17-19 Gladding Place

P O Box 76 134
Manukau City,
New Zealand

Phone: +64-9-262 2885

Fax: +64-9-262 2856
Email: structural@hera.org.nz

No. 67 April/May 2002

The author(s) of each article in this publication are noted at the The effort and input of all reviewers of the articles herein is
beginning of the article. greatly appreciated.

Introduction The development and details of the Circular Bolted

Flange Annulus (CBFA) design procedure are
given in DCB No. 65, with a design example in
The DCB is published bi-monthly, so a small
DCB No. 66.
change is being made in the date associated with
each issue.
The form of the connection to which this procedure
is applied is shown in Fig. 67.1. The key feature of
The reason for introducing that change at this time
the bolted flange plate used in this connection is
relates to the second article covered herein. That
that it does not extend past the inner face of the
comprises matters arising from the very successful
column wall for other than a short distance,
seminar series on the design and construction of
sufficient for fabrication and erection purposes.
composite steel and concrete floor systems that
This leaves the inside of the column open through
was held at five venues around New Zealand
the connection region.
during the second half of May.
The need for this type of connection arises from
The first article is the part 3 of a series on Circular
applications where this open access is critical, eg.
Bolted Flange Annulus connections. The first two
for passage of services and/or personnel or for
parts of this topic have been presented in DCB
subsequent filling with concrete to form a
Nos. 65 and 66, covering the design procedure
composite concrete filled steel tubular (CFST)
and a design example, respectively. This part
column. Such applications are becoming more
presents details of the Finite Element Analysis
common.
(FEA) verification of the design procedure.
As noted in DCB No. 65, all of the existing design
In This Issue Page procedures for circular bolted flange connections
Design of Circular Bolted Flange 1 from the literature, and available through HERA,
Annulus Connections: Part 3 of 3 require the flange plates to be effectively
continuous across the end of the member. To
Matters Arising From the Seminars 16 cater for the situation where this isn’t the case, the
on Composite Design and CBFA design procedure has been developed by
Construction HERA, through the adaptation of existing design
guidance.
References 21
The principal two publications used in this
development have been the Steel Structures
Design of Circular Bolted Flange Standard, NZS 3404 [1], and SCI Publication
P207/95 [2], Joints in Steel Construction: Moment
Annulus Connections: Part 3 of Connections.
3: FEA Verification Study
As described in section 6 of DCB No. 65
This article has been written by G Charles Clifton, HERA (Commentary to the Design Method) a number of
Structural Engineer and Nandor Mago, HERA Finite Element significant modifications to the procedures of [2]
Analyst.
have been required to accommodate the CBFA
expected modes of failure, along with the
1. Scope and Purpose of Study
establishment of load transfer paths through the
connection for transmitting axial forces (direct or
moment-induced) and shear forces.

HERA Steel Design & Construction Bulletin Page 1 No. 67, April/May 2002
 N* cw M* N* tw

A D

M*
Compression Tension
Quadrant Quadrant
Due to Due to
Moment Moment

B C

Compression side Tension side of

of connection connection

Fig. 67.1
Stiffened Circular Bolted Flange
Annulus Connection Between
Two Steel Sections

