Advances in Steel Structures - part 10 ppt

70
T. Usami et al.
TABLE 3
PARAMETERS OF CANTILEVER COLUMNS WITH UNSTIFFENED BOX SECTIONS
Specimen
UU1
UU6
UUll
U45-2513]
U45-4013]
U70-2513]
U70-4013]
h a B D t Material
(mm) (mm) (mm) (mm) (mm) k RI P/PY Number
762 306 157.5 120.0 4.51 0.362 0.664 0.2 I
1035 394 202.5 154.0 4.51 0.381 0.854 0.2 I
853 312 171.5 127.0 10.5 0.406 0.297 0.2 I
485 278 144.0 108.8 5.91 0.254 0.448 0.2 II
781 278 145.0 108.8 5.91 0.404 0.451 0.2 II
786 434 222.0 167.8 5.91 0.262 0.701 0.2 II
1217 434 222.9 167.8 5.91 0.406 0.704 0.2 II
Notes: refer to Table 5 for details of material numbers.
TABLE 4
PARAMETERS OF CANTILEVER COLUMNS WITH STIFFENED BOX SECTIONS
Model
B1
B2
B3
B4
B5
B6

B7
B8
B14
B16
h
(mm)
4311
7543
7559
10776
3264
5712
5551
5777
3403
5712
B t b,
ts
(mm) (mm) (mm) (mm) RI A, A u P/P,
1344 20 121 20 0.46 0.51 0.20 1.0 0.15
1344 20 121 20 0.46 0.51 0.35 1.0 0.15
1344 20 121 20 0.46 0.28 0.35 0.5 0.15
1344 20 121 20 0.46 0.51 0.50 1.0 0.15
1023 20 105 20 0.35 0.21 0.20 0.5 0.15
1023 20 105 20 0.35 0.21 0.35 0.5 0.15
1023 20 179 20 0.35 0.23 0.35 1.0 0.15
1023 20 70 20 0.35 0.33 0.35 0.5 0.15
882 9 80 6 0.56 0.63 0.26 1.0 0.12
1023 20 105 20 0.35 0.41 0.35 1.0 0.15
Notes" D = B; refer to Table 5 for details of material numbers.

Material
Number
III
III
III
III
III
III
III
III
IV
III
members. Besides, the computed results are compared with those reported in previous studies (Usami
1996; Nishikawa, et al. 1996; Gao 1998; Gao et al. 1998b; Nishikawa et al. 1999), which are analyzed
under cyclic lateral loading through experimental or numerical techniques. The ABAQUS program
(1998) and a kind of beam element B21 are employed for the pushover analysis.
Cantilever Box Columns with and without Longitudinal Stiffeners
Recently extensive experimental study have been carried out to survey the behavior of steel cantilever
box columns with and without stiffeners, which are subjected to cyclic lateral loading as well as a
constant axial load (see Fig. 7(a)). A detailed summary of these studies has been reported in the literature
(Usami 1996). According to this reference, the local buckling is observed to occur near the column base
in the range of about 0.7B (B is the width of the flange) or between the transverse diaphragms, if any.
And the mode shapes of the global and local buckling are found in the form of half sine-waves. These
Ductility Issues in Thin-Walled Steel Structures
observations are in good
agreement with the effective
length assumed above and the Material E
buckling modes occurred in the Number (GPa)
analyses of stub-columns. By I 197
introducing an index of ductility, II 216

8 95 / 8y (895
is the top lateral
displacement corresponding to III 206
95% of the maximum lateral load IV 206
after peak and 8y is the yield V 206
lateral displacement), empirical
equations of ductility related to VI 206
some main parameters has been VII 206
developed for both the columns
VIII 206
with and without longitudinal
stiffeners, which are defined as
follows (Usami 1996):
TABLE
5.
MATERIAL PROPERTIES OF EXAMPLES
v (MPa) E/gst g
st /[~y
0.269 266 21.0 11.3
0.270 282 32.4 16.9
0.300 314 30 7.0
0.300 379 30 10.0
0.300 235 40 10.0
0.300 290 40 14.0
0.300 269 40 14.0
0.300 294 40 10.0
Notes: Refer to Tables 3, 4, 6, and 7. for material numbers
71
Unstiffened columns: 69s
0.0670

