STRUCTURAL STEEL EDUCATIONAL COUNCIL

TECHNICAL INFORMATION & PRODUCT SERVICE

March 2001
Large Seismic Steel
Beam-to-Column Connections
by
Egor P. Popov, Professor
Shakhzod M. Takhirov, Ph.D.
Pacific Earthquake Engineering Research Center (PEER)
University of California, Berkeley
Professor Egor Paul Popov standing next to his last steel connection test specimen
(February 2001).
Egor Paul Popov, Professor Emeritus of Civil and Environmental Engineering at the
University of California, Berkeley, passed away Thursday, April 19, 2001, after a brief
illness.
Professor Popov began his engineering studies at the University of California, Berkeley,
and continued with graduate work at the California Institute of Technology, the
Massachusetts Institute of Technology, and Stanford University, where he obtained his
doctorate degree in 1946. In his lengthy and illustrious career, he was called upon by
NASA for his engineering expertise and played a key role in the structural analyses of the
Alaskan Pipeline and the Oakland-San Francisco Bay Bridge. Popov joined the
Department of Civil Engineering at UC Berkeley in 1946, and was active in teaching and
research there for more than 50 years. His research interests covered a wide spectrum of
topics in earthquake engineering, including cyclic testing and modeling of structural
members; the development, research, and application of the eccentrically braced frame;
research on the seismic resistance of steel connections and the development of improved
connection details; and the development of friction devices to retrofit existing structures.
He was appointed the first chairman of AISCD's Committee on Seismic Provisions for
Steel Buildings and served in this position for several years. Elected to the National
Academy of Engineering in 1976, Professor Popov was honored in 1999 with the
Earthquake Engineering Research Institute's highest honor, the George W. Housner
Medal.
Table of Contents
Abstract...........................................................................................ii
1 Review of the Previous Research
1.1 Introduction.. ... ... ... ... ... ... ... ..............................................................1
1.2 Overview .........................................................................................1
1.3 Tension Tests on A490 1 Bolts............................................................1
1.4 Test Specimen Design and Detailing.......................................................2
2 Connection Design and Estimate Calculations
2.1. Test specimens based on the proposed design...........................................2
2.2. Basic Parameters Used in the Connection Calculation................................2
2.2.1. Calculation of Plastic Hinge Location in the Beam.....................................3
2.2.2. Calculation of Probable Plastic Moment at the Hinges.................................3
2.2.3. Beam Shear Calculation... ... ..............................................................3
2.2.4. Calculation of the Moment at the Centerline of the Column..........................4
2.2.5. Check for Strong Column - Weak Beam Condition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.3. The Connection Details Calculations......................................................5
2.3.1. Calculation of the T-section Stem Thickness at the Weakest Section near the
Column Face. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.3.2. Calculation of the T-section Flange Size. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.3.3. Calculation of Bolt Size Between the T-section and the Column Flanges (During
the Design 1 inch High-Strength Bolts were Chosen)......................................7
2.3.4. Calculation of Weld Size to Beam Flanges for Both Specimens (3/4 inch Fillet
Weld Was Used). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3 Experimental Program
3.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.2 Test Specimens, Test Setup and Instrumentation.......................................9
3.2.1 Test Setup. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.2.2 Instrumentation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
5.2.3 Data Acquisition......................................................................................9
3.2.4 Loading History. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.2.5 Data Processing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.3 Test Results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.3.1 Specimen 1.....................................................................................11
3.3.2 Specimen 2.....................................................................................12
4 Experimental Results and Conclusions
4.1 Experimental results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
4.2 Conclusions: advantages and disadvantages of proposed connections............13
4.2.1 Advantages. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
4.2.2 Disadvantages. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4.2.3 Future Research Directions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
i
ABSTRACT
Two large bolted steel moment-resisting connections were studied by experiments. These
connections were single-sided beam-column assemblies that are representative of exterior
beam-column connections, and they were composed of W36xl50 Grade 50 beams and
W14x283 Grade 50 columns. T-sections were cut from W40x264 sections of Grade 50
steel. The T-section stems were welded to the beams and pre-stressed by bolts to the
beam flanges in the shop. Final beam-to-column assembly required no additional
welding: the T-section flanges were bolted to the column and the column shear tab was
bolted to the beam web. The specimens had two symmetrically located T-sections with
the difference in web geometry: the Specimen 1 had rectangular shape of stems, whereas
the Specimen 2 had U-shaped stems. During the cyclic testing the beam deformation was
minimal due to active participation of the T-section flanges: a separation between T-
section flanges and the column flanges were observed. This separation was occurred due
bending plastic deformation in the T-section flanges. This phenomenon allowed energy
dissipation and prevented the beam flanges and beam web from severe buckling.
ii
1 Review of the Previous Research
1.1 Introduction
The generally accepted detail of attaching steel beams to columns in seismic applications
consists of shear tabs attached to the column and direct welding of beam flanges with or
without cover plates to column flanges. Numerous tests on this type of a connection were
supported by NSF with many specimens donated by the fabricators. The testing of such
specimens was organized by SAC Joint Venture.
The moment capacity of such connections depends on cyclic endurance of flange welds
in both tension and compression. Under these conditions numerous tension weld failures
were observed both in the laboratory and the field. SAC has proposed six connections to
avoid future weld failures. We propose another connection that avoids weld failures and
it is field bolted and shop welded.
1.2 Overview
An attempt at the above approach on several end-plate connections was made by K.C.
Tsai and E.P. Popov at University of California, Berkeley (1988 and 1990). An example
of this kind of a connection is shown in Fig. 1-1. It is of interest to note that direct
welding of a beam to a column stub shown Specimen 9 in Fig. 1-1 results in good
behavior, but the erection is not generally practical. Specimen 10 in Fig. 1-1 with no ribs
over the beam flanges did not give satisfactory results. Specimen 10R with ribs over the
beam flanges at the column stub behaved very well under cyclic loading as can be seen in
Fig. 1-2. Note the required large thickness of the end plate (a connection based on the
design of Specimens 10 and 10R may require shims during assembling).
The above approach recently was also pursued by T.M. Murray and his associates in
2000 at Virginia Polytechnic Institute (VPI) with good results. They achieved a number
of successful tests with W36xl50 beams. It appears that for larger or heavier beams the
use of ribs over beam flanges at columns would be required.
An extensive excellent study of bolted connections has been done at Georgia Institute of
