failure mode of beam-

where: f y,wc⫽yield stress of the column web; A vc⫽shear area of to-column connections.
the column; M pl,fc,Rd⫽plastic moment of a column flange;
M pl,st,Rd⫽plastic moment resistance of a stiffener; h s ⫽distance
between the centerlines of the stiffeners; h b , h c , t f b , L b , and Specimen Geometry and Details
L c ⫽beam depth, column depth, beam flange thickness, beam, and A total of 18 beam-to-column fully welded joints (3 series
column length, respectively. ⫻6 specimens) have been designed, fabricated, and tested up to
The plastic moment capacities of the beam (M pb), panel zone failure under different loading histories. The specimens are
(M p,PZ), and column (M pc), computed for the three specimens on T-shaped beam-column subassemblages, consisting of a 1,000
the basis of both the nominal and actual measured yield stress mm long beam and a 1,800 mm long column. The beam flanges
values, are reported in Table 4. In the sixth column of Table 4, the have been connected to the column flange by means of complete
summation of the plastic moment capacity of columns framing in joint penetration 共CJP兲 double bevel vee groove welds, while fil-
the joint configuration considered in the experimental tests, i.e., let welds have been applied between both sides of the beam web
⌺M pc⫽2 M pc , is also provided. This value, in fact, can be di- and the column flange 共Fig. 2兲. The manual metal-arc welding
rectly compared to the beam and to the PZ moment capacities for 共MMAW, also known as SMAW兲 process has been used for the
establishing the weaker element in the beam-column subassem- specimens 共with E7018-1 electrodes兲. All welds have been made
blage, thus predicting the location where major inelastic deforma-
tions will occur. It can be observed that for all three specimens the
weakest component of the joint configuration is either the beam
or the PZ, which have plastic capacity values always quite close.
In particular, considering the nominal plastic capacity values, the
weakest element is the PZ for the BCC5 specimen, the beam for
the BCC8 specimen, while quite similar values of beam and PZ
capacity can be derived for the BCC6 specimen. With reference to
the actual values of plastic capacities, similar considerations can
be argued, with the exception of the BCC6 specimen, for which
the beam is slightly weaker than the PZ, due to the scatters be-
tween minimum specified and actual yield stress.
On the basis of relative strength values of column, beam, and
PZ, all specimens are expected to experience inelastic deforma-
tions both in the beam and in the PZ, and in particular, inelastic
deformation are likely to occur: mainly in the PZ for the BCC5,
both in the PZ and in the beam for the BCC6, mainly in the beam
for the BCC8.
Similar predictions could be derived by computing the values
of the column-to-beam strength ratio 共CBsr兲 and of the column-
to-PZ strength ratio, as usually suggested in the seismic provi-
sions for verifying the strong-column-weak-beam design. For ex-
ample, the UBC’94 共UBC 1994兲 provides
CBsr⫽⌺Z c 共 f yc ⫺ f a 兲 /⌺Z b f yb (6)
CPZsr⫽⌺Z c 共 f yc ⫺ f a 兲 /1.25⌺M pPZ (7)
where Z c and Z b ⫽plastic modulus of the column and beam sec-
tions, respectively; f yc and f yb⫽yield stress values of column and
beam section, respectively; and f a ⫽stress value in the column
due to axial load.
By considering that, for the experimental tests presented in
this paper f a ⫽0, the above equations provide results similar to
what was previously anticipated.
In conclusion, the choice of the specimens, and in particular
Fig. 3. 共a兲 Experimental setup; 共b兲 General view of test setup
the relative column-beam-PZ strength values of the three speci-

1304 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / OCTOBER 2003

J. Struct. Eng. 2003.129:1301-1311.

 Table 5. Loading Histories
Downloaded from ascelibrary.org by Universita Studi Napoli Fed Ii on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

Note: C. type⫽cyclic constant amplitude loading history; S.I. type⫽cyclic stepwise increasing amplitude loading history; M. type⫽monotonic loading
history.

