J. Construct.

Steel Research 10 (1988) 133-162

Recent Developments for the Design of Moment End-Plate

Connections

Thomas M. Murray
Department of Civil Engineering, Virginia Polytechnic Institute and State University.
Blacksburg, Virginia 24{)61, USA

ABSTRACT

Moment end-plate connections are used in the United States primarily in

pre-engineered, industrial buildings. Their use in multi-story, moment-
resistant frame construction is becoming more common because of ad-
vancements in design methods and fabrication techniques. This paper
reviews past literature and design methods for both flush and extended
end-plate configurations, including column side limit states. Recommended
design procedures, which are based on recent analytical and experimental
research in the United States, are then presented. Design procedures for two
flush end-plate configurations (two and four bolts at the tension flange) are
based on the yield-line method, but include tension bolt force predictions.
Procedures for two extended configurations (four and eight bolts at the
tension flange) are the result of finite element studies. Bolt force predictions
are also included in these procedures. Methods to evaluate the three column
side limit states (web yielding, web buckling and flange bending) at moment
end-plate connections are also presented.

INTRODUCTION

The pre-engineered metal building industry has pioneered the use of bolted
end-plate m o m e n t connections in the United States. The use of such con-
nections in multi-story, moment-resistant frame construction is becoming
133
J. Construct. Steel Research 0143-974X/88/$03-50 © 1988 Elsevier Science Publishers Ltd,
England. Printed in Great Britain
134 T. M. Murray

(a) (b)

(c) (d)

Fig. 1. Flush end-plate configurations. (a) Two-bolt unstiffened; (b) four-bolt unstiffened:
(c) four-bolt stiffened with web gusset plate between the tension rows of bolts; (d) four-bolt
stiffened with web gusset plate outside the tension rows of bolts.

m o r e c o m m o n because of advancements in design methods and fabrication

techniques, both of which have resulted in decreased costs.
A typical end-plate m o m e n t connection is composed of a steel plate
welded to the end of a beam section with attachment to an adjacent m e m b e r
using rows of pre-tensioned high-strength bolts. The connection may be
between two beams (splice plate connection) or between a beam and a
column. End-plate m o m e n t connections are classified as either flush or
extended with or without stiffeners and further classified depending on the
n u m b e r of bolts at the tension flange. A flush connection is detailed such
that the end-plate does not appreciably extend beyond the beam flanges and
all bolts are located between the beam flanges. A n extended end-plate is one
which extends beyond the tension flange a sufficient distance to allow the
location of bolts other than between the beam flanges. Extended end-plates
may be used with or without a stiffener. Flush end-plate connections are
typically used in frames subject to light lateral loadings or near inflection
points of gable frames. Extended end-plates are used for beam-to-column
m o m e n t connections.
Four types of flush end-plate connection configurations are shown in Fig.
 Design of moment end-plate connections 135

)
(a) (b)

) )
(e} (d)

Fig. 2. Extended end-plate configurations. (a) Unstiffened (four bolts); (b) stiffened (four
bolts); (c) four bolts wide; (d) stiffened (eight bolts).

1. Figure l(a) and (b) shows unstiffened flush end-plate configurations with
two and four bolts near the tension flange. Figure l(c) and (d) shows
stiffened flush end-plate configurations with four bolts near the tension
flange. In Fig. 1(c), a web stiffener plate is located on both sides of the web
between the two tension rows of bolts, while in Fig. l(d) the web plates are
located outside the tension rows of bolts. For both configurations, the
stiffener plates are welded to both the end-plate and the beam web.
Four extended end-plate configurations are shown in Fig. 2. The configur-
ation shown in Fig. 2(a) is probably the most commonly used in multi-story
frames and a design procedure is found in Ref. 1. Assuming a maximum bolt
diameter of 1½in (maximum practical size because of tightening considera-
tions) and A325 bolts, this connection is limited to use with less than
one-half of the beam sections available, assuming full beam moment capa-
city is to be utilized. The connection strength can be increased by adding a
stiffener (Fig. 2(b)), or increasing the number of bolts per row to four (Fig.
2(c)). Formal design procedures are not available for the former and the
latter requires a wide column flange. The stiffened, eight bolt connection
shown in Fig. 2(d) is capable of developing most of the available beam
sections using A 325 bolts even if bolt diameter is limited to 1" in.
136 T.M. Murray

FAB RICATIOI~I/ERECTION CONSIDERATIONS

As with any connection, end-plate connections have certain advantages and

disadvantages. The principal advantages are:
(a) The connection is suitable for winter erection in that only field bolting
is required.
(b) All welding is done in the shop eliminating field welding associated
problems.
(c) Because of the few bolts required and no field welding, the erection
process is relatively fast.
(d) If fabrication is accurate, it is easy to maintain plumbness of the
frame.
The principal disadvantages are:
(a) The fabrication techniques are somewhat more stringent because of
the need for accurate beam length and 'squareness' of the beam end.
(b) Column out-of-squareness can cause erection difficulties but can be
controlled by fabricating the beams ~-~ in short and providing ~finger'
shims.
(c) The cost of fabrication of the connection can be greater than other
types of moment connections, however, the installed cost is usually
less.
(d) The bolts are in tension which can result in prying forces.
A number of designers and fabricators in the United States have success-
fully used moment end-plate connections for building frames up to 30 stories
in height and it is believed that, in spite of several disadvantages, moment
end-plate connections can provide economical solutions for rigid frame
connections.

REVIEW OF LITERATURE

End-plate design
Research starting in the early 1950s and continuing to the present has
resulted in refined design procedures for both flush and extended end-plate
connections. The earlier design methods were based on simple statics and
assumptions concerning prying forces. These methods resulted in thick
end-plates and large diameter bolts. Other studies have been based on
yield-line theory. The more recent studies have used the finite element
 Design of moment end-plate connections 137