Technology by R. Leon and his associates in 2000. The work is very comprehensive, but
is limited to small and medium size members.
The newly developed and tested connection at University of California, Berkeley is
somewhat related to the end plate connection but is more versatile as it is more readily
adaptable to a larger range of heavier beams. The new connection depends on the use of
A490 1 bolts in tension in oversize round holes (as in the column flange and in the T-
section flange as well) and shop fillet welds.
1.3 Tension Tests on A490 1 Bolts
l
In order to conduct ductility study on the A490 1 bolts two tests were performed. For
the first test, a special device was built to test simultaneously the shank of the bolt and the
threaded part of the bolt below the nut. The actual failure occurred in the threaded region.
The remarkable ductility of A490 bolt was clearly demonstrated and the load versus
elongation diagram is presented on Fig. 1-3. Another experiment on a specimen of
constant diameter machined from A490 bolt also showed excellent ductility. Figure 1-4
shows stress versus strain diagram for this test.
1.4 Test Specimen Design and Detailing
One of newly developed connections using A490 1 bolts is shown in Fig. 1-5. The
details for the two specimens are shown in Fig. 1-6 and 1-7. In both cases the attachment
of a beams to a column is made using structural tees cut from W shapes (T-sections). A
large choice of such sections is available. By rotating the beam all fillet welds can be
done in the shop in a down-hand position. Generous rounded fillets occur in all cases
between a beam flange and the stem of the T-sections. Shop experience in fabricating
these two specimens was very encouraging.
2 Connection Design and Estimate Calculations
2.1. Test specimens based on the proposed design.
The beams were fabricated from a W36xl50 section of A572-Gr.50 steel, the columns
were fabricated from a W14x283 section of A572-Gr.50 steel. The T-sections were made
from W40x264 section of A572-Gr.50 steel. The geometrical properties of the beam
sections used in the specimens are presented in Tables 2-1 though 2-3. They represent
widely available data for standard rolled I-beams with W- shapes (see, for instance, AISC
1995a).
The global dimensions and geometry of the specimens are shown in Fig. 1-5. Figures 1-6
and 1-7 show the design details for the Specimen 1 and the Specimen 2. The material
properties of the connections from mill certificate data are presented in Table 2-4. The
specimens had two symmetrically located T-sections with the difference in the geometry
of the stems: the Specimen 1 had rectangular shape of the stems, whereas the Specimen 2
had U-shaped stems. The T-section stems were welded to the beams in the shop, and later
the T-section flanges were bolted to the columns.
2.2. Basic Parameters Used in the Connection Calculation.
Nonlinear deformation of a building frame is typically accommodated through the
development of inelastic flexural or shear stains within discrete regions of the structure.
At large deformations these regions can develop into plastic hinges, which can
accommodate significant concentrated rotations without significant changes of the load,
as shown in Fig. 2-1. This behavior is accompanied by significant energy dissipation,
particularly by members involved in plastic deformation. The formation of plastic hinges
in the beams is extremely desirable.
2
2.2.1. Calculation of Plastic Hinge Location in the Beam.
The probable location for the formation of the plastic hinges is a basic parameter for the
connection calculations. Figure 2-2 shows the suggested location of the plastic hinge. The
location depends on the type of connection and our design is close to a cover-plate type,
therefore the plastic hinge can be developed at the following distance from the face of the
column (FEMA 267a):
L
h
= d
sts
+ d
b
/4,
Where:
d
b
is a depth of the beam, the value of the beam depth db is presented in the Table 2-1;
d
sts
is a total depth of the T-section. This parameter is 28.75 inches for the Specimen 1
and 20 inches for the Specimen 2, as it is shown in Table 2-3.
Therefore the plastic hinge is located at the following distance from the face of the
column:
L
h
= 37.7 inches, for the Specimen 1 and
L
h
= 29.0 inches, for the Specimen 2.
2.2.2. Calculation of Probable Plastic Moment at the Hinges.
The probable value of the plastic moment, M
pr
, at the location of the plastic hinge should
be calculated from the equation, proposed in FEMA-267b:
M
pr
=1.1Z
b
F
yb
.
Where:
F
yb
is the actual yield stress of the beam material, as identified from mill test reports in
the Table 2-4,
Zb is the plastic modulus of the beam section determined from the Table 2-1.
Therefore for the proposed design and for chosen material properties Z
b
= 581 inch
3
, and
F
yb
= 56.6 ksi, the probable value of the plastic moment is as following:
M
pr
= 36173 kip*inch.
2.2.3. Beam Shear Calculation.
The shear in the beam, at the location of the plastic hinge should be determined. The
length of the arm at plastic hinge location is calculated from total beam length, L
b
, minus
the distance of the hinge location, Lh. Therefore the shear at the plastic hinge Vp can be
determined from the formula:
V
p
=M
pr
/(L
b
-L
h
).
3
Therefore the values of the shear at the location of the plastic hinge are
V
p
= 375.7 kips, for the Specimen 1, and
V
p
= 344.4 kips, for the Specimen 2.
2.2.4. Calculation of the Moment at the Centerline of the Column.
The moment at centerline of the column can be calculated as following, as it shown on
Fig. 2-3:
M
c
= M
pr
+ V
p
(L
h
+ d
c
/2).
This expression is modified by using previous expression for the shear and the formula
for the moment calculation has the following form:
M
c
= V
p
(L
b
+ d
c
/2).
In our particular case the values of the moment at the centerline of the column are
M
c
= 53485 kip*inch, for the Specimen 1, and
M
c
= 49030 kip*inch, for the Specimen 2.
2.2.5. Check for Strong Column - Weak Beam Condition.
Buildings with the plastic hinges in the beams dissipate more energy than buildings with
plastic hinges in the columns. Therefore, the connection in the building has to develop the
plastic hinge in the beam rather than in the column. To determine if the desired strong
column - weak beam condition exists, the connection assembly should be checked for the
following condition (FEMA 267b):
Where:
Zc - is the plastic modulus of the column section above and below the connection,
Fyc - is the maximum specified yield stress for the column above and below,
f
a
- is the axial stress in the column above and below,
- is the sum of the column moments at the top and the bottom of the panel zone,
respectively, resulting from the development of the probable beam plastic moments, M
pr
.
In our case the column moment at the top of the panel zone, M
ct
, and the moment at the
bottom of panel zone, M
cb
, are the same:
M
cb
= M
ct
= (M
C
/2)(L
C
/2 - d
c
/2)/( L
c
/2) .
For the proposed connection the last formula produces the following values for both
specimens:
4
Z F f M
c yc a c
( ) / . . >10
M
c
M
cb
= M
ct
= 23351 kip*inch for the Specimen 1 and
M
cb
= M
ct
= 21406 kip*inch for the Specimen 2.
The axial stress in the column is calculated from the shear load acting in the column
divided by the effective area of the column cross section:
f
ab
= f
ab
= V
p
/A
c
.
In case of the proposed connection the axial stress is calculated as following:
f
ab
= f
ab
= 4.51 ksi for the Specimen 1 and
f
ab
= f
ab
= 4.13 ksi for the Specimen 2.
Therefore the main criteria for the strong column - weak beam condition will be satisfied
for both specimens, because of the following results:
2.3. The Connection Details Calculations.
2.3.1. Calculation of the T-section Stem Thickness at the Weakest Section near the
Column Face.
The weakest cross section near the column face is located at the K-line of the T-section.
This section is at a distance, which includes the flange thickness of the T-section, t
sts
and
radius of the fillet in the K-line. This value is equal to 3 inches. The location of the
section is presented on Fig. 2-4. The moment at this location is calculated as following:
M
ws
= V
p
(L
b
3).
The numeric values for the moment in case of the proposed connection are calculated as
following:
M
ws
= 49214 kip*inch for the Specimen 1 and
M
ws
= 45114 kip*inch for the Specimen 2.