- + 2.60 (S = 1.09) (16)
8y [(1
+ e / Py
)RI ~~ ] 3"~
Stiffened columns: 89s
0.0147
= + 4.20 (S = 1.40) (17)
6y [(1 + P / Py
)gf~~ 3"s
where S is the standard deviation; and ~ is the column slenderness ratio parameter given by
t
_ ___2h 1 ,/o, (18)
r ~: V E
Here h is the column height and r is the radius of gyration of cross section. Equations (16) and (17) were
fitted corresponding to the average curve for test data (i.e., the M curve plotted in Fig. 8 by the solid line)
and the lower bound curves were also proposed as Eqs. (16) and (17) minus the standard deviation (S), as
the
M-S
curve shown in Fig. 8 by the dashed line.
Several specimens in the form of both unstiffened and stiffened columns reported in the reference
(Usami 1996) are adopted here to demonstrate the validity of the ductility evaluation method proposed in
this paper. The parameters of the columns are presented in Tables 3, 4 and 5. The computed ductility
estimations (8,,/8y) are presented in Fig. 8 compared with the empirical curves (Eqs. (16) and (17)), of
which
89s/dy
is denoted by 8u/dy for the accordance. For unstiffened columns (Fig. 8(a)), it is observed
that the proposed method gives the ductility predictions very close to the lower bound curve
(M-S
Curve), which has been recommended for the practical use considering the required safety
(Interim

1996). In Fig. 8(b) which is for stiffened columns, good agreement of computed results with the test
curves is also observed. Aswell, the previous method proposed by Usami et al. (1995), where the failure
strain equations based on isolated plates (Eqs. (11) and (12)) are used, is also applied to these examples
and the obtained results are included in Fig. 8. It can been seen that the previous method underestimates
the column ductility for most cases.
Cantilever Columns with Pipe Sections
The behavior of thin-walled steel cantilever-typed columns with pipe section has been investigated by
some researchers (e.g., Nishikawa et al. 1996; Gao et al. 1998b). In the cyclic test on such columns by
Nishikawa et al. (1996), the so-called elephant foot bulge mode was found to occur in the range of about
3.0\/R t (R is the radius of the pipe section) from the column base (Nakamura 1997). This range is
72
20
15
T. Usami et al.
Present
]
O Previous I
(Usami et
al. 1995)
I
I
,
Cyclic
Test (Usami 1996)
I
M curve [
t M-S
curve I
o _.
i , i i i , i , . i

0.2 0.4 0.6 0.8
(I+P/Py)R~ ~
(a) Unstiffened columns
20
] 9
Present
.~ I O
Previous
15 i\
I
(Usami et al. 1995)
~ ]Cyclic
Test (Usami
1996)
~.I0 \~,
0 l
M-S

5-
~~
0 , I , I , , ,
0.] 0.2 0.3 01.4 01.5 01.6
(I+P/Py) 9 Rr" ~ o.s
(b) Stiffened columns
Figure 8 Ductility estimations of cantilever-typed columns with box sections
almost as same as the effective failure
length assumed in this study (Eq. 15).
Through numerical cyclic analyses, some
researchers (Gao et al. 1998b) proposed an
empirical equation for the ductility of

cantilever-typed columns with pipe
sections, which is given by
~9 __ff_5 _.
0.24 (19)
~iy (1
+ P / Py )2,3-~,3Rt
Nine such columns are investigated here,
the parameters of which are presented in
Table 6. The computed ductility
estimations are plotted in the Fig. 9 by
comparison with Eq. (19). It is found that
all the points corresponding to the results
of the present study lie in the vicinity of the
equation curve. Thus, the applicability of
the proposed method to steel columns with
pipe sections is also verified.
One-story Rigid Frame
Although the behavior of thin-walled steel
cantilever columns has been extensively
investigated by researchers, available
research findings on the thin-walled steel
frames are too limited to supply sufficient
information on the ductility evaluation
(Nishikawa et al. 1999). The proposed
method is expected to be a simple but
efficient ductility evaluation tool for such
structures.
A one-story rigid frame, which has been
Speci
-men

P1
P2
P5
P8a
P8b 1897 891
P8-15 4391 891
P10 3303 580
Pll 4391 891
P12 4391 891
Notes: see Table 5 for details of material properties
TABLE 6
Parameters of Cantilever Pipe Columns
h d t
(mm) (mm) (mm)
Rt ~ e/ev
MaterialNumber
3403 891 9.00 0.110 0.26 0.12 VI
4391 891 7.32 0.115 0.30 0.15 V
4391 891 8.41 0.100 0.30 0.15 V
2598 891 11.2 0.075 0.18 0.15 V
12.6 0.067 0.13 0.15 V
11.2 0.075 0.30 0.15 V
20.0 0.031 0.37 0.09 VII
9.61 0.088 0.30 0.15 V
16.8 0.050 0.30 0.15 V
12 \ "9 P t I
10 ~ " Empirical Curve I
, {Gao et al. 1998b~l
8
4

