82

BEHAVIOUR OF STEEL BEAM-COLUMN CONNECTIONS,

MADE USING BOLTED END PLATES.
N.D. Johnstone and W.R. Walpole
1 2

Summary:

Simulated earthquake loading was applied to four cruciform

specimens, that represented part of a multistorey frame,
including the half-beams and half-columns framing into a
typical interior joint. This was made by welding plates on to
the ends of the beams and bolting these plates to the column
flanges using high strength proof loaded bolts. The joints
were designed so that recommendations, previously developed
elsewhere for monotonic loading, together with the rules given
by New Zealand Standard 3404:19 77, could be studied. When the
existing rules were followed ductile specimens were obtained;
however, when some of the joints were deliberately designed to
be understrength, failure modes were obtained that were not
predicted by the existing simple design methods.

1. INTRODUCTION: and four macalloy bars. This system did

not allow P-delta effects to b e generated
Bolted end plate steel beam to column from relative interstorey lateral displace-
connections can facilitate quick structural ments as indicated in Figure 1.
erection and provide reasonable strength,
stiffness and ductile characteristics. In
the past there has been a significant amount
of r e s e a r c h ^ ' > , into the behaviour
1 , 2 3 , 4 , 5

of these connections especially with respect

to medium and smaller sized beam members in
raonotonically loaded situations, for which
design recommendations have been developed.
This paper describes some recent tests
which were carried out to determine the
applicability of these recommendations when
the connections were subjected to reversing
cyclic loads. Plastic analysis was used
for assessing the strengths of the connect-
ions and their associated beams and columns.

NZS 3 4 0 4 : 1 9 7 7 Clause 12.4.7 ( 6 )

requires that the columns be overstrength

to discourage the formation of plastic Figure 1 - Arrangement of Test Specimens.
hinges in them, and Clause 12.4.8 requires
that the design of connection should allow
for moment gradient and strain hardening
effects in the beams. A 25% increase in
the plastic moment of the beams was allowed 2. DESIGN OF SPECIMENS:
above the experimental yield moment when
designing the connection of specimens one The 310 UB 4 6 beam size was used for
and two. The connections of specimens all specimens and has a flange outstand to
three and four were under-designed so that flange thickness ratio of 6.7 3 which is
their behaviour could be studied more reasonably squat compared to the limit,
closely at failure. 136//F required by NZS 3404:1977
Clause 10.8.1 for a ductile plastic hinge.
It was not intended that this joint The column sizes, shown in Table 1, were
should be used as part of a three dimen- chosen for specimens one and two so that a
sional frame. The connection tested here moment 25% in excess of the experimental
can resist significant actions from members yield moment could be reached in the beam,
in one plane only. Inertia forces normal together with an axial load in the column,
to this plane must be resisted by a which was fixed so that the column was
structural system constructed in a normal close to forming a plastic hinge. The
plane. The form of the experiment was column size was reduced for specimen three
chosen so as to be reasonably represent- so that a column flange mechanism would
ative of a beam-to-column connection in a develop. This size was maintained for
multi-storey plane frame under earthquake specimen four. The reduced plastic moment
loading as shown in Figure 1. The beam allowing for axial load was calculated from
loads were applied at their ends by means NZS 3404:1977< ) Clause 10.5.2 for pre-
6

of a hydraulic jack through a tie rod liminary design:

system, and the column axial compression
load was applied by two hydraulic jacks
M - 1.18 1 -
City Engineers Department, Wellington pc
City Council.
Department of Civil Engineering,
University of Canterbury, Christchurch.
B U L L E T I N O F T H E NEW Z E A L A N D N A T I O N A L S O C I E T Y FOR E A R T H Q U A K E E N G I N E E R I N G , V O L . 15, N O . 2, J U N E 1982
 83