method and regression analysis to develop design equations. Accurate

solutions can be developed using the latter technique, however, the pro-
cedure is expensive and the resulting design equations usually involve terms
to odd powers which virtually eliminates 'structural feel' from the design.
Early attempts (prior to about 1975) to develop design criteria for
moment end-plate connections were based on the 'tee-stub' analogy. An
exhaustive review of methods, suggested by Douty and McGuire, 2 Nair et
a l . , 3 and Kato and McGuire 4 among others, is contained in Fisher and
Struik 5 and will not be repeated here. All of the cited studies resulted in
design procedures which predict a high degree of 'prying action' resulting in
large end-plate thicknesses and large bolt diameters. The method developed
by Douty and McGuire 2 is contained in Ref. 6.
More recently, methods based on refined yield-line analyses have been
suggested. Zoetemeijer 7used the yield-line theory for analysis ofT-stub and
end-plate connections with four bolts. In his study, a design method for the
tension side of the statically loaded connection is developed. The design
methodology is based on the analysis of two different collapse mechanisms.
Tests were performed to ensure that the developed design rules would lead
to connections that would satisfy the limit state of deformations as given in
the European regulations for Constructional Steel Work. The test results
showed a satisfactory agreement with the proposed design rules.
Packer and Morris 8 used curved yield lines to predict end-plate capacity
with testing directed at the failure of the column flange. Due to limited
experimental data, the study did not give any conclusive results. Sub-
sequently, Phillips and Packer 9 conducted a series of tests to determine the
influence of the end-plate thickness on moment rotation characteristics and
the end-plate collapse mechanism. They also studied flush end-plate con-
nections with two rows of tension bolts in order to study the influence of the
second row on the stiffness of the connection. They concluded that flush
end-plates with two rows of two bolts in the tension region are suitable for
semi-rigid construction, and that the second row of tension bolts is effective
but to a much lesser extent than previously estimated. They also concluded
that the French design specifications best predict the connection moment
capacity when the bolt proof load can be reached before plate failure.
Mann and Morris ~°reviewed previous research programs and proposed a
design procedure for the extended end-plate connection. They suggest that
in plastic design, connections must fulfill two basic requirements. First they
must be strong enough to withstand hinge moments and second they must
provide adequate hinge rotation while sustaining these moments. Their
proposed end-plate design method is based on the assumption that some
form of yield line pattern will develop. The recommended design procedure
includes methods to determine both bolt size and end-plate thickness.
138 T. M. Murray

Zoetemeijer" analyzed an infinitely long plate bounded by two fixed

edges and one free edge and loaded with a concentrated force using yield-
line theory. As a result of this analysis, a chart was developed by which the
ultimate load of a stiffened column flange or a flush end-plate can be
determined once the distance from the bolt to the beam flange and web are
known. However, he notes that the analysis was only an approximation.
Srouji ~2presents methodology, based on yield-line analysis, for the design
of four types of end-plate configurations. The four types of end-plates
considered are: flush with two bolt rows in tension region, flush with four
bolt rows in tension region, unstiffened extended with two bolt rows on
either side of tension flange, and stiffened extended with two bolt rows on
either side of tension flange. He developed prediction equations for the bolt
forces including prying action. Experimental verification of the yield line
analysis was conducted by Srouji for the two-bolt flush end-plate and the
four-bolt flush end-plate connections. He recommended the design
methods for these connections using both strength and stiffness of the
connection. Since stiffness data was only available for the two-bolt and the
four-bolt flush end-plate connections, complete design procedures for the
extended end-plates were not recommended in this study.
Hendrick ~3continued the work of Srouji and presented unified yield-line
based, design procedures for four types of flush end-plate configurations:
two-bolt unstiffened; four-bolt unstiffened; four-bolt stiffened with a web
gusset plate placed between the two rows of tension bolts; and four-bolt
stiffened with a web gusset plate placed outside the tension rows of bolts.
The resulting end-plate design procedures are based on yield-line analyses
and are consistent with regard to philosophy and assumptions among the
four connection types. Experimental verification of the end-plate design
equations was conducted for all configurations. A comparison among
configurations was made based on a strength criterion, i.e. required end-
plate thickness and resulting connection capacity, and a stiffness criterion,
i.e. moment capacity for a stiffness limit suitable for rigid framing
connections. Based on the experimental results, design methods were
recommended using strength criteria.
Morrison ~4 has furthered the work of Srouji and Hendrick for the ex-
tended, four-bolt stiffened and unstiffened, extended with multiple rows
below the tension flange configurations.
Finite element methodology for the analysis of end-plates was first de-
veloped by Krishnamurthy.15'~6 An exhaustive analytical study of four-bolt,
unstiffened, extended, end-plates along with a series of experimental in-
vestigations lead to the development of the design procedure found in Ref.
1.
Tarpy and Cardinal ~7carried out a study on the behavior of unstiffened
 Design of moment end-plate connections 139

beam-to-column end-plate connections. On the basis of an elastic finite

element study and experimental verification, a design methodology was
proposed. It was noted that the end-plate and column flange thickness were
two different parameters unlike most previous studies.
Maxwell et al. 18studied the end-plate and other connections using various
methods: simple bending theory, yield-line theory, the finite difference
method and the finite element method. They concluded that the finite
element method was the best feasible alternative for analyzing such a
problem, because it will significantly reduce the number of full scale tests
normally required to establish behavior with this type of investigation, and
also the finite element method results are close to experimental results.
Based on the finite element method and the experimental tests, they de-
veloped prediction equations for ultimate moment of the connections and
the moment-rotation relationships.
Ahuja 19 was the first to investigate stiffened end-plate moment connec-
tions with two rows of bolts on each side of the tension flange. Basically this
study was an extension of the work done by Krishnamurthy. 16 The finite
element analyses used in the study were based on a hybrid 2D-3D model
where the end-plate and bolt shanks and bolt heads were modeled as 3D
finite elements and the beam and stiffener components were modeled as 2D
finite elements. Only elastic material properties were considered. Using
results from the finite element analyses, a feasibility and sensitivity para-
metric study was conducted to establish pertinent variables (geometry and
force related) that govern the connection behavior. The ranges of the
variables were restricted to practical ranges. For 30 selected cases, finite
element analyses were conducted and the results regressed to develop
prediction equations for end-plate behavior. Because of the limitation of
elastic material properties, design rules developed from the prediction
equations were found to be excessively conservative.
Ghassemieh 2° extended the work of Ahuja by including inelastic end-
plate and bolt material properties. He then conducted finite element
analyses for selected cases and used regression analysis to develop predic-
tion equations for both strength and stiffness. Based on comparison of the
analytical results with experimental data a design procedure for A36 steel
and A325 bolts was presented.

Beam-to-end-plate weld design

Only one paper known to the writer has addressed beam-to-end-plate weld
design. Griffiths 21 suggests that either full penetration welds or fillet welds
sufficient to develop the beam flange be used to connect the end-plate to the
beam. This recommendation holds even if the full capacity of the beam is not
140 T. M. Murray

being utilized because of the large local deformations that occur along the
end-plate. A similar recommendation is made for the beam web-to-end-
plate weld design in the vicinity of the tension bolts.