2
0 t
I
i
I t I i I
0 0.01 0.02 0.03 0.04
(l+P/Py)Rl'S~, ~
Figure 9: Ductility estimations of cantilever-
typed columns with pipe sections
Ductility Issues in Thin-Walled Steel Structures
TABLE
7
PARAMETERS OF ONE-STORY RIGID FRAME
B
Element Plate
(mm)
Column Flange 600
Web 600
Beam Flange 600
Web 600
t b~ t~ a
(mm) (mm) (mm) (mm) Rf 2~,
6 60 6
600 0.497 0.422
6 60 6
8 80 8
600 0.497 0.314
6 60 6
73
9 -

-Material
No.
VIII
VIII
tested in a recent study (Nishikawa et al.
1999), is analyzed through proposed
method. The general layout and some
pertinent parameters of the frame are
presented in Fig. 10 and Table 7. In the
beam-column connection parts, all the
panels of both beam and column
sections are strengthened by doubling
the thickness. The flame was tested
under cyclic lateral loading with the
constant vertical loads of P =0.12Py at
the top of the frame.
The afore-mentioned method is applied
to this structure, where it should be
noted that in a flame system, the axial
force of the columns varies with the
change of lateral load. And this makes
the trial and error method required for
calculating the failure strains (see Eq.
(7)). For this frame, the critical parts are
the regions marked by (~), (~), (~), @, (~)
and (~ in Fig. 10. And the place where
the average compressive strain first
reaches the corresponding failure strain
is found at part (~). Figure 11 illustrates
the normalized lateral force versus

displacement curve from the pushover
analysis compared by the normalized
hysteretic curve from the cyclic test
(Nishikawa et al. 1999). Both of the
points corresponding to the maximum
strength (6,.) and 95% of the maximum
strength after peak on the test envelop
curve (69s) are used to represent the
Figure 10 General layout of the frame
Figure 11 Force-displacement relation curve of the flame
ultimate state of the flame. The failure points are denoted by different marks in Fig. 11. It is observed
that the computed ductility (ru/ry) of the proposed method is close to the 6~/~ from the cyclic test,
whereas the ductility prediction based on previous method (Usami et al. 1995) is too conservative. In the
light of safety required in practical design, the ductility capacity predicted by the proposed procedure is
satisfactory.
74
CONCLUSIONS
T. Usami et al.
The ductility of thin-walled steel stub-columns with and without longitudinal stiffeners was investigated
through extensive parametric analyses. The key parameters affecting the ductility of box stub-columns
are found to be the flange width-thickness ratio, magnitude of the axial force, and the stiffener's
slenderness ratio. The effects of the cross-sectional shape and the columns aspect ratio were also
investigated and found insignificant for the ductility of stub-columns. On this basis, empirical equations
for the ductility in terms of the failure strain were developed. Besides, empirical equations of failure
strains proposed for isolated plates and short cylinders in previous studies (Usami et at. 1995; Gao et al.
1998a) were also presented in this paper.
Moreover, an evaluation procedure has been proposed to employ the ductility equations into the
ductility estimation of practical steel structures composed of thin-walled box or pipe sections. A
simplified pushover analysis was utilized and a failure criterion was defined. The procedure can be used
to evaluate the ductility of thin-walled steel structures in the form of not only cantilever-typed columns

but also framing structures. The proposed method was used to successfully evaluate the ductility of some
cantilever-typed columns and a one-story rigid frame. By the comparison with ductility estimations
obtained from cyclic tests or numerical analyses reported in the literature, the reliability of the proposed
method was verified.
References
ABAQUS/Standard User's Manual. (1998). Ver. 5.7.
"DIN 4114, Blatt2."
(1953).
Stahbau, Stabilitatsfalle (Knickung, Kippung, Beulung),
Berechnungsgrundlagen, Richtlinien,
Berlin, Germany (in German).
Fukumoto, Y., ed. (1997).
Structural stability design- steel and composite structures.
Elsevier Science
Ltd., Oxford.
Galambos, T. V., ed. (1998).
Guide to Stability Design Criteria for Metal Structures,
5th Ed., John Wiley
& Sons, Inc., New York.
Gao, S. B., Usami, T., and Ge, H. B. (1998a). "Ductility of steel short cylinders in compression and
bending." a r.
Engrg. Mech.,
ASCE, 124(2), 176-183.
Gao, S. B., Usami, T., and Ge, H. B. (1998b). "Ductility evaluation of steel bridge piers with pipe
sections." a r.
Engrg. Mech.,
ASCE, 124(3), 260-267.
Nakamura, H. (1997). "Formulae for evaluating shear-bending buckling strength of steel piers with
circular cross section and applicability of the numerical buckling analysis method."
Proc. of

