L11R - 05 - DCB 56 Gusset and BasePlates
L11R - 05 - DCB 56 Gusset and BasePlates
Manukau City
P O Box 76 134
Manukau City, New Zealand
Phone: +64-9-262 2885
Fax: +64-9-262 2856
Email: structural@hera.org.nz
HERA Steel Design & Construction Bulletin Page 1 No. 56, June 2000
Revised Design Guidance for for maximum compactness without introducing
additional eccentricities.
Proportioning Design Actions
from the Braces into the This means that the system and member design
can proceed on the basis that the centrelines of
Supporting Members of the brace/beam/column intersect at the WP. The
Brace/Beam/Column connection can then be designed for a compact,
efficient layout, even if this involves some brace
Connections relocation at the connection, without needing to
This article has been written by G Charles Clifton, HERA
Structural Engineer
recheck the supporting members for additional
moments from increased eccentricities generated
Introduction and Background by the revised layout.
DCB Issue No. 47 presented an article on design
guidance for brace/beam/column connections, Prior to presenting the Thornton method from [4]
incorporating gusset plate(s). That article has in DCB Issue No. 47, the HERA Structural
three parts, namely: Engineer undertook some trial designs to check
its application for a range of connection
(i) design of a gusset plate itself configurations. These designs all used an initial
(ii) proportioning actions from the brace into brace orientation (through the WP) of 45º, with a
the supporting members range of values to the connector centroids in the
(iii) calculating column bending moments due beam and column. The method worked as
to joint eccentricities. anticipated, generating an equilibrium condition
consistent with that shown in Fig. 47.2 (a).
The design of the gusset plate itself is covered by
sections 10.2.2.1 and 10.7.2 of HERA Report The second article in this current issue is to
R4-80, Limit State Design Guides Volume 1 [3] present design concepts for brace/beam/column
and the first part of the DCB No. 47 article bolted endplate connections of the general type
summarised the various failure modes and the shown in Item 39, HERA Report R4-58 Manual of
relevant sections of [3] that cover each mode. Standard Connection Details for Structural
That advice remains unchanged and is not the Steelwork, Second Edition [5]. In preparing for
subject of this revision, except that the limited this second article, the HERA Structural Engineer
guidance on flexible endplate design and applied Thornton's method, as detailed in DCB
checking of unstiffened column flange capacity Issue No. 47, to proportioning actions for a
given on page 5 of DCB No. 47 is superseded by brace/beam/column in which the initial orientation
the more detailed guidance on that topic given in of the brace was at 53º to the horizontal, which is
the next article herein. at the upper end of the recommended range of
brace angles to apply in practice. The result was
Axial forces from the brace are transferred (via the not as expected; the resultant lines of action of
gusset plate) into the supporting beam and forces through the brace to beam and the brace to
column members. There are various equilibrium column connections did not intersect the
models available for this transfer, with the extension of the brace centreline through the WP
simplest involving the column resisting the vertical at all, nor were the resultant actions close to being
component of brace axial force and the beam "in balance" with the horizontal and vertical
resisting the horizontal component. This model component of the brace actions as determined
introduces moments into the beam and column from the initial angle of brace orientation through
members from the eccentric transfer of axial the WP. Consideration of further cases where the
forces and these moments must be considered in initial brace orientation varied from 45º showed
design of the beam and column members. that the proportioning method presented in the
DCB Issue No. 47 equations 46.1 to 46.6
A slightly more complex model has been becomes unconservative, to an increasing extent,
developed by Thornton [4] and was the model as the orientation varies from 45º.
presented in DCB Issue No. 47. The basis of this
model is to achieve equilibrium about the Work However, a simple revision to that method can be
Point (WP - ie. the intersection of the brace/beam/ made, which not only removes that
column centrelines) with linear actions only, as unconservatism but also allows the effect of
was illustrated in DCB Fig. 47.2. noding eccentricity to be eliminated. The revised
details are presented shortly, followed by the
A significant advantage claimed [4] for this calculation of bending moments from connection-
method is that it takes account of the centroid of induced eccentricities into the supporting
the connectors between the brace/gusset plate members. First, however, the sections of DCB
and the beam and column, when proportioning Issue No. 47 that are superseded by this advice
actions into the supporting members, thus are stated.
allowing the connection layout to be determined
HERA Steel Design & Construction Bulletin Page 2 No. 56, June 2000
Fig. 56.1
Revised Linear Equilibrium Model for Compact Brace/Beam/Column Connection Design
In Fig. 56.1
∗
N = design axial force in brace
∗
V, H = vertical and horizontal components of N
∗
R = design shear force from beam
Point A (Fig. 56.1(a)) is the centroid of the connectors between the gusset plate and the beam
Point B (Fig. 56.1(a)) is the centroid of the connectors between the gusset plate and the column
HERA Steel Design & Construction Bulletin Page 3 No. 56, June 2000
eB = distance from beam centreline to beam downwards and N ∗ is tension. When R ∗
face includes a seismic component, its sign must
= dB/2 for a uniform, doubly symmetric be determined from the relevant section of
section [6]. Note that with eccentrically braced
ec = distance from column centreline to
frames, the seismic component of R ∗ always
column face acts in opposition to the vertical component
= dc /2 for a uniform, doubly symmetric
section of brace force; hence RE∗ is opposite in sign
to VB∗ for an EBF.
The revised proportioning method ensures that
the resultant of the brace actions into the
supporting members always intersect at the Work The relative magnitudes of HB∗ , H ∗c , VB∗ and Vc∗
Point (WP). for a brace at an angle of approximately 45° are
shown in Fig. 56.1. As expected, HB∗ and Vc∗
The actions to be carried into the supporting
have the greatest magnitude, H ∗c and VB∗ the
members to achieve this are as follows:
lowest.
r = (α + e c )2
+ (β + eB )
2
(56.1)
In Fig. 56.1, the beam design shear force, R ∗ , is
shown carried only by the beam to column
HB∗ = horizontal force carried into the beam connection. (R ∗ is termed Vbeam in [6].) If this
α + ec ∗ force is large and acts in the same direction as
= N (56.2) VB∗ , it may be desirable to distribute it into the
r
gusset plate to column connection as well. In this
case, the gusset plate serves as a haunch and the
VB∗ = vertical force carried into the beam
vertical force between the gusset plate and beam
=
eB ∗
N (56.3) ( VB∗ ) must be increased by the component of R ∗
r that is carried through the gusset plate to column
part of the connection.
Hc∗ = horizontal force carried into the column
ec ∗ The final point to re-emphasise in this part of the
= N (56.4) article is that the above method for proportioning
r
the actions from the brace into the supporting
beam and column members, and the guidance in
Vc∗ = vertical force carried into the column the following part of this article relating to
β + eB ∗ calculating the column moments due to noding
= N (56.5) and beam eccentricities, are both applicable to
r brace/beam/column connections involving other
than gusset plates; eg. to the Item 39 connection
N∗ = design axial force in brace given in HERA Report R4-58 [5].
H∗ = horizontal component of N ∗
V∗ = vertical component of N ∗ Calculating Bending Moments Introduced
into the Columns by Connection
The connection between the beam and the Eccentricities
column must resist: In general, there are two sources of column
moments from connection eccentricities, namely:
(i) A horizontal force equal to Hc∗
(R )
(i) potential moments from noding eccentricity
∗
(ii) A vertical force equal to +VB∗ or (ii) moments from eccentric transfer of beam
(R ∗
−VB∗ ) shear force, R ∗ .
where:
Potential moments from noding eccentricity arise
R ∗ = design shear force from collector when the brace centreline does not meet at the
beam, eg. as calculated from HERA intersection of the beam and column centrelines,
Report R4-76 [6], Seismic Design of as shown in Fig. 56.2 (a) and (c).
Steel Structures.
(R )
In the overall system analyses, the initial brace
∗
+VB∗ will apply when R ∗ is acting angle (θbrace,initial) will have been such as to give
downwards and N ∗ is compression; zero noding eccentricity (Fig. 56.2(b)). In the
(R ∗
−VB∗ ) will apply when R ∗ is acting
quest for a compact gusset plate solution, the final
brace angle may have been altered, as is shown
HERA Steel Design & Construction Bulletin Page 4 No. 56, June 2000
Fig. 56.2
Definitions of Noding Eccentricity (from [6])
in the connection detail in Fig. 56.3, where the Design Concepts for Brace/
brace end at the connection is moved slightly to
the left compared with its position so as to give Beam/Column Connections in a
zero noding eccentricity. Braced Steel Frame Seismic-
However, the actions into the supporting members Resisting System
are then proportioned using equations 56.1 to
56.5 to bring the resultants back through the WP. Introduction and Scope of Article
This is done on the basis of the final as-proposed This article presents design concepts for
connection detail. brace/beam/column connections of the type
shown in Fig. 56.3. This form of connection is
The benefit is that it allows the columns to be well suited to eccentrically braced frame (EBF) or
designed for no additional moment due to noding to V-braced concentrically braced frame (CBF)
eccentricity, irrespective of the final connection systems, where the assemblage comprising beam
layout. The downside is that the connection and two braces is fabricated in the shop and
design must incorporate the force components assembled on site by site bolting. Examples of it
VB∗ and H ∗c , which impact on the brace to are shown in Fig. 19.3 of HERA Report R4-76 [6]
and as Item 39 in HERA Report R4-58 [5]. In
column, brace to beam and beam to column
such instances, it is used at one end of the brace,
connections. However, these components will with the other end typically shop welded to the
often not have a significant effect on the size and
beam. In an X-braced CBF, it can be used at both
strength of the bolts and endplates required and
ends of the brace.
this is offset by the more straightforward column
design.
This article looks first at the performance
requirements for the connection, then at the
The column bending moment from eccentric determination of design actions on the connection
transfer of the beam shear force, R ∗ , is and the connection components. This is followed
calculated in accordance with NZS 3404 Clause by guidance on the initial selection of connection
4.3.4.2; see also the relevant section of [6], such components, followed by a listing of the principal
as section 11.2, step 8.3 (b), for category 1 and 2 design checks to be made and guidance on the
EBFs. carrying out of these design checks.
Having calculated the column design bending As mentioned in the introduction to this Bulletin,
moments from eccentric transfer of beam shear the aim of this article is to present design
force, R ∗ , these are transferred into the column concepts, both for information to designers and to
above and below a given level in accordance with encourage feedback.