Bolt design

Early end-plate design procedures 2-4 all involved the calculation of bolt
prying forces based on various assumptions. The assumed location of the
prying force was at or near the edge of the end-plate. References 7-11 have
all included prying action forces in their yield-line based design procedures.
The recommendations range from rather complicated procedures to a
simple increase in bolt force over the applied tension. 10
Krishnamurthy ~5 argues that even though prying action is present, it is
overly conservative to assume it to be acting at the edge of the plate as this
normally results in thicker than necessary end-plates. His studies explain the
prying force as a pressure bulb which is formed under the bolt head due to
the pretensioning of the bolt and shifts towards the edge as the beam flange
force increases. For any given loading the pressure bulb is located some-
where between the edge of the end-plate and the bolt head. He states, for
service load conditions, that when the beam flange loads are small, it is more
towards the bolt head than towards the edge and the plate moments are
much smaller than those predicted by prying force formulae. Consequently,
in his design procedure for four-bolt, extended, unstiffened end-plates,
prying forces are ignored, that is, the bolt size is determined directly from
the force delivered by the beam flange.
Kennedy e t a l . 22 have presented a method for calculation of prying forces
as a function of plate 'thickness' relative to applied load. They identified
three types of end-plate behavior. The first case, to which all end-plates
belong under low applied loads, is identified by the absence of plastic hinges
in the end-plate and the end-plate is said to be 'thick'. The upper limit on this
behavior occurs at a load which causes yielding in the end-plate at the beam
flange. Once this load is exceeded, a plastic hinge is formed at the flange and
the end-plate is said to be of 'intermediate' thickness. As the load is in-
creased, a second plastic hinge forms at the bolt lines. At this load, the
end-plate is considered to be a 'thin' plate. Next, they consider bolt force to
be the sum of a portion of the flange force and prying forces. They identify
three stages of prying action corresponding to the three phases which the
end-plate goes through. For 'thick' plates, the prying force is assumed to be
zero. When the end-plate is considered as 'thin', the prying force is at its
maximum. For 'intermediate' plates, the prying forces are somewhere be-
tween zero and the maximum value.
Kennedy e t al. 22 suggest a design method based on their findings, how-
 Design of moment end-plate connections 141

ever, only two tests were used to verify the method. They suggest that for
ideal design, the end-plate should be 'thick' under service loads, 'inter-
mediate' under factored loads and function as a 'thin' plate at ultimate loads.
Srouji, 12 Hendrick 13 and Morrison 14 have modified the Kennedy et al. 22
approach for use with two- and four-bolt flush end-plates; four-bolt, stiffened
extended end-plates; and extended end-plates with multiple bolt rows below
the tension flange. Each researcher has presented experimental evidence to
verify the prediction equations.
A h u j a 19 and Ghassemieh 2° have presented finite element/regression
analysis equations to predict bolt forces above the pretension level for
eight-bolt, stiffened, extended end-plates. Ahuja's results are based on
elastic material properties and Ghassemieh's results include inelastic
material properties.

Column side design

Relative to end-plate research, the amount of effort devoted to the column

side of end-plate m o m e n t connections is quite limited. Only a few papers
have been published which suggest design guidelines for the three failure
modes: column web yielding, column web buckling and column flange
bending failure.

W e b yielding
The critical section for column web yielding is at the toe of the column web
fillet. For design of welded connections, the present AISC Specification 23
criterion is based on a load path which is assumed to vary linearly on a 2½:1
slope from the beam flange through the column flange and fillet. If the stress
at this critical section exceeds the yield stress of the column material, a
column web stiffener is required opposite the beam compression flange.
For the case of end-plate moment connections, the width of the stress
pattern at the critical section may be considerably wider due to the insertion
of the end-plate into the load path.
Mann and Morris ~° reviewed the results of several research programs
pertaining to column webs at end-plate connections and proposed design
criteria. It was stated in their report that the 1977 edition of the 'European
Convention of Constructional Steelwork Recommendations for Steel Con-
struction' considers the stress to be distributed on a 1:1 slope through the
end-plate and on a 24:1 slope through the column. The same failure criterion
as in the AISC Specification is also specified.
Hendrick 24conducted a number of column compression region tests using
both stiffened and unstiffened end-plates and concluded that the slope of
the stress path through the end-plate can be taken as 1:1 and that in the
142 T. M. Murray

column as 3:1. This recommendation is also found in Hendrick and

Murray. 25 Hendrick's recommendations, except for the 3:1 slope, are also
found in the AISC Engineering for Steel Construction manual 26where 2~: 1 is
used.

Web buckling
Newlin and Chen 27 recommend that an interaction equation be used to
check combined web yielding strength and web buckling. Possibly antici-
pating resistance to such form, they also provided a simple check for web
buckling. This latter provision was adopted by AISC in their 1978 specifica-
tion revision.
Witteveen et al. zs conducted tests in the Netherlands in an attempt to
develop design rules to compute the moment capacity of unstiffened welded
(no end-plate) and bolted (end-plate) connections. Both full connection
tests and simulated compression flange tests were conducted. They recom-
mend, as a lower bound solution for the 'failure load (buckling, crippling or
yielding of the web in compression)', a solution similar to Hendrick's except
with a 2~: 1 stress distribution slope in the column.

Flange bending
Fisher and Struik 5 state that the problem of end-plate to column connection
is extremely complex and that 'no satisfactory design approach is available at
the present time (1973).' However, they suggest criteria based on welded
beam-to-column connections, pending further research, and which result 'in
a conservative design' since 'the concentrated forces are more localized in
welded connections.'
Fisher and Struik 5 recommend the use of European criteria found in
E C C S 29 for stiffener requirements regarding the column flange in the
tension region. The column flange is considered adequate if the moment
induced by the end-plate connection on the flange over an effective length is
within certain limits. They suggest column stiffeners should be proportioned
only to carry the excess concentrated force that the column web and flange
are unable to resist. It is further recommended that for beam-to-column
connections where only one beam frames into the column or where the
moment from one beam is much greater than the other beam, the column
web should be checked for shearing stresses.
Witteveen et al. 28found three modes of failure for the column flange. The
first mode prevails when the column flanges are thick when compared with
bolt diameter. The second failure mode is when the stiffnesses of the bolts
and flanges are such that prying forces can develop. Yield lines form in the
flange near the fillet between the flange and both the flange and the bolts fail.
The third failure mode occurs when yield lines form in the flanges near the
 Design of moment end-plate connections 143