Specimen 1 3 4
2
number

Beam size 310 UB 46 310 UB 46 310 UB 46 310 UB 46

Column size 250 UC 89 250 UC 89 250 UC 73 250 UC 73

End plate 32 32 20 16
thickness mm

High strength 30 30 24 24
bolt diameter mm

Doubler plates Adjacent to On column On column On column

position column web flange tips flange tips flange tips

Doubler plates
2 @ 16 2 @ 16 2 § 12 2 @ 12
thickness mm

Beam flange Single bevel Pair of 10 mm Single bevel Single bevel

weld to end plate butt weld fillet welds butt weld butt weld

Beam web weld Pair of 6 mm Pair of 6 mm Single bevel Single bevel

to end plate fillet welds fillet welds butt weld butt weld

TABLE 1 - COMPARISON OF SPECIMEN COMPONENT SIZES

For final selection of the axial load, the

moment capacity was checked from the more
accurate formula given on Page 40 of the
AISC Safe Load T a b l e s ^ . The column size
7 5

chosen must have a flange thick enough to

resist the tension bolt forces. Many of + 2v + 2w - D' 1
the UC sections rolled were found to be
adequate for this purpose, but most UB
m J
sections were noted to be too thin and
where w = (m(m + n - 0.5 D'))^,
would require flange reinforcing plates.
The tension force, P in the beam flange
m s
F^ = Yield stress of the column flange
to form the yield pattern shown in Figure 2 c

for a column flange with horizontal stiff- T = Thickness of the column

c
eners as shown in Figure 3, was calculated
for specimen one from the equation recom- & v,m,n,D' are defined in Figure 2.
mended by Packer and M o r r i s t ) 4

E d g e of
fillet

(b)

Figure 3
• Sagging yield lines

Hogging yield lines (a) Horizontal column flange stiffeners,

both sides.
(b) Doubler plates welded to column flange
Figure 2 - Yield line pattern for P with tips both sides.
horizontal stiffeners.
84

Doubler plate
The tension force, P in the beam m r

flange to form the yield pattern shown in

Figure 4 was also given by Packer and
Morris. The smaller of P and P was m r m s

taken to give the capacity of the horizont-

ally stiffened column flange.

F T - + i [ (n - 0.5 D ) E

"mr yc c
v m !

E d g e of
ec 2
tan" 1
^ In ^
m v S
7T m fillet

Sagging yield lines

Hogging yield lines

Figure 5 - Yield line pattern for P,

mp

Doubler p l a t e

Figure 4 - Yield line pattern for with

horizontal stiffeners.

For specimens two, three and four where the E d g e of

doubler plate was butt welded to the column fillet
flange tips, and there were no horizontal
stiffeners, as shown in Figure 3b, the
tension force, P p in the beam flange was
m

derived(8) for the yield line pattern shown

in Figure 5 and calculated from the follow-
ing equation:

2m + 2n D f
2g + c D'
F T +

' P
m
yc c

(1 + r)(2g + c) - 2D'
+
2n

S a g g i n g yield lines
mn (2m + 2n - D ) 1

where g 55

2n + (1 + r)m Hogging yield lines

When the vertical distance, c, between Figure 6 - Yield line pattern for P,
adjacent bolts becomes large, then individ- mq
ual mechanisms occur around each of the
bolts, as shown in Figure 6 for which the
flange force, P g was d e r i v e d .
m

P = 2 F T 2m + 2n D 1

mq yc c f 3 d x
g

+ g(r + 1) - 0.5 D 1
+ 2g_ 0.5 D' This assumes that the tension flange loads
are distributed equally to the four bolts
near the flange, together with an allowance
where "g" is the same as for P of 33.3% for prying forces. The beam shear
mp was assumed to be carried equally by all
eight bolts in the connection. The effect
The tensile load in each bolt was of shear in the bolts was checked using
calculated using the equation recommended the equation from AS 1511:1973 ( ) 10

by Surtees and Mann (9)