Fig. 67.2
Finite Element Model Showing Main Features of the Geometry

HERA Steel Design & Construction Bulletin Page 2 No. 67, April/May 2002
The extent of these modifications have been such point load at the top. This subjects the joint to
as to require verification of the newly developed bending and shear in similar proportion to many
procedure. This verification has been by finite applications in practice.
element analysis. A brief overview of the FEA
study has been given in section 7 of DCB No. 65; As part of FEA work on HERA’s semi-rigid sliding
the article herein and reference [4] presents hinge joint (SHJ) research project, the sliding
details of that study. component of that joint has been analysed in
detail. (See details in DCB No. 64, pp. 24-33).
The scope and purpose of the FEA study has For this work, the property class 8.8 bolts to
been to: AS/NZS 1252 [3] were modelled as solid
elements, with the bolt head/nut bearing surfaces
(1) Determine how a stiffened CBFA assumed to be circular. The details of this model
connection behaves at levels of loading are shown on page 27 of DCB No. 64 and this
from 0.70 to 1.39 times the design capacity accurately represents what happens in practice.
of the CBFA connection and compare this
with the behaviour assumed in the design However, that approach is computationally and
procedure. numerically demanding and the application of this,
even on the half model CBFA joint, comprising of
(2) Determine how an unstiffened CBFA 16 M30 bolts, has been beyond the available
behaves under levels of loading up to 2.29 hardware resources. (Currently, HERA is in the
times the design capacity, and compare this phase of a significant hardware upgrade to a
with the behaviour assumed in the design HPX4000 Workstation, which is going to give
procedure. much greater capability to tackle these types of
problems). The solid element bolt modelling was
(3) Determine the influence on connection also unnecessary in this study, as sliding of the
behaviour if the high strength structural two flange plates across the interface will not
bolts are snug tightened (/S mode to occur in practice and so does not need to be
NZS 3404 [1] Table 9.3.1) instead of fully included in the model. For those reasons, the
tensioned (/TB mode). bolts in this study were represented by one
dimensional beam elements, which have top and
Fig. 67.2 shows the joint being modelled in the bottom ends (ie. nodes) rigidly linked to the upper
study. The joint has been modelled as occurring and lower bearing surfaces of the flanges,
1 metre above ground on a 5.96 metre high respectively.
cantilever tower, subject to design bending
moment and shear force generated by an applied

Fig. 67.3
Modelling of the Actual Bolt to Plates Interfaces With Solid Modelled Fasteners

Notes: The bearing surface radius is R = 23.35 mm on the flanges, while the bolt head/nut radius is 25 mm. The slight difference is to
provide a smooth formulation between contact pairs on the contact surfaces. Bolt hole diameter = 33 mm.

HERA Steel Design & Construction Bulletin Page 3 No. 67, April/May 2002
 Fig. 67.4
Surfaces Where the Beam (Left) and Solid (Right) Modelled Bolt
Force is Transferred into the Flange

Fig. 67.5
Loading Cases for the Force Transfer Studies With the Beam Element Modelled Fastener

Notes:
1. On the left are the support conditions and representation of bolt pretensioning
2. On the right is the location of the subsequently enforced displacements