bolts and the fillet. Design provisions for each failure mode are presented as
well as test results to verify the analytical work.
Mann and Morris ~° present complete design procedures for the column
side of end-plate connections which are based primarily on the work of
Packer and Morris. 8 In this work, only the case where the column
flange was much less stiff than the end-plate was studied. For this situation,
three possible failure modes were found to exist. If the flange is very stiff
there are no prying forces and the failure occurs when the bolts rupture. The
second failure mode occurs when the column flange is less stiff which results
in a combination of bolt fracture and flange yielding near the column web.
The third failure mode is when a mechanism occurs with yield lines forming
so as to cause double curvature in the flange plate. Provisions to estimate the
column flange capacity for each of the failure modes are provided. If the first
failure mode governs, the total bolt force is equal to the applied flange force.
For the second failure mode, prying forces are accounted for by limiting bolt
capacity to 80% of tensile capacity. Mann and Morris ~° do not provide
methods to estimate prying forces if the third failure mode governs.
Granstrom 3° extended tee-hanger results to include column flanges. The
procedure to determine the required column flange thickness is the same as
that used for tee-hanger flange thickness except that an effective column
flange length is used. Granstrom does not consider prying action effects.
Tarpy and CardinaP 7 conducted tests on unstiffened beam-to-column
flange end-plate connections. The end-plate thickness ranged from approxi-
mately 1.5 to 4 times the column flange thickness and sections which would
require stiffeners using the then current (AISC) design specifications were
tested. A column axial load, equal to approximately one-third of the column
yield load, was maintained throughout the tests.
A n elastic analytical finite element model for the full depth of the con-
nection was then developed coupling the end-plate and column flange. The
displacement of the column flange and the stress at the toe of the fillet were
d e p e n d e n t variables while the connection geometry parameters were the
independent variables. Standard multiple linear regression analysis tech-
niques were used to derive the prediction equations for the two independent
variables. The resulting end-plate/flange thicknesses were found to be very
large by Hendrick. 24
Hendrick 24 conducted a limited series of tests to evaluate the methods
suggested by Mann and Morris, ~0Witteveen e t a l . 28 and Granstrom 3° for use
with North American rolled sections. He concluded that the method pro-
posed by Mann and Morris 1° is the most suitable for the evaluation of
unstiffened column flanges in the tension region of four-bolt, unstiffened
end-plate connections. He also modified the Krishnamurthy 15procedure for
end-plates by introducing an effective column flange length to obtain the
144 7". M. Murray

same results as found with the Mann and Morris equations. Finally, he
developed a 'rule of thumb', found in AISC Engineering f o r Steel Con-
struction manual, 26which states that, under certain limitations, the column
flange is adequate if its thickness is greater than the required bolt diameter
from the Krishnamurthy end-plate design procedure.
Curtis 3~ has proposed design rules for column flange strength in the
tension region of eight-bolt, stiffened end-plate connections. His method is
based on the Ghassemieh 2° end-plate design procedure with an effective
column flange length.

C o l u m n stiffeners
Procedures for the design of column web stiffeners to prevent web yielding
or buckling have been suggested by Hendrick and Murray 25 and have the
same form as found in the current AISC specificationz3 for welded beam-
to-column connections.
Mann and Morris ~°have presented methods to estimate the resistance of
column flanges stiffened as shown in Fig. 3(a). Zoetemeijer 7 has recom-
mended the form of stiffening shown in Fig. 3(b) which has become known
as 'flange washer plate stiffeners.' He has also provided design rules for the
four-bolt unstiffened end-plate configuration, however, confirming tests
were not reported. Curtis 3~ reported extensive analytical (yield-line) and
experimental work on washer flange stiffening at both four-bolt unstiffened
and eight-bolt stiffened, extended end-plates.

R E C O M M E N D E D DESIGN P R O C E D U R E S

Flush end-plates including bolt design

Srouji ~2 evaluated the design methods suggested by Douty and McGuire, 2

Zoetemeijer, H Kennedy et al. 22 and those in the French (Norme Franqaise
Enregistrre 32) and German (DSTV/DAST 33) specifications for the two-bolt
flush configuration and found that the variation between the methods in
required end-plate thickness was over 100%. He concluded that 'it is diffi-
cult to justify any of the methods and that further research is required.'
Srouji subsequently developed procedures for the two- and four-bolt
flush configurations (Fig. l(a) and (b)), using the yield-line method to
determine end-plate thickness and a modification of the procedure sug-
gested by Kennedy et al. 22 to predict bolt force magnitude including prying
effects. He then conducted series of tests for each configuration and ob-
tained excellent agreement for both end-plate strength and bolt force
magnitude.
 Design of moment end-plate connections 145

--J Web
1 ~._~ffener
Flange
Stiffener
\
(a) (b)
Fig. 3. Types of column stiffeners. (a) Web stiffener; (b) column flange plate stiffener.

T h e controlling yield-line patterns are shown in Fig. 4. The required

end-plate thicknesses are:
(a) two-bolt flush configuration (Fig. l(a))

tp = (h - pt)[(bf/2)(1/pf + 1/s) + (pt + s)(2/g)] (1)

where tp = required plate thickness,

s = 1/2x/(bfg)

and all other variables are defined in Fig. 4(a).

(b) four-bolt flush configuration (Fig. l(b))

Mu/Fpy } '/2 (2)

tp = ( h - P t ) [ ( b d 2 ) ( 1 / P f + 1/u) + (2/g)(pf + Pb + U)] -- btpb/( 2u)

where

and all other variables are as defined in Fig. 4(b).

Srouji 12c o n d u c t e d seven two-bolt flush and six four-bolt flush tests, using
A572 G r 50 plate material and A325 bolts to verify the above equations. T h e
ratio of actual failure moment-to-predicted failure m o m e n t for the two-bolt
tests varied from 0.94 to 1.08 and for the four-bolt tests from 0.97 to 1.06.
Typical m o m e n t versus plate separation plots are shown in Fig. 5.
146 T. M. Murray

bf bf

'Pt "C)- _ f L ;,,''" "'" ~u p~bf

b
- / ---5", "!
""I ~'(

O O O O

(a) (b)
Fig. 4. Yield-line patterns for flush end-plate configurations. (a) Two-bolt flush end-plate;
(b) four-bolt flush end-plate.

Kennedy et al. 22 consider the bolt force to be the sum of a portion of the
flange force and prying forces. The amount of prying action corresponds to
the stage of behavior the end-plate is within. When the end-plate is 'thick'
the amount of prying action is negligible and is taken as zero. When the
end-plate is 'thin' the prying force is at its maximum value. For 'inter-
mediate' behavior, the prying force is somewhere between zero and its
maximum value. The basic analytical model is shown in Fig. 6.
The analytical model of Fig. 6 was modified by Srouji '2 as shown in Fig. 7
for the two-bolt, flush end-plate and as shown in Fig. 8 for the four-bolt,
flush end-plate. The two-bolt model is essentially one-half of the original
analytical model. The force in each bolt is one-half of the flange force plus
prying action forces. The four-bolt model is similar to the two-bolt model
with the addition of a second row of bolts. The force in the second row of
bolts is also unknown which produces an indeterminate problem. In order to
obtain the bolt forces for this particular model, assumptions had to be made
concerning the force in the inner row of bolts. The force in the inner bolt
(B2), was assumed to be a function of the flange force (Ff), depending on the
stage of end-plate behavior. Srouji t2 assumed that for 'thick' end-plates, B2
is zero, for 'intermediate' end-plates, B2 = Ff/10, and for 'thin' end-plates,
B2 = Fd6. These values were determined from experimental results. The
force in the outer bolt was then taken as the remaining flange force plus the
force due to prying.
 Design of moment end-plate connections 147

125

100

~ 7s

50

25

0
0 5 10 15 20 25
ROTATION, RADIANS x 10 .3
(a)

200

175

150

125

100
Beam Line
75 • Type I
.... Test
50

25

0
5 lO 1; 2;
ROTATION, RADIANS x 10 .3

(b)

Fig. 5. Typical moment versus rotation relationships from flush end-plate tests. (a) Two-bolt
flush connection test; (b) four-bolt flush connection test.