The chosen cross section consists of the cross sections of T-section stems and the cross
section of the shear plate, as it is shown in Fig. 2-5. The out of scale picture of the chosen
cross section with the dimensions used during the calculation is presented on Fig.2-5.
The connection design allows a plastic deformation in this cross section. Therefore the
total moment in this cross section will consists of two moments. The fist moment, M
1
, is
calculated with the plastic modulus of the shear plate, Z
1
, and has the following
expression:
M
1
=F
ysp
* Z
1
,
5
Z F f M
c yc a c
( ) / . . = > 110 10
Z F f M
c yc a c
( ) / . . = > 121 10
for the Specimen 1 and
for the Specimen 2.
where Z
1
=(0.75*(24)
2
) /4.
The corresponding numerical value (for F
ysp
= 56.6 ksi) is calculated as following and it
is the same for both specimens:
M
1
=6113 kip*inch.
The second moment, M
2
, is calculated with the plastic modulus of stems sections, Z
2
:
M
2
=F
yts
*Z
2
,
where Z
2
=2*18.4*16*0.96.
Therefore the numerical value of this moment (with F
yts
= 64 ksi) can be calculated as
M
2
=36176 kip*inch.
The total moment at this cross section calculated with the assumption of plastic
deformation of the section is:
M
designed
=42289 kip*inch.
The difference between the designed moment, M
designed
, and the expected one at this
weak section, Mws, is within 14 percent. This difference is acceptable, because the weak
cross section has extremely small length (less then 3 percent of the beam depth) and it is
followed by a very strong cross section. The strong cross section has very high value of
the elastic section modulus and allows only elastic deformation up to the following value
of the moment:
M
stong
=58415 kip*inch.
2.3.2. Calculation of the T-section Flange Size.
In order to find the thickness of the T-section's flange the flange was modeled as a fixed-
end beam shown in Fig. 2-6. The concentrated load acts at the midpoint of the beam and
represents the stem's puling force. The maximum force developed in the stem can be
calculated as following:
P = F
yts
A
stem
,
where the cross section area of the stem is A
stem
=16*0.96=15.36 inch
2
. Therefore the
value of the load is as following:
P = 983 kips.
The statically indeterminate beam presented in Fig. 2-6, can be solved for the reaction
forces and moments and they have the following expressions (AISC 1995a, page 4-195):
R
a
= R
b
=P/2,
6
M
a
= M
b
=PL/8.
Therefore the reaction moment applied at the bolt location has the following maximum
value:
M
a
= M
b
=584 kip*inch
The connection design assumes that the plastic deformations can be developed near the
bolt location. Therefore the required plastic modulus of the flange cross section is as
following:
Z
required =
M
a
/F
yts
,
or Z
required
= 9.1 inch
3
.
The chosen flange of the T-section has to have the plastic modulus not lower then the
required one, and it's value for the chosen flange thickness is:
Z
designed
= 16*(1.73)
2
/4 =11.97 inch
3
.
The elastic section modulus for the rectangular cross section of the flange is 1.5 times
less then the plastic one and is equal to
S
designed
= 7.98 inch
3
.
Therefore the design allows the flange yielding but without developing a plastic hinge
near the bolt location.
2.3.3. Calculation of Bolt Size Between the T-section and the Column Flanges (During
the Design 1 inch High-Strength Bolts were Chosen).
As it was discussed before the model for the T-section flanges is shown in Fig. 2-6.
According this model the total axial force acting in each row (of 4 bolts) is:
R
a
=R
b
=P/2=492kips.
The corresponding axial force acting in one bolt is:
R
bolt
=123 kips.
For the chosen 1 inch high-strength bolts the design tensile strength is (AISC 1995b)
R
bolt(LFRD)
=104 kips.
The conducted tests on the high-strength bolts (see section 1.3 of this report) show very
high ductility of the used bolts, the bolt specimen started to yield at 132 kips with the
ultimate tensile load of 150 kips (see results presented on Fig. 1-3). These results explain
7
the chose of the 1 inch high-strength bolts for the column and the T-section
connection.
2.3.4. Calculation of Weld Size to Beam Flanges for Both Specimens (3/4 inch Fillet
Weld Was Used).
The total length of the weld is different for the specimens, the length per the beam flange
is designed as following (including weld in U-shaped part for Specimen 2):
l
weld
- 2*23.5=47.0 inch for the Specimen 1 and.
l
weld
= 2*14.75+2*9.25=48.0 inch for the Specimen 2.
The maximum force in the beam flange will be equal to the product of the flange cross
section area, A
bflange
, and the specified yield stress for the beam, F
yb
:
F
max
= A
bflange
F
yb
.
For the numerical parameters of the proposed connection, the value of this force is equal
to
Fmax = 638 kips for both specimens.
The load per inch of the fillet weld is calculated as following:
F
per\inch
=13.3 kips/inch for the Specimen 1 and
F
per\inch
=13.6 kips/inch for the Specimen 2.
The required leg length is calculated the widely used expression
l
leg
= F
per\inch
/(0.707*1.7*21 ksi).
The required leg length for the designed specimens is determined as following:
l
leg
=0.54 inch for the Specimen 1 and
l
leg
=0.53 inch for the Specimen 2.
Based on the results of these calculations the 3/4 inch weld was used during the
manufacturing the specimens.
3 Experimental Program
3.1 Introduction
This section summarizes the results of cyclic testing of two full scale beam-column
bolted connection specimens. The specimens were designed by Prof. E.P. Popov and
were fabricated by Stoltz Metals, Inc. The tests were carried out in the Structural
8
Research Laboratory of the Pacific Earthquake Engineering Research Center, University
of California at Berkeley.
3.2 Test Specimens, Test Setup and Instrumentation
3.2.1 Test Setup
The specimens were tested in the Structural Research Laboratory of PEER, UC Berkeley.
The test setup was designed to accommodate specimens with columns in vertical
position, as shown in Fig. 3-1. The specimens were attached to horizontal and vertical
frames. The horizontal steel frame was pre-stressed to the strong floor. The columns in
the test specimens were attached to the horizontal frame and the vertical reaction frame
using short segments of W 14x311 to achieve near pinned boundary conditions.
The load was applied to the cantilever beam end by a 400-kip hydraulic actuator, through
a clevis bolted to the beam end plate. The testing setup had displacement capacity of
7.75 inches and load capacity of 350-kip. No axial load was applied to the column.
The test was conducted using the beam end displacement control. The beam end was at a
distance of 134 in from the column face. To prevent out of plane movement of the beam,
a vertical bracing system was provided near the beam end. The photograph in Fig. 3-2
shows a view of a test in progress.
3.2.2 Instrumentation
Many sensors were used to monitor the response of the specimens during the test in order
to understand the specimen behavior. Figure 3-3 shows the location of displacement
measuring instruments on the specimens. The imposed displacement at the end of the
beam was measured by LVDT (Linear Variable Differential Transformer). This
displacement is denoted by , a load cell in-line with the actuator measured axial force P.
The DCDT (which is a LVDT with built-in solid state oscillator and phase-sensitive
demodulator) displacement transducers were used to provide the remaining displacement
measurements. The deformation of the beam panel zone was calculated from readings at
and DCDT locations. Global deflection shape of the column was measured by
displacement transducers. The amplitude of gap opening between the T-section
flanges and the column flanges was measured by two displacement transducers and
Strain gages and rosettes were glued at critical locations to investigate local response.
Figures 3-4 shows these locations on the Specimen 1, the strain measuring
instrumentation for the Specimen 2 was the same, excluding some minor changes in gage
locations. Thirty-eight channels of data were used during testing.
3.2.3 Data Acquisition
The test control and the data acquisition system were run by a PC Windows-based control
and acquisition program called Automated Testing System (ATS) developed by SHRP
Equipment Corporation of Walnut Creek, California. This Program is capable of signal
9

