STEELCOMMITTEEOFCALIFORNIA

TECHNICALINFORMATION & PRODUCTSERVICE

DECEMBER, 1990
Design of single Plate
Shear ConnectS,ohs.
by
Abolhassan Astaneh
Steven M. Call
Kurt M. McMullin
with discussion by
Ralph M. Richard
DISCUSSION*
Design of Single Plate Shear Connections
Paper by ABOLHASSAN ASTANEH, STEVEN M. CALL and KURT M. McMULLIN
(lst Quarter, 1989) See page 7 of This Publication.
Discussion by Ralph M. Richard
The paper develops a design procedure for single plate shear
connections based upon the results of a shear-rotation de-
vice (shown in Fig. 4 of the original paper). The claim is
made that in previous studies "...the shear connectors have
been subjected to moment and rotation or only direct shear
without rotation." This is not true.
This writer developed a design procedure for single plates
based upon stub beam tests and full scale beam tests that
included realistic connection shears. Shown in Figs. 13 and
14 of this writer's paper are moment-rotation curves which
show the effect of shear and given on page 45 of that paper
is the analytical moment-rotation curve which indeed includes
the effect of shear. It was found, however, that for practi-
cally all single plate designs the ratio, e/ h, (eccentricity
divided by bolt pattern depth), was 0.5 or greater and as
shown in Fig. 13, the moment-rotation relationship is not sig-
nificantly affected by the connection shear. The reason for
this is that the maximum moment in single plate shear con-
nections occurs at about 1.5 times the service load. This is
shown for a three and a five bolt connection in Figs. 1 and
2, respectively, of this discussion paper and is in agreement
Ralph M. Richard is professor, Department of Civil Engineer-
ing and Engineering Mechanics, University of Arizona, Tuc-
son, Arizona.
with Astaneh's observation that "...based on observations
made during the tests, it appears that shear' tabs go through
three distinctive phases of behavior. At the very early stages,
a shear tab acts as a short cantilever beam with moment be-
ing dominant. As load increases, the shear tab acts as a deep
shear beam with the shear yielding effect dominant." Had
Astaneh performed a full scale test, he would have observed
that the shear tab does not begin the shear yielding phase
of action before application of 1.5 times service load. This
linear connection action is shown in the shear-rotation plots
of Fig. 9 in Astaneh's paper. Moreover, consider Astaneh's
Design Example 1. His design procedure results in a 21 in.
x 1/2 in. x 41/2 in. plate with a shear of 102 kips service
load. At 1.5 times service load, the shear stress in this 3 in.
long and 21 in. deep cantilever beam is approximately 15
ksi which is less than the shear yield stress of 21.6 ksi for
A36 steel. In his Design Example 2, he uses a 12 in. x IA
in. 41/2 in. plate with a service shear load of 33 kips. The
shear stress in this plate at 1.5 times service load is 16.5 ksi
which again is well below the yield stress of 21.6 ksi for A36
steel.
The research at the University of Arizona, based upon stub
beam tests, full scale beam tests, and inelastic finite element
analyses that used experimentally determined bolt-deforma-
tion results, found thaf the maximum connection moment
0
0
Moment
Uniform ' ..-
IO
" x ./' ..... ,s 'X
,span L=20 'x
' / C ......... .3,3/4 A325 ' x
/ Plate : 5/t6 x
' " l " l h l l l l l l l l h I H H ' ' ' I ' ' H ' ''1 'J'Lcltlll' '[ " '' ' ' ' ' ' l ' ' "' ' ' H I ' "' ' H' 0
lO 20 30 40 50 60 70 80
Uniform Load. If, Kips
Fig. 1. Single plate moments and eccentricities.
ca
c
400
3'50
300
3,90
190
100
50
O
'm rT?T?T? rn r TrrlTrTr' [.........I.........f .........I'" crrrr
't
, Loed Uniform '
Ir' Benm w 18 X ,' 1,5 '
, Span L=20
w Connection 5 . 3/4' A325
. ' Plate 5/16'
. , w.,.. 46I Kips
d
uLt u r[ uuaxt tlu&xu h i i ,, ,, i, I. . . . . . . , h . , , , ,
lo 30 30 40 50 60 70 80 90
Unilorm Iond. W, Kips
Fig. 2. Single plate moments and eccentricities.
3O
20 >
tO
5
o
'k Reproduced from AISC Engineering Journal, Volume 27, No. 3, 3rd Quarter 1990
occurred near or above 1.5 times working load as shown in
Figs. I and 2 of this paper. The structural engineering pro-
fession requires that structural elements (connections, beam,
etc.) must be designed to have the strength to resist the max-
imum value of the envelope of forces the element is subjected
during loading. For the single plate shear connection, the
maximum value of the moment the weld is subjected is at
about 1.5 times the service load. Beam end rotations at these
loads are of the order of 0.006 to 0.014 radians which are
well below the 0.030 test values used by Astaneh. For uni-
formly loaded beams, it is noted that in Design Example 1,
the end rotation of this beam is 0.0055 radians at service
load and for Design Example 2 it is 0.0046 radians. How-
ever, Astaneh's recommended test and design procedure
which is based upon shear yielding of the plate, used rota-
tions four to six times these values.
Because of the significant difference in the design eccen-
tricities recommended by Astaneh and those of this writer
for the design of the single plate welds, this writer strongly
recommends that a minimum of three full scale tests with
beams subjected to a factored uniform load of 1.5 times the
service load be performed by an independent laboratory to
evaluate the moment generated by the single plate shear con-
nection before this design procedure is recommended to the
structural engineering profession. This writer has found that
these connections generate significantly larger moments than
double framing angles subjected to the same beam shear. 2
Because the bolts of the single plate are in single shear,
whereas these are in double shear for double framing an-
gles, the single plate is twice as deep and therefore much
stiffer.
REFERENCES
1. Richard, R. M. , P. E. Gillett, J. D. Kreigh, and B. A.
Lewis, "The Analysis and Design of Single Plate Fram-
ing Connections," AISC Engineering Journal, Vol. 17
(No. 2): 38-52.
2. Richard, R. M., W. K. Hsia, and M. Chmielowiec, "Mo-
ment Rotation Curves for Double Framing Angles," In-
ternational Journal of Computers and Structures, Vol. 30
(No. 3): 485-94.
Addendum/Closure by A. Ast aneh, S. C. Call and K. M.
McMullin
The discussion by Professor Richard mainly compares the
research methodologies and design procedures developed by
researchers at the University of Arizona (UA Method) to
those developed by Authors at the University of California
at Berkeley (UCB Method). The UCB Method has formed
the basis of the methods currently in the 9th Edition of the
AISC Manual. u In order to make the closure of discussion
useful to the readers, the authors have responded to the state-
ments made in the above discussion and have provided a brief
comparison of the UA and UCB design methods in the fol-
lowing sections.
RESEARCH METHODOLOGY
In the paper, it was indicated that "... in the past in most
cases, the shear connections have been subjected to moment
and rotation or only shear without rotation instead of a realis-
tic combination of shear and rotation." This statement is par-
ticularly true with experiments conducted at the University
of Arizona. Figure I (reproduced from Fig. 3 of the paper)
shows representative shear-rotation relationship that existed
in the connections tested by Professor Richard and his re-
search associates (Lines OA and OB). Also shown in the fig-
ure are actual shear-rotation relationship in a shear connec-
tion (Line OCD) and shear-rotation relationship that existed
in UCB tests (Line OCE).
In the stub (cantilever) tests conducted at UA, the con-
nections were primarily subjected to rotations with very small
shear applied to the connection. The shear-rotation relation-
ship for these tests is represented in Fig. I by the line OA.
By comparing this shear-rotation line to the actual shear ro-
tation line (Line OCD), it is clear that the connections in
stub beam tests were subjected to unrealistically large rota-
tions with very small shear forces applied to the connection.
Since shear forces generated in stub (cantilever) beam tests
are small compared to actual shear forces in shear connec-
tions, failure modes are very unrealistic, therefore, unrealistic
tests should not be used to develop design procedures for
shear connections.
From published data apparently a total of four tests have
been conducted using the test set-up shown in Fig. 2. Simi-
lar test set-ups have been used in the past by several
researchers to apply large shear forces to the connection.
However, if the beam shown in Fig. 2 is not loaded to fail-
ure, the amount of rotation that will be developed in the con-
Q_
:f
O
0
z
z
0
0
Z
0
bO
ACTUAL SHEAR-ROTATIOb! CURVE D G4.
C Ac2uAL>-, ' '
100
0.02 0.04 0.06
ROTATION OF BEAM END, rod.
Fig. 1. Shear-rotation relationship in UA and UCB tests.
2
nection will be very small and will be limited to elastic end
rotations which are very small compared to realistic rota-
tions that will be imposed on the connection at the time of
beam collapse.