NZS 3404 Clause 4.3.4.3. The same advice is
also given in the relevant section of [6], eg.
equations 11.19 and 11.20 of [6] for category 1 or
2 V-braced EBF systems.
HERA Steel Design & Construction Bulletin Page 5 No. 56, June 2000
Fig. 56.3
Brace/Beam/Column Bolted Endplate Connection
(1) The gusset plate shown between the brace and the horizontal endplate is required to transfer horizontal force from brace to beam
∗ ∗
( HB from equation 56.2) and vertical tension force from brace to beam ( VB,tens from equation 56.3) when N∗ is tensile.
(2) Refer to the accompanying article for guidance on locating the brace and making the initial selection of components.
(3) The welds between the brace flange and the endplate bolted to the beam and between the brace flange and the endplate bolted
to the column may be not pre-qualified to AS/NZS 1554.1[7] due to the angle of intersection between the two members being
outside that allowed for pre-qualified joints. In this instance, specific weld procedures will need formulating for these welds in
accordance with one of the options (b) to (e) available from Clause 4.2; typically Clause 4.2(b) would be used. Clearance in the
brace web will be needed to lay the butt welds between the brace flanges and endplates; use eg. a 20 mm radius cope as shown
above.
(4) The connection between beam and column is a modified MEPS-4 connection from [1]. The modification involves the connection
having to resist H ∗c and R ∗ ± VB∗ and the endplate being flush with the top of the steel (ie. dimension af = 0). Start with an
MEPS-4 30/15 for sizing of components; use either a butt weld or a compound weld (partial penetration butt weld top side and
fillet weld bottom side between the beam top flange and endplate).
(5) The two endplates to the brace should overlap in the corner formed by the brace/beam intersection and be fillet welded on each
side, as shown in Fig. 56.3.
(6) It is desirable to use dimensions for bolt setout that are consistent with the provisions of R4-100 [1]. Suitable dimensions for ae1
and Sg on the vertical face are given in the detail drawings for the MEPS - 4 connection (see this drawing near the back of [1]).
Suitable dimensions for af on both faces and for Sg, ae1 on the horizontal face are given in the detail drawings for the MEPS-8
connection.
As with most steel connections, there are a present appropriate concepts and references to
number of ways in which the design can be detailed design guidance; readers with suggested
satisfactorily undertaken. The aim here is to improvements on these are encouraged to send
HERA Steel Design & Construction Bulletin Page 6 No. 56, June 2000
details to Charles Clifton, HERA Structural local column moment generated through
Engineer. The longer term aim is to present a connection-induced eccentricity might cause
connection design example for this type of either mode of failure and must be suppressed.
connection in seismic design seminars later in
2000 and to present tabulated connection Objective (4) comes from inelastic time-history
capacities for a range of brace/beam/column sizes analyses of EBFs and CBFs. These have shown
in the 2000/2001 year. that, when these systems are subjected to seismic
attack above that associated with the design level
Performance Requirements on Connection ultimate limit state event (eg. that associated with
the maximum considered event (MCE)), their
Design objectives response is improved when the
These are as follows: brace/beam/column connections are semi-rigid to
rigid, rather than pinned. See, for example, ref.
(1) To resist the capacity design derived design [5] from [6], for details relating to EBF analyses.
actions generated within the brace and to
distribute these to the supporting beam and Objective (5) is a requirement of NZS 3404
column members. Clause 12.9.7.3.2 and is intended to ensure that
the connection will not be subject to local failure
(2) To resist these actions without weld or bolt by inelastic out-of-plane brace behaviour, where
failure alone or failure of the gusset plate this may occur. Such behaviour is very unlikely in
being the limiting conditions on the EBFs, but is likely in V-braced CBFs and
connection. X-braced CBFs.
(3) To distribute brace actions into the Design Actions on Connections and
supporting members in such a way as to Connection Components
prevent failure of the supporting members
through crippling or fracture at the Design actions from the brace
connections. For category 1, 2 and 3 systems, these are the
capacity design derived design axial forces,
(4) To provide a connection which remains c
N brace , in conjunction with the appropriate axial
effectively rigid up to the level of the seismic
force from gravity loading, Nbrace, GQu.
design actions and functions as a semi-rigid
connection, with appreciable stiffness, under
These are determined from the capacity design of
higher levels of seismic loading on the
the system, for category 1, 2 and 3 CBF and EBF
structure.
systems.
(5) To accommodate any out-of-plane rotation of
For EBFs, the relevant sections of R4-76 [6] are
the brace member that may occur at the
as follows:
connection without compromising any of (1)-
(4) above.
§ section 11.2, steps 7.1 and 9 for category 1
and 2 V-braced EBFs
Reasons for these objectives
§ Section 11.4, steps 7.1 and 9 for category 1
Objective (1) is a requirement of the capacity
and 2 D-braced EBFs
design process of [2, 6], which is required for
§ see also section 13.1, especially section
category 1, 2 and 3 EBFs and CBFs.
13.1.4, for connection related
recommendations
Objective (2) incorporates the requirement of
NZS 3404 Clause 12.9.1.1.3 that the load path
Note that, if the proportioning method given in the
and strength hierarchy within the connection shall
previous article herein is used to distribute brace
be such as to avoid inelastic demand being
actions into the beams and columns, the noding
concentrated into the connectors or connection
eccentricity, en, given in section 13.1 of [6], is
components (except for column panel zones,
taken as zero.
which are not relevant to this type of connection
and system). This requirement is also expressed
For CBFs, the relevant sections of R4-76 [6] are
in the relevant sections of [6] and is embodied in
as follows:
all the connections presented in HERA Report
R4-100 [1] that form part of a seismic-resisting
§ section 19.1.2, which then references back to
system or may be subject to seismic-induced
the relevant part of sections 16-18 as
inelastic demand.
appropriate. This is because the calculation
of capacity design derived brace actions
Objective (3) is required to avoid an unexpected
depends on the type of system and the
failure of the supporting members, especially the
effectiveness of braces in compression, with
columns, due to crippling or fracture. Excessive
HERA Steel Design & Construction Bulletin Page 7 No. 56, June 2000
three possibilities (X-braced, braces effective Initial Selection of Connection Components
in tension and compression; V-braced, Making a realistic initial section of connection
braces effective in tension and compression; components is important towards achieving a
X-braced, braces effective in tension only) rapid connection design. The following guidance
will be of assistance:
Note that the compression slenderness reduction
factor for a buckled brace, α ′c , is now given by (1) Obtain an estimate of the number and size of
bolts required between the brace and column
NZS 3404 Equations C12.2.3(1) or C12.2.3(2),
and between the brace and beam from
instead of as specified by [6].
equations 56.6 and 56.7.
c
The earthquake generated axial forces, N brace , ∗
Vinitial,brace
are tension and compression. For EBFs, the n bolt,brace− colum ≥ (56.6)
tension and compression magnitudes are the φV f n
same, for CBFs they are different. Connection
design must take account of both directions of ∗
H initial,brace
axial force input from the brace. n bolt,brace− beam ≥ (56.7)
φV f n
Proportioning the design actions into the
supporting members where:
∗
This is undertaken in accordance with equations Vinitial,brace = vertical component of brace
56.1 to 56.5 herein; see page 4 herein.
design axial force, N ∗
∗
The bending moments introduced into the H initial,brace = horizontal component of
columns by the connection eccentricities are then ∗
brace design axial force, N
determined in accordance with the
φV f n = bolt design shear capacity,
recommendations on pages 4 to 5 herein. In this
regard, there are no design moments from noding threads included in shear
eccentricity. Column moments due to eccentricity plane, eg. from [8].
arise only from the eccentric transfer of the beam
If n bolt ≤ 4, look at using the layout from a 4
shear force, R ∗ , into the column.
bolt MEPS-4 type configuration from [1] as a
Location of Brace into Connection starting point.
In the initial analyses of the braced system, the
brace will have been located so that the If 4 < n bolt ≤ 6 , look at putting in an extra
centrelines of brace, beam and column intersect row of bolts; outside the top flange of the
at the Work Point (WP). These centreline brace into the column and/or inside the
locations are shown in Fig. 56.3. bottom flange of the brace into the beam, as
required.
In order to make the brace/beam/column
connection more compact and cost-effective, the (2) Having selected the initial bolt size, use the
final location of the brace should be offset so as to relevant dimensions from R4-100 [1] where
require gusset plate stiffening on one side only. shown in Fig. 56.3 for location of the bolts.
Which way to offset the brace from its initial
position for analysis will depend on the initial (3) For initial selection of endplate thickness, use
brace angle and the sizes of the beam, column the next available plate thickness above the
and brace members. For initial brace angles brace flange thickness for the endplate
above 45º, it is often preferable to offset it connected only to the brace mitred end and
upwards so that the majority of the mitred end of go to the next thickness higher for the
the brace frames into the column face; that is the endplate connected partially to the brace and
situation shown in Fig. 56.3. Where the beam partially to the gusset plate. The difference is
depth is greater than the column depth, this because the latter baseplate has a lower
upward offset will also be preferable for initial tension capacity, due to one edge being
brace angles below 45º. unsupported.
The brace location should be such that the mitred (4) Make the gusset plate thickness equal to the
face of the brace on the side not carrying the next available plate thickness above the
gusset plate can accommodate one flange and brace flange thickness.
sufficient web depth to accommodate the
connection dimensions shown in Fig. 56.3 on that (5) For the connection between the beam and
side (the column in that instance). The shorter column, use an MEPS-4 50/25 connection
mitred face is then strengthened with a gusset from [1] as a starting point.
plate for shear and tension transfer.
HERA Steel Design & Construction Bulletin Page 8 No. 56, June 2000
Principal Design Checks to Make below the level under consideration in
These are as follows: accordance with NZS 3404 Clause 4.3.4.
(1) The brace to endplate to column connection Guidance on Carrying Out These Principal
must be designed for Vc∗ (equation 56.5) and Design Checks
H ∗c (equation 56.4). Brace to endplate to column connection
Design for Vc∗ is straightforward. This shear force
(2) The brace to endplate to beam connection
is shared equally amongst the bolts connecting
must be designed for HB∗ (equation 56.2) and the brace endplate to the column. The endplate is
VB∗ (equation 56.3). also checked for this shear, although this check
will seldom be critical.