A second assumption made by Srouji 12 was in the location of the prying

action, the a distance. In the split-tee analogy of Fig. 6, the value of a was
limited between 2db and 3db because of a definite distance to the edge of the
end-plate. However, for flush end-plates, the location of prying action is
below the bolt line and is not limited by the edge of the end-plate. Srouji 12
suggested using a = tp if tp/db < 2/3 and a = 2tp, otherwise. By comparing
measured bolt forces to those predicted using the models shown in Figs 7 and
148 T. M. Murray
2F

Fig. 6. Kennedy etal. 22 analyticalmodel for bolt force prediction.

"°' ! a .t
O

Fig. 7. Modified Kennedy model for two-bolt flush end-plate.

8, H e n d r i c k ~3concluded that the location of the prying action force can be

approximated using

a = 3.682 (tp/d~) 3 - 0.085 (3)

H e n d r i c k also suggests that the assumption B2 = Fe/6 for 'thin' end-plates be

changed to F d 8 for the four-bolt unstiffened configuration (Fig. l(b)).
Using all of the above, the following steps are suggested to predict bolt
forces including prying action for two-bolt and four-bolt, unstiffened, flush
end-plate connections.
 Designof momentend-plateconnections

IL
149

Ff

Mb Mb

t. ., L ~° ± 3'
r" -g "g "1
B1 82 Q

Fig. 8. Modified Kennedy model for four-bolt flush end-plate.

(1) With a given end-plate moment for which the bolt forces are to be
determined, calculate the resulting flange force, Ff,

F, = M/(a - t,) (4)

(2) Find the thick plate limit, tt, using the following approximate
equation

t, = ~/(4.21 P~Ff/(bfFpy)) (5)

Then refine the value using the following exact iterative equation:

tl = 2 (6)

Once the thick plate limit is determined, the actual end-plate thickness (tp) is
compared to it. If tp > t~, then the prying force, Q, is taken as O, otherwise,
the next step is used.
(3) Find the thin plate limit, t~, using the following approximate equation

J {2(Ffpf-rcd3bFyb/16)]
tl~ = \ Fpy(O'85bf/2+ 0.8w')] (7)
150 T.M. Murray

where Fyb = yield stress of the bolt. Then refine the value using the exact
iterative equation:

When performing the iterations in the thin-plate stage, if the flange force
becomes large, a negative value can occur under the radical in eqn (8). If this
occurs then the end-plate is failing due to shear effects rather than yielding
due to bending. Thus, the following limitation should always be satisfied
before performing iterations with eqn (8):

Ff < 2w' tll Fpy/N/3 (9)

where tll is found from eqn (7). If eqn (9) is not satisfied, the beam capacity
must be increased so that shear failure does not occur. Again the end-plate
thickness is compared to/11. If tp > t., the plate is intermediate and one of
the two equations for prying action is used. To determine the prying force
for one-row, two-bolt, flush end-plate connections (Fig. l(a)), the following
equation is to be used.

Q_FP~a bft2p8a J ( F ~ y - 3 (2b-~o)2) rrd3bFyb32a (10).

where F = flange force per bolt = Fd2 and a is found from eqn (3). The
following is used for the two-row, four-bolt flush end-plate connections
(Fig. l(b)):

Q- (a+pb) 8(~-+--pb)~/F~y-3 \b--~p) 16(a+pb)

where F2 = Fd2 and Pb = pitch between bolt rows.

The bolt force (B), for two-bolt, flush end-plate connections is equal to

B = FJ2+Q (12)

but B must be at least equal to the pre-tension force. The outer bolt force in
the four-bolt, flush end-plate connections is given by

B, = FJ2.5 + Q (13)
 Design of moment end-plate connections 151

(4) If tp < tn the end-plate is said to be thin and the prying force is at its
maximum:

Qm,x - 4a F~y-3 (14)

where F' is the lesser of the following;

F,i=, = ~ Fp,(0.85bf/2 + 0.S0w') + ,,d~ F,b/8
4p, (15)

or F=., = (F0=.x/2

50 i
i
i

40 /
M y b / /
re 30 Proof Load ~ ~ " ~
©
u.

.J
© 20
tn
. Kennedy
Kennedy • ........ Test
10

o.... ;o ;o 6'o ;o 1~0

MOMENT
(a)

50

40
Proof Load Myil ,/~
30
n-
O
LL
p-
20 Kennedy ] ....
.J __ Kennedy
o¢D ........ InnerBolt
10 - --- --OuterBolt

i i. i

0 25 50 7'5 1O0 125 150

MOMENT
(b)
Fig. 9. Typical bolt force versus moment relationships from flush end-plate tests. (a) Two-
bolt flush connection test; (b) four-bolt flush connection test.
152 T. M. Murrav

The bolt force for two-bolt, flush end-plate connections is then given by
B = Ff/2 + Qma× (16)
For four-bolt, unstiffened, flush end-plate connections, the outer bolt
force is then given by
B1 = 3Ff/8+ Qm~x (17)
Figure 9 shows typical analytical/experimental results from the tests con-
ducted by Srouji. ~2
Moment-rotation curves for the tests conducted by Srouji are presented
by Hendrick. 13 Figure 10 shows two typical curves with the test beam
beam-line also plotted. Hendrick concluded from his work that the full
moment capacity, which can be calculated from eqns (1) or (2) for a given
end-plate thickness, can be used only in Type III (semi-rigid framing)
construction as defined in the AISC specification. 23For Type I construction
(rigid or continuous framing), the capacity should be limited to 80% of that
given by eqns (1) and (2).
The design procedures are based on specific yield-line mechanisms. Signi-
ficant changes in the geometric relationships can affect the mechanism
configuration and thus the predicted capacity. For these reasons, the
following limitations apply to the above design procedures:
pf -< 2.0 in. tp/db <-- 1"0
g--< 4"0 in. bf/g <- 2.25
Fy <- 50 ksi A325 bolts