 85

2 where A was taken to be the area of the

'0.6 P ' column the cross-sectional area and A to
f
[0.6 v' w
< 1.0 be the cross-sectional area of the column
V
ob web and doubler plates.
and was found to be minimal. The values of Where no horizontal stiffeners are
P t oand V used were taken from the
o b provided the column web and doubler plates
maximum permissible applied loads given in should be checked for crushing under the
AS 1 5 1 1 , Table 2. beam flange forces, by the equation
recommended by Hogan and Thomas ( H )
The end plate thickness for specimens
one and two was calculated using the p < F t (T + 5 k + 2T)
(1) TB yc c BD
c
equation recommended by Mann and Morris

They should also be checked for bucklinc

T = under these forces, the capacity being
B F determined from the safe load for an
yp unstiffened web from Clause 5.13.2.1 of
This equation may be derived by regarding NZS 3404:1977 divided by 0.6.
the beam flange and end plate as a T-stub.
The additional strength from the connection
of the end plate to the web had been
earlier taken into acount by Surtees and
Mann (9) who suggested the yield line
pattern shown in Figure 7, but if the
thickness of an end plate is based on their
equation below, the end plate will contri-
bute to the flexibility of the structure. © 0
M.
B
T -
2B
C A YP

For comparative purposes the loads required

to form the yield line pattern of Surtees
and M a n n ^ ' were calculated and these were
9
© ©
compared to the load to reach the capacity
of the other joint components, in Table 2. V / / 7 7 T 7 ]
Specimens three and four had weak end
plates.
©
The panel zone was checked by
Clause 12.4.9 of NZS 3404:1977:

Figure 7 - Yield line pattern for the end plate.

1.0
A F 0.55 A F
s yc
w yw

I Specimen
1 2 3 4
number

Beam hinge 1.00 1.00 1.17 1.67

Column hinge 1.30 1.30 1.21 1.70

Mann & 0.99 beam A 0.94 A - load

1.38 1.54 0.79 A + load
End Morris ^ 1.18 beam B
0.82 B ± loads
plate
mechanism Surtees
2.12 2.12 1.13 1.00
k Mann( ) 9

Yielding of
panel zone(6) 1.10 1.10 1.00 1.41

Column flange
mechanism(8/9) 1.13 1.56 1.08 1.60

(Specified) proof
load of bolts with 1.37 1. 37 1.00 1.43
33% prying

TABLE 2 - RELATIVE EXPERIMENTAL STRENGTHS OF SPECIMEN COMPONENTS

 86

adjacent to the column flange face. The

The deflections of the specimens one
above procedure was followed with specimens
and two, under loading sufficient to cause
one and two. For specimens three and four
yield moments in the beams , gave a sway
the imposed deflections were essentially
angle of 0.0135 which exceeded the ratio of
similar to those used for specimens one and
0.010 recommended in Clause 3.8.3 of
two, except that higher deflections were
NZS 4 2 0 3 : 1 9 7 6 < ) with separated non-
12

achieved, with some extra cycles of loading.