HERA Steel Design & Construction Bulletin Page 4 No. 67, April/May 2002
The modelling of that bolt/plate connection the installed bolt tension to be lost. Furthermore,
required careful consideration, in order to under subsequent design tension force across the
accurately represent the real condition. Therefore, bolt line, the plates will undergo further premature
a fundamental correlation study has been yielding, leading to local and therefore global
performed, with the aim being to identify the load-deflection behaviour being inaccurately
correct FE modelling approach for the beam represented.
element modelled fastener case. The goal has
been to obtain approximately identical bolt forces Fig. 67.5 shows the connection model used in the
(as for the solid modelled bolt case) at the various basic bolt study. This model was used to
loading levels. determine the influence of bolt pre-tensioning (left)
and subsequent application of a similar lead
The results of this basic one bolt study are pattern as in the CBFA. To determine this, the
presented first, in section 2. This is followed, in following non-linear steps were applied to the
section 3, by the details of the complex joint model shown in that figure:
models studied.
Step 1 : The bolt length is adjusted to achieve the
The results of that study are given in section 4, fully tensioned condition (modelled by
followed in section 5 by a comparison of the FEA shortening the bolt within the depth of the
predicted behaviour with that from the design connection by a specified amount
procedure. corresponding to the shortening that will
be generated by applying the part
2. Basic Bolt Study turn method of NZS 3404 [1] Clause
15.2.5.2). The base of the model is fully
2.1 Reasons for study and options covered fixed for all steps.
As outlined in section 1, the size of the model has Step 2: The bolt length is fixed at the final
meant that solid modelling of each bolt and the position at the end of step 1 (Time 1-2).
supporting plate are beyond the scope of our
system’s current capabilities. Instead, the bolt has Step 3: The top of the model is pulled through an
had to be represented by a beam element, with the enforced displacement regime; causing
bolt forces transferred to a number of plate increased tension force across the bolted
elements via rigid links. This is shown in the left joint. The response of the bolt to this
hand side of Fig. 67.4. enforced displacement is monitored
(Time 2 – 3).
In practice, when solid modelling of bolt and plates
are used and the bolt is fully tensioned, the very The behaviour of two bolt/plate models was
high tension forces developed in the bolt are determined in this study. These models were:
transferred to the plate by bearing across the
contact surfaces (Fig. 67.4, right). Provided that (1) The solid model bolt bearing on the plate.
the bolt and plate dimensions are accurately This formed the benchmark case, as the
modelled, the mesh is sufficiently fine and the model was derived from the actual
contact surfaces are correctly established, the dimensions of bolt and plate and the nominal
model will accurately convey what happens in material properties. In this model, the
practice during the tensioning of an initially snug transfer of bolt tension forces into the ply
tight bolt. That process causes the bolt shank to and the ply deformation under these forces
plastically stretch, while the plate bearing area is realistically modelled, with normal mesh
under the bolt head and the nut undergoes a small density.
amount of permanent deformation in compression.
Soft springs were added to the bolt in this
When the bolt is modelled as a beam element model to prevent rigid body motions; the
(beam or truss element), the tension force is influence of these on the result was
transferred into the plate by means of rigid links negligible, as described in [4].
between the nodes at the top and the bottom of the
bolt element and the adjacent nodes in the top (2) The bolt modelled as a beam element.
surface of the top plate and the bottom surface of In this case, the bolt is modelled by a two
the bottom plate. This is shown in the left hand noded beam element in space with axial
side of Fig. 67.4. degrees of freedom (DOF) only. The bolt
translational degrees of freedom at the top
Establishing a sufficient area of plate over which and bottom nodes are made equal to the
the bolt tension force is to be transferred for the plate annulus displacements, through linear
bolt beam element model is very important. Too constraint equations. In this option,
small an area will cause local yielding of the plate determining the appropriate tributary area of
elements under the generated bolt pretension ply that will receive the beam tension force
forces, causing the plate to squash and much of (Fig. 67.4) is critical. If this area is made too

HERA Steel Design & Construction Bulletin Page 5 No. 67, April/May 2002
 Fig. 67. 6
Bolt Force Versus Radius for Beam
Modelled Bolt of Identical Mesh Density
(R = 21.3 mm gives closest agreement with solid model)

Fig. 67. 7
Deformed Plate (Scaled x 20) Due to External Loading
for Beam Modelled (Left) and Solid Modelled (Right)
Bolt Cases

HERA Steel Design & Construction Bulletin Page 6 No. 67, April/May 2002
 Fig. 67. 8
Bolt Force versus Load Steps (LS) for the Beam and Solid Model Bolt Options
LS1 = bolt pretensioned; LS3 = maximum extent of
enforced displacement (plates just under yielding)

small, then the local deformation of the ply Finally, the truss model gave bolt forces within
under the incoming force is too great. Also 0.3% of the beam model for the same tributary
as shown in Fig. 67.4, the tributary area is area (R = 21.3 mm).
that given by π (R – R0)2, where R0 is the
hole diameter and R is the outside line of Further details of this work are given in [4].
bearing. The radii R and R0 are the basis
for mesh generation, which means that 3. Details of Connection Model and
there are solid element nodes on the Options Studied
surfaces of these cylinders.
3.1 General details and material properties
The principal aim of this study was to run
the analyses for increasing R, until the bolt Fig. 67.9 shows a plan view of the half connection
force after pre-tensioning was very close to modelled, while Fig. 67.10 shows a side view.
that for the solid modelled bolt. Advantage was taken of symmetry to model only
2.2 Results of bolt study half the joint, with the axis of symmetry through
the centre. The load is applied at the top, along
As stated in section 2.1 (2), the principal purpose the axis of symmetry.
of the study was to determine the plate tributary
area required (keeping the same mesh density) in The dimensions of the connection elements are
order to obtain a very similar bolt force response given on page 30 of DCB No. 65.
to the solid modelled bolt case. The result,
The bolts were modelled with beam elements,
for R = 21.3 mm, gave an installed bolt tension of
with contact from bolt to plate via linear constraint
396 kN for the former, compared with 402 kN for equations, as described in section 2. Owing to
the latter. hardware limitations, the mesh used in the full
Fig. 67. 7 shows the comparison for R = 21.3 mm, connection model had to be made slightly coarser
where both FE models have the same mesh than that for the sub-assemblage studies shown in
density. The very similar deformed slopes and Fig. 67.3. The effect of this has been to give a 3%
bolt forces obtained throughout the load steps (ie. higher bolt force following pre-tensioning for the
as shown in Fig. 67. 8 justifies the idealisation same radius R; this difference is not significant.
strategy employed to model the effect of the bolt
that holds the flanges together.