in the horizontal position, no special access holes have been nec- the reaction wall and to the foundation by means of prestressed
essary to make the connection, and no weld runoff tabs have been bars. The lateral frame was designed to prevent specimens’ lateral
used to make the CJP welds. In the welding process, special care displacement and is located at the end of the beam 关Fig. 3共b兲兴. An
has been paid in order to minimize porosity and defects, thus automatic testing technique was developed to allow computerized
reducing the potentials of crack formation. The control of the control of the power jackscrew, of the displacement, and of all the
welds has been made through nondestructive tests using ultra- transducers used to monitor the specimens during the testing pro-
sonic equipment. The continuity of the connection through the cess 关Fig. 3共b兲兴. Specimens have been instrumented with electri-
column has been ensured by horizontal 10 mm thick plate stiff- cal displacement transducers 共LVDTs兲, which record the displace-
eners, fillet welded to the column web and flanges. ment 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
Experimental Setup and Loading Histories
three specimen types.
The test setup, represented in Figs. 3共a and b兲, mainly consists in Each set of specimens has been tested up to failure under
a foundation, a supporting girder, a reaction RC wall, a power several rotation histories. The different rotation histories applied
jackscrew, and a lateral frame 关Fig. 3共b兲兴. Due to the characteris- to the specimens can be grouped in the following sets: 共1兲 Mono-
tics of the test setup the column was the horizontal element while tonic; 共2兲 cyclic constant amplitude; and 共3兲 cyclic stepwise in-
the beam was the vertical one. The power jackscrew 共capacity creasing amplitude. This latter test type has been carried out ac-
1,000 kN, stroke ⫾400 mm兲 is attached to a specific frame, pre- cording to the basic loading history recommended in ECCS
stressed against the reaction wall, and designed to accommodate 共1986兲. Table 5 provides the test loading histories, defined in
the screw backward movement. The power jackscrew is con- terms of cyclic amplitude of the applied beam tip displacement
nected to the end of the beam through a pinned connection in 共d兲; cyclic amplitude of the applied beam tip displacement nor-
order to avoid the introduction of moments. The specimen is con- malized to theoretical value of the specimen yielding displace-
nected to the supporting girder through two steel elements 共A and ment d y (d/d y ); cyclic amplitude of the applied interstory drift
B兲. The connections between the ends of the column and the steel angle (d/H), i.e., d normalized to the distance H between the
elements 共A and B兲 are a pinned type. With this type of boundary beam tip and the column centerline. One of the loading histories
conditions it is possible to obtain approximately a null moment at has been applied twice on each specimen series 共see Table 5: Test
both ends of the column and represent half of the length of the B and BB, for both BCC5 and BCC6; Test B and E for the
column in a real steel frame. The supporting girder is fastened to BCC8兲.

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / OCTOBER 2003 / 1305

J. Struct. Eng. 2003.129:1301-1311.

Downloaded from ascelibrary.org by Universita Studi Napoli Fed Ii on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

Fig. 4. 共a兲 BCC5 specimens: Moment at column centerline versus total rotation; 共b兲 BCC5 specimens: Moment at column face versus beam
plastic rotation

Cyclic Behavior and Failure Modes As can be derived from the M -d/H curves, the cyclic behavior
of Specimen BCC5 is characterized by a great regularity and
Premise stability of the hysteresis loops up to failure, with no deterioration
In the following sections the experimental results obtained in the of stiffness and strength properties at an increasing number of
test program are provided. In particular the cyclic behavior and cycles. The very last cycle in each test presents a sudden and
the failure modes observed for the three sets of specimens are sharp reduction of strength, corresponding to the physical col-
described, and the moment rotation hysteresis loops obtained in lapse of the specimen. The collapse has always occurred in a
the tests are provided. In the moment rotation hysteresis loops brittle mode, due to a fracture close to the weld, initiated in the
presented hereafter, the rotation values have been calculated both beam flange and propagated also in the web. This collapse mode
as the ‘‘unprocessed’’ total rotation given by the applied interstory is related to the significant distortion of the joint PZ, which gives
drift angle d/H, and as the beam plastic rotation ⌽ b,pl , obtained rise to high stress and strain concentrations at weld locations.
from the measured LVTDs displacements at the beam reference While considerable distortion of PZ occurred during all tests,
section, by subtracting the contributions of the beam and column no large plastic deformation in the beam was observed.
elastic rotations as well as of the column inelastic deformation, In Fig. 5 a photo of Specimen A of the BCC5 series at the end
which was always localized in the PZ. Correspondingly, in the of the test is provided. This failure mode is representative of what
M -d/H experimental curves the moment M is evaluated at the
centerline of the column, while in the M b -⌽ b,pl curves the mo-
ment M b is evaluated at the column face. According to the AISC
共1997兲, ‘‘beam-to-column connection testing shall demonstrate a
flexural strength, determined at the column face, that is at least
equal to the nominal plastic moment of the beam at the required
inelastic rotation.’’ Thus in the M b -⌽ b,pl diagrams two horizontal
lines at M b ⫽⫾M pb ⫽⫾147.5 kNm, are reported for qualifying
the strength properties of the connection at high levels of defor-
mation.
For each specimen type, the experimental curves obtained in
three tests are reported, namely, a large constant amplitude rota-
tion test (d/H⫽7.5%), the increasing amplitude rotation test, and
a small constant amplitude rotation test (d/H⫽3.75%).