Extended end-plates including bolt design

Four-bolt, unstiffened
Design procedures for four-bolt, extended unstiffened end-plates (Fig.
2(a)), have been developed by Douty and McGuire, 2 Zoetemeijer, 7
Krishnamurthy 15 and Mann and Morris. ~° The procedure formulated by
Krishnamurthy is found in Ref. 1.
Krishnamurthy's procedure is allowable stress design based with regres-
sion analysis prediction equation reformated using a modified split-tee
analogy. The required end-plate thickness (tp), is found from

tp = ~/((6Me)/(bpFb)) (18)

where Fh = the allowable bending stress (0.75 Fy) and

Mr = ~mFfp~/4 (19)
 Design of moment end-plate connections 153

70 tt

60
e/

50
Lu
n-
O 40
LI.
I.-
o 30

20

F.E.M. (Near Bolt)

10 ........ T E S T (Near Bolt In)
~ ~ T E S T (Near Bolt Out)

I i I I I |

0 100 200 300 400 500 600

MOMENT
(a)

700

/
600

500

400
p-
Z
~ 300
o
200
- - F.E.M.
....... TEST
100

0 0.02
' 0.04
' 0.06
' 0.08
' 0.1
' 0. ',2
PLATE SEPERATION

(b)
Fig. IO. Typical analytical and experimental results from eight-bolt stiffened end-plate tests.
(a) Bolt force versus moment results; (b) moment versus plate separation results.

with

Pe = Pf - (db/4) - 0.707 t . (20)

am = Ca C b ( A f / A . ) °'333 (pe/db) °'~ (21)

where pe = effective pitch, t. = fillet weld size, Ca = a constant depending

o n t h e p l a t e y i e l d stress a n d bolt material, Cb = (bJbp) °s, A , = area o f
154 T. M. Murray

beam flange, Aw = area of beam web and be = beam flange width. Bolt
forces are calculated directly from the beam flange force (B = Fd4) without
consideration of prying action forces.
Krishnamurthy ~5 verified his design procedure with ten full scale tests.
However, the failure modes listed were associated with the beam (flange or
web buckling, torsional twisting) and not the end-plate, welds, or bolts. The
only ratios cited are maximum 'moment developed by beam in test-to-
ultimate m o m e n t capacity of beam'. The ratios varied between 0.58 and
1.04.
Krishnamurthy does not indicate what failure mode(s) are included in his
design procedure nor is it evident from his test results. Also, construction
type suitability is not discussed although analytical moment-rotation curves
are suggested. 34 From the discussion and examples in the AISC Manual,
Type I construction use seems to be implied.

Four-bolt, stiffened
Formalized design procedures for four-bolt stiffened (Fig. 2(b)), are not
found in the literature. Krishnamurthy ~5 has done some work, but a final
procedure was never published. Morrison ~4is currently doing an analytical/
experimental study. His design procedure will be based on yield-line end-
plate analysis and bolt force predictions using the methodology described in
a previous section.

Eight-bolt, unstiffened
Although no known work has been done for the eight-bolt unstiffened
configuration (Fig. 2(c)), the design procedure found in Ref. 1 implies that
the Krishnamurthy procedure can be used for this configuration without
modification.

Eight-bolt, stiffened
Using the work of Ahuja ~9 and Ghassemieh, 2° Murray and Kukreti 35 have
developed a design procedure for the eight-bolt, stiffened end-plate config-
uration (Fig. 2(d)). The procedure is based on finite element/regression
analyses and includes a check of maximum bolt force. Six eight-bolt,
stiffened tee-stub and six full-scale end-plate tests were conducted to verify
the method. Figure 10 shows typical correlation between predicted and
measured end-plate separation and bolt forces versus moment.
Pertinent geometry is shown in Figure I 1 and the design steps follow.
(1) Select beam size for the required moment and compute the factored
flange force Fu from

F. = M J ( d - trb) (22)
 Design of moment end-plate connections 155

where M, = the factored moment capacity, d = beam depth and

tfo= beam flange thickness.
(2) Estimate the maximum bolt force Tfrom
T = Fu/6"8 (23)

This relationship is based on the four near bolts being fully effective and
the four far bolts being effective to the pretension level, approximately
0.7 times the tensile strength of the bolt.

t5
--fV db

• H
f
Pb

"r Pf
// // tf

.! ,, Pf

tp-- 2 J Pb

bp

Fig. 11. Geometryof eight-boltstiffenedextendedend-plate.

(3) Establish end-plate geometry: g, pf, Pb, bp and ts (see Fig. 11). The
stiffener thickness ts should be approximately the beam web thickness
tw.
(4) Determine the required end-plate thickness tpl to satisfy a stiffness
criterion of less than 0.02 in plate separation at the flange at maximum
loading.

0"00553 p0.873 g0.577 R . 9 1 7

/pl ~ A0.924 ~,0.112 ~0,682 (24)
t,t b ts vp

(5) Determine the required end-plate thickness /p2 to satisfy the strength
criterion:

0"0037 1 p0.257g0-548Flu017
tP 2 = A0.719 ~,0.162 ~0,319
o~ b ts vp
(25)
156 T. M. Murray

(6) Select an end-plate thickness tp greater than to1 and/p2 o

(7) Calculate m a x i m u m bolt force from:

2-305 x 1 0 - s p °'59LF2u583
T : ~.885d~.909~.327vph °'965 + e'r (26)

where PT is the bolt pre-tension force.

(8) C h e c k bolt size so that

T < Tu (27)

(9) Redesign if necessary.

All variables in the above equations must be in inches and kips. The
p r o c e d u r e is applicable only for A36 end-plate material and A325 bolts.
T h e stiffness requirement in the design procedure permits the use of the
connection in Type I construction.

Beam-to-end-plate welds

It is r e c o m m e n d e d that Griffiths' suggestions 21 be used to design beam-to-

end-plate welds. That is, the full tension flange capacity should be de-
veloped using either full-penetration or fillet welds, as well as the web
capacity in the vicinity of the tension bolts, even if the full bending strength
of the b e a m is not used.