(3) The gusset plate unsupported edge should
be checked against compression buckling Design for H ∗c is more complicated. This involves
using equation (10.16) from [3]. determining the design tension capacity, φH c , of
the endplate/bolts and ensuring that this capacity
(4) The gusset plate must be designed to resist
is greater than H ∗c . The design tension capacity
the tension force from VB∗ and the
is based on yieldline determination and involves
component of shear force, HB∗ , that is not use of a yieldline procedure which covers the
transferred through the mitred face of the arrangement of endplate/bolts/stiffeners shown in
brace. Fig. 56.3.
(5) A combined shear and tension check must An appropriate procedure is given in step 1A,
be undertaken on: section 2.8 of SCI Publication No. 207/95, Joints
§ the cross sectional area of the brace (and in Steel Construction, Moment Connections [9].
gusset plate, when appropriate) that is in Brief details of the procedure are given in
contact with the endplate which is bolted Appendix A56 herein. That procedure covers all
to the beam. practical combinations of endplate/bolts/stiffeners
§ the cross sectional area of the brace (and with two bolts per row, including those for this
gusset plate, when appropriate) that is in brace to column connection. It involves the
contact with the endplate which is bolted following:
to the column).
(1) Calculating the effective length of yieldline in
(6) The welds between brace and endplates and an equivalent T-stub, where this effective
between gusset plate and brace, gusset plate length is determined for each bolt row using
and endplate must be designed to resist the Tables 2.4 to 2.6 of [9].
applied actions. The brace flanges are
typically butt welded; brace web and gusset (2) Determining the design tension capacity
plate are double-sided fillet welded. These φH c , which is termed Pr in [9], from step 1,
welds can conservatively be sized to develop pages 17-18 of [9]. In making this
the design tension capacity, in kN/mm run, of determination, note the following:
the web or gusset plate, as appropriate. This
will cover both overall and local effects on § the strength reduction factors for steel,
these welds. φ = 0.9, and for bolts in tension, φ = 0.8,
need to be incorporated into the
(7) The beam to column connection must be equations for each failure mode
designed for H ∗c and for a vertical force § mode 3, which involves bolt failure alone,
equal to R ( ∗
+ VB∗
or R ) ( ∗
− VB∗
, depending ) is not permitted by NZS 3404 for seismic-
resisting connections. The bolt capacity,
on the sign of each component.
ΣφN tf (which is termed ΣPt ′ in [9]), must
(8) The column flange must be able to resist be greater than the minimum tension
∗ capacity calculated for modes 1 and 2
H c,tens , preferably without stiffening.
§ because the endplates are relatively thin,
Likewise the beam flange must be able to mode 1 will often govern (ie. will give the
∗
resist VB,tens without stiffening. lower design tension capacity of the two
modes).
(9) Additional moment into the column arising
from the eccentric transfer of the beam shear (3) The equations for modes 1 and 2 from [9]
both implicitly incorporate prying effects on
force, R ∗ , must be determined and
the bolts, however these effects are
proportioned into the column above and
HERA Steel Design & Construction Bulletin Page 9 No. 56, June 2000
potentially overestimated by the simple tee- connecting the brace endplate to the beam. The
stub approach. There are two options for endplate is also checked for this shear, although
considering the prying effect on the bolts, this check will seldom be critical.
these are:
The gusset plate is used to extend the cross-
(i) If H c∗ ≤ 0 .5 Pr,mode1 , then prying effects section of brace in contact with the endplate so
can be neglected, as the yieldline under that it can carry the combined tension, VB∗ , and
the bolts is not required to be developed shear HB∗ . The cross-section used for resistance
at this level of design force.
of these forces comprises the full area of brace
(ii) If H c∗ > 0 .5Pr,mode1 , then the additional plus gusset plate in contact with the endplate.
tension force on the bolt induced by The length of gusset plate, if sized in accordance
prying, N p∗ , can be determined from with the bolt layout parameters from R4-100 [1]
given in Fig. 56.3, should give a cross-section
NZS 3404 Equation CM2.4. The sufficient to resist these actions; if not, then
endplates are relatively thin and the extend it further along the beam.
additional prying force can, without
undue conservatism, be set at the The weld between the brace web and the
maximum, from that equation, of endplate should be a double-sided balanced fillet
N p∗ = 0 .4 N tf∗ . weld, sized to resist the design tension capacity of
the web, φt w, brace fyw,brace . Likewise for the weld
The bolts must be checked for combined shear between the gusset plate and endplate and
and bending to Clause 9.3.3.3 of [1], where: between the gusset plate and brace, which is
sized to resist the design tension capacity of the
Vc∗ gusset plate, φt gp fy ,gp .
V∗ = , n b = no. of bolts (56.8)
nb
H ∗c Design of the endplate and bolts for VB∗ uses the
N∗ = + prying component N p∗ , if
nb procedure from step 1A, section 2.8 of [9]. This is
required (56.9) similar to design of the endplate that is bolted to
the column to resist H ∗c and the guidance given in
The welds between the brace and endplate must items (1) - (3) of the previous subsection applies,
be designed to transfer the design actions substituting VB∗ for H ∗c . In this case, however,
( )
Vc∗ , H ∗c from brace to endplate and also to allow there is generally no stiffener on the gusset plate
for localised increase in tension actions generated unsupported edge away from the column, which
by the yieldlines around the bolts. The weld must be accounted for in developing Leff for the
between the brace flange and endplate will endplate; Leff for the row of bolts adjacent to this
typically be a butt weld, with due allowance for edge is considerably lower than for the next row of
weld access (see Note 3 to Fig. 56.3). The weld bolts and is given by the second case down, right
between the brace web and endplate should be a hand side, Table 2.6 of [9].
double-sided balanced fillet weld, sized to resist
the design tension capacity of the web, The gusset plate unsupported edge should be
φt w, brace fyw,brace , as this covers the global design ∗
checked against local buckling under N brace, comp ;
actions and any localised increase in tension from equation 10.16 of [3] this requires
action generated around the bolts.
bgp 250
≤ 40 , where b is the length of the
The cross-section of brace in contact with the t gp fy gp
endplate must be able to resist the combined
unsupported edge (see Fig. 56.3).
tension, H ∗c , and shear Vc∗ . The shear capacity
should be taken as non-uniform, using a non- Check on column flange, beam flange tension
uniformity factor of 0.83 as determined from capacities
NZS 3404 Clause 5.11.3. The cross-section for The unstiffened column flange tension capacity to
resistance of Vc∗ and H ∗c comprises the full area ∗
resist H c,tens , is given by step 1A, section 2.8 of
of brace that is welded to the endplate (including a [9]; see Table 2.6 of [9] for determination of Leff.
gusset plate, if used). When the column is subject to very high levels of
design axial force, this tension capacity may be
Brace to endplate to beam connection and reduced by the axial force. Use the factor η in
gusset plate NZS 3404 Equation 12.9.5.3 (5) to account for this
Design for HB∗ is straightforward. This shear reduction; that gives η = 0.95 when N ∗ = 0.5φNs .
force is shared equally amongst the bolts
HERA Steel Design & Construction Bulletin Page 10 No. 56, June 2000
A similar approach is used for the beam tension design; this is especially applicable in the
∗ selection of initial components.
flange capacity to resist VB,tens .
Introduction and Scope of Article (2) To allow for inelastic action to develop in the
This article presents design concepts for moment- column base adjacent to the connection for
resisting column baseplate connections of the columns in category 1, 2 or 3 seismic-
type shown in Fig. 56.4. That figure shows a resisting systems.
connection with an unstiffened baseplate and
such an option should first be considered, within (3) To distribute the internal actions from (1) and
the recommended limitations on baseplate (2) into the reinforced concrete supporting
thickness and grade given herein to suppress member, without causing significant crushing
brittle fracture of the baseplate. These are given of the supporting concrete, yielding of the
on pages 12-14 under the heading Limits on hold-down bolts, weld failure or endplate
baseplate thickness and grade to suppress brittle fracture. Inelastic demand in the endplate
fracture. They will restrict application of this type itself is acceptable, provided that this
of column base connection, especially where the demand is minimal. The connection must
column is subject to axial tension and moment. retain most of its pre-earthquake rotational
rigidity throughout a severe event.
The design concepts given herein are based
around the detailed guidance on column (4) To resist horizontal slip between steel and
baseplate connections given in section 6 of [9] concrete under seismic-induced shear forces.
and make reference to that publication for the
detailed requirements. Where appropriate,
modifications to the requirements of [9] are made,
eg. for consistency with NZS 3101 Concrete
Structures Standard [10]. Guidance is given on
using R4-100 [1] where possible to facilitate rapid
HERA Steel Design & Construction Bulletin Page 11 No. 56, June 2000
Fig. 56.4
Moment-Resisting Column Baseplate Connection With Unstiffened Baseplate (Item 2a from [5])
Reasons for these objectives Finally, the connection must be able to resist
In category 1, 2 or 3 multi-storey seismic-resisting seismic-induced shear without horizontal sliding,
systems which are subject to capacity design, the as this would significantly effect inelastic demand
columns will be secondary elements and will be on adjacent column members, concentrating this
subject to capacity design derived design actions demand into columns where shear slippage did
generated by the system response. In all not occur and increasing the potential for column
category 1, 2 or 3 seismic-resisting systems, the failure there.
columns will still be subject to potential inelastic
action at their base, when connected by moment- Limits on Baseplate Thickness and Grade to
resisting connections into the foundation system. Suppress Brittle Fracture
The connection must be able to resist the internal Background
actions generated by these conditions; hence As advised in section 6.2 of [9], the baseplates in
objectives 1 and 2. unstiffened moment-resisting column baseplate
connections are very thick, by comparison with
Unless special design and detailing end plates of steel to steel connections. When
considerations such as ring-springs under the subject to earthquake loading conditions, thick
hold-down bolts or other specialist devices are baseplates may be vulnerable to brittle fracture.
used, the deformation demand on the connection An example of this is, from the 1994 Northridge
components during a severe earthquake must be Earthquake, is shown in Fig. 56.5.
limited. Localised yieldline action can be
developed in the baseplate, although this is This earthquake and the 1995 Kobe (Great
limited by the relatively thick baseplates used. Hanshin) earthquake caused a considerable
Inelastic stretching of the hold-down bolts and/or number of failures in steel connections, including
crushing of the concrete must be suppressed to many examples of brittle fracture. The damage
the extent where the rigidity of the connection has been well documented and a large amount of
itself would be significantly reduced during the research has been undertaken from 1994 to the
earthquake, as this increases inelastic demand on present day to identify causes for these brittle
the system being supported. That is the reason fractures and develop design, detailing provisions
for suppression of the first two failure modes given to suppress them. Two publications which cover
for objective 3. The reasons behind suppression both these areas are [11, 12]. More general
of the last two failure modes is given in the account of steel building performance in the Kobe
following section. Earthquake is given in DCB Issue Nos. 8, 9 and
10.