structural elements under code loading
increased by the modification factor given
The relationship obtained for specimen
by Clause 3.8.1. Hence this requirement
one between the load at the right-hand end
would govern over strength requirements
of the beam and the deflection of that end
with this code and as a result ductility
is shown in Figure 1 1 . The relationship
demands are likely to be reduced.
at the left-hand end was practically
identical. After ductilities of four had
The details of the specimens tested been reached, the beams started to bend and
are shown in Figures 8 9 and 1 0 . f
twist, out of the plane of loading, and
this was only partially restrained by the
3. TEST RESULTS: bracing. The load carrying capacity
continued to rise, with strain hardening
The loading sequence followed the until ductility five, after which the load
method generally used at Canterbury carrying capacity dropped a little, with
University, whereby the deflections at further lateral and local buckling. The
ductilities ± 0.75 are determined experi- end plate did not deform significantly,
mentally, by applying forces to cause 75% although some yielding occurred. Similarly
of the experimental beam yield moment, shear deformations in the panel zone were
based on coupon tests for yield stress and small although strains well in excess of
a known geometric section shape. The yield were obtained, as would be expected
following cycles are deflection controlled, from Table 2 where it is noted that only
with gradually increasing peak displacement 10% extra beam load was required to cause
magnitudes, which are multiples of the yield in the panel zone above that required
average deflections in cycle one. to form a plastic hinge in the beam, with-
out strain hardening. Elongation and
The member displacement ductility bending of the bolts was minimal.
factors quoted in this paper follow the
definition of Clause C12.2.1 of NZS 3404: For specimen two the relationship
1977 i.e. "the ratio of the maximum trans- obtained between the load at the left-hand
verse deflection developed in a member to end of the beam and the deflections at this
its yield deflection". The "yield deflect- end is shown in Figure 1 2 . The relation-
ion which produces yield point stresses in ship at the right-hand end was practically
the extreme fibres, multiplied by the shape identical to the left-hand end and the
factor for the section". These definitions performance of this specimen without
modify those given in NZS 4203 :1976 ( ) , and 12
horizontal stiffeners was similar to speci-
are intended to cover the situation where a men one.
plastic hinge is formed in the beams

sA-
50 i U0 50
570 x 2 / , 0 x 3 2 m m FL

7 5 x 16mm E s t i f f e n e r s 320 holes

b o t h s i d e s of c o l u m n
+
Butt weld

+
SECTION (A)

SECTI0N ( § )
16mm doubler FL's
b o t h s i d e s of column

Figure 8 J o i n t Details f o r Specimen One

 87

50 U0

Figure 9 J o i n t Details f o r Specimen T w o

630 x 2 2 5 x 12mm R.
(both sides)
130 45
f s o f ' M 2£ H . S . B o l t s 517 x 2 0 ft . N2 3
8.8X/TB ' 531 x 16 d , HSA

2 6 ei h o l e s

4 4

2 z:

m
4 +
LO LO
I V V SV \ V M
LTI fM

4 4

SECTION © SECTION (?)

Figure 10 J o i n t Details f o r Specimens T h r e e and F o u r

 88

Member Displacement Ductility

-4 -2 0 2 4
Column a x i a l l o a d = 1440 k N

-50 -

-100 0 100 -300 -100 0 100

Beam Deflection at B (mm)'

(a) Left-hand beam

Figure 11 - Load-deflection relationship for
Specimen 1. C o l u m n a x i a l l o a d = 1440 k N

Member Displacement Ductility

-6 - 4 - 2 0 2 4
i J ~— h-
100

.50

-200 -100 0 100 300

-50
(b) Right-hand beam
•100
\ I Figure 13 - Load-deflection relationship for
-200 -100 0 100 200 Specimen 3.
Beam Deflection at A (mm)

ii B e a m Column a x i a l l o a d = 1440 k N
Figure 12 - Load-deflection relationship for 100 " L o a d
Specimen 2. (kN)
—"—•~& 7

f )l 1)
The plots of load versus deflection 50 —
for the ends of the left-hand and right-
hand beams for specimen three are shown in
Figure 1 3 . Out of plane bending and twist-
ing in the b e a m s , together with local
f / / fl
buckling was of less importance in specimens
three and four, because the spacing of the
-50
lateral bracing was reduced to half that
used for specimens one and two, and also
/ @
T tJ^
the stresses in the beams were lower for Beam D e f l e c t i o n at A (mm)
specimen three and four. At a beam end -100 f 1 1 1 — » .