BCC5 Specimen
In Fig. 4共a兲 typical moment-total rotation (M -d/H) experimental
curves of the BCC5 specimens are plotted, while in Fig. 4共b兲 the
corresponding moment-beam plastic rotation (M b -⌽ b,pl) curves
Fig. 5. BCC5A specimen after test
are plotted.

1306 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / OCTOBER 2003

J. Struct. Eng. 2003.129:1301-1311.

Downloaded from ascelibrary.org by Universita Studi Napoli Fed Ii on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

Fig. 6. 共a兲 BCC8 specimens: Moment at column centerline versus total rotation; 共b兲 BCC8 specimens: Moment at column face versus beam
plastic rotation

was observed for all the BCC5 specimens, i.e., visible PZ distor- In Fig. 7 a photo of Specimen D of the BCC8 series at the end
tion; brittle failure mode with cracks either at or close to the of the test is shown. This photo clearly shows the formation of a
beam-to-column weld locations, and no plastic hinge in the beam. plastic hinge in the beam, which has been observed for all the
It is worth pointing out that such behavior was the expected one, BCC8 specimens, giving rise to a very ductile behavior. The pre-
on the basis of the relative strength values provided in the section diction of a major contribution of the beam to the inelastic defor-
‘‘Specimen Properties.’’ mation mode of the connection, which has been made in the sec-
tion, ‘‘Specimen Properties,’’ through the comparison of beam,
column, and PZ strength capacities, has been confirmed by the
BCC8 Specimen
experimental observations.
In Fig. 6共a兲 the moment-total rotation (M -d/H) experimental
curves of the BCC8 specimens are plotted, while in Fig. 6共b兲 the
corresponding moment-beam plastic rotation (M b -⌽ b,pl) curves,
are plotted.
The hysteresis loops obtained from the tests on the BCC8
specimens 共except the one obtained in the BCC8C test兲 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 progressive deterioration of the flexural
strength of the connection is related to the appearance of local
buckling in the beam flanges, when cyclically subjected to com-
pression, and to the consequent spreading of the buckling to wider
zones of the flanges and to the web. A well defined plastic hinge
in the beam formed in all the tested specimens. In Specimens
BCC8 the contribution of the PZ deformation was not significant,
and the plastic deformation mainly took place in the beam. In Test
C, where the specimen has been subjected to a large constant
amplitude rotation history, with peak cyclic rotation equal to 7.5%
rad, an unstable behavior of the specimen was observed, with
multiple buckling occurring in the beam flanges starting from the
first plastic cycle, and a sudden failure occurring at the third plas-
tic cycle due to the fracture in the beam flange along the weld.
The final collapse of Specimens BCC8A and BCC8D was due
to fracture of the beam flange in the buckled zone, at approxi-
mately 10–15 cm from the weld section. In Tests BCC8B,
BCC8C, and BCC8E the collapse of the specimens occurred due
to fracture in the beam, starting along the weld or very close to
Fig. 7. BCC8D specimen after test
the weld line.