-
.
generation, four-channel servo-actuator command, and sixteen-channel data acquisition.
For the tests the ATS system was used to monitor and control the displacement and force-
feedback signals.
Other data were monitored and recorded using an AutoNet data acquisition system with a
capacity of 64 channels. Pacific Signal Conditioners were used to amplify the transducers
and the strain gages signals and to remove frequencies above 100 Hz from the analog
signal.
3.2.4 Loading History
The testing program was based on the ATC-24 document "Guidelines for Cyclic Seismic
Testing of Steel Structures". The specimens were tested under displacement control,
following a loading history consisting of stepwise increasing deformation cycles. At
certain stage of plastic deformation of the specimens a few cycles with small amplitude
were imposed. Each loading step was defined by the peak beam end displacement and by
the number of cycle. Table 3-1 presents the testing program for the Specimen 1 and the
Specimen 2.
3.2.5 Data Processing
The specimen behavior was characterized by the following parameters: applied load,
beam end displacement, total plastic rotation of the connection, panel zone shear
deformation, column deformation, deformation in the T-section flange, and beam
deflection. A test specimen layout, the corresponding measurements, and the chosen
positive direction of applied load, and measured displacements are shown on Fig. 3-3.
Total displacement of the beam end is caused by rigid body motion of the
connection, the deformations of the beam itself, column, panel zone, and deformation in
the T-section flange. The rigid body motion was possible due small flexibility in the
vertical reaction frame. This part of the displacement was not too large, but it could not
be neglected. Therefore the beam end relative displacement was calculated from the
total one by subtracting of the rigid body displacement. As a result of the column and
panel zone deformations, the panel zone rotates trough an angle and changes its initial
configuration. Four displacement measurements ( and ) were used to compute
the connection rotation due column deflection and panel shear deformation . The
total beam rotation can be separated into four components: rotation due deformation of
the beam itself , rotation caused by rigid connection rotation , the contribution from
the panel zone , and the rotation due gap opening in the T-sections . These values
were determined as follows:
Total beam end displacement:
Relative beam end displacement: The remainder of the
calculation was done using this value of the displacement; where H is a distance from
pin to pin along the column, and L is the distance from the beam end to the center line
of the column
Total rotation:
10