deflection of 152 mm in the fourth cycle, -300 -200 -100 0 100 200 300
for the right-hand beam of specimen three,
a cleavage fracture started to spread
through the end plate just outside the edge (a) Left-hand beam
of beam flange butt weld. This fracture
caused a sudden drop in the load carried
as shown in Figure 13b, because only the C o l u m n a x i a l l o a d = 1440 k N
inner bolts now carried load. It was
found that the column flange deformed
significantly under the action of the
tension bolt loads. This was particularly
marked for the two bolts carrying all the
flange load after the end plate fracture;
and eventually this local overload and
deformation caused the doubler plate to
column flange tip butt weld to fail. Yield- Beam D e f l e c t i o n at B(mm)
ing of the column flanges and webs was J I L
occurring just beyond the edge of the -300 -200 -100 0 100 200 300
doubler plates.
(b) Right-hand beam
The load-deflection performance for
the left-hand and right-hand beams of Figure 14 - Load-deflection relationship for
specimen four are shown in Figure 1 4 . It Specimen 4.
 89

was found that specimen four was initially Specimens one and two produced well-formed
about 30% more flexible than specimen three. ductility loops and it would appear that
It was also found that there were percept- connections designed in this way, or
ible differences in stiffness between the possibly even less conservatively, would
two beams of specimen four and also differ- produce the ductilities required by the
ences for either beam between stiffness in commentary of NZS 4203:1976. However it is
each direction. There were noticeable unlikely that structures, which are as
differences in the dimension b between all flexible as these specimens, would require
four beam flanges. This was compensated to the ductilities as high as suggested by that
some extent by the differences in (c - b ) . commentary. The cycling of specimens one
This problem is unavoidable because of and two was discontinued when the deflect-
rolling tolerances on the beam depths, ion of the specimens reached the limits of
unless the end plate hole positions are the rig at that time. Significant twisting
fixed after measuring the beam depth. Also of the beams occurred although the bracing
the weld reinforcement or splay varied complied with the rules of NZS 3404:1977;
significantly between flanges. The left- however the connection showed no signs of
hand beam was noticeably weaker than the distress even at the high ductilities
right-hand beam throughout the test, in reached.
both the initial elastic region and the
plastic region. Lower loads were required The connections in specimens one and
to reach the same deflection. The differ- two were strong enough to transmit beam
ence in behaviour became particularly pro- actions and to encourage most of the in-
nounced after cycle 5, with the formation elastic deformations to occur in the beams.
of cracks at the edge of the beam flange to In specimen three the left-hand side of the
end plate butt welds, and the eventual connection behaved very well with respect
spread of this crack through the end plate. to its design strength. The right-hand
side was strong enough to-surpass its design
The deformation of the end plate into strength but it did not have the ductile
double curvature was evident near the end of characteristics in the end plate, and
cycle 4 at beam deflections of - 81 mm. possibly in the column flange stiffening,
Half-way through cycle 5 at a deflection of to carry the ultimate load to as high a
150 mm, all the significant plastic deform- ductility as the other side. In specimen
ation was occurring in the end plate. The four where the end plates were the weakest
waviness in Figure 14 for cycle 6, is caused elements of the specimen, it was found
by loss of strength due to cracking, then again that the end plates did not provide
strain hardening followed by further crack sufficient ductility, and cleavage fractures
penetration and more strain hardening. through the end plates eventually occurred.
Just before reaching the maximum deflection
in cycle 6 the crack near the edge of the In specimens one and two, lateral
butt weld between the beam flange and the bracing was provided for the beams at a
end plate penetrated right through the end distance of 2130 mm from the end plate.
plate at deflection of - 169 mm. On re- This length may be compared with the code (6)
loading in the positive direction the right- requirements (Clause 10.9) of 960 a r / i / F ~
y

hand beam exhibited significantly less which was calculated to be 2370 mm for ^
stiffness in cycle 7 with the crack at the these specimens, with a as one, r as 3 9 . 0 m m
y

edge of weld between the end plate and the and F as 250 MPa. Significant lateral
y