HERA Steel Design & Construction Bulletin Page 7 No. 67, April/May 2002
 Fig. 67.9
Plan View of Half Connection Showing Applied Force at Top of Column
and Bolt Designations

Fig. 67.10
Side View of Connection With Contact Pairs (Between Flanges)
and the Position of the Linear Constraint Equations (Bolts) Highlighted
and the Multi-Point Constraints (Rapid Mesh Transition) Also Shown

HERA Steel Design & Construction Bulletin Page 8 No. 67, April/May 2002
The column steel, flange plate steel, stiffeners, • The upper limit of serviceability limit state
and bolts were modelled as tri-linear elastic- conditions
plastic. The material properties used for the yield • Design ultimate limit state conditions
stress and tensile strength were the nominal • Extreme overload conditions
(minimum specified) in each instance; the strain
hardening stress-strain behaviour used was To achieve this, the following loading regime was
derived for the material strain/stress relationship used for each connection option studied:
being linear from (ε y, fy) to (0.75 ε u, fu), then fu kept
constant. In practice, inelastic demand did not Step 1: The bolt length is adjusted to either of
exceed 0.75 ε u at any location. the fully tensioned condition (0.88 mm
shortening) or the snug tight condition
Full tensioning of the bolts was achieved through (0.03 mm shortening), as required.
shortening their length by 0.88 mm.
Step 2: The bolt length is fixed at its position at
Compression bearing between the flange plates the end of step 1.
was modelled with contact surfaces capable of
transmitting compression and shear friction, using Step 3: 70% of the ultimate limit state design
µ = 0.35. The contact surfaces were modelled as capacity of the stiffened connection is
100% in contact prior to full tensioning (as would applied (114% of the unstiffened
be effectively the case in practice when the snug connection capacity).
tightening phase was correctly applied).
Step 4: 92% of the ultimate limit state design
For the connection with snug tightened bolts, the capacity of the stiffened connection is
bolt pre-tension was set to 0.03 mm. This gave a applied (153% of the unstiffened
bolt pre-tension force of 80 kN, which would be connection capacity).
reasonable to achieve through the properly
applied snug-tightening process (see sections Step 5: The load is increased to 116% of the
1.5.4 and 3.3.2 of [5] for details). stiffened connection capacity (191% of
the unstiffened connection capacity).
The upper and lower part of the column and the
16 mm thick stiffeners are built from Step 6: The final increase to 139% of the
ABAQUS/Standard general-purpose four noded ultimate limit state design capacity of the
reduced-integration (S4R) shell elements, while stiffened connection is applied (229% of
the annulus connection is constructed from the unstiffened connection capacity).
8-noded reduced-integration solid elements
(3D8R) [6]. For the stiffened and the unstiffened steel to steel
connection, as designed in section 2 of DCB
3.2 Connection Options No. 66, the applied loads are (on half of the FEM):

Three CBFA connection options have been Step 2: 278 kN

studied, namely; Step 3: 370 kN
Step 4: 463 kN
1. Stiffened connection with fully tensioned Step 5: 555 kN
bolts
These loads represent the following percentages
2. Unstiffened connection with fully tensioned of the design capacity of each connection:
bolts