HERA Steel Design & Construction Bulletin Page 12 No. 56, June 2000
(paper from [11]). However, there are no reported
examples, that the HERA Structural Engineer is
aware of, of baseplate brittle fracture (as shown in
Fig. 56.5) occurring in the Kobe (Great Hanshin)
earthquake. This is despite the earthquake
generating high ground accelerations, velocities
and occurring at the end of a winter's night with
o
the air temperature just below 0 C.
HERA Steel Design & Construction Bulletin Page 13 No. 56, June 2000
this section of the article ends with brief (4) Shear studs will resist both column tension
suggestions of an alternative column base detail and compression, thus the baseplate does
which does not require thick endplates and avoids not have to transfer high compression load
the risk of brittle fracture that may be associated into the concrete.
with these.
(5) Stud spacing should be to NZS 3404 [2]
Recommended limits on baseplate thickness Clause 13.3.2.3(f).
and grade for column baseplate connections
(6) Shear stud capacity is to NZS 3404
These are as follows: Equation 13.3.2.1 incorporating φsc = 0.80.
(1) Steel grade to be 250 or 300 or equivalent; Design Actions on Connections and
see NZS 3404 Table 2.6.4.4. Connection Components
These are given in section 4, pages 22-25 of DCB
(2) Maximum baseplate thickness to be: Issue No. 50. Those requirements extend the
coverage of HERA Report R4-76 [6] to giving
(2.1) 50 mm for columns subject to moment plus design actions at the column bases.
axial compression.
For MRFs, design actions for fixed bases and for
(2,2) 32 mm for columns subject to moment plus pinned bases are given. A large number of MRF
axial tension. time-history analyses undertaken as part of
HERA's semi-rigid connection development have
The more stringent limit for columns subject to
quantified the rotation demand expected on a
moment plus axial tension is for two reasons,
column base in both instances and the provisions
namely:
in that issue for pinned bases are designed to
§ the more severe loading condition, with deliver the rotation demand required from the
regard to brittle fracture potential, of axial connection. These analyses have also shown the
tension and moment considerable influence that the choice of
§ The more severe consequences if brittle connection rigidity (fixed or pinned) has on the
fracture does occur MRF response and the importance of maintaining
rigidity and moment resistance in a fixed
The above limits are recommended for both connection.
unstiffened and stiffened baseplates. The latter
When applying the provisions of section 4.2.1,
are only briefly covered by this article. If stiffeners
design actions for MRF fixed bases, the design
are used, the load path to get actions from column
to stiffener to baseplate should be carefully moment is taken as φMr for design of the
considered, with stiffeners positioned so as to baseplate and determination of stress block into
the concrete (eg. as shown in Fig. 56.6) and
reduce baseplate thickness for both tension and
compression actions. φoms Mr for "design of the fillet welds, bolts, shear
studs." In the case of the column onto unstiffened
Alternative column base connection detail baseplate shown in Fig. 56.6, this means the
An alternative detail involves placing the column following:
in a pocket within the concrete base and using
§ calculation of baseplate thickness on the
welded, headed shear studs to enact transfer of
axial load from each element of the column into compression and tension sides uses φMr
the concrete. This concept is shown as
§ calculation of hold-down bolt size uses
Connection Item 30 in HERA Report R4-58 [5].
φ oms M r for fixed base MRF connections; ie.
Some brief design and detailing points regarding use the tension forces as determined for the
this connection are as follows. φ
baseplate thickness calculations x oms
φ
(1) Use a BPP from [1] for the endplat e and
§ calculation of hold-down bolt size uses φM r
bolts.
for EBF and CBF connections
(2) Design the shear studs to resist
overstrength actions from the column § if using fillet welds between the column
flanges, design actions from the column flanges and baseplate, these are sized to
web. resist the overstrength tension capacity from
the column flange
(3) The shear studs should be located within
the cage of transverse and longitudinal § if using fillet welds between the column web
reinforcement in the reinforced concrete and baseplate, these are sized to resist the
section.
HERA Steel Design & Construction Bulletin Page 14 No. 56, June 2000
design tension capacity from the column nut and bearing plate can be omitted and
flange. should be if access is difficult. If tack welds
are used, they must be applied between
Initial Selection of Connection Components bolt and nut only on the unloaded face of
Making a realistic initial selection of connection the nut). Nut and hardened washer supply is
components is an important step towards to AS/NZS 1252 [14]; see section 5.5(2), pp.
achieving a rapid and effective connection design. 25, 26 of DCB Issue No. 50.
The following guidance will be of assistance:
(3) Use a specified 28 day compression strength
(1) Starting with the details given in HERA ( )
for concrete and mortar grout fc′ , fm′ of 30
Report R4-100 [1] for a MEP Cat 3 MPa for initial design; only go stronger if
connection detail; necessary.
§ use the general dimensions given therein
for connection layout (4) Take advantage of the increase in concrete
§ use the bolt diameter, numbers for the bearing strength possible from confinement
hold-down bolt initial estimate of the concrete under the baseplate given by
§ make the endplate approx. 50 mm wider NZS 3101 Clause 8.3.5.2; typically the
each side than the cleat width, b i, given maximum value of 2 can be used.
therein
§ start with the next thickest plate up from Design Procedure
the cleat thickness, ti, given therein
§ use the weld details between column An iterative approach is required; this is
flange and web and the endplate that are described in detail in section 6.9 of [9] and
given therein. summarised in Fig. 56.6.
(2) The bolts used will be fabricated from For outer columns in a MRF, two cases will need
AISI 4140 bar with an ISO cut thread to consideration, namely moment in conjunction with
AS 1275 [13]. A nut will typically be provided maximum compression force and moment in
under an end bearing plate which is cast in conjunction with minimum compression force.
the concrete, as shown in Item 3d of [5].
(The tack welds shown therein between the
Fig. 56.6
Distribution of Forces and Equations for Equilibrium Based on Compression Stress Block Centred
Under Column Compression Flange
Note to Fig. 56.6
The notation C, T, a, b, X, h are the same as those used in step 1, section 6.9 of [9].
HERA Steel Design & Construction Bulletin Page 15 No. 56, June 2000
These are given as section 4.2.1(1) and 4.2.1(2), A2
pp. 22, 23, DCB Issue No. 50. The second case = enhancement factor due to confinement,
is called "maximum tension force". However, this A1
axial force will, in practice, be either compression from NZS 3101 Clause 8.3.5.2
or tension depending on the relative magnitude of
the gravity compression force from (G + Qu) and The grout packing placed under the baseplate
the seismic tension force. This second case will must have a specified 28 day compression stress,
deliver the maximum moment. fm′ ≥ fc′ .
Guidance on Carrying Out The Design When using the design procedure from [9], fcb
Procedure from equation 56.12 is used instead of 0.6fcu.
Use of partial strength reduction factors in
determining design capacity directly START
Several steps in the design procedure involve
determination of internal actions/design capacities
based on both concrete and steel contributions. Choose trial base
One such case is the calculation of baseplate dimensions and bolt
thickness, which is a function of concrete design positions
bearing stress and hold-down bolt tension (see
Fig. 56.6). In such instances, strength reduction
factors appropriate to the material are applied in Choose appropriate
calculating the internal actions/design capacities concrete bearing stress
Amend sizes
directly, rather than using nominal capacities
if necessary
(based on minimum specified material properties
STEP 1
or characteristic material properties) and a global Determine extent of
strength reduction factor. This is termed use of compression stress block
partial strength reduction factors and is consistent and bolt tension forces
with the approach used in longitudinal shear
checks on composite beams to NZS 3404 [2]
Clause 13.4.10, where concrete and steel STEP 2
contribute to both the internal actions and design Determine baseplate
resistances. thickness
HERA Steel Design & Construction Bulletin Page 16 No. 56, June 2000
This involves solving the quadratic equation 6.6 X tf
from [9] to obtain the depth of compression stress e= − − 0 .8 t w (56.16)
2 2
block, termed X. That equation is based on the
compression stress block applying from the where:
compression side of the baseplate back for a fy,p = nominal yield stress of baseplate
distance X. The procedure specified on page 92 φ = 0.9
of [9] is straightforward to apply, but involves one e = cantilever length from edge of
substitution, which is to replace the concrete
compression stress block to centreline
bearing stress 0.6fcu, specified by [9], with fcb from of "compression" flange.
equation 56.12 above.
The thickness of the baseplate on the
For moment-resisting column baseplate compression side for the more general unstiffened
connections to MRF seismic-resisting systems, baseplate case is determined from step 2(a), page
the design moment is high relative to the axial
93 of [9]. That text also mentions the
load. This typically means that the depth of requirements for stiffened baseplates.
compression stress block is relatively small Thicknesses for stiffened baseplates for a range
compared with the total baseplate depth. In this
of boundary conditions can also be determined
case, the compression stress block can be using Formulas for Stress and Strain [13], by
assumed to be centred under the "compression" th
Roark & Young. In the 5 edition [13], table 11 on
flange side of the column, as shown in Fig. 56.6.
page 399 is useful. However, the solutions given
The rotation required from the baseplate to in [13] are all for elastic plate bending;
achieve this is only 1-2 milliradians and readily thicknesses so determined need to be multiplied
achievable, especially given the baseplate
thickness limits on pages 13 and 14 herein. Also 4
by = 0.82 to account for a plastic distribution
the tension forces are relatively high and will 6
typically require four bolts (two rows of two bolts) of stress, on which the provisions of NZS 3404
placed about the column "tension" flange, also as and [9] are based.
shown in Fig. 56.6. This means that the centroid
of action of the tension forces is close to the Baseplate thickness required on the tension
centroid of the column "tension" flange. side
This is given by step 2(b), section 6.9 of [9]; see
Following the same approach as given in step 1, page 94 therein. This equation applies for either
section 6.9 of [9] allows the depth, X, to be two bolts positioned outside the "tension" flange,
calculated directly as: as shown in Fig. 6.8 of [9], or four bolts placed
symmetrically about the "tension" flange, as
X =
1
f b (d − t )
∗
( )
M + 0.5N ∗ (d c − t f c )
shown in Fig. 56.6 herein.
mc = 0.5f cb e
φ
T ∗ = oms
(
C − N ∗ ) (56.17)
φ
2
(56.15)
C = f cb bp X (56.18)
HERA Steel Design & Construction Bulletin Page 17 No. 56, June 2000
where: A design friction coefficient of 0.3 is
fcb = concrete/grout design bearing stress recommended by [9], thus:
from equation 56.12
φVcon = 0 .3C (56.20)
X = depth of compression stress block,
either from equation 6.6 of [9] or where:
equation 56.13 herein C = internal compression force generated
N* = design axial force on column (positive within the connection; see equation
if compression) 56.18
φ oms
= allowance for overstrength moment
φ The shear resistance available from friction is
∗ ∗
action from column, where this is adequate if φVcon ≥ Vcon , where Vcon is given by
required from DCB Issue No. 50 (eg. DCB Issue No. 50.
for MRF columns from section 4.2.1)
C = internal compression force In practice, the shear capacity available through
friction will be adequate for most MRF column
This tension force is resisted evenly by the bolts. baseplate connections; it is only for EBF or CBF
connections with high tension uplift where further
Design solutions for the hold-down bolts are given mechanical means such as a shear key will be
in step 3, sections 6.9 of [9]. It is recommended required. The maximum baseplate thickness of
that anchor plates be used, with the tension force 32mm recommended on page 14 for these will
in the threaded rod which forms the hold-down limit use of this connection for EBF or CBF system
bolt transferred into the plate via a nut complying column base connections.
with AS/NZS 1252 [14].