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / OCTOBER 2003 / 1307

J. Struct. Eng. 2003.129:1301-1311.

Downloaded from ascelibrary.org by Universita Studi Napoli Fed Ii on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

Fig. 8. 共a兲 BCC6 specimens: Moment at column centerline versus total rotation; 共b兲 BCC6 specimens: Moment at column face versus beam
plastic rotation

BCC6 Specimen end of the test is shown. The photo shows one of the two typical
failure modes observed for the BCC6 specimens, depending on
In Fig. 8共a兲 the moment-total rotation (M ⫺d/H) experimental
the applied loading history. It is worth underlining that, as pre-
curves of the BCC6 specimens are plotted, while in Fig. 8共b兲 the
dicted on the basis of the relative strength values of beam, col-
corresponding moment-beam plastic rotation (M b -⌽ b,pl) curves,
umn, and PZ, an intermediate behavior between the BCC5 and
are plotted.
BCC8 specimens was observed for this specimen series.
Throughout the test program, two different kinds of cyclic
behavior have been observed for the BCC6 specimens. In some
cases 共Tests C and D兲 the behavior of the specimens is close to
the behavior observed for the BCC5 type, with almost no dete-
rioration of the mechanical properties up to the last cycle, during
which the collapse occurred. For the other tests 共A, B, and BB兲 a
gradual reduction of the peak moment at an increasing number of
cycles is evident, thus indicating a progressive deterioration of the
flexural strength of the connection. In these cases, starting from
the very first plastic cycles, local buckling of the beam flanges
occurred, with a consequent decrease of the moment carried by
the beam. As the buckling zones extend and spread to the beam
web, a well defined plastic hinge formed in the beam. In Tests C
and D, buckling in the beam flanges was less significant and, even
if some plastification in the beam occurred, the plastic hinge was
not as evident as in the A, B, and BB tests.
In all the BCC6 specimens the contribution of the PZ defor-
mation was not as significant as in the BCC5 specimen type.
However at the end of Tests A, B, and BB, permanent distortions
of the PZ, respectively, equal to 0.02, 0.035, and 0.01 rad, were
registered.
The collapse of Specimens BCC6A and BCC6B was due to
fracture of the beam flange in the buckled zone, at approximately
10–15 cm from the weld section. Specimen BCC6BB failed due
to fracture in the beam flange along the weld line. Also for Speci-
mens BCC6C and BCC6D a crack formed in the beam flange
close to the weld, and propagated through the flange and the web
of the beam, driving the specimens to collapse.
Fig. 9. BCC6BB specimen after test
In Fig. 9 a photo of Specimen BB of the BCC6 series at the

1308 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / OCTOBER 2003

J. Struct. Eng. 2003.129:1301-1311.

 is emphasized that although weak-panel design of the beam-to-
column welded connections can substantially reduce beam plastic
rotation demands and effectively contribute to global connection
ductility, the stress state arising at high levels of applied rotation
increases the potentials for brittle fracture collapse mode of the
connection. Additional results, based both on experimental tests
and FEM theoretical studies are provided by Lu et al. 共2000兲
where it is found that weak column panel specimens 共with PZ
contribution equal to 70% of the total plastic rotation of the speci-
men兲 show brittle failure modes, with no previous deterioration of
Fig. 10. Maximum rotation values from increasing amplitude tests
the strength capacity, due to fracture in the beam web weld and
beam bottom flange.
Evaluation of Connections Performance
Downloaded from ascelibrary.org by Universita Studi Napoli Fed Ii on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