beam flange propagating through the end bending and twisting had occurred when the
plate of beam B at a deflection about tests one and two were stopped, at a beam
+ 260 mm together with the failure of one end deflection of 238 mm, because of the
of the inside bolts before a deflection of magnitude of the end deflections, caused by
+ 30 2 mm was reached. On completing cycle 7 in-plane beam bending. The beams were still
to deflection - 30 2 mm, there was signifi- able to carry high loads at this stage
cant loss of stiffness and a further inside despite their lateral twist and local
bolt failure on the right-hand beam. Also buckles. To reduce these out-of-plane
during cycle 7 a cleavage crack started to deformations, additional braces were placed
spread through the end plate of the left- on the beam at 9 80 mm from the end plate,
hand beam at the edge of the flange butt for specimens three and four.
weld, at a deflection of about - 220 mm.
Only the left-hand beam was loaded in When proof loading bolts, with the 32 mm
cycle 8 and at a deflection of about 220 mm thick end plates in specimens one and two,
a cleavage crack again started to spread it was difficult to bring the joint surfaces
through the end plate and the inside bolts into contact without a very high bolt load.
failed. On reloading in the negative In these circumstances it is better to use
direction a cleavage crack spread through shims to pack the joint. In specimen three
the end plate with only the inside bolts the loads in the left—hand beam bolts
carrying significant load until they failed, exceeded their design capacity, based on
leaving only a small amount of load to be specified proof stress, without failure by
carried by the outside bolts. a factor of 1.54. After fracture had
occurred in the end plate of the right-hand
4
* DISCUSSION OF RESULTS beam the bolts just inside its tension
flange had to carry a very large load. The
In these tests all four specimens load on these bolts exceeded their design
were generally well behaved, although the capacity by a factor of 1.7, ignoring pry-
specimens three and four, where the con- ing. For this type of connection it appear-
nections were deliberately underdesigned, ed that the load was shared equally among
produced degrading hysteresis loops as the four tension flange bolts, because
higher ductilities were achieved. Hence whilst the bolts inside the beam flanges
this type of design would only be accept- carry more of the tension flange load, the
able where low ductilities were required. bolts outside it appear to be subjected to
90