Check welds between baseplate and column
Recommended lengths of hold-down threaded rod shaft
(bolt) and anchor plate thickness are given in The design actions for these have already been
Table 6.3 of [9]; for M36 use 750 mm length and given and suitable weld sizes can be obtained
30 mm thick anchor plates. A combined anchor from the MEP-8 tables of [1] for the appropriate
plate incorporating the four bolt group will be column size.
required for the detail shown in Fig. 56.6, using
the concept given on page 96 of [9] to size the These weld design actions are more severe than
plate. the provisions of step 5, section 6.9 of [9],
because of the seismic actions imposed on the
The hold-down bolts/anchor plate must be connection and adjacent column.
contained within the longitudinal and transverse
reinforcement of the supporting concrete member. Detailing and Constructability Issues
In addition to the anchorage plate resisting
bearing, the concrete should be checked for General
punching shear failure. There are no appropriate This article ends with some guidance/discussion
provisions in NZS 3101 [10], so the provisions on of important detailing and constructability issues.
page 96 of [9] should be used, in accordance with
the following details: Hold-down bolt sizes, layouts
Having determined the number, size and position
(1) The equation for design concrete shear of hold-down bolts required in the connection
stress, vc , incorporates a partial strength design, the bolt holes in the baseplate must
1 comply with NZS 3404 Clause 14.3.5.2.2. This
reduction factor of
1.25 allows the hole diameter to be up to 6 mm greater
(2) The equation for vc in [9] uses concrete cube than the bolt (threaded rod) diameter. Note the
strength, fcu. New Zealand practice uses requirement for a plate washer at least 4 mm thick
which has minimum dimensions (diameter or side
cylinder strength, fc′ , which must be
length = hole diameter plus 8 mm) when the hole
converted to cube strength by: diameter is more than 3 mm and up to 6 mm
fcu = 1.25 fc′ (56.19) larger than the bolt diameter.
(3) All other terms are applied as stated in [9]. The tolerances on anchor bolt location given in
NZS 3404 Fig. 15.3.1 are written for compatibility
Check adequate shear transfer to concrete with the 6 mm diameter oversize baseplate holes
This follows step 4, section 6.9 of [9]; page 97 and should be specified in the contract
therein. documents.
HERA Steel Design & Construction Bulletin Page 18 No. 56, June 2000
Prior to the fabricator drilling the baseplate holes, sleeves (necessary to prevent corrosion) and for
the contractor (or whomever is responsible for thoroughly filling the space under the baseplate.
placing the hold-down bolts into the concrete) It also makes a reasonable allowance for levelling
should obtain a surveyor's report that reports tolerances.
either the hold-down bolts are within the specified
tolerances or notes their actual locations. The A non-shrink, pourable grout with specified 28 day
fabricator should then fabricate to these positions. compression strength, fm′ ≥ f c′ , should be used.
This is covered by Clause 4.5.1 and the
specification checklist item 4.2 of the HERA
In baseplates of size 700 mm x 700 mm or larger,
Specification, Report R4-99 [16].
50 mm diameter holes should be provided to allow
trapped air to escape and also for inspection [9].
Various means of building in tolerance 2
A hole should be provided for each 0.5 m of base
adjustments within the cast-in hold-down bolts are
area. If it is intended to place grout through these
available, but are not covered further herein. The
holes the diameter should be increased to
use of templates to position the heads of the bolts,
100 mm. The hole should be located near the
in conjunction with the anchor plate which also
centre of the baseplate; ie. away from regions of
locates the base, plus accurate positioning of the
high design tension or design compression.
bolt group by survey should mean that the
NZS 3404 tolerances will be sufficient to allow the
Tightening of bolts once grout/concrete has
bolts to be cast directly into the concrete.
cured
It is important that the moment-resisting column
Anchor bolt lengths should be made say 15mm
baseplate connections in a seismic-resisting
longer than the minimum required to allow for the
system are as rigid as practicable, as the rigidity
bolts being set too low. Similarly with the
of this connection has a significant effect on the
threaded lengths at the top and bottom of the
behaviour of the seismic-resisting system(s) and
bolts.
the overall building.
Debonding of concrete over top length of bolts
Tightening of the hold-down bolts to achieve a
As described below under Tightening of bolts
minimum level of bolt pre-tension eliminates a
once grout/concrete has cured, the hold-down
potential source of undesirable connection
bolts should be tightened to achieve an installed
flexibility.
bolt tension of around half the yield strength in the
bolt. This tightening will involve some extension
Over-tightening, however, adds extra cost for no
of the bolts relative to the concrete and it is
long-term benefit, as the concrete will creep under
important that the bolts are debonded from the
sustained long-term compression load from pre-
concrete and grout for a sufficient distance into
tensioning, reducing this to a lower value over
the supporting member to allow the extension to
time.
occur.
It is therefore recommended to base the
It is recommended that the top half length of bolt
tightening of the hold-down bolts on achieving an
be debonded. Debonding can be achieved by use
initial bolt pre-tension of approximately half the
of duct tape, plastic wrap or similar.
bolt yield strength.
Levelling of column baseplate
One option for achieving this force is to use direct
Because the hold-down bolts are tightened as
tensioning equipment, similar to that used for
described opposite, it is very important not to
post-tensioning steel bars. However, a more
place levelling nuts on these bolts under the
easily applied option is to use the part turn
baseplate, but to instead use shims or similar to
concept of NZS 3404, with the bolt extension
achieve the required tolerances from Clauses
associated with achieving an (elastic) extension of
15.3.2.2 and 15.3.3 of NZS 3404 [2] for level and
0.5 x yield force (0.5 As fyf ) calculated and applied
plumb of the baseplate and attached column
as a specified turn of nut past snug tight.
respectively. Use of levelling nuts on these bolts
under the plate would prevent the compression
This is achieved as follows:
induced on the baseplate by bolt tensioning being
transferred to the concrete, as is intended. It
Step 1 Calculate the required bolt extension, as:
would also cause the bolts to be loaded into the
fyf
plastic range under the calculated rotation from ∆ 0.5y = 0.5 L (56.21)
equation 56.22, which is not intended, nor Ef
desirable, nor readily achievable.
where:
Grouting between baseplate and concrete
fyf = yield stress = 795 MPa for AISI 4140 bar
A bedding space of at least 50 mm is normal.
Ef = 205,000 MPa
This gives reasonable access for grouting the bolt
HERA Steel Design & Construction Bulletin Page 19 No. 56, June 2000
L = length of hold-down bolt, measured from this design moment is exceeded. The joint is then
top of anchor plate to free end of bolt designed and detailed to withstand the expected
inelastic rotation associated with the design level
Step 2 Calculate the required rotation past snug ultimate limit state earthquake with negligible
tight in degrees, as: damage, such that minimum or no repair is
necessary when the MRSF is subjected to that
∆ 0.5y magnitude of earthquake. Finally, the joint is
θ 0.5y = x360 o (56.22) expected to be able to withstand greater levels of
ρ
inelastic rotation, associated with more severe
events, without catastrophic failure, instead
where: undergoing at worst a gradual loss of capacity
ρ = specified pitch for the hold-down bolt with increasing cyclic rotation demand beyond the
diameter, from AS/NZS 1252 Table 2.1. design severe seismic level.
Step 3 Round the required rotation to the Of the five joint types that have been researched
nearest proportion of turn consistent to date, two details have emerged as preferred
with the values used in NZS 3404 Table options for the beam to column connections of
15.2.5.2, ie: MRSF. These are the Standard Flange Bolted
Joint (SFBJ) and the Sliding Hinge Joint (SHJ).
1 turn = 90º
4
This article briefly outlines the current status of the
1 turn = 120º SFBJ, which is the simpler of the two joints to
3
fabricate and erect. Fig. 56.8 on page 22 shows
1 turn = 180º
2 one of the large-scale SFBJ test specimens at the
2 start of testing. This joint is intended for low
turn = 240º
3 ductility demand applications, with µ design = 2 for
3 turn = 270º MRSF systems using this joint. It delivers high
4
moment capacity, possesses low design inelastic
Step 4 Specify tightening to the part turn rotation capacity (ie. up to the point beyond which
method of NZS 3404 Clause 15.2.5.2, repair is required) but can sustain significant
additional inelastic rotation demand while
using the amount of turn specified from
Step 3, above instead of that from maintaining its shear and moment integrity and
NZS 3404 Table 15.2.5.2. being repairable.
This is to be applied once the grout and This article focuses on the results from the small-
concrete have reached their specified scale SFBJ component testing using the M30
HSFG bolts, which was undertaken following the
curing period (eg. 28 days).
last research update, published in November
Step 5 Inspect that this tightening has been 1999 in DCB Issue No. 52. It then highlights the
relatively minor changes to the design and
applied.
detailing requirements that were presented in that
DCB Issue No. 52 article and have been
HERA's Semi-Rigid Standard incorporated into a spreadsheet program on the
Flange Bolted Joint Research; joint design.