In the full-scale tests conducted at the University of Ari-
zona, the amount of maximum shear applied to the connec-
tions is unexplainably very low. A representative of the shear-
rotation relationship applied to the connections in UA full-
scale tests is shown in Fig. 1 as Line OB. Due to application
of very low shear to the connection in these full-scale tests,
no realistic failure mode has been observed or reported and
apparently only some minor yielding of bolt holes and defor-
mation of bolts have been observed.
It is unfortunate that full-scale tests conducted at the iJni-
versity of Arizona have not been loaded to failure. Appar-
ently, the loading was not even enough to cause significant
yielding in the connections. If the tests were destructive,
several failure modes observed by us as well as by other
researchers -6' might have been observed and invaluable
data on strength of connection could be obtained. The rea-
son for stopping the loading at such a low level apparently
was a decision to load the specimens up to 1.5 times yield
capacity of the beam. From published information, it is not
clear why strength of the connections were studied under such
an arbitrary and unrealistically low load level. Therefore,
in our view, full-scale tests conducted at the University of
Arizona were incomplete and have not provided informa-
tion regarding strength and failure modes of the connections.
The details of full-scale tests conducted at the University
of Arizona and the results are not published. However, from
published data, it appears that the objective of full-scale tests
at the University of Arizona may have been to study move-
ment of point of inflection of the beam and moment-rotation
behavior. Since these full-scale tests have been non-
destructive and no connection failure modes have been ob-
served, it is not clear how the information obtained from
loading of specimens in elastic range could be used to de-
velop design procedures concerning failure modes and the
corresponding shear strength capacities.
The inelastic finite element program used in UA studies
is an analysis program and could only provide useful infor-
{top and bottom)
W Section
Load Cell I ,
/2 /2
i
Fig. 2. Test setup used in UA tests (Ref 8).
mation on the state of the strain and or stress. The program
is not capable of predicting failure modes and strengths such
as weld fracture, bolt fracture, fracture of net section or frac-
ture of the edge distance. Apparently, the finite element pro-
gram is used to simulate moment-rotation response. Again,
similar to full-scale tests, in the finite element analyses the
maximum load was about 1.5 times service load of the beams.
As far as behavior of the connection is concerned, the max-
imum load of 1.5 times service load of the beam used in UA
tests and finite element analyses is very small. For exam-
ple, the connection studied in Fig. 2 of the Discussion is
loaded up to about 50 kips shear force (100 kips total beam
load) whereas according to information obtained from our
destructive tests of similar connections and by using well
established design concepts, the shear capacity of the con-
nection is about 130 kips (260 kips total beam load). It
appears that the University of Arizona studies were limited
to the initial stage of loading where beam and connection
are almost elastic. Then the results of these studies are applied
to full range of loading up to the failure. Since the problem
is highly nonlinear, the validity of this extrapolation is
questionable.
To remove the above difficulties, the authors have devel-
oped and used a test set-up that has enabled them to apply
realistic combinations of shear and rotation to the connec-
tion until the connection fails. The shear-rotation relation-
ship used by the authors is shown in Fig. 1 as Line OCE.
The details of test set-up as well as authors' methodology
are given in several references (1 to 6) and are not repeated
here. The experimental work has resulted in establishing
realistic failure modes and corresponding design procedures
as reported in the paper.
COMPARISON OF UCB DESIGN PROCEDURES
WITH UA PROCEDURES
The destructive tests conducted by a number of researchers
including the authors have indicated that single plate shear
connections have six failure modes as follows:
a) shear yielding of plate
b) bearing failure of bolt holes
c) failure of edge distance
d) shear fracture of net section
e) bolt failure
f) weld failure
The following sections provide a discussion of each fail-
ure mode and corresponding design equations in UCB
Method and UA Method. In summarizing UA Method, the
authors have used the available published information. TM
a. Shear Yielding of Plate
In UCB method, this failure mode, which is very ductile and
desirable, is intentionally made to be the governing failure
mode.
3
The equation to be used to calculate the ultimate shear
strength of connection for this failure mode is:
R, = (L)(t)(0.6F,.) (1)
In UA method, this failure mode is not recognized.
b. Bearing Failure
In the UCB studies, t 6 bearing failure was observed in some
specimens. In the corresponding design procedures bearing
failure mode is recognized and equations that already exist
in the AISC Specification u are used to predict bearing fail-
ure capacity of the connections.
In UA method, this failure mode is not considered. Using
UA method, since there is no lower limit on the thickness
of the shear tab, it is quite possible that designer unknow-
ingly can use a thin plate with relatively large diameter bolt
and cause bearing failure to be governing without ever notic-
ing it.
The UCB design procedures as well as UA method recog-
nize the beneficial effects of limited bearing yielding at the
bolt holes. As a result both methods have an upper limit of
thickness of plate relative to the bolt diameter. In UCB
method the limit is dr,/2 + h6 inch and in UA method the
limit is d,/2. The limited bearing yielding provides rota-
tional ductility and causes release of moment in the con-
nection.
c. Shear Fracture of Net Area
I n U C B method this failure mode is fully recognized and
the following design equation is recommended to be used
to predict ultimate shear capacity of the net area:
R,,/, = [L - N(V2)(dl, + 6)l(t)(0.6F,) (2)
In a conservative approach, Eq. 3 which reflects the phi-
losophy used in the AISC Specification u for shear failure
of net area can be used.
R,a, = [L - N(dh + 6)](t)(0.6F,) (3)
The UA method apparently does not consider this failure
mode. Again, similar to bearing failure mode, it is possible
that by using thin plates, net section failure can govern with-
out the knowledge of the designer.
d. Edge Distance Failure
As a result of experiments conducted by the authors at UCB,
it was realized that due to dominance of shear, the vertical
edge distance below the lowest bolt is the most critical edge
distance and should not be less than 1.5db nor 1.5 in. In
UCB design method, it is recommended that this limitation
be applied to all edge distances (see Fig. 3a).
In UA method, it is recommended that horizontal edge dis-
tance should not be less than 2do (see Fig. 3b). Apparently
this recommendation is derived from results of cantilever
+o* .... hra la,,mc or cnlaioptcl tn lrof rntation and small
shear forces. In our tests, the horizontal edge distances did
not show signs of being critical whereas vertical edge dis-
tances particularly the lower vertical edge distance proved
to be very important and critical.
e. Failure of Bolts
In UCB method, bolts are designed for the combined effects
of direct shear and bending moment along the bolt line. Our
tests indicated that as beam is loaded, connections yield and
bending moment in the connection continuously is released
to the midspan of the beam. As a result, point of inflection
of the beam continuously moves toward the connection and
is stabilized at a distance of el, from the bolt line. The value
of et, can be obtained from the following equation.
el, = (n - a - 1)(1.0) (4)
Therefore, in UCB method, bolts are designed to resist
combined effects of shear reaction of the beam and a mo-
ment equal to reaction multiplied by et,.
In UA method, bolts are designed for direct shear only.
This implies that bolt line is the location of point of inflec-
tion of the beam where moment is zero and only shear ex-
ists. Our experiments, as well as other tests conducted in
Canada, have clearly indicated that some moment de-
velops along the bolt line.
Figure 4 shows variation of shear force and bending mo-
ment in a typical shear tab connection. The connection used
to plot the curves is the same used in Fig. 2 of the Discus-
sion. Figure 4 shows an experimental curve, UA finite ele-
ment results and design equations according to UCB and UA
methods. It should be mentioned that test results shown in
Fig. 4 are plotted using test results for exactly similar speci-
men but with Y8 in. thick plate rather than 6 in. The test
results for 3/8 in. plate are multiplied by Y6to adapt them to
h6 in. plate and then are plotted in Fig. 4.
It is not known why UA's design method neglects the mo-
ment that exists along the bolt line. Even the finite element
Critical
)
Critical
Critical ical
Fie. 3. Ed.ee distance requirements in UCB and UA methods.
4
analysis given by Professor Richard in Figs. 1 and 2 of the
Discussion shows that considerable moment is present along
the bolt line. In our view, based on seven tests conducted
so far by us and several other tests by other researchers on
the shear tabs, neglecting moment along the bolt line is not
justifiable and can result in unconservatively overestimat-
ing shear capacity of the bolts.
f. Weld Failure
In UCB method welds are designed for the combined effects
of direct shear and a moment due to the eccentricity of the
reaction from the weld line, e,.. The eccentricity e,, is given
by the following equation. The equation is based on results
of tests.
e,. = n(1.0) (5)
In UA method welds are designed for combined effects
of shear and moment, however, the moment that is estab-
lished for design of the welds is unrealistically very large.