Column side

Web yielding
T h e current A I S C criterion 23 for column web yielding was developed to
p r e v e n t local failure of H-shaped columns when beam flanges or m o m e n t
connection plates are welded directly to the column flange. T h e assumed
stress distribution is as shown in Fig. 12. Tests conducted by Hendrick 24 to
simulate the condition of end-plate connections showed that the distribu-
tion a s s u m e d for weld connections is conservative for end-plate conditions
and that the web yielding strength of an unstiffened column can be estimated
from

Pma, = Fyctwc(teo + 6 k + 2tp + 2tw) (28)

where P,,ax = m a x i m u m force the column web is capable of resisting,

Fyo = yield stress of column, two = column web thickness, tro= beam flange
thickness, k = column 'k' distance, tp = end-plate thickness, and tw =
 Design o f moment end-plate connections 157

\
\ ~ Flange
\
\
if Pbf
+

A
/I _ 2.5
/__J
/1

Fig. 12. Web stress distribution at welded connections.

b e a m flange to end-plate fillet weld leg dimension. Thus, in the form of the
current A I S C specification, the required stiffener area is

As~ = Pmax- Fyctwc(teb+ 6k + 2tp + 2t~) (29)

Fyst
where As, = required stiffener area if greater than zero and Fys, = yield
stress of stiffener material. This criterion is a significant liberalization of the
current criterion for welded, beam-to-column m o m e n t connections.
A l t h o u g h the criterion was developed for the compression region of the
c o l u m n web, it can be conservatively used at the tension region of the
c o l u m n web.
T h e above criterion is also found in Ref. 26 except that 6k is con-
servatively replaced by 5k for consistency with welded connections. Either
criterion can be used for any of the end-plate configurations shown in
Figs 1 and 2.

Web buckling
C o l u m n web buckling requirements at end-plate connections can be
c h e c k e d using the current AISC Specification, 23 that is stiffeners are re-
quired if

dc -< 4100 ?w<~/(F~<)lPm.~ (30)

where d c = column web depth clear of fillets.

158 T. M. Murrav

Column flange bending

To the writer's knowledge, design rules for column flange bending at flush
end-plate connections have not been formulated. Conservatively, the
column flange thickness should be equal to or greater than the end-plate
thickness if a column web stiffener of type shown in Fig. 3(a) is opposite the
beam tension flange.
Hendrick, 24 after reviewing the work of many investigators and con-
ducting a number of tests, recommends the following criterion based on the
work of Mann and Morris, ~° for determining column flange strength in the
tension region of four-bolt, unstiffened moment connections:
Fma = 4(0.8) T u

Fcol = m i n
~ (
Fmb = GcFyc 3 " 1 4 + (m+n------~
°5c I + 4(0"8)Tun
(re+n)
(3l)

Fmc= ~cFyc{3"14+(2n+c-d,)/m}
where Fco~= the column flange capacity, Tu = tensile bolt capacity,
Fyc = yield stress of the column steel and all other terms are defined in Fig.
13. In addition, the usable capacity of the bolts is to be limited to 80% of the
published tensile strength.
Hendrick 24 has also shown that the modification of the Krishnamurthy ts
end-plate design procedure can be used to evaluate column flanges at
four-bolt unstiffened end-plates, i.e. eqns (18)-(21). Hendrick found that if
Af/Aw in eqn (21) is taken as 1.0, pf is taken as l/2(g - two) in eqn (20), and
the end-plate width bp is replaced by an effective flange width equal to 3.5
times the bolt pitch, results equivalent to those from the Mann and Morris ~°
procedure are obtained, assuming an appropriate factor of safety is used.
Hendrick recommends using the AISC specification23 allowable stresses for
bolt design.

~ m
¢1 i Ill rl
Ii

oi',o
........
~ c

In'

O ©

Fig. 13. Geometry for Mann and Morris ~° flange strength procedure.
 Design of moment end-plate connections 159

Both Mann and Morris 1°and Hendrick 24state that, for a reasonable gage,
if the flange thickness is greater than the bolt diameter, the column flange is
adequate. In the AISC Engineering f o r Steel Construction manuaF 6 it is
suggested that this rule applies only if the horizontal distance from the toe of
the column web fillet to the center line of the hole is less than 2 in.
If stiffeners are needed to develop necessary column flange strength and
the type shown in Fig. 3(a) are used, the resistance of the stiffened flange can
be calculated from Mann and Morris: 10

Fms ---- ~cFyc v + (2m+2n'--dh)-~ (32)

m

w h e r e x = { m ( m + n' - 0.5dh)} 1/2, and v is as defined in Fig. 2(a).

If flange washer plate stiffening is used (Fig. 3(b)), the resistance of the
stiffened flange is given by: 7'1°

Fmz = Fyc + s___~t (33)

m

where ts = stiffener thickness and Fys = stiffener yield stress.

Recommendations for determining column flange strength at eight-bolt
stiffened end-plate connections cannot be made at this time.

SUMMARY AND RECOMMENDATIONS

FOR FUTURE RESEARCH

Relative to other types of beam-to-column m o m e n t connections, an

e n o r m o u s amount of effort has been devoted to the study of m o m e n t
end-plate connections. However, most of this effort has been concerned
with the behavior of the end-plate itself or in the determination of bolt forces
including possible prying action effects. Relatively little work has been done
on the behavior of the column side of the connection and only a few tests of
the complete joint with the column loaded in compression have been
reported.
However, complete design procedures for the four-bolt, extended un-
stiffened configuration, including the column side, are available. Thickness
criteria for flush end-plate and bolt force determination are also available,
however, criteria for the column side has not been suggested. Similar
statements can be made for the large capacity, eight-bolt, extended stiffened
configuration.
It is r e c o m m e n d e d that future research be directed at unifying the various
160 T. M. Murray

procedures using complete joint tests which include column axial load, at
investigating the suitability of end-plate connections for use in seismic areas,
at developing criteria for other configurations, such as the multiple row flush
end-plate configuration, and at developing moment-rotation data for the
various types of connection configurations.

ACKNOWLEDGEMENTS

Much of the information developed for this paper is the result of research
conducted at the University of Oklahoma under the sponsorship of the
American Institute of Steel Construction and the Metal Building Manu-
facturers Association. Their sponsorship, along with the works of former
University of Oklahoma graduate students, Vipul Ahuja, Larry Curtis,
Mehdi Ghassemieh, Alan Hendrick, David Hendrick, Ali Mazroi, Scott
Morrison and Ramzi Srouji is greatly appreciated. The effort and contribu-
tions of the writer's colleagues, Professor Abolhassan Astaneh-Asl and
Anant R. Kukreti, to this research are also sincerely appreciated. Finally,
the contributions of John D. Griffiths and R. H. R. Tide, who reviewed the
manuscript and offered valuable suggestions, are acknowledged.