Mid-2000 Update on Research This is followed by coverage of two
Programme and Design constructability and detailing issues.
Provisions The article ends with a brief review of the current
(This article has been written by G Charles Clifton, HERA
Structural Engineer, who is responsible for this research (mid-2000) status of research into the other joint,
programme). the Sliding Hinge Joint (SHJ).
General Introduction and Scope of Article Overview of Results from M30 Bolts
HERA and the University of Auckland are Component Testing
engaged in a long-term research project aimed at As mentioned on page 17 of DCB Issue No. 52,
developing innovative new forms of semi-rigid the aim of the component tests has been to
joints for moment-resisting steel framed seismic- investigate the performance of the most critical
resisting systems (MRSFs). These joints are joint component, namely the beam flange to
intended to remain rigid up to the design level flange cleat connection, under a range of bolt
ultimate limit state earthquake moment, eg. as sizes, cleat thicknesses and strengths. The initial
derived from NZS 4203 [17], then to allow rotation results from the M24 bolt tests were used in
to occur between the beam and the column, when writing the article presented in that issue.
HERA Steel Design & Construction Bulletin Page 20 No. 56, June 2000
During the first half of 2000, small-scale joint capacity observed at the typical inelastic
component tests were undertaken on the rotation demand level of loading expected in
remaining cleats using M24 bolts and on all the a design level severe earthquake (see Fig.
cleats using M30 bolts. Fifteen tests have been 56.11). Furthermore, negligible bolt damage
undertaken, covering the following range of occurs at the maximum level of inelastic
parameters: rotation demand expected for a design level
event (see Fig. 56.12). Cleat fracture would
§ bolt sizes M24 and M30 not then be expected, in practice, until
§ ratio of (design shear capacity of bolt group to around 1.5 x this level of rotation demand,
cleat axial strength) varying from 0.9 to 1.25 with the joint then maintaining its moment
§ ratio of cleat thickness to bolt diameter capacity under extreme levels of rotation
varying from 0.5 to 0.83 demand via the horizontal rows of web bolts.
§ difference between M24 8.8 bolts and M24
4.6 bolts (one test) (5) As expected, in one of the tests involving a
§ effect of repeated loading event on bolts and cleat strength to bolt strength ratio > 1.0,
cleat failure was through bolt shear fracture under
§ effect of repeated loading event, after 4 week an imposed rotation of 1.25 x maximum
delay, on cleat strength, ductility capacity. design rotation. There was only minor
elongation of the bolt holes, so the
Fig. 56.9 shows the test set-up for Test No. 1.8.1, opportunity was taken to repeat the test using
which involved two M30 bolts, 20 mm thick cleat, M24 4.6/S mode bolts.
ratio of (design shear capacity of bolt group to
cleat strength) of 1.25 and one test through the As expected, in the repeat test, failure was
full loading regime. The ratio of bolt to cleat through fracture of the bolts, at a peak
strength was such that cleat fracture was the strength 0.52 x that developed by the
expected failure mode, with minimal bolt damage. property class 8.8 bolts. However, the very
The test results confirmed this. Fig. 56.10 shows interesting point was that the ductility
the force-rotation curves for this test from all capacity (shear displacement prior to
cycles of loading. Fig. 56.11 shows the force- fracture) was the same for both property
rotation curves for the three cycles to the typical classes. This is consistent with informal
inelastic rotation demand expected under a reports from Japanese tests, which show
design level [17] severe earthquake, while similar ductility from both bolt grades is
Fig. 56.12 shows the curves for loading to the achieved under seismic-dynamic conditions,
maximum inelastic rotation demand expected although the strengths obtained are different.
under a design level severe earthquake. (The
force-rotation curves measure the following: force Final Design Requirements for the SFBJ
through the cleat versus rotation about the It was noted, in presenting the design
centroid of the bearing pin shown in Fig. 56.9. requirements in DCB Issue No. 52, that the
The rotation is the lateral movement of the cleat provisions presented therein would be evaluated,
relative to the beam stub divided by the vertical once the component testing programme was
distance between the cleat/beam stub interface completed, and any necessary changes made.
and the centroid of the bearing pin).
While the remaining four component tests
The results of these tests, all of which were involving M16 bolts have yet to be undertaken, the
undertaken at seismic-dynamic rates of loading, consistency of results between the M24 and M30
have shown the following: bolt sizes has allowed preliminary details within
the design procedure, as presented in DCB Issue
(1) Using a ratio of φN cleat ≤ 0.8ΣφVf ensures No. 52, to be finalised in the expectation that the
that cleat fracture will occur prior to general recommendations will also apply to the M16 bolts.
bolt shear yielding and loss of installed bolt
tension, for all ratios of cleat thickness to bolt Also, as had been hoped in writing that article,
diameter tested. there are few changes required to the design
procedure as presented in Issue No. 52. For this
(2) Provided that (1) is met, the ratio of cleat reason, the changes are presented with reference
thickness to bolt diameter is not a significant to that article, using the same set of headings.
variable, at least up to the maximum tested
value of 0.83. Details are as follows:
(3) Within the limits of (1) and (2), bolt size is not 1. Changes to detailing requirements for
a significant variable. connection components
(4) Within the limits of (1) and (2), there is no No change from DCB Issue No. 52.
visual sign of damage nor deterioration in
HERA Steel Design & Construction Bulletin Page 21 No. 56, June 2000
Fig. 56.8 Fig. 56.9
SFBJ Large-Scale Test Specimen Showing Flange Cleat for SFBJ With M30 Bolts in Test Rig
Joint Components Prior to Loading
Fig. 56.9
Fig. 56.10
Force-Rotation Curves for Test 1.8.1, All Cycles of Loading
HERA Steel Design & Construction Bulletin Page 22 No. 56, June 2000
2. Calculate the element design action 4. Determine the bottom flange bolts and
reduction factor, φ ∗r bottom flange cleat details required
where:
where:
ΣφVfn,wb = n wb φV fn,wb
φVf,tfb = design single shear
capacity of bolt, typically with threads
included (φVf n), from NZS 3404 Clause This eliminates the use of φVbhe from the SFBJ
9.3.2.1, or as listed in [8]. procedure; the latest series of component tests
have shown use of φVbhe as originally proposed is
The 1.3 comes from 1/0.8, which is the new not justified and adds extra work for no
limit given in step 3.8, rounded up slightly to improvement in performance.
account for the cleat thickness selected (step
3.5) typically being greater than that required 6. Calculate sizes of welds required between
(step 3.4). column flange and cleats.
The limit on plate thickness associated with a As for DCB Issue No. 52
given bolt size is now significantly relaxed
7. Calculate tension/compression stiffener
over that previously specified; details are
requirements
given in step 3.6 herein.
As for DCB Issue No. 52.
3.3-3.5 As for DCB Issue No. 52, with step 3.2
above used. 8. Adequacy of joint panel zone
3.6 Check that the plate thickness used ≤ 0.85 df , As for DCB Issue No. 52.
where df = bolt diameter. The limiting plate
thicknesses are now: Spreadsheet Program Available
The spreadsheet program developed for the ten
§ 16 mm for M20 bolts storey building SFBJ design that is mentioned on
§ 20 mm for M24 bolts page 20 of DCB Issue No. 52 has been modified
§ 25 mm for M30 bolts to incorporate the changes detailed above. It is
§ 30 mm for M36 bolts produced on Microsoft Excel for Office 97 and a
copy is available free-of-charge on a "use at your
3.7 As for DCB Issue No. 52. own risk" basis.
3.8 Check that the ratio of bolt group shear Those wanting a copy, which will be sent via e-
capacity to cleat axial capacity given by mail, contact Charles Clifton at:
equation 56.24 is met.
structural@hera.org.nz
φN t , t f p and φN c , t f p ≤ 0.8ΣφVf , t f b (56.24)
One of the benefits of this spreadsheet is to show
where: the typical bolt sizes and numbers that would be
ΣφVf , t f b = n tfb φVf , t f b applicable.
n t f b = no. of top flange bolts from section 3.2 Allowance for Manufacturing Tolerances in
the Supported Beams and Inclusion of
If this is not initially satisfied, add an extra Decking Support Shim
pair of top flange bolts and recheck. As shown in Fig. 56.8, the SFBJ involves bolted
connections to each beam flange, via a flange
HERA Steel Design & Construction Bulletin Page 23 No. 56, June 2000
cleat, and bolted connections to the beam web, side(s) of the beam support(s) steel decking and
via a web cleat. The clear distance between the provides a support to the decking during
(inside faces of the) flange cleats must be construction. An example is shown in Fig. 56.8; it
sufficient to accommodate manufacturing is also detailed in Item 35 of HERA Report R4-58
tolerances in the supported beams, plus the [5]; see especially Item 35c in the context of this
inclusion of a decking support shim, where used. article. It is an extra component to consider in
fabrication and erection but one which greatly
The offset required to accommodate rolling facilitates laying of the decking around the
tolerances can be determined from the material connection. Note also the 3 mm plate extensions
supply standards; tolerances given in welded onto the top tension/compression
AS/NZS 3679 [18] are representative of general stiffeners in Fig. 56.8 for the same purpose.
good practice and are used in formulating the
recommendations below. Tightening of Large Diameter HSFG Bolts
The SFBJ connections will routinely require the
The tolerances must include allowance for use of fully tensioned M30 high strength structural
variation in height and flange out-of-square. bolts and occasionally the use of fully tensioned
According to [18], these can be as much as: M36 HSFG bolts. They share this need with
bolted moment end-plate connections, which, for
§ + 6 mm on specified depth for hot-rolled larger beam sizes, require [1] use of up to M36
sections up to 610 UB bolts. Two such examples of MEP connections
§ + 8 mm on specified depth for welded are presented in DCB Issue No. 51.
sections above 610 UB and up to 1200 WB.
Achieving full tensioning of these sizes is beyond
The recommendations for manufacturing the scope of a standard impact wrench. Instead,
tolerance allowance in the SFBJ are as follows: a suitable torque wrench for this was described on
page 21 of DCB Issue No. 51.
(1) The allowances are provided as an offset of
each flange cleat away from the specified As part of the HERA/University of Auckland semi-
centreline position of the beam. rigid test programme, the two organisations jointly
purchased a smaller version of the torque wrench
(2) The up offset for the top flange cleat is as mentioned in that Issue. The Alkitronic E/A PG
follows: 280/A analogue torque wrench was chosen, being
able to fully tension up to M30 bolts, in theory,
§ 3 mm for beam depths up to 610 mm with only the as-supplied bolt lubricant required by
§ 4 mm for beam depths above 610 mm Clause 3.2.5 of AS/NZS 1252 [12]. At the time of
§ 3mm is added to all the above to writing DCB Issue No. 51 it had been used to fully
accommodate a decking support shim, tension up to M24 HSFG bolts quickly and easily.
where used.