Figure 5 shows shear and moment variation along the weld
line for the same shear tab shown in Fig. 2 of the Discus-
sion. Similar to bolt design, the figure shows test results,
UA finite element analysis (adapted from Fig. 2 of the Dis-
cussion) as well as design equations according to both
methods. The plots clearly shows that if one follows UA
method in design of welds, the design point will be some-
where in the vicinity of point A where moment is much larger
and shear force is smaller than the realistic values that actu-
ally occur in the connection (test curve).
The reason UA method results in using very large and un-
realistic moment in design of welds is the use of large ec-
centricity. Notice that in Fig. 5, slope of lines drawn from
the origin (such as aA and OB) represent values of constant
eccentricity. In the Discussion Professor Richard indicates
that connection should be designed for maximum possible
values of shear and moment. This statement is correct, but
in UA's method rather than designing connection for maxi-
mum combination of shear and moment, the connection is
designed for shear corresponding to 1.5 times service load
of the beam and an eccentricity of shear that exists at the
point of 1.5 times service load of the beam. What this actu-
ally means is that as beam is loaded, eccentricity moves to-
ward the support and when shear force exceeds a value cor-
responding to 1.5 times service load of the beam, the
eccentricity remains constant. This is shown in Fig. 5 by
Line CA. This is not realistic. As Fig. 5 indicates in actual
loading shown by test curve, after onset of the bolt slip and
yielding in the connection (Point D), eccentricity decreases
continuously and stabilizes at much smaller value than the
eccentricity corresponding to Point C. This can easily be seen
by comparing slope of Line CA (e,,. = 13 in.) and Line EB
(e, = 5 in.).
In summary, tests conducted at the University of Arizona
were not destructive and thus cannot be used to establish fail-
ure modes and design procedures. And, furthermore, the cor-
responding design procedure considers only bolt failure and
weld failure which are only two of the six failure modes that
actually should be considered. In addition, the design equa-
tions suggested for the bolt failure appear to be unconserva-
rive whereas equations proposed for weld design are baed
on unrealistically large moment and a small shear.
The design procedures proposed by the authors are only
a step in direction of improving the design methods by us-
ing more realistic test results and failure modes. Much work
needs to be done in this area particularly with respect to cy-
clic behavior of these connections.
700
i
600
5OO
400
E
o 300
.__ 2oo
_J
0 100
CD
0
i I '
[ - eb=O
I ' , " , " T . i F - / , : , I ' '
20 40 60 80 100 200
....... UA'sAnalysis
- ,; ; UA'sDesign Line
. . . . UCBTest Result
UCB's Design Line
I I I I i I I
300
Uni f orm Load, W , Kips
400
1500
1400 --
1300
' i 1200
1300
1000
900
800
700
_ 600
500
400
2 300
- 200
100
. . . . . . . UA'sAnalysis
- - - - - UA'sDesign Une
Ultimate Strergth / . . . . UCBTest Result
of P/ot... / - - UCB'sDesign Line
Y i e ' d S t r e n g t h / m
I I I I I I I I I I' I I I I I I I'
20 40 60 80 100 200 300
Uni f orm Load, W , Ki ps
4OO
Fig. 4. Variation of shear and moment along the bolt line. Fig. 5. Variation of shear and moment along the weld line.
5
ACKNOWLEDGMENTS
The opinions expressed in this closure are those of the authors
and do not necessarily reflect the views of the AISC or the
University of California at Berkeley. The words "UCB
method" and "UA method" are only used to refer to the
methods developed by the authors and by the researchers
at the University of Arizona respectively.
sign of Single Plate Framing Connections with A307
Bolts," AISC Engineering Journal, Vol. 19 (No. 4).
10. Wyss, U., "Single Plate Connections for Steel Beams,"
Thesis presented to University of British Columbia,
April 1967.
11. American Institute of Steel Construction, Manual of
Steel Construction, 9th ed., Chicago: AISC, 1989.
NOTATION
a Distance between bolt line and weld line, in.
dt, Diameter of bolts, in.
et, Eccentricity of beam reaction from bolt line, in.
e,, Eccentricity of beam reaction from weld line, in.
L Length of shear tab, in.
M, Moment along bolt line, kip-in.
n Number of bolts.
R, Reaction of the beam causing yielding of shear tab, kips.
R,, Reaction of the beam causing fracture of net section,
kips.
t Thickness of shear tab, in.
W Total load carried by the beam, W = 2R, kips.
REFERENCES
1. Astaneh, A., "Experimental Investigation of Tee-
Framing Connection," Progress Report, submitted to
American Institute of Steel Construction, April 1987.
2. Astaneh, A., "Demand and Supply of Ductility in Steel
Shear Connections," Review in Journal of Steel Con-
' struction Research, March 1988.
3. Astaneh, A., K. M. McMullin, and S. M. Call, "De-
sign of Single Plate Framing Connections," Report No.
UCB/SEMM-88/12, Department of Civil Engineering,
University of California-Berkeley, July 1988.
4. Astaneh, A., K. M, McMullin, and S. M. Call, "De-
sign of Single Plate Shear Connections," AISC Engi-
neering Journal, Vol. 26 (No. 1).
5. Astaneh, A. and M. Nader, "Behavior and Design of
Steel Tee Framing Connections," Report No. UCB/
SEMM-88/ll, Department of Civil Engineering, Uni-
versity of California-Berkeley, July 1988.
6. McMullin, K. M., and A. Astaneh, 'Analytical and Ex-
perimental Investigations of Double-Angle Connec-
tions;' Report No. UCB/SEMM-88/14, Department of
Civil Engineering, University of California-Berkeley,
August 1988.
7. Richard, R. M., "Single Plate Framing Connection
Designs," Steel Tips, Steel Committee of California,
December 1986.
8. Richard, R. M., P, E. Gillett, J. D. Kriegh, and B. A.
Lewis, "The Analysis and Design of Single Plate Fram-
ing Connections," AISC Engineering Journal, Vol. 17
(No. 2).
9. Richard, R. M., J. D. Kriegh, and D. E. Hormby, "De-
AISC Commentary on Design of Shear Tabs
AISI and AISC sponsored research on single shear plate con-
nections (shear tabs) at the University of Arizona in the late
1970s and early 1980s. At the request of the AISC Commit-
tee on Manuals and Textbooks and the ASCE Committee on
Steel Building Structures, additional research was conducted
at the University of California-Berkeley in 1988-89. In each
case, the project scope and limit state criterion was suggested
by AISC and followed by the researchers. Because the limit
state was different in the two cases, the design procedure
resulting from each research effort is different. This is evi-
dent by the two preceding discussions in this issue of the
Engineering Journal, AISC assumes responsibility for these
changes in the context of a natural evolution of research and
improved understanding of shear tab behavior.
In the University of Arizona case, AISC directed the limit
state to be a maximum connection rotation in this initial re-
search on shear tab connections. Because AISC did not re-
quest tests to destruction, none were made. On this basis,
tests and analytical studies were made and a design proce-
dure appearing in several AISC publications was developed.
In the recent University of California-Berkeley case, the
limit state was changed to ultimate load, to be determined
by testing to destruction. Based on this work and previous
research, a different design procedure was then developed
by applying a conservative factor of safety.
The AISC Committee on Research and the AISC Com-
mittee on Manuals and Textbooks determined that the ulti-
mate load criterion given to the University of California-
Berkeley was more realistic and better represented the be-
havior traditionally assumed for steel connections. The ASCE
Committee on Steel Building Structures concurred in this
judgment.
AISC feels that both shear tab design procedures include
an adequate factor of safety and either can be safely used.
Because of the simpler nature of the new University of
California-Berkeley method, and because its strength limit
states are considered to be more complete and realistic, that
method was adopted for inclusion in the Ninth Edition of
the Manual of Steel Construction. Additional research on this
method to expand its applicability to other detailing condi-
tions is in progress.
AISC expresses its appreciation to both Professor Richard
and Professor Astaneh for their contributions to the solu-
tion of this vexing design problem.
6
Design of Single Plate Shear Connections*
ABOLHASSAN ASTANEH, STEVEN M. CALL AND KURT M. McMULLIN
INTRODUCTION
Single plate shear connections, often referred to as shear
tabs, have gained considerable popularity in recent years
due to their efficiency and ease of fabrication. Shear tab
connections are primarily used to transfer beam end reac-
tions to the supporting elements. The connection consists
of a plate welded to a support at one edge and bolted to
a beam web. Figure 1 shows typical applications of single
plate shear connections. This paper presents the summary
of a research project on the behavior and design of single
plate shear connections. Based on experimental and aha-
lyrical studies, a new design procedure is developed and
presented.
The AISC-ASD 5 as well as AISC-LRFD 6 specifica-
tions have the following provisions with regard to shear
connections:
"Except as otherwise indicated by the de-
signer, connections of beams, girders, or
trusses shall be designed as flexible, and may
ordinarily be proportioned for the reaction
shears only.