higher prying effects. One feature which length of end plate, where sharp curvatures
was evident in these tests, was the reduct- occurred. The simple theory used to analyse
ion in the prying effects due to the deform- the end plate behaviour does not take
ation of the column flanges. If these account of the high stresses which must
column flange deformations had not occurred occur in transferring the beam flange
then the bolt prying effects would have stresses into the bolts.
been larger. The beam shear force in all
four specimens was transferred across the In these four specimens two different
interface between the end plate and the types of column flange stiffening were
column flange predominantly by friction. used. The first and more conventional type
This was deduced from the fact that there used in specimen one resulted in some
was no significant relative sliding move- significant welding distortions although it
ments . would probably have still been acceptable
in practice. The second type of stiffening
In specimens one to three the portion tested here in specimens two to four was
of the end plate outside the beam tension more easily fabricated and resulted in only
flange did not deform in double curvature to relatively small welding distortions. The
the extent that was initially assumed. In column flanges of specimens one and two
specimen three the deformation of the column both had adequate strength and deformed
flange relaxed the constraints that encour- very little. In specimens three and four
age the end plate outstand to develop a significant plastic deformations occurred
reverse curvature in itself. Subsequently in the column flange under tension bolt
most of the inelastic deformation, that loads.
occurred in the end plate, was concentrated
in the -area immediately adjacent to the
beam tension flange. The end plate inside 5. CONCLUSIONS
the beam flanges seemed to behave generally
in the way that was predicted by Surtees 5.1 Four bolted end plate beam to column
and M a n n ' ^ . This was perhaps most obvious connections were tested here. The
after fracture had occurred in the right- first two were designed to comply
hand beam end plate of specimen three. with the NZS 3404:1977. This was
After this fracture had formed the bolts done by using the equations set out
and end plate immediately inside the beam in Section 2 of this paper and
tension flange were able to support a beam assuming beam strain hardening of 25%
moment of 70% of what it was immediately above the actual yield stress. The
prior to the fracture event. This decrease third and fourth connections were
in load carrying capacity is shown in underdesigned so that their behaviour
Figure 13b and indicates that the larger at high ductilities could be more
portion of the beam tension flange force closely studied.
was carried by the end plate immediately
inside that flange. Close inspection of 5.2 In the first two tests the connections
that end plate in the region between the were able to support the ultimate
beam flanges after testing showed that it plastic hinge beam loads in a rela-
had deformed in much the same way as that tively stiff manner with significant
assumed by Surtees and Mann (9). In part- ductilities and strain hardening.
icular the deformation in the end plate
due to the restraining effect of the beam 5.3 In the third and fourth tests the
web w e n t to zero at the mid depth of the connections underwent larger ductile
beam. One variable that Surtees and Mann's deformations at relatively higher
equation, for end plate thickness does not loads without any strength degradation
include is the thickness of the beam flange until later in the tests when failures
and web and the thickness reinforcement or occurred in the end plates and bolts.
splay of the w e l d s , e.g. fillet welds or Although each of these individual
butt w e l d s . It was apparent throughout the components provided only limited
tests that the end plate deformations ductility, the overall beam behaviour
around the beam tension flange were quite was still ductile, but some degrad-
sensitive to the magnitude of "b", see ation occurred.
Figure 7.
5.4 The tests indicated that the beams
The fracture in the specimen three needed laterally restraining at
right-hand beam end plate occurred in an closer intervals than those required
area of high stress concentration. These by the Standard AS 1250:1975 if higher
stresses were due to bending in the end ductilities were to be achieved in the
plate and tension from the beam tension beams.
flange. The magnitude of these stresses
was thought to have been increased by the 5.5 The tests carried out here have not
close proximity of the bolts to the beam been able to cover a wide range but
tension flange. Their close proximity may their results indicate that these
have prevented the stresses from dispersing connections can be designed and fabri-
more evenly across the end plate width. cated to reliably withstand reversed
The curvature ductility demand was greater cyclic loads from yielding b e a m s .
in the end plate immediately outside of the
beam tension flange then it was immediately 5.6 Further research is required to
inside it. Similar cleavage fractures establish a theory which predicts the
occurred in all the end plates of specimen formation and spread of cleavage
four. There, the end plate was significant- fractures through the end plates.
ly understrength and most of the plastic
deformations were concentrated into a short
 91

5.7 Further research is required into 7. "Safe Load Tables for Structural
other configurations of end plated Steel", Third Edition, Australian
connections, such as these utilizing Institute of Steel Construction,
eight bolts per flange, haunched beams, Sydney, 19 75.
and other forms of column flange
stiffening. 8. Johnstone, N.D. and Walpole, W.R.,
"Bolted End Plate Beam to Column
5.8 This type of connection could be used Connections Under Earthquake Type
for other member sizes; but with Loading", Research Report 8 1 - 7 ,
heavier sections more bolts will be Department of Civil Engineering,
required to be accommodated in a limited University of Canterbury, Sept. 1981.
space or else large bolts will be
required. Difficulties can arise in 9. Surtees, J.O. and M a n n , A . P . , "End-
proof loading M 30 or larger bolts Plate Connections in Plastically
without special equipment. Designed Structures", Proceedings of
Conference on Joints in Structures,
5.9 It is hoped that standardised connect- Vol. 1, Paper 5, University of
ions can be developed, because they Sheffield, England, July 1970.
would avoid the trial and error
process of design. 10. AS 1511, "SAA High Strength Struct-
ural Bolting Code", Standards Associ-
6. ACKNOWLEDGEMENTS ation of Australia, Sydney, 1973.

The first three specimens were tested 11. Hogan, T.J., arid Thomas, I.R.,
by the first author during project work for "Standardised Structural Connections,
his Master of Engineering s t u d i e s ( 1 3 )
u n der Part B;Design M o d e l s " , Australian
the supervision of the second author, who Institute of Steel Construction,
tested specimen four. Sydney, 1978.