Nonlinear Numerical Analysis and Seismic Design of Steel Bridge Piers, JSCE,
37-42. (in
Japanese)
Nishikawa, K., Yamamoto, S., Natori, T., Terao, O., Yasunami, H., and Terada, M. (1996).
"An
experimental study on improvement of seismic performance of existing steel bridge piers." 3'.
of
Struct. Engrg.,
42A, 975-986 (in Japanese).
Nishikawa, K., Murakoshi, J., Takahashi, M., Okamoto, T., Ikeda, S., and Morishita, H. (1999)
"Experimental study on strength and ductility of steel portal frame pier." a r.
Struct. Engrg., JSCE,
45A, 235-244 (in Japanese).
Usami, T., ed. (1996).
Interim guidelines and new technologies for seismic design of steel structures.
Committee on New Technology for Steel Structures (CNTSS), JSCE (in Japanese).
Usami, T., Suzuki, M., Mamaghani, I. H. P., and Ge, H. B. (1995). "A proposal for check of ultimate
earthquake resistance of partially concrete-filled steel bridge piers."
Struct. Mech./Earthquake
Engrg., JSCE,
508/I-31, 69-82 (in Japanese).
HIGH-PERFORMANCE STEEL STRUCTURES:
RECENT RESEARCH
L.W. Lu, R. Sause and J.M. Ricles
Department of Civil and Environmental Engineering, Lehigh University
Bethlehem, PA 18015-3176, USA
ABSTRACT
Much effort has been devoted in the recent years to the development of high-performance structures
for civil and marine construction. Emphasis of this effort has been on the use of high-performance
steels and innovative structural concepts to improve performance and reduce life-cycle cost. The

paper first gives a summary of the properties of high-performance steels available in the market.
This is followed by a description of research exploring application of such steels to I-girder bridges
and critical elements in building structures within the framework of the current construction
practice. Three innovative structural concepts are then presented: a post-tensioned connection for
building frames resisting seismic forces, use of high performance dampers for dynamic response
control, and unidirectional double hull structure for ships. Their potential applications are also
discussed.
KEYWORDS
High-performance structure, high-performance steel, building, bridge, ship, weldability, fracture
toughness, connection, seismic resistance.
INTRODUCTION
What is a high-performance structure? Presently, there is not a universally accepted answer to this
question and different people are likely to provide different answers which will depend on the types
of structures the individuals having in mind, the desired levels of performance and the performance
of structures built according to the present practice. No attempt, therefore, will be made to define
"high-performance." The following criteria are often used to judge the overall quality of a structure:
(1)
(2)
(3)
Performance under service load,
Performance under overload, and
Life-cycle cost.
75
76 L.W. Lu et al.
For a structure to be considered as a high-performance structure it should have one or more
improvements related to these criteria. Different approaches may be adopted to achieve the desired
improvements. This paper is concerned with (1) the use of high-performance materials and (2) the
development of innovative structural concepts to enhance overall performance and to reduce life-
cycle cost.
HIGH-PERFORMANCE STEELS AND THEIR PROPERTIES