"Flexible beam connections shall accommo-
date end rotations of unrestrained (simple)
beams. To accomplish this, inelastic action in
the connection is permitted."
Steel shear connections not only should have sufficient
strength to transfer the end shear reaction of the beam but
according to above provisions, the connections should also
have enough rotation capacity (ductility) to accommodate
the end rotation demand of a simply supported beam. In
addition, the connection should be sufficiently flexible so
that beam end moments become negligible. Thus, like any
shear connection, single plate shear connections should be
designed to satisfy the dual criteria of shear strength and
rotational flexibility and ductility.
Shear-Rotation Relationship in a Shear Connection
To investigate the behavior and strength of a shear con-
nection, it is necessary that realistic shear forces and their
corresponding rotations be applied to the connection. In
Abol hassan Astaneh is assistant professor, University of Califor-
nia, Berkeley.
Steven M. Carl is graduate research assistant, University of
Cafifornia, Berkeley.
Kurt M. McMullin was graduate research assistant, University of
Cafifornia, Berkeley.
an earlier research project? the shear-rotation relation-
ship for the end supports of simply supported beams was
studied. A computer program was developed and used to
simulate increased monotonic uniform loading of the
beams supported by simple connections until the beams
collapsed. 1,2
The studies indicated that the relationship between the
end shear and end rotation is relatively stable and depends
primarily on the shape factor Zx/Sx of the cross section,
L/d of the beam and the grade of steel used. Figure 2
shows a series of curves representing shear forces and cor-
responding rotations that will exist at the ends of simply
supported beams. The curves correspond to beams of A36
steel having cross sections from W16 to W33 and L/d ra-
tios of 4 to 38. Also shown in Fig. 2 is a tri-linear curve
"abcd" suggested to be a realistic representative of the
shear-rotation curves. The tri-linear curve "abcd" is pro-
posed to be used as a standard load path in studies of shear
connections. Curve "abcd" is used instead of the more
conservative curve "aef" because it is felt that curve
"abcd" represents a more realistic maximum span-to-
depth ratio for most steel structures. For special cases of
very large span-to-depth ratio or high strength steels, the
rotational demand may be greater than that of curve
"abcd". For such cases special care must be taken to as-
sure the rotational ductility demand of the beam is sup-
plied by the connection.
CONCRETE SUPPORT COLUMN
(o) (b)
(d) (e)
BEAM BEAM
(c)
Fig. 1. Typical Single Plate Shear Connections
'k Reproduced from AISC Engineering Journal, Volume 26, No. 1, 1st Quarter 1989
7
The shear-rotation curves plotted in Fig. 2 are estab-
lished based on the assumption of elastic-perfectly-plastic
bending moment capacity for the beam. To include the ef-
fect of strain hardening, the segment "cd" in curve "abcd"
is included.
The behavior of shear connections has been studied in
the past by several investigators. 8342 However, in most
cases, the shear connections have been subjected to mo-
ment and rotation or only direct shear without rotation in-
stead of a realistic combination of shear and rotation. Fig-
ure 3 shows the shear rotation relationships that existed
in several studies including this research project.
EXPERIMENTAL RESEARCH
In order to identify limit states of strength and to verify
the validity of the design procedures that were developed
and proposed, five full scale beam-to-column connection
assemblies were tested. A summary of the experimental
studies follows. More detailed information on the research
project can be found in References 3 and 6.
Test Set-up
The test set-up shown in Fig. 4 was used to apply shear-
rotation relationship of curve "abcd" in Fig. 2 to the speci-
mens.
The main components of the test set-up were a com-
puter based data acquisition and processing system, two
actuators R and S and support blocks. Actuator S, which
was close to the connection, was force controlled and pro-
vided the bulk of the shear force in the connection. Actua-
tor R, which was displacement controlled, provided and
controlled the beam end rotation.
Test Load Path
The proposed standard shear-rotation relationship
shown as curve "abcd" in Fig. 2 was applied to the connec-
tions in all of the test specimens. To establish the curve,
coupon tests of the plate material were conducted prior
to connection tests and the yield point and ultimate
strength of the plate material were obtained. The shear
yield capacity of the single plate in each test specimen was
calculated by multiplying the von Miess criterion of shear
yield stress, 1/X/3Fi, by the shear area of the plate. The
shear yield capacity of the plate, denoted as Ry, was taken
as equal to the shear at point "c" of curve "abcd" in Fig.
2. Thus the shear yield capacity of the shear tab was as-
sumed to occur when the moment at midspan was equal
to flip. As a result, a corresponding Mp can be calculated
for each connection to be equal to RyL/4. The end rota-
tion of the beam when midspan moment reached Mp was
set equal to 0.03 radians.
To establish point "b" in curve "abcd", the shear at this
point was set equal to 4My/L and the rotation was set
equal to 0.02 radian. This implies that when beam
midspan moment reaches My, the end rotation will be
equal to 0.02 radian. The value of My, the end rotation
will be equal to 0.02 radian. The value of My for each spec-
imen was calculated by dividing Mp by the shape factor.
A shape factor of 1.12 was used in all specimens.
Segment "cd" in Fig. 2 corresponds to strain hardening
of the beam and the increased moment at beam midspan
which results in increased shear at the beam ends. To es-
tablish "cd", it was assumed that when the midspan mo-
ment reaches a value of (F,/Fy)Mp, the beam end rotation
will be equal to 0.1 radian.
In summary, load path "abcd" in Fig. 2 reflects the be-
havior of the beam and its effect on connection shear and
rotation. Segment "ab" corresponds to the elastic behav-
ior of beam. At point "b", midspan moment of the beam
reaches My and the beam softens. Segment "bc" corre-
sponds to inelastic behavior of the beam. At point "c", the
midspan moment reaches Mp. Segment "cd" represents
extra beam capacity that can develop due to beam strain
hardening.
!6
!41 d
LO=4 C
' 'f(: /,',,. //',//, ( / / o Legend:
-I/J w3o - -
06 _...___d' ./,' /g' Y' ?' w2r
,,, . , ,,, , , , .,. L . o . 3 8
W24 - - -
( / W 8 ....
02
-1 a Vy=4SxFy L
oo !Ir I I I
I t I
O0 001 002 003 004 005 006 007
END ROTATION, r a d .
Fig. 2. Shear-Rotation Relationship for Ends of Simple Beams
g
Z
9
F-
O
Z
z 100
0
0
Z
O
Ud
fao
F/g. 3.
,- Ref 8
sEAR
11 ROI 12 .[ . . . . . . . . . . .
J ,4,
0.0 002 0 04 006
ROTATION OF BEAM END, r a d .
Shear-Rotation Relationship used in Several Studies
8
Table 1.
Properti es of Test Specimens
TEST TEST NO. OF DIA. OF TYPE OF PLATE EDGE ACTUAL BEAM PLATE
GROUP NO. BOLTS BOLTS BOLTS* DIMENSIONS DI STANCE W E L D MATERI AL MATERIAL
SIZE
in. in. x in. x in. in. in.
ONE 1 7 3/4 A325-N 21 x 3/8 x 4-Y4 1-V2 1/4 A36 A36
2 5 % A325-N 15 x % x 4-Y4 1-V2 '/4 A36 A36
3 3 % A325-N 9 x % x 4-Y4 1-V2 % A36 A36
TWO 4 5 3/4 A490-N 14-V3 x % x 3-% 1-Va %2 Gr. 50 A36
5 3 3/4 A490-N 8-% x 3/8 x 3-% 1-V8 %2 Gr, 50 A36
*Al l bolts were tightened to 70% of proof load. In all specimens diameter of bolt hole was 6 inch larger than nominal diameter of bolt. indicates that
i n al l specimens threads were included i n shear plane.
.1:Size of all welds was speci fi ed as inch.
Test Specimens
Each test specimen consisted of a wide flange beam
bolted to a single plate shear connection which was welded
to a column flange as shown in Fig. lb. The properties of
the test specimens were selected in consultation with a
professional advisory panel. These properties are given in
Table 1. The bolt holes in all specimens were standard
round punched holes. All bolts were tightened to 70% of
proof load using turn-of-the-nut m e t h o d ? ' 4 All shear
tabs were cut from a single piece of steel. The yield stress
and ultimate strength for material of shear tabs were 35.5
ksi and 61 ksi respectively. The condition of faying sur-
faces was clean mill scale. The electrodes were equivalent
of E7018.
The bolt spacing in all specimens was 3 in. The edge dis-
tance in the horizontal as well as vertical direction for
specimens 1,2 and 3 was 1-1/2 in. (two times diameter of
bolt) and for specimens 4 and 5 was l-lA in. (1.5 times di-
ameter of bolts).