The financial support and interest of 12. NZS 4203, "Code of Practice for Gen-
the New Zealand Heavy Engineering Research eral Structural Design and Design
Association (Inc) is gratefully acknowledged. Loadings for Buildings", Standards
Association of New Zealand,
Thanks are due to the University of Wellington, 1976.
Canterbury for the provision of laboratory
technicians, equipment and space. 13. Johnstone, N.D., "Bolted End Plate
Beam to Column connections Under
Ajax GKN Ltd kindly supplied all bolts Earthquake Type Loading", Unpublished
for specimens and rig construction. M . E . Thesis, University of Canterbury,
Feb. 1981.
Professors R. Park, T. Paulay and
I.R. Wood are thanked for their interest 8. NOTATION
and encouragement.
A = Horizontal distance between bolt
7. REFERENCES holes (gauge)

1. Sherbourne, A.N., 'Bolted Beam to = Area of column section

Column Connections", The Structural
Engineer, Vol. 39, June 1961, = Cross-sectional area of column web
pp. 203-210. and doubler plates

2. "Plastic Design in Steel - A Guide B = Width of end plate

and Commentary", ASCE Manual No. 41,
American Society of Civil Engineers, b = Distance from centre of outer
New York, 19 71. bolt hole to edge of b e a m flange
weld
3. Fisher, J.W. and Struik, J.H.A.,
"Guide to Design Criteria for Bolted = Vertical distance between bolt
and Riveted Joints", John Wiley and holes (pitch)
Sons, New York, 19 7 4 .
D' = Bolt hole diameter
4. Packer, J.A. and Morris, L.J., "A
Limit State Design Method for the = Distance between the centres of
Tension Region of Bolted Beam/Column the beam flanges
Connections", The Structural Engineer,
N o . 1 0 , V o l . 5 5 , Oct. 1 9 7 7 , = Yield stress of the beam
p p . 446-458.
= Yield stress of the column
yc
5. M a n n , A . P . and Morris, L.J., "Limit
Design of Extended End-Plate Connect- = Yield stress of the doubler plate
ions", Journal of Structural Division,
American Society of Civil Engineers, = Yield stress of the end plate
YP
ST3, March 1979, pp. 511-526.
= Weighted average of the yield
yw stress of the column w e b and
6. NZS 3404, "Code for Design of Steel
Structures", Standards Association doubler plates
of New Zealand, Wellington, 19 77.
 = Distance from bolt centre line to
sagging yield line for minimum
collapse load

= Thickness of the column flange

plus web root fillet

= Overstrength yield moment of the

beam i.e. probable yield moment
increased by strain hardening

= Plastic moment capacity of the

column

= Plastic moment capacity of the

column, allowing for axial load

= Distance from the bolt hole centre

to the edge of the root fillet

= Distance from the bolt hole centre

to the edge of the doubler plate

•= Axial load in the column

= Tensile load in the bolt

» P P
m r ' m p ' mq
= Tension force in the beam flange
to cause the yield line patterns
of Figures 2, 4, 5 and 6 respect-
ively
= Compression flange force in the
beam

= Maximum permissible tensile

force in the bolt from Table 2
AS 1511 - 1973

= Squash load of column i.e. area

of column multiplied by the
column yield stress

= Ratio of the moment capacity of

doubler plate to the column
flange i.e. F „ T-, /F T
yd d 7
yc c
= Radius of gyration of the beam
about its minor axis

= Thickness of the end plate

= Thickness of the beam flange

= Thickness of the column flange

= Thickness of the doubler plate

= Thickness of the column web

= Shear force carried by the bolt or

by the panel zone from the result-
ant of beam flange forces and the
column shear force

= Maximum permissible shear force

on the bolt from Table 2
AS 1511 - 1973

= Distance from the centre of the

bolt hole to the edge of the weld
to the horizontal stiffener

= Distance from the centre of the

bolt hole to the yield line for
the minimum value of the collapse
load