total

1

2

4

5

total

, , ,
= - -
total
L H ( ) / .
6 3
= / L

c
Connection rotation due column deflection: Where d is a distance
between continuity plates
Panel zone shear deformation: . Where a and b are the
dimensions of the rectangular panel zone area (distance between targets in horizontal
and vertical directions)
Rotation due gap opening (and deformation) between T-section flange and column
flange:
Total plastic rotation: Where M=PL
0
(L
0
is a distance from center-line
of the actuator to the face of the column) is the moment at the face of the column and
is the elastic stiffness determined from M versus curve. The unloading path of
one of the elastic cycles below the reverse point was used to estimate this stiffness, to
avoid the influence of initial imperfections, clearances, hysteresis, etc.
A set of programs for the MATLAB 5.3 environment was created to process data and to
plot results in accordance with the procedure described above.
3.3 Test Results
3.3.1 Specimen 1
Testing of the first specimen was conducted on July 30, 2000. The specimen sustained all
loading steps up to and including the 5.69" beam tip displacement cycles without
significant damage. Testing was stopped because the maximum load for the test setup
was reached. Photo of the specimen's close up side view after the test is presented on Fig.
3-5. During the last set of the load reversals a slight buckling in the beam web and
flanges was observed. The residual buckling in the beam flange and web is shown on Fig.
3-6 and Fig. 3-7. During the test energy was dissipated by cyclic yielding of the flanges
of the T-sections, the gap between the T-section and column flanges was opening and
closing periodically. The residual gap in the upper T-section is shown on Fig. 3-8.
Table 3-1 presents the loading protocol for the both specimens. The first row in the table
presents total beam end displacement (in other words: beam end total vertical
deflection). The loading history for Specimen 1 is plotted in Fig. 3-9.
The plot of applied force versus beam tip displacement response is presented on
Fig. 3-10. The values of the displacement were obtained directly from the LVDT reading.
The relative displacement ( ) was calculated from previous displacement by subtracting
the specimen's displacement as rigid body. Because of some small flexibility in the
vertical reaction frame occurred this displacement could not be neglected. The plot of
applied force versus relative beam tip displacement is presented on Fig. 3-11.
Based on the values of the relative beam tip displacement the total beam rotation is
calculated. The imposed moment versus beam total rotation is presented on Fig. 3-12.
Figure 3-13 shows the applied moment versus the beam plastic rotation. The deformation
of the column panel zone is presented on Fig. 3-14.
11