In this instance, the stiffened connection % Design % Design

geometry was used, however the elastic Capacity of Capacity of
modulus of the stiffener material (E) was Stiffened Unstiffened
reduced by a factor of (1/103), thus making Connection Connection
the stiffener contribution negligible
Step 2 70 114
3. Stiffened connection with snug light bolts.
Step 3 92 153
3.3 Loading regime
Step 4 116 191
The aim of this FEA study was to determine the
Step 5 139 229
behaviour of the connection at load levels
corresponding to:

HERA Steel Design & Construction Bulletin Page 9 No. 67, April/May 2002
 Fig. 67.11
Magnified Deformed Shape (By Factor of 20) of Connection at the Tension Side for the Unstiffened (Left)
and the Stiffened (Right) Connection
Left hand side: load on unstiffened connection (229% of design capacity)
Right hand side: load on stiffened connection (139% of design capacity)

Fig. 67.12
Bolt Forces in the
Stiffened and Unstiffened Connection
Notes:
(1) See Fig. 67.9 for location of the bolts: bolt 1 is on the compression side, bolt 8 in the shear quadrant and bolt 16 on the tension
side.

(2) Time 1 ≡ pre-tensioning completed

Time 2 ≡ applied loading commences
Time 3 ≡ 70%/114% of design load applied (stiffened/unstiffened)
Time 4 ≡ 92%/153% of design load applied
Time 5 ≡ 116%/191% of design load applied
Time 6 ≡ 139%/229% of design load applied

HERA Steel Design & Construction Bulletin Page 10 No. 67, April/May 2002
 Fig. 67.13
History of the Bolt Forces (Bolt 1, 8 and 16) in the Snug Tight Modes and the Full Tensioned Modes

Fig. 67.14
Force Versus Deflection at the Top of the Column
for the Three Connection Options Analysed

Notes:
1. For the unstiffened connection, 75% design load = 244 kN
100% design load = 325 kN

2. For the stiffened connection, 75% design load = 302 kN

100% design load = 403 kN

HERA Steel Design & Construction Bulletin Page 11 No. 67, April/May 2002
 Fig. 67.15
Magnified Deformed Shape at 70% Design Capacity on the Tensional Side of
the Stiffened Connection: Snug Tightened Bolts (Right) and Fully Tensioned Bolts (Left)

Fig. 67.16
Minor Plasticity Developed Around Bolt Holes
at 70% Design Capacity for Stiffened Connection,
Tensioned Bolts, Magnified Deformed Shape (20x)

HERA Steel Design & Construction Bulletin Page 12 No. 67, April/May 2002
 Fig. 67.17
Local Plasticity Developed Around Bolts, Stiffened Connection, 92% of Design Capacity,
Magnified Deformed Shape (20x)

Fig. 67.18
Plasticity Developed Around the Bolt Holes at 114% Design Capacity
for the Unstiffened Connection,
Fully Tensioned Bolts