CONTROL
COMPUTER GRAPHICS ANALOG
i ' , PLOTTER
ACTUATOR ACTUATOR PRI NTER
CONTROLLER CONTROLLER
['"'"'"3
FEEDBACK
TOI NSTRUMENTAI ON
^OTUATOR' :V !7
SINGLEPLATE 41:-i::! :: :-:-ii:.:]
SHEARCONNECTION / :i-i!i;!i :.- ]
COLUMN / li:.i-:.!:! i <: : ]
REACTION BLOCK j
Fig, 4. Test Set-up Used in Experiments
Behavior of Test Specimens
The experiments were conducted in two groups as indi-
cated in Table 1. The main differences of specimens in
these two groups were the type of bolt (A325 or A490),
material of beam (A36 or grade 50) and edge distance (2db
or 1.5db). The behavior of specimens in the two groups
is summarized in the following sections.
Behavior of Specimens 1,2 and 3 (Group One)
Specimens 1,2 and 3 showed very similar behavior
throughout the loading. The most important observation
was the significant inelastic shear deformations that took
place in all three specimens as shown in Fig. 5.
All test specimens failed due to sudden shear fracture
of the bolts connecting the single plate to the beam web
as shown in Fig. 6a. The examination of bolts after failure
indicated that the A325 bolts in these specimens had de-
veloped significant permanent deformations prior to frac-
ture as indicated in Fig. 6b. In these three specimens the
welds did not show any sign of yielding other than in speci-
men 3 which showed minor yielding at the top and bottom
of welds prior to fracture of bolts.
A study of the bolt holes after the completion of tests
1,2 and 3 indicated that permanent bearing deformations
had taken place in the plate as well as in the beam web.
The magnitude of the deformations in the plate and beam
9
bolt holes were almost equal but in opposite directions.
The deformations of the plate bolt holes, drawn to scale
are shown in Fig. 7. The arrows indicate the direction of
the movement of the bolts which is expected to be approx-
imately the direction of the applied force due to shear and
moment. It is interesting to note that nearly vertical orien-
tations of arrows indicate the presence of a large vertical
shear accompanied by a relatively small moment in the
connections.
Behavior of Specimens 4 and 5 (Group Two)
The behavior of specimens 4 and 5 was similar to the
previous three tests. However, shear yielding of the plate
was more apparent. Specimen 4 failed due to shear frac-
ture of bolts in a manner similar to previous tests shown
in Fig. 6a. In addition, minor yielding was observed on the
weld lines of this specimen. Specimen 5 failed by almost
simultaneous fracture of weld lines and bolts as shown in
Fig. 8. It appears that at the time of failure, weld lines
started to fracture first while bolts were on the verge of
fracture. When sudden fracture of welds occured the re-
sulting shock caused fracture of the bolts which appeared
to be almost simultaneous with weld fracture. Bolts in
specimens 4 and 5 were A490 bolts. An examination of the
bolts after fracture showed less permanent deformations
in these bolts than the A325 bolts used in previous three
tests (see Fig. 6b).
Study of bolt holes in the shear tabs of specimens 4 and
5 indicated that significantly larger bolt hole deformations
had occured in these two specimens compared to speci-
mens 1,2 and 3. However, the bolt holes in the beam web
in specimens 4 and 5 had only minor permanent deforma-
tions.
In summary, based on observations made during the
O
i
I
Iq)
O
<Y
43
+
TEST ONE TEST TWO TEST THREE
Fig. 5. Plate Shear Deformations in Specimens 1,2,3
tests, it appears that shear tabs go through three distinc-
tive phases of behavior. At the very early stages, a shear
tab acts as a short cantilever beam with moment being
dominant. As load increases, the shear tab acts as a deep
shear beam with the shear yielding effect dominant (as in
specimens i through 4). If bolts and welds do not fail dur-
ing the shear phase, because of large deformations, the
shear tab acts similarly to the diagonal member of a truss
and carries the applied shear by a combination of shear
and diagonal tension effects (as in specimen 5),
Experimental Data
The results of experiments at the time of failure are
summarized in Table 2.
DISCUSSION OF EXPERIMENTAL RESULTS
Shear Yielding of Single Plate
The yielding of the single plate was primarily due to
shear stresses and was quite ductile. It was evident that
considerable shear yielding occurred in the plate between
the bolt line and weld line. The shear yielding was almost
uniformly distributed throughout the depth of the plate as
measured by strain gages that were attached to the
plates. 3'0 Therefore, in the proposed design procedure dis-
cussed later, the shear capacity of plate is calculated by
multiplying gross area of plate by uniformly distributed
shear stresses.
In specimen 3, at later stages of loading and after signifi-
cant shear yielding, the bottom portion of the shear tab
showed signs of minor local buckling as shown in Fig. 6a.
This local buckling was attributed primarily to loss of stiff-
ness of plate material due to shear yielding. Until this phe-
A325 Bolts
(15 Bolts Tested)
Fig. 6.
A490 Bolts
(8 Bolts Testedl
(a) (b)
Typical Bolt Failure of Test Specimens
10
Table 2.
Results of shear strength Tests
Specimen Observed Connection Response
Failure Mode $ $ Maximum
Test Test No. of Shear Shear BeamEnd Moment at Moment at Moment at
Group No. Bolts Displacement Force Rotation Bolt Line WeldLine Weld Line
in. kips rad. kip in. kip in.
(1) (2) (3) (4) (5) (6) (7) (8) (9) (1O)
1 7 Bolts Fractured 0.27 160 0.026 306 745 1028
One 2 5 Bolts Fractured 0.34 137 0.054 314 691 734
3 3 Bolts Fractured 0.46 94 0.056 20 279 350
Two 4 5 Bolts Fractured 0.35 130 0.053 273 631 686
5 3 Welds and Bolts Fractured 0.52 79 0,061 -47 170 237
*In some cases like these, moment decreased as shear and rotation increased.
Positive moments cause top of connection to be in tension.
nomenon is studied thoroughly, it is suggested that local
buckling be avoided. To prevent local buckling, it is rec-
ommended that the distance between the bolt line and the
weld line be less than 1/2 of the plate length.
Fracture of Net Area of Plate
In the single plate specimens that were tested, the net
area of the plate did not fracture. Only specimen 5 showed
signs of approaching fracture of net section. Nevertheless,
this failure mode has been observed in similar cases in sev-
eral experiments on tee framing connections. 4'5 The stem
in a tee framing connection behaves similarly to a shear
tab. The formula currently used in calculating net area in
shear fracture is: ]5
Ans = Avg-n(db +l/16)tp (1)
The studies of tee connections indicated that the shear
fracture occurred consistently by fracture of net section
along the edge of the bolt hole and not along the
centerline of bolts. It was suggested that 4,s the net area ef-
fective in shear be equal to the average of net area along
the bolt centerline and the gross area. Using the suggested
method to calculate net area in shear, the effective net
area in shear can be written as:
Anse = Avg-(n/ 2)(db +V]o)tp (2)
Shear-Rotation Behavior
Figure 9 shows the actual shear-rotation relationship
that was recorded during each test. It is observed that the
rotational ductility of the connections increased as the
Fig. 7. Plate Bolt Hole Deformations after Tests Fig. 8. Failure of Welds and Bolts in Specimen 5
11
number of bolts decreased. The rotational ductility of the
connection in specimen ! with 7 bolts was 0.026 radians
which was about half the rotational ductility of the connec-
tions in specimens 2, 3, 4 and 5 with three or five bolts,
all of which were able to reach rotations in excess of 0.05
radians.
Movement of Point of Inflection
Figure 10 shows movement of point of inflection of the
beam toward the support as the shear force was increased.
Even under relatively small load, in all specimens, the
point of inflection moved toward the support and re-
mained almost stationary for the remainder of each test.
Using experimental data, the following empirical equa-
tion was developed to define the location of the point of
inflection for test specimens.
e = (n-I)(1.0), in. (3)
where n is the number of bolts used in the connection, and
e is the distance of point of inflection from the support
(i.e. from the weld line).
It is important to realize that in the experiments re-
ported here, the columns were fixed to supports and rigid
body rotation of the connections was prevented. If due to
frame action or other causes, the support to which a shear
tab is connected rotates, due to rigid body rotation, the
location of point of inflection may be affected. However,
the concurrent values of shear and moment acting on the
shear tab at any given time cannot exceed the values ob-
tained from plasticity conditions (interaction curves) of
plate for shear and moment.
Behavior and Design of Bolts
In all specimens, an examination of bolts and bolt holes
after failure indicated that bolt shanks had experienced
considerable shear deformations before failure.