The BCC6 specimens have displayed a behavior sometimes

close to the BCC5 specimens 共Tests BCC6C and BCC6D兲, some-
Effect of Panel Zone on Specimen Rotations times close to the BCC8 specimens 共Tests BCC6A, BCC6B, and
BCC6BB兲, depending on the applied loading sequence. Also with
Fig. 10 compares the rotations experienced by the three speci- regard to the final collapse of the specimens, in the former cases
mens in the increasing amplitude test 共BCC5C, BCC6C, and it involved fracture in the beam starting at or close to the weld
BCC8D兲. In particular for each specimen series the bars reported location, while in the latter cases it was due to cracking in the
in the chart provide the total (elastic⫹plastic) rotation globally buckled zones of the beam flanges. These aspects are more deeply
experienced by the beam-column subassemblage d/H tot ; the plas- analyzed in the next subsection.
tic rotation globally experienced by the beam-column subassem-
blage d/H pl ; the beam plastic rotation; and the PZ rotation. As
can be seen from this figure, the total rotation capacity d/H tot in Effect of Loading History on Failure Modes
the increasing amplitude tests reach values of 0.064 rad for the No significant effect of the loading history on the failure mode of
BCC5 specimen, 0.053 rad for the BCC6 specimen, and 0.046 rad the BCC5 specimens has been observed throughout the experi-
共at a loss of strength equal to 10% of the maximum strength兲 for mental program: All BCC5 specimens have shown a sudden col-
the BCC8 specimen. Comparing these values to the default value lapse, corresponding to a sharp decay of the load carrying capac-
of interstory drift capacity required for special moment resisting ity of the connection and associated to a brittle fracture, generally
frames 共FEMA 2000兲, equal to 0.04 rad 共reported as a horizontal developed in the proximity of the weld 共brittle failure mode兲.
line in the chart兲, the performance of the three specimens can be For the BCC8 specimens a ductile deformation and failure
judged fully satisfactory. However these quite large values of total mode has been observed, with final collapse occurring due to
rotation correspond to limited plastic rotations in the beams, re- fracture either in the buckled beam flange at the plastic hinge
spectively, equal to 0.0057, 0.0175, and 0.0242 rad for the three location 共Test D at increasing amplitude, Test A at 5% constant
specimens, thus confirming that large rotations can be experi- amplitude兲 or in the beam flange close to the weld location 共Tests
enced thanks to column web panel deformations. B and E at 3.75% constant amplitude兲. However in all cases a
The contribution of the PZ deformation to the global rotation well defined plastic hinge formed in the beam and the failure of
of the specimens, throughout the experimental program, has been the connection occurred after a slow degradation of the strength
remarkable 共in average equal to the 80% of the total imposed capacity of the specimens was observed at an increasing number
rotation兲, in the specimens having the smallest column section of cycles 共ductile failure mode兲, thus a real load effect cannot be
共BCC5兲; less significant 共in average equal to the 65% of the total established for this specimen series.
imposed rotation兲 for the specimens with intermediate column On the contrary, for the BCC6 specimens two different kinds
section 共BCC6兲; minor 共40–50% of the applied rotation兲 in the of inelastic behavior and collapse mode have been observed in the
specimens characterized by the largest column section 共BCC8兲. cyclic tests, depending on the applied loading history: for the test
Consistently, the plastic rotations registered in the beam are minor at increasing stepwise amplitude 共C兲 and Test D, at the smallest
for the BCC5, comparable to the PZ rotations in the BCC6, larger constant amplitude (d/H⫽3.75%), a cyclic behavior similar to
for the BCC8 specimens. the one exhibited by the BCC5 specimens, with sudden decay of
the strength properties of the connection and brittle fracture
Effect of Panel Zone on Cyclic Behavior and Failure mode, has been observed. For Tests A, B, and BB, at larger con-
Mode stant amplitude (d/H⫽5 and 7.5%兲, the behavior has been simi-
lar to the one of the BCC8 specimens, with formation of a plastic
The BCC5 specimens, even though able to experience high plas- hinge, slow degradation of the strength capacity, and ductile frac-
tic deformation levels, have shown sudden failure modes in all ture mode. Thus for the BCC6 specimen series, the loading his-
cyclic tests, with hysteresis loops practically overlaid and no deg- tory has significantly affected the failure mode.
radation of the flexural strength up to the very last cycle, where a
sharp decay of the load carrying capacity occurred due to fracture,
generally developed in the proximity of the weld. On the contrary Effect of Loading History on Hysteretic Behavior
the BCC8 specimens have exhibited a typical ductile behavior, In order to assess the effect of the loading history on the hyster-
with formation of a well defined plastic hinge in the beam starting etic behavior of the specimens, different performance parameters
from the first plastic cycles, and a slow decrease of the peak can be computed and examined. In this paper both the cyclic
bending moment at increasing number of cycles up to the col- energy and the cumulated plastic rotation have been used for this
lapse. purpose.
These results confirm some major findings reported in El- In Fig. 11 the normalized cyclic energy, i.e., E i /E el,pl , defined
Tawil et al. 共1999兲 where, on the basis of FEM analyses results, it as the ratio between the absorbed energy in the single plastic

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / OCTOBER 2003 / 1309

J. Struct. Eng. 2003.129:1301-1311.

 stress levels and large plastic strains, thus the final collapse of the
specimen occurs due to sudden failure at the weld locations.