HERA Steel Design & Construction Bulletin Page 13 No. 67, April/May 2002
Results of the Analyses exhibit slight inelastic action up to the design
ultimate limit state and show dependable load-
Important results from the analyses are presented deflection behaviour at higher loads.
in Figs. 67.11 to 67.18; more details are in [4].
Fig. 67.14 shows that these criteria are all met for
Fig. 67.11 shows the magnified deformed shape both types of connection. It also shows that the
of the tension side of the connection at 139% of lower design capacity of the unstiffened
the design moment capacity for the stiffened connection is appropriate, given its earlier
connection and at 229% of the design moment departure from linear elastic behaviour.
capacity for the unstiffened connection. (From the
design example, in section 2.3 of DCB No. 66, Fig. 67.14 also shows that, for the stiffened
pp. 14 and 15, the design moment capacity, connection, there is very little difference in the
which is governed by the tension quadrant, is overall column load-deflection behaviour between
3681 x 424/392 = 3981 kNm for the stiffened snug tight and fully tensioned bolts (provided that
connection and 3681 x 141/216 = 2403 kNm for the flange plates are brought fully into contact by
the unstiffened connection). the snug tightening process, which has been
assumed in the analyses and is required by
Fig. 67.12 shows the bolt force versus loading NZS 3404).
history for the bolt on the extreme compression
side (Bolt B1), the bolt on the extreme tension 5.2 Bolt forces for the different connection
side (Bolt B16) and the bolt at the centreline (ie. options
near the neutral axis (Bolt B8)). The influence of
the stiffener on the history of bolt forces is These forces are shown in Fig. 67.12 and
demonstrated when all bolts are fully tensioned. Fig. 67.13. Note that the vertical scale, which
gives the bolt forces, are different in the two
Fig. 67.13 shows the bolt forces for each of these figures, which gives an apparent greater variation
three bolts in the stiffened connection, with the in bolt force with loading history in the first figure.
different bolt modes (snug tight and fully
tensioned). For the compression bolts (bolt 1) that are initially
fully tensioned, the maximum bolt tension force is
Fig. 67.14 shows the load-deflection behaviour at reached during pre-tensioning and this slowly
the top of the column for the three connection reduces under increasing applied moment on the
options studied. This figure also shows the 75% joint, however the reduction is not large. For the
and 100% design load level for the unstiffened fully tensioned bolts, there is negligible difference
and the stiffened connections. in the compression bolt force variation with
loading between the stiffened and unstiffened
Fig. 67.15 shows the magnified deformed shape connection (Fig. 67.12). For the stiffened
at the tension side of the stiffened connection for connection with snug tight bolts, the small amount
the two bolt modes (/S and /TB), at the upper limit of bolt tension induced by the snug tightening is
of serviceability loading. Fig. 67.16 shows the not reduced during subsequent loading of the
local plasticity in the plate developed at that level joint.
of loading.
As expected, the bolt in the shear quadrant (bolt
Fig. 67.17 shows the same features for the 8) exhibits little change in loading following pre-
stiffened connection at 92% of the design load. tensioning. It is interesting to see that the bolt 8 in
Rapid changes in the contour colours indicate the the unstiffened connection does show an increase
necessity for a finer mesh in those regions, to in bolt force over load steps 5.5 to 6 (see
more clearly show the yieldlines. Fig. 67.12). This is due to the tension effects
propagating around the connection as far as the
Finally, Fig. 67.18 shows the plasticity and midpoint of the shear quadrant under this high
deformed shape in the plate around the bolt holes level of applied moment (229% of the unstiffened
of the unstiffened connection at 114% of the joint design moment capacity).
design load for that connection.
The tension bolts (bolt 16) that are pretensioned
5. Comparison of FEA Behaviour With initially undergo a very slight drop in force under
Design Example the applied load. This is due to negative bending
moment (tension on the top face) developing in
5.1 Load-deflection behaviour at the plate at the bolt position. This reduction is a
serviceability and ultimate limit states feature of both the bolt beam model and bolt solid
model, as shown in Fig. 67. 8. The reduction is
The load-deflection behaviour of the column with minor and the bolt force returns to the original pre-
this connection should remain nearly elastic up to tensioned level in the latter stages of loading. The
the maximum serviceability limit state loading, bolt force does not increase above that level for