Studies on the behavior of single bolts in shear have
indicated that for A325 bolts and A36 plate, if the thick-
ness of the plate is not greater than 1/2 times the diameter
of the bolt, considerable but tolerable bolt hole deforma-
tions will take place. The limited bolt hole deformations
are desirable since they increase rotational flexibility and
ductility of the connections. In studies of tee connec-
tions4'5 in three specimens, V2 in. thick tee stems were
used with 7/8 in. diameter bolts. The behavior of these tee
specimens indicated that even when thickness of stem was
equal to db + V6 in., desirable bearing deformations took
place in the bolt holes. Therefore, based on these studies,
and to obtain flexible and ductile single plate connections,
the thickness of the plate is recommended to be less than
or equal to 1/2 of the bolt diameter plus 1/16 in.
An examination of the deformations of bolts and bolt
holes at the completion of the tests indicated that the bolts
were primarily subjected to direct shear accompanied by
a small moment (see arrows in Fig. 6a).
As Fig. 10 indicates, the point of inflection for test spec-
imens was almost stationary, fluctuating between an ec-
centricity of n and n - 1 in. At the time of failure of the
bolts in all specimens, the location of the point of inflec-
tion was close to n - 1 in. Therefore, it is recommended
that bolts be designed for combined effects of direct shear
and a moment equal to the shear multiplied by the eccen-
tricity of the bolt line from point of inflection given by:
eb = ( n - 1 ) ( 1 . 0 ) - a (4)
where,
a = distance between the bolt line and weld line,
eb = distance from the point of inflection to the bolt
line.
Behavior and Design of Welds
Table 2 gives values of shear and moment at failure for
each test. The fillet welds mainly experienced a direct
200-
180;
1601
03 1401
Cl..
" 1201
a/ '
,< 100.
LU
"1- 80.
60
40
20-'
0'
TEST 1 0 V
f (7-Bolt)
/ :::':4 . . . . . .
. . . . . . . . .
/ / / _---Tost3
%" ; " ' " ". . . . . . . ,.. . . . . . . ( 3 B o l t )
0.02 004 006
ROTATION AT BOLT LINE, rad.
g
of
<I:
W
OO
200-1 7" ,
. . . . Test1(7A325Bolts)
I -- - - - Test2(5A325Bolts)
I I I - - - - - Test3{3A325 Bolts)
150 1 q" J(I Test4(5.A490Bolts)
lOO. Li(',
0 . . . . . . . . . . . . . . . . . ,, , - - ' ; -
fewf
v
Point of
Inflection
I -"mll
Momenl
10 20 30
DISTANCE OF PT. OF INFL. FROM WELD, in.
4O 5O
Fig. 9. Shear-Rotation Curves for Test Spectmens Fig. 10. Movement of Point of Inflection
12
shear accompanied by a relatively small moment. The
strain measurements adjacent to the welds also supported
this conclusion. 3'6 Therefore, fillet welds are recom-
mended to be designed for the combined effects of shear
and a small bending moment.
The main goal of the proposed design procedure is to
ensure yielding of shear tab prior to failure of welds. In
order to achieve this goal the welds should be designed to
be stronger than the plate. Thus, the design shear force
acting on the welds is recommended to be equal to the
shear capacity of the plate and not the applied shear force.
Therefore, the maximum shear force acting on the weld
is equal to 1/V'-3 FyLptp. In Allowable Stress Design, the
design shear force for welds is equal to 0.40FyLvtp. The
moment acting on the weld is equal to shear force multi-
plied by the eccentricity of the point of inflection from the
weld line. To be conservative, it is recommended that the
eccentricity of the point of inflection from the weld line
be equal to n inches,
ew = (n)(1.0) (5)
Since the design of welds in the proposed method is a
capacity design, it is not necessary to use welds that can
resist forces much greater than the plate capacity. As part
of phase two of this investigation, a study was conducted
to establish minimum and maximum weld requirements to
develop the strength of single plate. The study indicated
that for A36 plate and E70 electrodes the weld size need
not be more than 0.75tv and should not be greater than
tv. The upper limit of tv on the weld size was imposed to
prevent excessive welding of the plate which will be costly
and might cause heat damage to the plate without achiev-
ing extra strength in the connection.
Moment-Rotation Curves
Moment-rotation curves for the test specimens are
shown in Fig. 11. Moments and rotations were measured
1200-
6
lO00-
Z
.-I
800-
,,-I
LLI
600-
I--
l,,-, 400-
Z
LIJ
200-
0
0
0
SpecimenswithA325Bolts
Specimens with A490 Bolts
Test 1
(7-Bolt) . Test 2
. . . . . . . %' Test 4
(5-Bolt)
Test 3
(3-Bolt)
/ / . . . . . . . . . . . . . . . . . . . .' - Test5
- - - (3-Bolt)
' o . b 2 ' & 4 ' 0.6 '
ROTATION OF BEAMEND, rad.
Fig. 11. Moment-Rotation Curves for Test Specimens
along the bolt line. As the plots indicate, connections with
fewer bolts developed smaller moments and exhibited
larger rotational ductility. During the elastic range of be-
havior, moment increased with shear. As the load in-
creased, due to connection deformations, rotational stiff-
ness and bending moment decreased and then gradually
increased at a much smaller rate. The decrease is attrib-
uted to slips and inelastic deformations in the connections
and the increase is attributed to strain hardening.
PROPOSED DESIGN PROCEDURE
The following design procedure is based on the analyses
of the experimental results and the information available
on the actual behavior of shear connections?6'9
General Requirements
The single plate framing connections covered by these
procedures consist of a plate bolted to a beam web and
welded to a support on one edge of plate.
In design of a single plate framing connection, the fol-
lowing requirements should be satisfied:
1. The cortnection has only one vertical row.of bolts
and the number of bolts is not less than 2 or more
than 7.
2. Bolt spacing is equal to 3 in.
3. Edge distances are equal to or greater than 1.5db.
The vertical edge distance for the lowest bolt is pre-
ferred not to be less than 1.5 in.
4. The distance from bolt line to weld line is equal to
3 in.
5. Material of the shear plate is A36 steel to facilitate
yielding.
6. Welds are fillet welds with E70xx or E60xx elec-
trodes.
7. Thickness of the single plate should be !ess than or
equal to db/2 + 1,/16.
8. The ratio of Lp/a of the plate should be greater than
or equal to 2 to prevent local buckling of plate.
9. ASTM A325 and A490 bolts may be used. Fully
tightened as well as snug tight bolts are permitted.
The procedure is not applicable to oversized or long
slotted bolt holes. Standard or short-slotted punched
or drilled holes are permitted.
Consideration of Limit States in Design
The following limit states are associated with the single
plate framing connections.
1. Shear failure of bolts.
2. Yielding of gross area of plate.
3. Fracture of net area of plate.
4. Fracture of welds
5. Bearing failure of beam web or plate.
13
Shear Failure of Bolts
Bolts are designed for the combined effects of direct
shear and a moment due to the eccentricity eb of the reac-
tion from the bolt line. The eccentricity eb for single plate
connections covered by these procedures can be assumed
to be equal to 3 in., which is the distance from bolt line
to weld line. The value is conservative when the single
plate is welded to a rigid support. The value is more realis-
tic when the supporting member is a relatively flexible ele-
ment.
More realistic values for eb can be calculated from the
following equations:
if single plate is welded to a rotationally rigid element, eb
is obtained from:
eb = (n- 1)(1.O)-a (6)
if single plate is welded to a rotationally flexible element,
eb is larger value obtained from:
eb = Max
(n- 1)(1.0) - a (7a)
a (7b)
where,
n = number of bolts
a = distance from bolt line to weld line, in.
eb = eccentricity, in.
By using methods outlined in Reference 7 including
using Tables X of the AISC-ASD Manual 3 the bolts are
designed for the combined effects of shear R, and moment
equal to R%.
Yielding of Gross Area of Plate
The equation defining this limit state in allowable stress
design (ASD) format is:
fry < Fy (8)
where,
fry = R / Avg (9)
Fry = 0.40 Fy (10)
Avg = Lp tp (11)
Fracture of Net Area of Plate
The equation defining this limit state in allowable stress
design (ASD) format is:
fy. -< Fu (12)
where,
fuu = R /Ans (13)
Fvu = 0.30/:, (14)
Ans= [Lp -- n(db + 1/16)]tp (15)
If the beam is coped, the block shear failure of the beam
web also should be considered as discussed in the AISC-
ASD Specification. 5
Weld Failure
The welds connecting the plate to the support are de-
signed for the combined effects of direct shear and a mo-
ment due to the eccentricity of the reaction from the weld
line, ew. The eccentricity ew is equal to the larger value
obtained from:
ew = Max
(n)(1.0) (16a)
a (16b)
where,
n = number of bolts
ew = eccentricity, in.
a = distance from bolt line to weld line, in.
By using methods outlined in Reference 7 including
using Tables XIX of the AISC-ASD Manual, 13 the fillet
welds are designed for the combined effects of shear equal
to R and moment equal to Re,,.