c
d = - ( )/ .
5 4
a b ab = - + ( ) / ( )
5 4
2 2
2

T
d = - / ( ) .
8 7

pl
M K = - / .
K

pl

total

total

During the test the visible opening between the T-section flanges and the column flanges
was observed. The amplitude of the opening between flanges was measured in the
following way, the installed DCDT shows the relative displacement between targets
located at the center plane of the column and the T-section flanges (see details in Fig.3-
3). Therefore during mutual compression of two flanges this displacement is negative,
whereas the tension in T-sections web increases this distance and it becomes positive.
This relative displacement between the flanges is called as "gap opening" in the report.
Figure 3-15a and 3-15b shows these values during the test. The gap opening between T-
section flange and column flange for the upper T-section is presented on Fig. 3-15a. The
same value for the lower T-section is presented on Fig. 3-15b. The beam rotation due
these openings in the T-sections is presented on Fig.3-16.
The imposed force versus beam rotation due panel zone rotation is presented on Fig. 3-
17. The relative beam rotation calculated by subtracting rotation of the panel zone,
rotation due gap opening in T-sections and the panel zone deformation is presented on
Fig.3-18.
3.3.2 Specimen 2
Testing of the second specimen was conducted on July 20, 2000. The specimen sustained
all loading steps up to the 5.69" beam tip displacement cycles and failed at the first ramp
of the last cycle. The fracture was caused by crack in the web of the lower T-section. The
crack line started at the end of the weld and went through the hole for 1 inch bolt. Testing
was stopped after the finishing this cycle. Photo of the specimen's side view after the
testing is presented on Fig. 3-19.
During the test some energy was dissipated by cyclic yielding of the T-sections, the gap
between the T-section and the column flanges was open and closed periodically. The
residual gap in the top T-section is shown on Fig. 3-20.
At the end of the test a slight buckle in the beam web and flanges was observed. The
residual buckling in the beam flanges is shown on Fig. 3-21 and Fig. 3-22.
The crack in the stem of the bottom T-section is shown on Fig. 3-23 and 3-24. Fig. 3-23
presents the view of the location of this crack on the stem of the T-section. The location
was close to the K-line of the T-section and it was parallel to it. The crack started from
the end of the fillet weld, continues through the nearest hole for the 1 in bolt and ends at
the next bolt hole. The close view of the crack is presented on Fig. 3-24. The arrows trace
the crack line.
The loading protocol for the Specimen 2 is presented in Table 3-1. The loading history is
plotted in Fig. 3-25.
The layout of the displacement measuring instrumentation was identical for both
specimens given in Fig. 3-3. The displacement at the beam tip was measured by LVDT,
whereas the remainder of displacement measurement was done using DCDT.
12
The plot of applied force versus beam tip total displacement is presented on Fig. 3-26.
The values of the displacements were obtained directly from the LVDT reading. The
relative displacement was calculated from previous displacement by subtracting the
specimen's displacement as a rigid body. The flexibility of the reacting frame was taken
into account. The plot of applied force versus relative beam tip displacement is presented
on Fig. 3-27.
Based on the values of the relative beam tip displacement the total beam rotation is
calculated. The imposed moment versus the beam total rotation is presented on Fig. 3-28.
Figure 3-29 shows the applied moment versus the beam plastic rotation. The deformation
of the column panel zone is presented on Fig. 3-30.
During the test the visible opening between the T-section and column flanges was
observed. The values of the gap opening were measured by DCDT. Figure 3-3la and 3-
31b shows these values during the test. The gap opening between the T-section flange
and column flange for the top T-section is presented on Fig. 3-3la. The same data for the
bottom T-section is presented on Fig. 2-31b. The beam rotation due these openings in the
T-sections is presented on Fig. 3-32.
The imposed force versus beam rotation due panel zone rotation is presented on Fig. 3-
33. The relative beam rotation calculated by subtracting rotation of the panel zone,
rotation due gap opening in the T-sections and the panel zone deformation is presented on
Fig.3-34.
4 Experimental Results and Conclusions
4.1 Experimental results
A brief summary of experimental results and key parameters characterizing the
performance of tested specimens is presented in Table 4-1. The beam end displacement
corresponds to the relative beam end displacement .
4.2 Conclusions: advantages and disadvantages of proposed connections
4.2.1 Advantages
The design and performance of the proposed beam-to-column connections shows the
following advantages:
- all welding work can be done in a welding shop, in convenient welding positions
final assembling with bolts is relatively easy procedure and does not require a
rigorous quality assurance inspection (in order to achieve the required clamping force
between the column and the T-section flanges the widely available torque multiplier
from WRIGHTTOOL: Model 9S393A was used; the device does not produce any
noise and has an accuracy of 5%)
13