HERA Steel Design & Construction Bulletin Page 14 No. 67, April/May 2002
the pre-tensioned bolts, which is consistent with The analysis also shows that the yieldline CF
the deformed shape showing no separation of the between the bolt pairs (Fig. 65.13) does not
plates at the bolt position (Fig. 67.16, Fig. 67.17 undergo a constant rotation along its length (see
and Fig. 67.18). This is not fully consistent with Fig. 67.11, right hand side). The behaviour along
the design procedure, as explained in section 5.3. that yieldline is closer to mode 2 than mode 1.
In contrast, the tension bolt that is initially snug Further mesh enhancement would be expected to
tightened shows a significant increase in show this more clearly.
bolt tension force during the applied loading
(Fig. 67.13). The change in slope of that force The magnified deformed shape also shows the
corresponds to the bolt yield in tension being hinging of the column wall, as expected from
reached, at load step 3. Figs. 65.17 or 65.18. Details are in Fig. 67.11,
right hand side; the effect only shows clearly,
5.3 Comparison of modes of behaviour even at 20x magnification, for the highest applied
between design and analysis, stiffened load in each instance.
connection
5.4 Comparison of modes of behaviour
The postulated yieldline pattern for the flange between design and analysis;
yielding between stiffeners in a stiffened flange unstiffened connection
connection is shown in Fig. 65.13 of DCB No. 65. The postulated yieldline pattern is shown in
The magnified deformed shape of the stiffened Fig. 65.14.
connection at 70% of the design load and above,
as shown in Fig. 67.16 and Fig. 67.17, is The actual yieldline pattern at the highest level of
consistent with that shape. loading for the unstiffened connection is shown in
Fig. 67.11, left hand side.
When calculating the stiffened flange
tension capacity, from section 3.3 of DCB From the design example for the unstiffened
No. 65, pp. 21-22, two modes must be checked flange capacity, section 2.3.3, page 15, DCB
for this yieldline shape. The first mode is No. 66, the governing failure mode is mode 1.
associated with complete flange yielding, as This involves no separation at the bolt line and no
shown in Fig. 65.17. The second mode is increase in bolt force due to prying.
associated with bolt elongation and flange
yielding, as shown in Fig. 65.18. The first mode is Fig. 67.12 shows that the tension bolt does not
associated, in theory, with no separation of the experience tension forces greater than the pre-
plates at the bolt location and hence no increase tensioning force, which is consisted with the mode
in bolt force above the pre-tension force due to 1 behaviour.
prying under load. The second mode is
associated with separation and additional prying However, Fig. 67.11, left hand side, shows that
forces. The lesser of the two modes gives the plate separation along the yieldline CD
calculated design tension capacity. (Fig. 65.14) does occur, giving a combination of
modes 1 and 2 action along this yieldline.
For the connection element sizes and
properties used in the design example of DCB 6. Conclusions
No. 66, the calculated design tension capacity,
φNtw, is 424 kN for mode 2 and 601 kN for mode 1 The key results from the modelling are given in
(DCB No. 66, p.14). This means that mode 2 section 7.3 of DCB No. 65. Although they were
governs the calculated design capacity. written from the results of preliminary analyses
only, they are supported by the results of the full
The FEA studies do show a slight separation of analyses, as presented herein and in more detail
the plates at the bolt line, for the fully tensioned in [4].
bolt stiffened connection (See Fig. 67.17).
However, there is no increase in bolt force due to The FEA study has been an essential tool towards
prying. Thus the observed behaviour from the verifying the design procedure presented in DCB
FEA is intermediate between the two modes. The No. 65. The analyses have shown that the design
lack of bolt force increase due to prying procedure is appropriate in terms of the
appears to be due to the plate undergoing performance generated at the serviceability limit
increased localised plasticity around the bolt (see state and at the ultimate limit state. The analyses
Fig. 67.15 and Fig. 67.17) to accommodate the have also shown that the influence of bolt modes
increased deformation. This may be a function of (/S and /TB) on the behaviour of the stiffened
the method of attachment of the bolt and plate, connection is minor, provided that the snug tight
masking an increase in bolt force that would occur bolts have been properly snug tightened.
in practice. However any increase in bolt force
due to prying in practice would be minor in this
instance.