Bearing Failure of Plate or Beam Web
To avoid reaching this limit state, it is recommended
that the established rule of horizontal and vertical edge
distances equaling at least 1.5 the bolt diameter be fol-
lowed. The bolt spacings should satisfy requirements of
the AISC-ASD Specification. 5 The bearing strength of
connection can be calculated using the provisions of the
AISC-ASD Specification. 5
Summary of Design Procedure
The following steps are recommended to be taken in de-
sign of single plate framing connections:
1. Calculate number of bolts required to resist combined
effects of shear R, and moment R% using Table X of the
AISC-ASD Manual. 13
If the single plate is welded to a rotationally rigid sup-
port eb is the value obtained from Eq. 6.
If the single plate is welded to a rotationally flexible ele-
ment, eb is the value obtained from Eq. 7:
2. Calculate required gross area of plate:
Avg R / 0.40Fy (17)
Use A36 steel and select a plate satisfying the following
requirements:
a. lh and l,, 1.5db.
b. Lp--> 2a
c. t, -< db/2 + V6
d. ti, --> Ava/Lp
e. Bolt spacing =3 in.
(18)
(19)
(20)
(21)
14
3. Check effective net section:
Calculate allowable shear strength of the effective net
area:
Rns=[tp-n(db+Vl6)]()(o.3eu) (22)
and satisfy that R. -> R.
4. Calculate actual allowable shear yield strength of the
selected plate:
Ro = Lptp (0.40Fy) (23)
Design fillet welds for the combined effects of shear Ro
and moment Roewusing Table XIX of the AISC Manual?
ew is given in Eq. 16 as:
ew = Max ]
i
(n)(1.0) (16a)
I a (16b)
The weld is designed for a capacity of Ro, and not for
the applied R, to ensure that the plate yields before the
welds. However, for A36 steel and E70 electrodes the
weld size need not be larger than 3/4of the plate thickness.
5. Check bearing capacity of bolt group:
(n)(t)(db)(1.2Fu) > R (24)
If the bolts are expected to resist a moment (as they nor-
mally would), this calculation should reflect the reduced
strength as determined by Table X of the AISC Manual3
as demonstrated in the following examples.
6. If the beam is coped, the possibility of block shear fail-
ure should be investigated.
Application to Design Problems
The following examples show how the design procedure
can be implemented into the design of steel structures.
Design Example 1
Given:
Beam:
Beam Material:
Support:
Reaction:
Bolts:
Bolt Spacing:
Welds:
W27 x 114, t = 0.570 in.
A36 steel
Column flange (Assumed rigid)
102 kips (Service Load)
7/8 in. dia. A490-N (snug tight)
3 in.
E70XX fillet welds
Design a single plate framing connection to transfer the
beam reaction to supporting column.
Sol ut i on:
1. Calculate number of bolts:
Shear = R = 102 kips
Let us assume M = 0, (will be checked later)
n = R/rv= 102/16.8 = 6.1
Try 7 bolts
The distance between the bolt line and the weld line
a is selected equal to 3 in.
Check moment:
e0 = ( n - 1 ) l . 0 - a = 7 - 1 - 3 = 3.0 in.
Moment = 3 x 102 = 306 kip-in.
Using Table X of the AISCASD Manual 3 with eccentric-
ity of 3 in., a value of 6.06 is obtained for effective number
of bolts (7 bolts are only as effective as 6.06 bolts).
Therefore,
Rbolt = 6.06 x 16.8 = 101.8 - 102 kips O.K.
Use: Seven in. dia. A490-N bolts.
2. Calculate required gross area of the plate:
Avg = R / 0.40Fy
Avg = 102/(0.40 x 36) = 7.08 in.2
Use A36 steel and select a plate satisfying the following
requirements:
a. lh and l > 1.5db
lh = lv = 1.5(7/8) = 1.32 in.
W = a +Ih = 3 + 1.32= 4.32; useW=
4V2 in.
b. Lp/a -> 2.0
Lp = 2 x 1.32 + 6 x 3.0 = 20.6 in.; use Lp =
21 in.
Check:Lp/a= 21/3 = 7 > 2 O.K.
c. tv < db/2 + 1/16
tp -- (7/8)/2 + 1/16 = V2in.
d. tv = Aug/L.
tv = 7.08/21 = 0.337 in.
Try PL 21 x 3/8 x 4-1/2
3. Calculate allowable shear strength of the net area:
Rns = [Lp-n(db + IA6)](tp)(O.3Fu)
R, = [21-7(% + 1/16)](8)(0.3 58) = 94 < 102
kips N.G.
Try V2 in. thick plate:
Rs = [21-7(7/8 + 1A6)](I/2)(0.3 x 58) = 125 > 102
kips. O.K.
Use: PL 21xx4%, A36 Steel.
4. Calculate the actual allowable yield strength of the se-
lected plate:
Ro = Lptp (0.40Fy)
Ro = 21 x 0.5 x 0.40 x 36 = 151 kips
Design fillet welds for the combined effects of shear
and moment:
Shear = Ro = 151 kips
[ n(1.0) = 7(1.0) = 7 in.
ew Max
I a = 3in.
Therefore, ew = 7.0 in.
Moment = Roe = 151 x 7 = 1057 kip-in.
Using Table XIX AISC Manual 3
1 5
a = 7/21 = 0.333
C1 = 1.0
C = 1.07
Dl6 = Ro/CC/Lp = 151/(1.0 x 1.07 x 21). = 6.72
Since weld size need not be greater than 0.75tp,
Use: % in. E70 Fillet Welds.
5. Check bearing capacity:
For plate:
rv = drip (1.2Fu) = .875 x .5 x 1.2 x 58 = 30.45
Rbrg = 6.06(30.45) = 184.5 kips > 102 kips. O.K.
Since the beam web is thicker than the plate, the web
will not fail.
6. Beam is not coped, therefore, there is no need for con-
sideration of block shear failure.
Design Example 2
Given:
Beam:
Beam Material:
Support:
Reaction:
Bolts:
Bolt Spacing:
Welds:
W16x31, tw = 0.275
A572 Gr. 50 steel
Condition of support is unknown
33 kips (Service Load)
3/4 in. dia. A325-N or A490 (snug tight)
3 in.
E70XX fillet welds
Design a single plate shear connection to transfer the
beam reaction to the support.
Solution:
1. Calculate number of bolts:
Shear = 33 kips
Let us assume M = 0, (will be checked later)
Try A325-N bolts with 9.3 kips/bolt shear capacity:
n = R/r, = 33/9.3 = 3.5
Try 4 bolts.
The distance between bolt line and weld line a is
selected equal to 3 in.
Check moment:
Since condition of support is not known, the sup-
port is conservatively assumed to be flexible for
bolt design. Therefore eb is equal to 3 in.
Moment = 3 x 33 -- 99.0 kip-in.
Interpolating from Table X13, C 2.81
Rat/= 2.81 x 9.3 = 26.1 kips<33 N.G.
Which indicates 4 A325 bolts are not enough. Let
us try 4 A490-N bolts:
Ra//= 2.81 x 12.4 = 34.8 kips>33 O.K.
Use: Four % in. dia. A490-N bolts.
2. Calculate required gross area of plate:
Avg = R / 0.40Fy
Avg = 33/(0.40 x 36) = 2.29 in.2
Use A36 steel and select a plate satisfying the following
requirements:
a. lh and Iv -> 1.5db.
Ih = Iv = 1.5(3/4) = 1.125 in.
W= a + lh = 3 + 1.125 = 4.125 in.
Use: W = 41/2 in.
b. Lp/a -> 2.0
Lp = 3 + 3 x 3 = 12 in.
Check: Lp/a = 12/3 = 4 > 2 O.K.
c. tp -< db/2 + 1/16
tp < (3/4)/2 + 1/16 = ?/16 in.
d. tp = A vg/mp
tl, = 2.29/12 = 0.19 in.
Use: PL 12xx41/2, A36 Steel.
3. Calculate allowable shear strength of the net area:
R,s = [Lp-n(db + 1A6)](tp)(O.3F)
R,s = [12 - 4(3/4 + V16)](I/4)(0.3 x 58) = 38.1 kips
R,, -> R is satisfied.
4. Calculate actual allowable yield strength of the selected
plate:
Ro = Lpt, (0.40Fy)
Ro = 12 x 0.25 x 0.40 x 36 = 43.2 kips
Design fillet welds for the combined effects of shear
and moment:
Shear = Ro = 43.2 kips
(n)(1.0) = 4(1.0) = 4 in.
ew = Max
a = 3.0
Therefore, ew = 4.0 in.
Moment = Roe,, = 43.2 x 4 = 172.8 kip-in.