We are now able to report that it handles full
(3) The down offset for the bottom flange cleat is tensioning of M30 HSFG bolts with equal ease.
as follows: Furthermore, provided the bolts are relubricated, it
§ 1 mm for beam depths up to 310 mm has also easily retensioned M30 bolts which have
§ 2 mm for beam depths from 311 mm to been through two previous tension cycles to the
610 mm part-turn method. Retensioning of HSFG bolts
§ 3 mm for beam depths above 610 mm. takes them further along the stress-strain curve
with higher bolt tensions generated for the
In practice these will lead to a gap existing specified rotation, making them harder to tighten.
between the beam flange and cleat in most
instances; this gap is readily closed by the bolt Further details on this or other appropriate
tightening. wrenches are available from a specialist supplier,
such as:
The smaller offset on the bottom is to facilitate
positioning of the web bolts when the beam is Hydraulic Tool Hire Ltd
made exactly to or slightly below the specified Phone: 0-9-274 0121
depth. Fax: 0-9-274 5192
E-mail: hytc@voyager.co.nz
The web cleat must be offset from the column
flange centreline by an amount equal to half the Very Brief Review of Sliding Hinge Joint
beam web thickness plus 0.5 mm. Research
Finally, in this article, a very brief review of the
Finally, note mention of the decking support shim. status of research into the second semi-rigid joint,
This is made from 3 mm thick plate, extends namely the Sliding Hinge Joint (SHJ).
50 mm beyond the top flange cleat on whichever
HERA Steel Design & Construction Bulletin Page 24 No. 56, June 2000
The concept behind this joint is described in DCB steel framed car parking buildings that were either
Issue No. 52; it is designed to deliver high built, using the membrane option, at the time or
ductility capability with minimal joint degradation have been built since then, using either option. A
and need for repair. brief update on experiences with both is now
given.
As reported in DCB Issue No. 52, component
tests on the beam flange to flange cleat sliding Update
component, using the M24 bolts and conducted at There are a number of waterproofing membranes
both pseudo-static and seismic-dynamic rates of which are resistant to vehicular traffic. An
loading, have exhibited excellent performance. example of one (the Vulken membrane) is given in
Stable sliding, without joint deterioration, was DCB Issue No. 54, in an example of an innovative
achieved for inelastic rotation demands of over car park solution written by Graeme Stubbing of
± 50 milliradians. Speedfloor New Zealand Ltd.
The first full-scale tests were conducted in The concrete waterproofing admixture described
January/February 2000 on a joint, including floor in the DCB Issue No. 50 article has also been
slab, of the same size as the SFBJ shown in used in a number of car parking buildings, either
Fig. 56.8. on its own or with another system, to prevent
leakage. Examples are in the Pineridge
The results showed that the full-sized joint should Condominiums and Sturdee Street Caparks, both
be able to deliver the potential indicated by the in Auckland, plus it is being used in two very large
component tests, provided that three issues are shopping centres currently under development in
addressed. These relate to accommodation of Auckland.
construction tolerances, clearances between
beam and column and providing effective pinning Readers interested in more details on the
capacity at the top flange/flange cleat level. performance of these systems can follow up with
either Graeme Stubbing in the former case or
Solutions to these issues have been proposed, a
Demden Ltd in the latter case. Contact details are
second full-sized test specimen designed,
given in the DCB Issue Nos. 50, 54.
fabricated, installed and it is currently being
prepared for testing in early August. Preliminary
The effectiveness of the concrete waterproofing
results from these tests will be published in the
admixture is dependent on maximum crack width
next issue of the DCB.
developed in the slab; the supplier guarantees it
to prevent leaks for cracks up to 0.5 mm in width
Waterproofing of Car Park and states that it is likely to effectively seal cracks
Floor Slabs: Update against leakage that are up to 1.0 mm in width.
Limiting the maximum crack width to 0.5 mm
requires a level of reinforcement associated with
Background "strong crack control" as described on page 10 of
Pages 7-14 of DCB Issue No. 49, April 1999, DCB Issue No. 50. This increases the typical
present a comprehensive article on the durability minimum reinforcement content considerably. For
of concrete slabs in steel framed car parking example, the 663 mesh for a 90 mm thick slab
buildings, including recommendations to minimise shown in Fig. 49.6 of DCB Issue No. 49 would
or prevent leakage through the slab. need increasing to a 333 mesh to achieve "strong
crack control"; double the area of steel/m width in
That article makes the point that, where leakage the mesh. However, the combination of additional
must be dependably prevented and not just mesh and waterproofing admixture is still
suppressed, a proprietary waterproofing considered to be cost-competitive with a
membrane which is resistant to vehicular traffic is membrane system resistant to vehicular traffic
required. placed over a slab with minimum specified
reinforcement.
That article attracted the attention of a supplier of
an additive to the concrete mix, which it was
claimed, when used in addition to the good
Design of Steel/Concrete Floor
practice requirements from the Issue No. 49 Systems for Satisfactory In-
article, would dependably prevent leaks without
the addition of a membrane, at a cheaper overall
Service Floor Vibration
cost and less construction time. Details were Response
presented in a short article in DCB Issue No. 50,
June 1999. Introduction
Steel/concrete floor systems, comprising a
Although it is only a year since the latter issue concrete slab on steel deck and supported on
was published, there have been quite a number of steel secondary beams and primary beams, have
HERA Steel Design & Construction Bulletin Page 25 No. 56, June 2000
been growing in popularity for several years, response has to be considered on a slab
particularly in car parking and apartment panel basis, involving the interaction of slab
buildings. They are now starting to be used and beams
increasingly often in office buildings. § not only are the critical parameters very
different to other design strength and
Some reasons for this increasing use are their: serviceability limit state criteria, but they differ
§ light weight depending on the proposed use of the floor.
§ ability to span long distances Thus the criteria to control in-service floor
§ ability to accommodate building services vibration response in an office building are
§ flexibility to accommodate irregular shaped different to those for a car park and different
buildings in plan without significant cost or again for a shopping mall.
time penalty
§ speed of construction. Fortunately, excellent design guidance is
available, as introduced in the next part of this
These advantages are being assisted by article.
technological developments in design and
construction; developments which have been Appropriate Design Procedures
highlighted in various DCB Issues.
Scope of application required
Properly designed, these floor systems provide The scope of application of any design
excellent performance. This design must include procedures must cover:
addressing serviceability considerations, the two
most important of which are design to appropriate (1) The range of floor systems used in practice,
deflection limits and design for satisfactory in- including:
service floor vibration response.
§ concrete slab on profiled steel deck,
The former is addressed through NZS 3404 supported on secondary beams which span
Clause 13.1.2.6 and Commentary Clause into primary beams
C13.1.2.6, plus an article on pages 8-13 of DCB § concrete slab on profiled steel deck, propped
Issue No. 32, along with an update on page 9 of during construction and spanning relatively
DCB issue No. 52. long distances onto widely spaced secondary
beams or directly onto primary beams
The latter has been covered, since 1989, by § closely spaced light-weight joist systems
Appendix B13 of HERA Design Guides Volume 2 (such as Speed-Floor), spanning onto primary
[19]; however that procedure has a narrow scope beams
of application and is now supersceded by more § precast floor systems such as double-Ts
recent procedures which are wider in scope and spanning onto primary beams.
give better predictions of the actual in-service
vibration behaviour. (2) The range of purpose group users
encountered in practice, including:
The role of this article is to introduce these
procedures and to give a very brief introduction § offices - open plan, electronic (minimum
into their use. storage) and partitioned
§ car parking floors
Prior to that, and to end this introduction, an § shopping centres
answer to the question; why is consideration of § sports centres
satisfactory in-service floor vibration response so § footbridges
important? The answers are because: § apartments
HERA Steel Design & Construction Bulletin Page 26 No. 56, June 2000
Of these publications, the AISC Design Guide [20] ap 1 1
= function of Po , fn, , (56.26)
β W
and Murray's paper [22] address the bulk of the
scope of applications required from (1) and (2)
g
above. It covers all the purpose group uses in (2)
in regard to concrete slabs supported on where:
secondary and primary beams on light-weight joist
systems. It is written on the basis that the Po = the specified excitation force, which
vibration will be dictated by the response of the varies according to the purpose
supporting beams and their tributary slabs, but group
that the slab itself will not contribute to the
vibration response. fn = the calculated fundamental natural
frequency of the floor panel
Where the slab spans a relatively long distance or β = the modal damping ratio, which
a precast concrete floor system is used, either the depends on the degree of services or
slab or a region involving the slab and supporting partitions installed and has a
beams may be critical. In this instance, the minimum value of 0.01 for a bare
inclusion of the slab effects are covered by [21]. slab floor with no services
W = the effective weight supported by the
The ATC publication [21] also covers floor floor panel
systems utilising wood decking and wood joists. It
includes a chapter on retrofitting of existing floor The effective weight, W, includes a small
systems which either show an in-service vibration proportion of the design live load and the
problem or are required to meet more stringent superimposed dead load.
criteria through a proposed change of use.
Details are given in [20 - 22] on selection of
The design criteria in [21] and [20] are compatible values for these variables. Further guidance will
in philosophy, procedure, principal equations and also be given in a SSAS seminar in September,
application. Both publications contain see details following.
comprehensive design examples. In [20], these
are presented in both SI and USC units, in that Some important points to consider
order. In [21], only USC units are used. (1) Readers who are using the HERA (modified
Murray) method in [19] should now switch to
It is not the aim of this article to present the design the Murray - Allen method of [20 - 22]. They
procedure; readers must consult [20 - 22] for this. need to read the new method carefully, as
However, in order to illustrate the point made in there are major differences between the two.
the introduction to this article about these design
requirements being very different to those for Note in particular the damping values are
other strength and serviceability limit state very different, as those in [20 - 22] are steady
conditions, a very basic introduction to the design state values, while those in [19] are transient
procedure is now given. values. The transient values are higher and
must not be used for the method of
Basic introduction to procedure [18 - 20].