-
after test disassembling of Specimen 2 shows that repairing and replacing beam with
new T-section is neither difficult nor expensive
the beam deformation is minimal due to active participation of the T-sections flanges
and the column flanges during cyclic input
with shims properly installed, the connection develops less residual strain
eliminating large quantities of field weld greatly helps the connecting work to keep
up with the steel erection.
4.2.2 Disadvantages
The chosen design and the failure of Specimen 2 show the following disadvantages and
suggested improvements:
steel along the K-line of the T-section must be carefully selected
1 inch bolts (as used in Specimen 2 to pre-stress the T-section web to beam flange)
requires a greater distance between the bolt and the end of the fillet weld.
Alternatively, it appears that the bolts can be omitted altogether
- steel material of 1 " bolts has to be high quality as used in the tested connections
- connection based on the proposed design require shims for field assembly
- beams with welded top and bottom T-sections require more shipping space during
transportation.
4.2.3 Future Research Directions
Based on the conducted tests and followed data analysis the following future research on
this type of connections is planned:
- conduct 3D finite element analysis (FEA) of the connection to explore the possibility
of exchanging the existing 1 inch bolts to clamps and to investigate the decision to
remove some or all of them
- conduct 3D FEA of the connection to evaluate the critical parameters at the column-
tee joint, including the T-section size, bolt diameter, the clamping load variation and
the prying action
- fabricate and test new specimens with an improved design based on the theoretical
research and results of the previous tests.
14
-
-
-
-
-
-
References:
1. Tsai, K.C. and Popov, E.P. 1990. Cyclic behavior of end-plate moment connections.
ASCE J. of Struct. Engineering, Vol.116, No.11.
2. Tsai, K.C. and Popov, E.P. 1988. Steel Beam-Column Joints in Moment Resisting
Frames. Report No. UCB/EERC 88/19, Earthquake Engineering Research Center,
University of California at Berkeley.
3. Murray, T.M. et al. 2000. Cyclic testing of bolted moment end plate connections.
Struct. And Materials Lab., Virginia Polytechnic Institute and State University.
4. Leon, Roberto et al. 2000. Tests on bolted connections. School of Civil and
Environmental Engineering, Report No. SEMM 00-02, Georgia Institute of
Technology.
5. FEMA-267. 1995a. Interim guidelines: evaluation, repair, modification and design of
welded steel moment frames. FEMA Report No. 267. Washington, D.C.: Federal
Emergency Management Agency.
6. FEMA-267. 1995b. Interim guidelines: advisory No. 1. supplement to FEMA 267.
FEMA Report No. 267. Washington, B.C.: Federal Emergency Management Agency.
7. AISC. 1995a. Manual of steel construction. Load & resistance factor design. Vol.1,
Structural members, Specifications & codes, Second edition. Chicago: American
Institute of Steel Construction, Inc.
8. AISC. 1995b. Manual of Steel Construction. Load & Resistance Factor Design.
Vol.2, Connections, Second edition. Chicago: American Institute of Steel
Construction, Inc.
15
Table 2-1. Dimensions of the W36xl50 beam.
Flange width,
b
fb
[inch]
12
Flange
thickness, t
fb
[inch]
0.94
Web
thickness,
t
wb
[inch]
0.625
Depth, d
b
[inch]
35.85
Section
modulus,
z
b
[inch
3
]
581
Area, A
b
[inch
2
]
44.2
Moment of
inertia, I b
[inch
4
]
9040
Length, L
b
[inch]
134
Table 2-2. Dimensions of the W14x283 column.
Flange width,
b
f
[inch]
16.125
Flange
thickness, t
fc
[inch]
2.06
Web
thickness,
t
f
[inch]
1.29
Depth, d
c
[inch]
16.74
Section
modulus,
Z
c
[inch
3
]
542
Area, A
c
[inch
2
]
83.3
Moment of
inertia, I
c
[inch
4
]
3840
Length, L
c
[inch]
136
Table 2-3. Dimensions of the T-sections cut fromW40x264.
Flange width,
f
ts
[inch]
12
Flange thickness,
[inch]
1.73
Stem thickness,
t
sts
[inch]
0.96
Total depth for Specimen 1
(Specimen 2), d
sts
[inch]
28.75 (20)
Length, L
ts
[inch]
16
Table 2-4. Material Properties.
No
1
2
3
Part of Connection
Beam
Column
T-section
Yield Stress, F
y
[ksi]
56.6
52
64
Ultimate Stress
[ksi]
74.4
66
79
Section Size
W36xl50
W14x283
WT40x264
Grade
Gr50
Gr50
Gr50
Table 3-1. Testing program for the both specimens.
Total beam end
displacement
[inch]
No of cycles
0.36
6
0.53
6
0.71
6
1.07
6
1.42
4
2.14
2
0.53
2
2.85
2
4.27
3
5.69
6*
* Only 2 cycles at this level were performed for Specimen 2.
16
t
fts
Table 4-1. Short summary of test results
Key parameters
Yield load [kips]
Beam end displacement at the yield point [inch]
Elastic stiffness of the connection [kips/inch]
Maximum beam end displacement
Beam end displacement at failure [inch]
Maximum imposed load [kips]
Maximum imposed moment at the column face [kips*inch]
Maximum connection rotation [ % ]
Maximum plastic connection rotation [ % ]
Maximum rotation due gap opening [ % ]
Maximum relative beam rotation itself [ % ]
Specimen 1
230
1.2
180
5.2
N/A
345
48645
4
2.5
1.0
0.6
Specimen 2
230
1.2
178
5.2
3.5
327
43164
4
3.3
0.7
1.5*
This value is high because it includes the beam rotation after the bottom beam flange failure
17
Figure 1-1. Design details of end-plate connections for Specimens 10 and 10R, and that
of direct welding to column, Specimen 9 (K.C. Tsai, E.P. Popov 1988, 1990).
L
o
a
d

(
k
i
p
s
)
Beam Rotation (%)
Figure 1-2. Cantilever beam load versus beam rotation for Specimen 10R (K.C. Tsai, E.P.
Popov 1988, 1990).
18
Figure 1-4. Stress versus strain for coupon test of A490 1 bolt material.
19
Figure 1-3. Load versus elongation for A490 1 bolt.
F
i
g
u
r
e

1
-
5
.

G
l
o
b
a
l

d
i
m
e
n
s
i
o
n
s

a
n
d

g
e
o
m
e
t
r
y

o
f

t
h
e

t
e
s
t
e
d

s
p
e
c
i
m
e
n
s
.
20
F
i
g
u
r
e

1
-
6
.

D
e
s
i
g
n

d
e
t
a
i
l
s

o
f

S
p
e
c
i
m
e
n

1
.
21
F
i
g
u
r
e

1
-
7
.

D
e
s
i
g
n

d
e
t
a
i
l
s

o
f

S
p
e
c
i
m
e
n

2
.
22
Figure 2-1. Desired plastic frame behavior with plastic hinges developed in beams.
Figure 2-2. Probable plastic hinge location.
23
Figure 2-3. Calculation of the moment at the centerline of the column.
Figure 2-4. The weakest cross-section of the beam near the column face.
24
Figure 2-5. Dimensions of the weakest cross section near the column face.
Figure 2-6. View of the column and T-section connection with the corresponding
mechanical model.
25
Figure 3-2. View of a test in progress.
26
Figure 3-1. Test setup for both specimens.
Figure 3-3. Reference dimensions and measurements for the test specimens.
Figure 3-4. Strain gages and rosettes location for Specimen 1.
27
Figure 3-6. Residual beam flange buckling (after the test).
28
Figure 3-5. Specimen 1 after the test (side view).
Figure 3-7. Residual beam web buckling (after the test).
Figure 3-8. Residual gap opening in the upper T-section (after the test).
29
Figure 3-9. Loading history for Specimen 1.
Figure 3-10. Imposed load versus total beam end displacement for Specimen 1.
30
T
i
p