Using Table XIX AISC Manualt3
a -- 4/12 = 0.33
C1 -= 1.0
C = 1.07
D/6 = Ro/CC/Lp = 43.2/(1.0 x 1.07 x 12) = 3.36
Since weld size need not be greater than 0.75tp,
Use: 6 in. ET0 Fillet Welds.
5. Check bearing capacity.
For plate:
nddp (1.2F,) = 2.81 x .75 x .25 x 1.2 x 58
= 36.7 kips > 33 kips.
and for beam:
nddw(1.2F) = 2.81 x .75 x .27 x 1.2 x 65
= 44.4 kips > 33 kips.
6. Beam is not coped, therefore, no need for considera-
tion of block shear failure.
CONCLUSIONS
Based on the studies reported here, the following con-
clusions were reached:
1. The experimental studies of single plate connections in-
16
dicated that considerable shear and bearing yielding
occurred in the plate prior to the failure. The yielding
caused reduction of the rotational stiffness which in
turn caused ielease of the end moments to midspan of
the beam.
2. The limit states associated with single plate connections
are:
a. Plate yielding.
b. Fracture of net section of plate.
c. Bolt fracture.
d. Weld fracture.
e. Bearing failure of bolt holes.
3. A new design procedure for single plate shear connec-
tions is developed and recommended. The procedure
is based on a concept that emphasizes facilitating shear
and bearing yielding of the plate to reduce rotational
stiffness of the connection.
4. To avoid bearing fracture, the horizontal and vertical
edge distance of the bolt holes are recommended to be
at least 1.5 times diameter of the bolt. The study re-
ported here indicated that vertical edge distance, par-
ticularly below the bottom bolt is the most critical edge
distance.
5. Single plate connections that were tested were very
ductile and tolerated rotations from 0.026 to 0.061 radi-
ans at the point of maximum shear. Rotational flexibil-
ity and ductility decreased with increase in number of
bolts.
ACKNOWLEDGMENTS
The project was supported by the Department of Civil
Engineering, the University of California, Berkeley and
the American Institute of Steel Construction, Inc. The
support and constructive comments provided by R. O.
Disque, N. Iwankiw and Dr. W. A. Thornton are sin-
cerely appreciated. Single plates used in the test speci-
mens were fabricated and supplied by the Cives Steel
Company. The assistance of R. Stephen, laboratory man-
ager, in conducting the experiments was essential and is
appreciated.
REFERENCES
1. Astaneh, A., "Experimental Investigation of Tee-
Framing Connection", Progress Report submitted to
American Institute of Steel Construction, April 1987.
2. Astaneh, A., "Demand and Supply of Ductility in
Steel Shear Connections", Journal of Steel Construc-
tion Research, 1989.
3. Astaneh, A., K. M. McMullin, and S. M. Call, "De-
sign of Single Plate Framing Connections," Report
No. UCB/SEMM-88/12, Department of Civil Engi-
neering, University Of California, Berkeley, July,
1988.
4. Astaneh, A., and M. Nader, "Design of Tee Framing
Shear Connections," Engineering Journal, American
Institute of Steel Construction, First Quarter, 1989.
5. Astaneh, A., and M. Nader, "Behavior and Design of
Steel Tee Framing Connections," Report No. UCB/
SEMM-88/ll, Department of Civil Engineering, Uni-
versity of California, Berkeley, July, 1988.
6. Call, S. M., and A. Astaneh, "Behavior of Single
Plate Shear Connections with A325 and A490 Bolts",
Report No. UCB/SEMM-89/04, Department of Civil
Engineering, University of California, Berkeley,
April 1989.
7. Iwankiw, N. R., "Design for Eccent'ric and Inclined
Loads on Bolts and Weld Groups," Engineering Jour-
nal, American Institute of Steel Construction, 4th
Quarter, 1987.
8. Lipson, S. L., "Single-Angle Welded-Bolted. Connec-
tions,'' Journal of the Structural Division, March,
1977.
9. McMullin, K. M., and A. Astaneh, "Analytical and
Experimental Investigations of Double-Angle Con-
nections'', Report No. UCB/SEMM-88/14, Depart-
ment of Civil Engineering, University of California,
Berkeley, August, 1988.
10. Patrick, M., I. R. Thomas, and I. D. Bennetts, "Test-
ing of the Web Side Plate Connection," Australian
Welding Research, December, 1986.
11. Richard, R. M., P. E. Gillett, J. D. Kriegh, and B.
A. Lewis, "The Analysis and Design of Single Plate
Framing Connections," Engineering Journal, Ameri-
can Institute of Steel Construction, 2nd Quarter,
1980.
12. White, R. N., "Framing Connections for Square and
Rectangular Structural Tubing, Engineering Journal,
American Institute of Steel Construction, July, 1965.
13. American Institute of Steel Construction, Manual of
Steel Construction, 8th Edition, Chicago, 1980.
14. American Institute of Steel Construction, Manual of
Steel Construction. LRFD, 1st Edition, Chicago,
1986.
15. American Institute of Steel Construction, Inc., Speci-
fication for the Design, Fabrication and Erection of
Structural Steel for Buildings, Chicago, November 13,
1978.
16. American Institute of Steel Construction, Inc., Load
and Resistance Factor Design Specification for Struc-
tural Steel Buildings, Chicago, September 1, 1986.
17
Arts
An.se
Ay s
C
c,
D6
VU
L
R
Rbolt
R.
Ro
Ry
NOMENCLATURE
Net area in shear, in.2
Effective net area of plate in shear, in.2
Gross area of plate in shear, in2.
Coefficient in the AISC Manual Tables X and XIX
Coefficient in the AISC Manual Table XIX
Number of sixteenth of an inch in fillet weld size
Specified minimum tensile strength of steel, ksi
Allowable shear stress for plate in yielding =
0.40Fy, ksi
Allowable ultimate shear strength = 0.30Fu, ksi
Specified yield stress of steel, ksi
Length of span, in.
Length of plate, in.
Plastic moment capacity of cross section = ZxFy
Yield moment of beam cross section, kip-in.
Reaction of the beam due to service load, kips
Allowable shear capacity of bolt group
Allowable shear fracture capacity of the net section
Allowable shear yield strength of plate, kips
Reaction corresponding to plastic collapse of beam,
kips
$x Section modulus in.3
V Shear force, kips
W Width of plate, in.
Zx Plastic section modulus, in.3
a Coefficient in the AISC Manual Table XIX
a Distance between bolt line and weld line, in.
d Depth of beam, in.
db Diameter of bolt, in.
e Eccentricity of point of inflection from the support
eb Eccentricity of beam reaction from bolt line, in.
ew Eccentricity of beam reaction from weld line, in.
fry Computed shear stress in plate gross area, ksi
fvu Computed shear stress in plate effective net area,
ksi
In Horizontal edge distance of bolts, in.
lv Vertical edge distance of bolts, in.
n Number of bolts
rv Allowable shear strength of one bolt, kips
tp Thickness of plate, in.
tw Thickness of beam web, in.
This publication expresses the opinion of the author, and care has been taken to insure
that all data and information furnished are as accurate as possible. The author and
publisher cannot assume or accept any responsibility or liability for errors in the data
or information and in the use of such information.
The information contained herein is not intended to represent official attitudes, recom-
mendations or policies of the Structural Steel Educational Council. The Council is not
responsible for any statements made or opinions expressed by contributors to this
publication.
18
THE STEEL COMMITTEE OF CALIFORNIA
Northern California
43 Quail Court, No. 206
Walnut Creek, CA 94596
(415) 932-0909
Southern California
9420 Telstar Ave. No. 207
El Monte, CA 91731
(818) 444-4519
SPONSORS
Ace & Stewart Detailing, Inc.
Allied Steel Co., Inc.
Artimex Iron Co., Inc.
Bannister Steel, Inc.
Baresel Corporation
Bethlehem Steel Corporation
C. A. Buehen Corporation
Butler Manufacturing Co.
Central Industrial Engineering Co., Inc.
Cochran-Izant & Co., Inc.
Dovell Engineering, Inc.
The Herrick Corporation
Hoertig Iron Works
Hogan Mfg., Inc.
Inland Steel Company
Junior Steel Co.
Kaiser Steel Corporation
Lee & Daniel
McLean Steel, Inc.
Martin Iron Works, Inc.
Nelson Stud Welding Co.
Palm Iron & Bridge Works
Pascoe Steel Corporation
PDM Stroeal, Inc.
Reno Iron Works
Riverside Steel Construction
H. H. Robertson Co.
Schrader Iron Works, Inc.
Stockton Steel
Stott Erection, Inc.
Verco Manufacturing, 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, estimated tonnage, cost comparisons, fabrication details,
and delivery schedules.
Funding for this publication provided by the California Field Iron Workers Administrative Trust.