Concluding Remarks and Design Implications

The analysis of the experimental data presented in this paper has
given us the opportunity of evidencing some aspects which gov-
ern the cyclic behavior of welded connections. In particular, for
three series of specimens, the effect of the column size, which
drives to different PZ properties, and of the loading history, on the
cyclic behavior, deformation capacity and failure mode of the
Fig. 11. Normalized cyclic energy connections has been assessed.
Downloaded from ascelibrary.org by Universita Studi Napoli Fed Ii on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

Concerning the effect of the column size and of the PZ prop-

erties, the following conclusions can be stated:
cycle E i and the energy that might be absorbed in the same cycle • For the BCC5 specimens, characterized by the smallest col-
if it had an elastic-perfectly plastic behavior E el,pl is plotted umn section and by the lowest PZ strength value, a ‘‘brittle’’
against the number of cycle. The curves in Fig. 11 refer to the failure mode has been observed in all the tests, due to the
three specimens BCC5, BCC6, and BCC8, subjected to the tests governing role of the PZ in the inelastic deformation mode,
at d/H⫽3.75% and d/H⫽5.0%. From the graph it is possible to which places severe demands to the beam-to-column welds;
derive a steady decrease of the normalized cyclic energy for the large scatters in the number of cycles to failure and in the
BCC8 specimens, which finally experienced ductile failure mode, values of performance parameters of the connections have
and, on the contrary, the sudden decay, due to brittle failure mode, been observed for this specimen series;
of the normalized cyclic energy for the BCC5 specimens. For the • On the contrary for the BCC8 specimens, characterized by the
BCC6 specimens, in the case of d/H⫽5.0% 共ductile failure strongest PZ and by occurrence of inelastic deformations
mode兲, the first kind of trend can be derived, while in the case of mainly in the beam, a typical ductile behavior has been always
d/H⫽3.75% 共brittle failure mode兲, the behavior is close to the observed throughout the experimental program, with forma-
one observed for the BCC5 specimens. tion of a well defined plastic hinge in the beam starting from
The above observations are also confirmed by examining the the first plastic cycles, and a slow decrease of the peak bend-
values of the cumulative plastic rotation, which is an important ing moment at increasing number of cycles up to the collapse;
deformation parameter for qualifying the performance of the and
welded connection, since it globally accounts for the plastic de- • For the intermediate size specimens 共BCC6兲, characterized by
formation occurring in specimens subjected to both constant and close values of beam and PZ plastic capacity and, conse-
increasing amplitude cyclic histories. In the bar chart of Fig. 12 quently, by occurrence of inelastic deformations both in the
the cumulated plastic rotation values are reported for the different beam and in the PZ, the experimental evidence suggests an
rotation histories applied to the specimens. The values appear intermediate behavior, with a strong dependence of the cyclic
highly variable for the BCC5 (d/H pl,cum⫽0.27– 0.65 rad) and behavior, of the performance parameters values and of the fail-
BCC6 (d/H pl,cum⫽0.30– 0.73 rad) specimens, while the BCC8 ure mode on the applied loading history.
specimens, excluding the test C 共7.5%兲, show quite similar values In order to assess the effect of the loading history, both in-
for all the tests (d/H pl,cum⫽0.48– 0.55 rad). creasing cycle amplitude and constant cycle amplitude 共large am-
Therefore it can be concluded that, in the case of ‘‘brittle’’ plitude and small amplitude兲 histories, have been applied to the
failure mode, the behavior of the different specimens show large specimens. On the basis of the observations, the following con-
scatters in the number of cycles to collapse and in the values of clusions can be stated:
the cycle energy at collapse, while in the case of ‘‘ductile’’ failure • Not a significant effect of the loading history on the inelastic
mode the various specimens show a similar behavior in all tests, behavior and failure mode has been observed for the BCC5
with absorbed energy steadily decreasing up to failure. In fact, the and BCC8 specimen series, since, in the former case the PZ,
formation of a plastic hinge in the beam reduces the stress state in and in the latter case the beam, were decisively the weakest
the welds, thus the failure of the specimen is achieved in the base element of the connection; and
material at the plastic hinge location and is strictly related to the • For the BCC6 specimens, which has quite close values of
deterioration of the mechanical properties of the beam. On the beam and PZ strength, a strong dependence of the hysteretic
contrary, when the specimen inelastic deformations do not occur behavior and of the failure mode on the loading history has
mainly in the beam and the plastic hinge is not so evident, the been observed.
welds between the beam and the column are subjected to high The quite high values of the maximum plastic rotations
(d/H pl) experienced by these European-type connections are re-
lated to the significant contribution of PZ deformation, which has
been observed throughout the tests for all three groups of speci-
mens. This last point could suggest the possibility of utilizing the
joint panel for providing energy dissipation and stable behavior of
the connections even at large number of cycles. Weak PZ, how-
ever, give rise to inelastic deformation mode of the beam-column
subassemblage which produces high stress concentrations and
large plastic strains at the beam-to-column welded zones, and
finally drives the specimen to brittle collapse mode. For this rea-
Fig. 12. Cumulated plastic rotation values
son the contribution of PZ distortion to the global plastic rotation