First, a word on terminology. Publications
[20 - 22] are written for USA practice, in which: (2) Note differences in terminology and not ation
between New Zealand and USA practice;
§ secondary beams are termed beams refer to the notation definitions in [20, 21, 2]
§ primary beams are termed girders when in doubt.
§ closely spaced, light-weight trusses of steel or
timber are termed joists. Finally, more aspects of the design for in-service
floor vibration using [20 - 22] will be covered in a
th th
The USA terms are used in the rest of this article. seminar series in the week of September 4 - 8 .
See further details below.
The design procedure involves calculating the
ratio ap/g for a floor panel and comparing this with Forthcoming Seminar Series of
a specified limit, ao/g, for the appropriate purpose
group use. The floor panel is deemed satisfactory Interest
when (ap/g) ≤ (ao/g). This is a notice to readers of a SSAS seminar
series to be held in the week of September
th th
The equation for calculating (ao/g) is equation 4.1 4 -8
of [18]. It involves the following variables.
Venues: Dunedin, Christchurch, Wellington,
Auckland and Taupo, in that order
HERA Steel Design & Construction Bulletin Page 27 No. 56, June 2000
There are two sessions to the seminar. Details
are as follows: 3. Clifton, G C; Structural Steelwork Limit State
Design Guides Volume 1; HERA, Manukau
Session 1: 2pm to 4pm, which will cover: City, 1997, HERA Report R4-80.
As the seminar is being run by HERA, registered 8. Design Capacity Tables for Structural Steel,
attendees will qualify for 2 hours of IPENZ CPD Second Edition, Volume 1: Open Sections
credits for each hour of attendance. (including Addendum No. 1); Australian
Institute of Steel Construction, Sydney,
See the accompanying seminar notice for more Australia, 1994/1997.
details and to register.
9. Joints in Steel Construction, Moment
Tips on Seismic Design of Steel Connections; The Steel Construction
Institute, Ascot, England, 1997 (with
Structures amendments), SCI Publication No: 207/95.
Charles Clifton recently made a presentation to
the Auckland and Christchurch Structural Groups 10. NZS 3101:1995 (incorporating Amendments
on this topic. The presentation, which was 1 & 2; 1997), Concrete Structures Standard;
accompanied with a reasonably detailed set of Standards New Zealand, Wellington.
notes [23], covered the following:
§ general principles for design of seismic- 11. Selected Articles and Document Excerpts on
resisting steel systems and connections Northridge and Hanshin Welded Connection
§ scope of NZS 3404 seismic design provisions Fractures; 1996.
§ scope, concept of R4-76 [6] seismic design
provisions 12. Kobe Earthquake Damage to Steel Moment
§ changes being introduced via Amendment Connections and Suggested Improvement;
No. 1 to NZS 3404 and the revision of R4-76 Japanese Society of Steel Construction,
§ results of research into rigid and semi-rigid Tokyo, Japan, 1997, JSSC Technical Report
connections No. 39.
§ selecting seismic-resisting system form and
connections. 13. AS 1275:1985, Metric Screw Threads for
Fasteners; Standards Australia, Sydney,
A copy of the notes, which comprise nineteen Australia.
double-sided pages and a title page, are available
from HERA; see details on the attached order 14. AS/NZS 1252:1996, High Strength Bolts With
form. Associated Nuts and Washers for Structural
Engineering; Standards New Zealand,
References Wellington.
1. Hyland, C; Structural Steelwork Connections 15. Roark, R J and Young, W C; Formulas for
Guide; HERA, Manukau City, 1999, HERA Stress and Strain (Fifth Edition); McGraw-Hill
Report R4-100. International, Tokyo, Japan, (1975).
2. NZS 3404:1997, Steel Structures Standard; 16. Clifton, G C; HERA Specification for the
Standard New Zealand, Wellington. Fabrication, Erection and Surface Treatment
HERA Steel Design & Construction Bulletin Page 28 No. 56, June 2000
of Structural Steelwork; HERA, Manukau For connections subject to design loads or effects
City, 1998, HERA Report R4-99. incorporating earthquake, NZS 3404 [2] Clause
12.9.1.1.3 requires the load path and strength
17. NZS 4203:1992, General Structural Design hierarchy within the connection to be such as to
and Design Loadings for Buildings; avoid inelastic demand being concentrated into
Standards New Zealand, Wellington. the connectors or connection components. This
means that mode 3 failure must be suppressed by
18. AS/NZS 3679:1996, Structural Steel; means of suitably strong bolts.
Standards New Zealand, Wellington.
The calculation of plastic moment capacity, Mp,
19. Design Guidelines for Control of In-Service uses an equivalent T-stub length, Leff. This is
Floor Vibration in Composite Floor Systems determined from Tables 2.4 to 2.6 of [9]. These
(Appendix B13 of HERA Design Guides cover:
Volume 2); HERA, Manukau City, 1989,
Appendix B13 from HERA Report R4-49. (1) Table 2.4 presents Leff for eleven potential
patterns of yieldline deformation for a bolt
20. Murray, T M et. al.; Floor Vibration due to row acting alone. These cover almost all
Human Activity; American Institute of Steel practical locations of a two-bolt row, with one
Construction, 1997, Steel Design Guide bolt positioned each side of a column or
Series 11. beam web and inside or outside of a beam
flange.
21. Allen, D E et. al.;Minimising Floor Vibration;
Applied Technology Council, Redwood City, (2) Table 2.5 then presents effective T-stub
USA, 1999, ATC Design Guide; 1. lengths from Table 2.4 to consider for bolt
rows acting alone in a range of endplate and
22. Murray, T M; Floor Vibration and the column flange situations.
Electronic Office; Modern Steel
Construction, August 1998, pp. 24-28. (3) Table 2.6 then presents Leff to consider for
bolt rows acting in combination with adjacent
23. Clifton, G C; Tips on Seismic Design of Steel bolt rows.
Structures; Notes from Presentations to
Structural Groups mid-2000; HERA, Examples from these tables are shown in
Manukau City, 2000. Fig. A56.2; these examples comprise only the
table heading and first row in each instance.
Appendix A56: Calculating the
Application of these provisions is made on a row
Tension Capacity of Bolt/Plate by row basis; with the capacity of a group of bolt
Combinations rows determined as the lesser of:
HERA Steel Design & Construction Bulletin Page 29 No. 56, June 2000
bolts used are in [9] equivalent in strength and df = diameter of bolt (termed d in [9])
size to those from AS/NZS 1252 [14]. fuf = tensile strength of bolt (termed Uf in [9])
fy,p = yield strength of endplate or column
In New Zealand practice, the potential resistance flange, as appropriate (termed Pyp for
Pr, calculated as the minimum of modes 1, 2 or 3 endplate in [9] and Pyc for column
(mode 3 for non-seismic applications only), is flange)
expressed as φN m or φN ms , as appropriate.
The modification only needs to be made when
Modification of Bolt Row Force Distribution both endplate and column flange exceed the
The method of calculating bolt/plate tension limiting thickness.
capacity given in steps 1A - 1C, section 2.8 of [9]
is based on a plastic distribution of bolt forces, as For Grade 300 columns, Grade 250 endplates
is the method given in Appendix M of NZS 3404 and Property Class 8.8 bolts, the limiting
[2]. thicknesses are 0.88df for the column flange and
0.96 df for the endplate.
The latter method is restricted in scope to use with
four or more bolts placed as symmetrically as Application of UK Provisions to Connections
practicable about the beam tension flange in two Subject to Inelastic Demand
rows, as shown in Fig. M1 of NZS 3404. Finally, in this appendix, some brief advice on
application of the UK provisions to connections
If there are only four bolts, then the force subject to inelastic demand. These will be,
distribution on each bolt can be considered equal principally, connections in category 1, 2 or 3
for design purposes, as the deformation required seismic-resisting systems connecting primary
to achieve this is well within the capability of the members to secondary members; typically beams
bolts to deliver. If there are more than four bolts to columns.
in the group (ie. more than two bolts per row) then
the effect of differential loading needs to be This advice is as follows:
considered and this is flagged in Clause M2.4.
(1) Mode 3 failure of any bolt row (see Fig. A56.1
The method given in section 2.8 of [9] is based on on page 31 herein) must be suppressed.
two bolts per row, with as many bolt rows as can
be practicably included in the connection. When (2) The tension capacity of the column
multiple bolt rows above or below the tension flange/bolts must be sufficient to resist the
flange are used, then bolt rows adjacent to a overstrength action generated by the
beam flange or tension stiffener will attract a incoming beam. For a beam subject to
higher load than bolt rows further away. For moment alone, this action is given by
combinations of thick endplates and smaller bolts, ∗ φ M
N fbt = oms s
the bolt rows in the stiffest parts of the connection (d b − t f b )
may not have the deformation capacity to allow
them to plastically deform to the extent necessary (3) The same situation applies for the
to develop the full plastic action in the bolt rows endplate/bolts when any row of bolts is
located in the more flexible parts of the governed by mode 2 capacity (see Fig.
connection. In such instances, the resistance A56.1). If mode 1 governs for all bolt rows,
available from the bolt rows in the more flexible then the action can be based on the design
parts of the connection may need to be reduced. section capacity generated by the incoming
beam. There is no simple, general criterion
Full details on how to do this are given in step 1C
that can be used in advance to predict which
of [9].
mode will govern, so, especially for design by
This modification needs to be considered for hand, the overstrength action could be used
combinations of thick endplates/column flanges in all instances.
and relatively small bolts, when more than one
row of bolts either above, or below or one row (4) The beam flange to endplate weld, if it is a
each side of the beam tension flange are used. double-sided fillet weld, must develop the
The limiting equations for endplate or column overstrength tension capacity from the beam
flange thickness are given as equations 2.5 and flange.
2.6 from [9]. The basic equation is as follows:
(5) The beam web to endplate weld, when it is a
double-sided fillet weld, must develop the
df fuf
t limit < (A56.1) design section capacity in tension from the
1.9 f y ,p beam web.
where:
tlimit = limiting thickness of endplate or column
flange, as appropriate
HERA Steel Design & Construction Bulletin Page 30 No. 56, June 2000
Fig. A56.1
Extracts from [9] Showing Modes of Failure and Associated Equations
HERA Steel Design & Construction Bulletin Page 31 No. 56, June 2000
Fig. A56.2
Extracts from [9] of Tables for Calculating the Effective T-Stub Length of Yieldline Development
Note: these extracts comprise the table heading and first row from each table
HERA Steel Design & Construction Bulletin Page 32 No. 56, June 2000