A high-performance steel is defined as a steel that has the combined characteristics of high strength,
good ductility, high toughness, good weldability and fabricability. These are the properties essential
for successful construction of high-performance structures in a civil infrastructure system. For
exposed structures, such as bridges and ships, good corrosion resistance is also necessary. From
metallurgical composition and processing point of view, a yield strength above 450 MPa is
considered as high strength. The fracture toughness, weldability and formability of the steels should
be significantly better than those of the conventional steels. The key issue is the control of the
amount of carbon and carbon equivalent (Lu, Dexter, Fisher, 1994).
The early attempts of using the traditional high strength steels in bridge and ship construction
produced some unsatisfactory results. These steels were found difficult to fabricate due primarily to
susceptibility to hydrogen cracking and the risk of brittle fracture associated with materials having
inadequate fracture toughness. Other problems include: 1) welding defects other than hydrogen
cracking, and 2) potential for stress-corrosion cracking. Many bridges fabricated in the 1960's and
early 1970's from ASTM A514/A517 (690 MPa yield) steel suffered from hydrogen cracks which
occurred during fabrication (Fisher, 1984). Many of these hydrogen cracks occurred in the
longitudinal web/flange joint of welded built-up box sections used as tie girders in tied arch bridges
(Anon, 1979, Fisher, Pense and Hausammann, 1982) as well as welded built-up plate girders. One
example is the Gulf Outlet Bridge near New Orleans. Some bridges have also experienced hydrogen
cracking in transverse groove welds, e.g. the 1-24 bridge over the Ohio River near Paducah,
Kentucky (Fisher, 1984). Hydrogen cracking was also observed in the Navy's Seawolf submarine in
the 120-S weld metal used with the 690 MPa yield strength HY-100 steel (Anon, 1991).
Hydrogen cracking is most effectively avoided by using steel and weld metal with microstructures
that are not susceptible. It has been shown that susceptibility to hydrogen cracking increases
significantly as the carbon content exceeds 0.1 percent (Graville, 1976). The susceptible
microstructures are typically martensite. The new high-performance steels with low carbon contents
are not susceptible to hydrogen cracking.
Microalloyed steels with low carbon content, high manganese levels and microalloy carbide and
nitride formers have been available for sometime for use in construction of structures that require
high strength, high fracture toughness, and good weldability. Over the past 15 years, low-carbon,
age-hardenable steels have gained increasing usage in shipbuilding, heavy-vehicle manufacturing,

and offshore structure construction because of their excellent weldability and fracture toughness.
These steels have become known as High-Strength Low Alloy (HSLA) steels although their total
alloy content is generally around four percent. Another method of increasing strength without
increasing carbon and alloy content is controlled rolling combined with on-line accelerated cooling,
i.e. thermo-mechanical controlled processing (TMCP).
These high-performance steels offer some clear benefits when compared with the traditional high
strength steels (Bolliger, et al, 1988). Most are virtually immune to hydrogen cracking in the heat-
affected zone (HAZ) of welds. This superior resistance to hydrogen cracking allows these steels to
High-Performance Steel Structures: Recent Research
77
be welded without the application of preheat in most situations. The low-carbon, fine-grained
microstructure that results from typical processing yields a very favorable combination of high
strength and high toughness. The excellent fabricability, strength and toughness make high-
performance steel very attractive for use in many applications. For bridges, these advantages may
allow consideration of lifting the onerous requirements for fabrication of fracture critical members
(FCM). FCM are members subjected to tension which if fractured will cause failure of the structure.
The following are some examples of the currently produced high-performance steels:
Low Carbon Age-Hardening Nickel-Copper-Chromium-Molybdenum-Columbian and
Nickel-Copper-Columbian Alloy Steels, ASTM A710.
High Yield Strength, Age-Hardening Alloy, Structural Steels (HSLA 80 and HSLA 100),
MIL-S-24645A.
Structural Steel for Bridges, ASTM A709 Grade HPS 485W.
There are several copper-nickel high-performance steels for bridge construction under development
at the ATLSS Center of Lehigh University (Gross, Stout, and Dawson, 1998).
APPLICATION OF HIGH-PERFORMANCE STEELS
A substantial number of studies have been carried out in the ATLSS Center and elsewhere to
explore the use of high-performance steels in bridges, buildings, offshore structures, and ships.
Brief descriptions of three of these studies are given below:
1-Girder Highway Bridges
The advantages of using high-performance steel in conventional I-girder highway bridges has been