1310 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / OCTOBER 2003

J. Struct. Eng. 2003.129:1301-1311.

 of the connection should be properly calibrated in the design of References
the connection, through the definition of adequate values of the
relative strength of beam and PZ. American Institute of Steel Construction 共AISC兲. 共1997兲. Seismic provi-
sions for structural steel buildings, Chicago.
Notation Calado, L., and Mele, E. 共2000兲. ‘‘Cyclic tests on bolted and welded
beam-to-column connections.’’ J. Earthquake Tech., 37共4兲, 65– 88.
The following symbols are used in this paper: Castiglioni, C. A., Bernuzzi, C., and Calado, L. 共2000兲. ‘‘Cyclic behavior
A vc ⫽ shear area of column; of steel beam-to-column joints with concrete slab.’’ Proc., 3rd Int.
B ⫽ section width; Conf. on Steel Structures in Seismic Areas (STESSA), A. A. Balkema,
Montreal, 147–154.
d ⫽ applied beam tip displacement;
El-Tawil, S., Vidarsson, E., Mikesell, T., and Kunnath, S. K. 共1999兲.
d y ⫽ theoretical yield displacement;
‘‘Inelastic behavior and design of steel panel zones.’’ J. Struct. Eng.,
d/d y ⫽ applied beam tip displacement normalized to 125共2兲, 183–193.
theoretical yield displacement;
Downloaded from ascelibrary.org by Universita Studi Napoli Fed Ii on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