REFERENCES

1. AISC, Manual of Steel Construction, Eighth edn, American Institute of Steel

Construction, Chicago, Illinois, 1980.
2. Douty, R. T. and McGuire, W., High strength bolted moment connections,
Journal of the Structural Division, ASCE, 91 (ST2) (1965) 101-28.
3. Nair, R. S., Birkemoe, P. C. and Munse, W. H., High strength bolts subject to
tension and prying, Journal of the Structural Division, ASCE, 100 (ST2) (1969)
351-72.
4. Kato, B. and McGuire, W., Analysis of T-stub flange-to-column connections,
Journal of the Structural Division, ASCE, 99 (ST5) (1973) 865-88.
5. Fisher, J. W. and Struik, J. H. A., Guide to Design Criteria for Bolted and
Riveted Joints, John Wiley and Sons, New York, 1974, Chapters 16-18.
6. AISC, Manual of Steel Construction, Seventh edn, American Institute of Steel
Construction, New York, NY, 1969.
7. Zoetemeijer, P., A design method for the tension side of statically loaded,
bolted beam-to-column connections, Heron, 20 (1) (1974) 1-59.
8. Packer, J. A. and Morris, L. J., A limit state design method for the tension
region of bolted beara-column connections, The Structural Engineer, 5 (10)
(1977).
9. Phillips, J. and Packer, J. A., The Effect of Plate Thickness on Flush End-Plate
Connections, Joints in Structural Steelwork, Proceedings of the International
Conference on Joints in Steelwork held at Middlesbrough, Cleveland, United
Kingdom Pentach Press, London, 1981, pp. 6.77-6.92.
 Design of moment end-plate connections 161

10. Mann, A. P. and Morris, L. J., Limit design of extended end-plate connec-
tions, Journal of the Structural Division, ASCE, 105 (ST3) (1979) 511-26.
11. Zoetemeijer, P., Semi-Rigid Bolted Beam-to-Beam Column Connections with
Stiffened Column Flanges and Flush End-Plates, Joints in Structural Steelwork,
Proceedings of the International Conference on Joints in Steelwork, held at
Middlesbrough, Cleveland, United Kingdom Pentach Press, London, 1981,
pp. 2.99-2.118.
12. Srouji, R., Yield-Line Analysis of End-Plate Connections with Bolt Force
Predictions, Master of Science Thesis, School of Civil Engineering and En-
vironmental Science, University of Oklahoma, Norman, Oklahoma, 1983.
13. Hendrick, D. M., Unification of Flush End-Plate Design Procedures, Master
of Science Thesis, School of Civil Engineering and Environmental Science,
University of Oklahoma, Norman, Oklahoma, 1985.
14. Morrison, S. J., Studies of Certain End-Plate Moment Connections, Master of
Science Thesis, School of Civil Engineering and Environmental Science, Uni-
versity of Oklahoma, Norman, Oklahoma (1986).
15. Krishnamurthy, N., Fresh look at bolted end-plate behavior and design,
Engineering Journal, American Institute of Steel Construction, 15 (2) (1978)
39-49.
16. Krishnamurthy, N., FEABOC (Finite Element Analysis of Bolted Con-
nections), Civil Engineering Dept., University of Alabama, Birmingham,
Alabama, 1981.
17. Tarpy, T. S. and Cardinal, J. W., Behavior of Semi-Rigid Beam-to-Column
End-Plate Connections, Joints in Structural Steelwork, Proceedings of The
International Conference on Joints in Steelwork, held at Middlesbrough,
Cleveland, United Kingdom Pentach Press, London, 1981, pp. 2.3-2.25.
18. Maxwell, S. M., Jenkins, W. M. and Howlett, J. H., Theoretical Approach to
the Analysis of Connection Behavior, Joints in Structural Steelwork, Pro-
ceedings of the International Conference on Joints in Steelwork, held at
Middlesbrough, Cleveland, United Kingdom Pentach Press, London, 1981,
pp. 2.49-2.70.
19. Ahuja, V., Analysis of Stiffened End-Plate Connections Using the Finite
Element Method, Master of Science Thesis, School of Civil Engineering and
Environmental Science, University of Oklahoma, Norman, Oklahoma, 1982.
20. Ghassemieh, M., Inelastic Finite Element Analysis of Stiffened-End-Plate
Connections, Master of Science Thesis, School of Civil Engineering and En-
vironmental Science, University of Oklahoma, Norman, Oklahoma, 1983.
21. Griffiths, J. D., End-plate moment connections---their use and misuse, AISC
Engineering Journal, 21(1) (1984) 32-4.
22. Kennedy, N. A., Vinnakota, S. and Sherbourne, A. N., The Split-Tee
Analogy in Bolted Splices and Beam-Column Connections, Joints in Structural
Steelwork, Proceedings of the International Conference on Joints in Steel-
work, held at Middlesbrough, Cleveland, United Kingdom Pentach Press,
London, 1981, pp. 2.138-2.157.
23. AISC, Specification for the Design. Fabrication and Erection of Structural
Steel for Buildings, American Institute of Steel Construction, New York, NY,
1978.
24. Hendrick, R. A., Column Strength at End-Plate Connections, Master of
Science Thesis, School of Civil Engineering and Environmental Science, Uni-
versity of Oklahoma, Norman, Oklahoma, 1983.
162 T. M. Murray

25. Hendrick, R. A. and Murray, T. M., Column web compression strength at

end-plate connections, AISC Engineering Journal, 21 (3) (1984) 161-9.
26. AISC, Engineering for Steel Construction, American Institute of Steel Con-
struction, Chicago, Illinois, 1984.
27. Newlin, D. E. and Chen, W. E., Strength and Stability of Column Web in
Welded Beam-to-Column Connections, Fritz Engineering Laboratory, Report
No. 333.14, Lehigh University, Bethlehem, Pennsylvania, May, 1971.
28. Witteveen, J., Stark, J. W. B., Bijlaard, F. S. K. and Zoetemeijer, P.,
Welded and bolted beam-to-column connections, Journal of the Structural
Division, ASCE, 108 (ST2) (1982) 433-55.
29. ECCS, European Guidelines on the Use of High Strength Bolts in Steel
Connections, European Convention for Constructional Steelwork, Document
CECM-X-70-8D, 3rd edn, Rotterdam, The Netherlands, April 1971.
30. Granstrom, A., Bolted End-Plate Connections, Stalbyggnads Institute SBI
Report 86.3, September, 1980.
31. Curtis, L. E., Unstiffened and Washer Plate Stiffened Column Flanges at
Moment End-Plate Connections, Master of Science Thesis, School of Civil
Engineering and Environmental Science, University of Oklahoma, Norman,
Oklahoma, 1985.
32. Norme Frangaise Enregistr6e, Metal Construction, Joining by Means of Bolts
Controlled Tightening, Structural Requirements and Checking of Jointing (in
French), NF, 1979, pp. 22--460.
33. DSTV/DAST, Moment End Plate Connections with HSFG Bolts (in German),
IHE 1, 1978.
34. Krishnamurthy, N., Huang, H., Jeffrey, P. K. and Avery, L. K., Analytical
M-O curves for end-plate connections, Journal of the Structural Division,
ASCE, 105 (ST1) (1979) 133-45.
35. Murray, T. M. and Kukreti, A. R., Design of 8-Bolt Stiffened End-Plate
Moment Connections, Papers, Third Conference on Steel Developments,
Australian Institute of Steel Construction, Melbourne, May 20-22, 1985, pp.
145-9.