investigated by Sause and Fisher (1995). The investigation involved redesign of recently
constructed highway bridges, using conventional steels with yield strengths of 250 MPa and 345
MPa, and using high-performance steels with yield strengths between 485 MPa and 825 MPa. The
normalized weight of minimum weight girder cross-sections designed for each steel is plotted
versus yield strength in Figure 1. The weight of the design using 345 MPa steel is taken as 100%,
and the weight of the designs using other steels are normalized by the weight of the 345 MPa
design. Three cases are considered: (1) design for strength and stability according to the AASHTO
specifications without considering fatigue, indicated by the dashed line with circles; (2) design for
strength and stability without considering fatigue and allowing the plastic moment to be used as the
nominal bending strength of compact girder cross-sections, indicated by the dashed line with
squares; and (3) design considering strength, stability, and fatigue indicated by the solid line with
solid boxes. As seen in Figure 1, if fatigue is not considered, a higher steel yield strength usually
results in a smaller weight per length. However, an exception occurs at 485 MPa because the
AASHTO specifications permit the use of the plastic moment as the nominal bending strength of
compact girder cross-sections only when the yield strength is no more than 485 MPa. As a result,
higher strength steel girder cross-sections must be designed with the yield moment (yield stress at
the extreme fiber) as the nominal bending resistance. This limitation in the design specifications is
based on concern about the ductility of structural members fabricated from high-strength steel.
Girders fabricated from high-performance steel may not require this limitation, although further
study of this issue is needed. The dashed curve with squares in Figure 1 represents the case when
the plastic moment is used as the nominal bending strength of all compact cross-sections. The solid
78
120
L. IV. Lu et al.
1 1 t I i
lOO
.s
-~ 80
60
,,

I I ! ! I
40 60
80 1
O0 120
(275) (415) (550) (690) (830)
Yield
Stress, ksi
(MPa)
Figure 1. Weight versus yield strength for minimum weight steel I-girder cross-sections
line shows that when fatigue of welds between transverse stiffeners and the web and flange plates is
considered in design, potential decreases in weight with increasing yield strength end at a yield
strength of 690 MPa, because of stress range limits for the details.
In addition to stability and fatigue, deflection under live load may also be a design constraint. The
elastic live load deflection of I-girder bridges designed using high-performance steel was
considered by Sause and Fisher (1995). AASHTO deflection criteria were applied to high-
performance steel bridge designs to investigate whether these criteria are constraints on the use of
high-performance steel. Live load deflections were calculated for minimum weight bridge designs
developed for each yield strength level, and plotted versus yield strength level in Figure 2. The
deflection limit is L/800. Figure 2 shows that the bridge designs at each strength level satisfy the
deflection limit. However, the assumptions made in computing the live load deflections according
to AASHTO may not be acceptable to many bridge engineers. With more conservative assumptions,
the computed deflections for bridge designs at the highest yield strength levels may exceed the
deflection limit.
75 I 3
E
E 50
C
0
25
I ! 1 I I

9 Design for Mp
[] Design for My
Deflection
Limit
0
I. I ! I I
40 60 80 100 120
(275) (415) (550) (690) (830)
2 .E
r-
.s
Vleid Stress, ksi (MPa)
Figure 2. Live load deflection versus yield strength for minimum weight bridge designs
Connections in Building Frames
The superior ductibility, toughness, and weldability of the high-performance steels make them ideal
High-Performance Steel Structures: Recent Research
79
material for critical elements in structural systems. Examples of critical elements, where such
properties are required, include connecting plates in beam-to-column connections, connectors or
connecting devices, shear links in eccentrically braced frames, tension members of structures in
severe service environments, etc. These steels are also attractive for large and complex structures
because of the possibility of requiring no pre- and post-weld treatment.
A study of the use of A710 steel plates as the flange connecting plates in a beam-to-column web
connection shown in Figure 3 was carried out. An identical connection, but with ASTM A572 (50)
steel plates, was tested at Lehigh University. It failed prematurely due to fracture of one of the
connecting plates. The fracture was predominately brittle in nature although there was evidence of
several crack arrests which indicate some ductility in the region. The factors contributing to the
fracture include: (1) a large amount of plastic strain imposed on the members, (2) strain
concentrations at design details, and (3) orientation of the plates with the applied strain in the least
fracture-resistant direction (the rolling direction was parallel to the fracture plane). A post-test

examination showed that the defects in the weldments were no greater in size than might be found
in typical structural welds. It is felt that this connection fractured in a brittle manner due to the large
applied tensile strain which was concentrated at the design detail.
Figure 3. Beam-to-column connection test details with A710 steel.
For the A710 tests, the flange plates was orientated so that the applied strain was not in the direction
of the least fracture resistant direction; the rolling direction of the plate was parallel to the applied
tensile strain. The strain concentrations at the design details were difficult, if not impossible, to
avoid in construction. The connection was, therefore, assembled as if in an actual construction
environment. This specimen behaved in a very ductile manner and the ultimate load exceeded the
calculated plastic limit load by about 20% (Lu and Fisher, 1990).
The ATLSS Center has developed a wedge and socket type joint for a beam-column connection in a