European Committee for Standardization 共CEN兲. 共1997兲. TC250/SC3-

d/H ⫽ applied interstory drift angle; PT9, ‘‘Joints in building frames.’’ Part 1.1. Revised annex J, Edited
d/H pl ⫽ maximum plastic rotation globally experienced by approved draft, January 1997, Eurocode 3, Brussels.
beam-column subassembly; European Convention for Constructional Steelwork 共ECCS兲. 共1986兲.
d/H pl,cum⫽ cumulated plastic rotation globally experienced by ‘‘Seismic design. Recommended testing procedure for assessing the
beam-column subassembly; behavior of structural steel elements under cyclic loads.’’ Technical
d/H tot ⫽ maximum total rotation globally experienced by Communication 1—Structural Safety and Loadings, TWG1.3—Rep.
beam-column subassembly; No. 45.
Federal Emergency Management Agency 共FEMA兲. 共2000兲. ‘‘Recom-
E i ⫽ energy absorbed in single plastic cycle;
mended seismic design criteria for new steel moment-frame build-
E el,pl ⫽ energy absorbed in elastic-perfectly plastic cycle; ing.’’ FEMA Rep. No. 350, Washington D.C.
f a ⫽ axial compressive stress; Japanese Society of Steel Construction 共JCSS兲. 共1997兲. ‘‘Kobe earth-
f y ⫽ yield stress; quake damage to steel moment connections and suggested improve-
f yb ⫽ beam yield stress; ment.’’ JSSC Technical Rep. No. 39, Tokyo.
f yc ⫽ column yield stress; Kunnath, S. K., and Malley, J. O., eds. 共2002兲. ‘‘Advances in seismic
H ⫽ distance between beam tip and centerline of design and evaluation of steel moment frames: Recent findings from
column; FEMA/SAC Phase II Project.’’ J. Struct. Eng., 128共4兲, 415– 419.
h b ⫽ height of beam cross section; Lu, L. W., Ricles, J. M., Mao, C., and Fisher, J. W. 共2000兲. ‘‘Critical
h c ⫽ height of column cross section; issues in achieving ductile behaviour of welded moment connec-
tions.’’ J. Constr. Steel Res., 55, 325–341.
h s ⫽ distance between centerlines of stiffeners;
Malley, J. O. 共1998兲. ‘‘SAC steel project: Summary of phase-1 testing
h w ⫽ web height; investigation results.’’ Eng. Struct., 20共4-6兲, 300–309.
I ⫽ strong axis section moment of inertia; Mahin, S. A., Hamburger, R. O., and Malley, J. O. 共1996兲. ‘‘An integrated
L ⫽ length; program to improve the performance of welded steel frame build-
L b ⫽ beam length; ings.’’ Proc., 11th WCEE, World Conf. Earthquake Engineering,
L c ⫽ column length; Elsevier Science Ltd., Acapulco, Mexico, Paper No. 1114.
M ⫽ moment at column centerline; Mele, E. 共2002兲. ‘‘Moment resisting welded connections: an extensive
M b ⫽ moment at column face; review of design practice and experimental research in USA, Japan
M pb ⫽ beam moment capacity; and Europe.’’ J. Earthquake Eng., 6共1兲, 1–35.
M pc ⫽ column moment capacity; Mele, E., Calado, L., and Pucinotti, R. 共1997兲. ‘‘Indagine sperimentale sul
comportamento ciclico di alcuni collegamenti in acciaio.’’ Proc., 8th
M pl,fc,Rd⫽ plastic moment of column flange;
National Conf. Earthquake Engineering, Associazione Nazionale di
M pl,st,Rd⫽ plastic moment resistance of stiffener; Ingegneria Sismica 共ANIDIS兲, Taormina, Italy, 1031–1040 共in Ital-
M p,PZ ⫽ panel zone moment capacity; ian兲.
t w ⫽ web thickness; Nakashima, M., Suita, K., Morisako, K., and Maruoka, Y. 共1998兲. ‘‘Tests
t f ⫽ flange thickness; on welded beam-to-column subassemblages. I: Global behavior.’’ J.
V PZ ⫽ column web shear strength; Struct. Eng., 124共11兲, 1236 –1244.
Z ⫽ plastic section modulus; Plumier, A., Agatino, M. R., Castellani, A., Castiglioni, C. A., and Chesi,
Z b ⫽ plastic section modulus of beam; C. 共1998兲. ‘‘Resistance of steel connections to low-cycle fatigue.’’
Z c ⫽ plastic section modulus of column; Proc., 11th ECEE, European Conf. Earthquake Engineering,
z ⫽ lever arm; Balkema, Paris.
Suita, K., Nakashima, M., and Morisako, K. 共1998兲. ‘‘Tests of welded
⌬V PZ ⫽ shear strength contribution of transverse stiffeners;
beam-to-column subassemblages. II: Detailed behavior.’’ J. Struct.
and Eng., 124共11兲, 1245–1252.
⌽ b,pl ⫽ beam plastic rotation. Tanaka, A., Kato, B., Kaneko, H., Sakamoto, S., Takahashi, Y., and Ter-
Subscripts aoka, M. 共1997兲. ‘‘Seismic damage of steel beam-to-column
a ⫽ axial; connections—evaluation from statical aspects.’’ Proc., 2nd Int. Conf.
b ⫽ beam; on Steel Structures in Seismic Areas (STESSA), Kyoto, Japan, 856 –
865.
c ⫽ column;
Taucer, F., Negro, P., and Colombo, A. 共1998兲. ‘‘Cyclic testing of the steel
f ⫽ flange;
frame.’’ JRC ELSA Special Publication No. I.98.160, Ispra, Italy.
p, pl ⫽ plastic; The Steel Construction Institute. 共1992兲. Steel designers’ manual, G. W.
s, st ⫽ stiffeners; Owens and P. R. Knowles, eds., Blackwell Scientific Publications,
tot ⫽ total; London.
w ⫽ web; and Uniform Building Code 共UBC兲. 共1994兲. ‘‘International Conf. of building
y ⫽ yield. officials.’’ Whittier, Calif.

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / OCTOBER 2003 / 1311

View publication stats J. Struct. Eng. 2003.129:1301-1311.