D
i
s
p
l
a
c
e
m
e
n
t

[
i
n
c
h
]
Time [sec]
A
c
t
u
a
t
o
r

f
o
r
c
e

[
k
i
p
s
]
Beam end total displacement [inch]
A
c
t
u
a
t
o
r

f
o
r
c
e

[
k
i
p
s
]
Beam end displacement [inch]
Figure 3-11. Imposed load versus beam end displacement for Specimen 1.
M
o
m
e
n
t

[
k
i
p
s
*
i
n
c
h
]
Beam Rotation [%]
Figure 3-12. Moment versus beam total rotation for Specimen 1.
31
M
o
m
e
n
t

[
k
i
p
s
*
i
n
c
h
]
Beam Plastic Rotation [%]
Figure 3-13. Moment versus beam plastic rotation for Specimen 1.
A
c
t
u
a
t
o
r

f
o
r
c
e

[
k
i
p
s
]
Panel zone shear deformation [%]
Figure 3-14. Imposed load versus deformation in panel zone for Specimen 1.
32
A
c
t
u
a
t
o
r

f
o
r
c
e

[
k
i
p
s
]
Relative displacement between flanges[inch]
Figure 2-15a. Relative displacement between column and top T-section flanges for
Specimen 1.
A
c
t
u
a
t
o
r

f
o
r
c
e

[
k
i
p
s
]
Relative displacement between flanges[inch]
Figure 3-15b. Relative displacement between column and bottom T-section flanges for
Specimen 1.
33
A
c
t
u
a
t
o
r

f
o
r
c
e

[
k
i
p
s
]
Rotation [%]
Figure 3-16. Imposed load versus beam rotation due gap opening in T-sections
(Specimen 1).
A
c
t
u
a
t
o
r

f
o
r
c
e

[
k
i
p
s
]
Rotation [%]
Figure 3-17. Imposed load versus panel zone rotation for Specimen 1.
34
A
c
t
u
a
t
o
r

f
o
r
c
e

[
k
i
p
s
]
Relative beam rotation [%]
Figure 3-18. Imposed load versus relative beam rotation for Specimen 1.
35
Figure 3-19. Specimen 2 after the test (side view).
Figure 3-20. Specimen 2: Residual gap opening in top T-section.
36
Figure 3-21. Specimen 2: top beam flange buckling.
Figure 3-22. Specimen 2: bottom beam flange buckling.
37
Figure 3-23. Specimen 2: crack line location.
Figure 3-24. Specimen 2: close view of the crack line.
38
Figure 3-25. Loading history for Specimen 2.
Tip Displacement [inch]
Figure 3-26. Imposed load versus total beam end displacement (Specimen 2).
39
F
o
r
c
e

[
k
i
p
s
]
T
i
p

D
i
s
p
l
a
c
e
m
e
n
t

[
i
n
c
h
]
Time [sec]
F
o
r
c
e

[
k
i
p
s
]
Tip Displacement [inch]
Figure 3-27. Imposed load versus beam end displacement (Specimen 2).
Beam Rotation [%]
Figure 3-28. Moment versus beam total rotation (Specimen 2).
M
o
m
e
n
t

[
k
i
p
s
*
i
n
c
h
]
40
M
o
m
e
n
t

[
k
i
p
s
*
i
n
c
h
]
Beam Plastic Rotation [%]
Figure 3-29. Moment versus beam plastic rotation (Specimen 2).
F
o
r
c
e

[
k
i
p
s
]
Panel Zone Shear Deformation [%]
Figure 3-30. Imposed load versus column panel zone deformation (Specimen 2).
41
F
o
r
c
e

[
k
i
p
s
]
Gap Opening [inch]
Figure 3-3la. Relative displacement between column and top T-section flanges
(Specimen 2).
F
o
r
c
e

[
k
i
p
s
]
Gap Opening [inch]
Figure 3-31b. Relative displacement between column and bottom T-section flanges
(Specimen 2).
42
F
o
r
c
e

[
k
i
p
s
]
Rotation [%]
Figure 3-32. Imposed load versus beam rotation due gap opening in T-sections
(Specimen 2).
F
o
r
c
e

[
k
i
p
s
]
Rotation [%]
Figure 3-33. Imposed load versus panel zone rotation for Specimen 2.
43
F
o
r
c
e

[
k
i
p
s
]
Rotation [%]
Figure 3-34. Imposed load versus relative beam rotation for Specimen 2.
44
STRUCTURAL STEEL EDUCATIONAL COUNCIL
P.O. Box 6190
Moraga, CA 94570
Tel. (925) 631-1313
Fax. (925) 631-1112
Fred Boettler, Administrator
Steel tips may be viewed and downloaded at www.aisc.org
S P O N S O R S
Adams & Smith
Bannister Steel, Inc.
Baresel Corp
Bethlehem Steel Corporation
Bickerton Industries, Inc.
Bostrum Bergen
California Erectors
Eagle Iron Construcion
Trade Arbed
Four Star Erectors
Gayle Manufacturing
The Herrick Corporation
Hoertig Iron Works
Junior Steel Co.
McLean Steel Inc.
Martin Iron Works, Inc.
Nelson Stud Welding Co.
Oregon Steel Mills
Eandi Metal Works
PDM Strocal, Inc.
Plas-Tal Manufacturing Co.
Reno Iron Works
SME Industries
Shollenbarger-Borello, Inc.
Templeton Steel Fabrication
Verco Manufacturing, Inc.
Vulcraft Sales Corp.
Western Steel & Metals, Inc.
The local structural steel industry (above sponsors) stands ready to assist you in determining the most
economical solution for your products. Our assistance can range from budget prices and estimated
tonnage to cost comparisons, fabrication details and delivery schedules.
Funding for this publication provided by the California Field Iron Workers Administrative Trust.