HERA Steel Design & Construction Bulletin Page 15 No. 67, April/May 2002
Matters Arising From the with copies of pertinent papers and DCB
articles.
Seminars on Composite Floor
(2) HERA Report R4-107DD [8] Draft for
System Design and Comment: Guide to Practical Aspects of
Construction Composite Floor System Design and
Construction, Including Concrete
Introduction Placement. This guide [8], which was
released for wider review and comment at
HERA, in conjunction with Dimond and New these seminars, is presented in 3 parts.
Zealand Steel, recently conducted a series of
seminars around New Zealand on the design and The first part, Practical Aspects for
construction of composite steel and concrete floor Designers, provides information on issues
systems. The seminars were very well attended, from deflection and ponding effects of
with over 200 attendees at the five venues. unpropped construction, to prop
construction issues and, through to
During the seminars, various issues arose, some information on economical and practical
relating to the material presented and others precambering.
relating to issues not covered by those notes, but
which should have been covered. This article The second part, Specification for
addresses these matters arising from the Measurement of Surface Finish Tolerances,
seminars, following a general overview of the presents a specification for the
seminars. measurement of concrete slab surface
finish tolerances, including details on how
General Overview to reliably quantify surface flatness and
levelness.
The seminar comprised a very full and intensive
day, covering the full range of topics associated The third part, Guidelines for the Placement
with composite floor system design and and Finishing of Concrete, provides a
construction, from an overview of floor system suggested concrete placement method to
design and construction, detailed coverage of achieve the required surface flatness and
composite design principles and formulae, levelness requirements, including set up of
demonstration of the new HiBond Design Wizard, control points, sequence of concrete
followed by sessions on economical design, placement, etc.
vibration design principles and software, shear
stud design and detailing, spandrel beams, etc. (3) HERA Report R4-112 [9] Report and User’s
The key topics covered were: Manual for NZFl_Vib1, Program for the
Analysis of Floor Vibration. This spread-
• Practical aspects of composite floor system
design and construction sheet based program and comprehensive
user’s manual incorporates the
• Concrete placement guide
USA/Canadian design for vibration
• Overview, principles, and detailed design procedures given in the AISC Design Guide
formulae for composite floor design
Series 11 [10] and ATC-01 [12] into a
• Demonstration of the new HiBond Design comprehensive design tool for floor system
Wizard in-service vibration design. The program
• Deflection design has been customised for all the composite
• Acoustics issues relating to composite steel-concrete floor systems used in New
floors Zealand, ie. Hi♦Bond, TrayDec,
• Control of cracking and leaks Speedfloor, Comflor and for conventional
• Vibration design reinforced concrete flat slabs.
• Economical design and construction
• Shear stud design, testing and other floor (4) HiBond Design Wizard [12]. This CD based
systems program is a comprehensive design tool,
• Spandrel beams covering the full gambit of composite design
• Design for fire with Hi♦Bond composite systems. The
software facilitates rapid design solutions
This material is covered in three publications, and and evaluation of alternatives, thus allowing
one CD, namely: the designer to focus on the all important
higher level design issues.
(1) HERA Report R4-113 [7] Notes Prepared
for a Seminar on Design and Construction
The program covers the complete
of Composite Floor Systems. These notes
composite design needs from slabs,
[7] presented either the detailed notes or a
secondary and primary beams, including
copy of the powerpoint presentations, along

HERA Steel Design & Construction Bulletin Page 16 No. 67, April/May 2002
 unpropped, propped, and pre-cambered With such a comprehensive range of material
design scenarios. Incorporated are presented and topics covered, it was expected
comprehensive strength and serviceability that various issues would arise, during the
checks right through to vibration design, presentation of the seminars, for which follow-up
with user control over full/partial shear coverage was desirable. These are now
connection, acceptable deflection criteria, presented, under the session headings in which
etc. they were covered. Not all sessions gave rise to
follow-up issues, which is why the sessions
Also incorporated in the program is a covered below are not the full range of sessions
powerful preliminary design mode, which presented at the seminars.
allows design, in just a few minutes, of a
complete slab-secondary-primary beam
scenario, with just six data inputs/choices.

Flooding of concrete
over this region

Fig. 67.19
Placement of Concrete by Flooding Over a Slab Panel Region

Session 1.2: Concrete Placement Guide placement operation, following which they would
screed to level over that panel.
The session introduced the draft concrete
placement guide [8] and invited critical review and As shown in Fig. 67.19 (which is based on Fig. 3.7
comment, as well as implementation of the from [8]), this involves flooding the slab panel A1-
recommendations. Quite a few comments were B1-B2-A2, thus depositing and levelling the
received, with the issues covered as follows: concrete over quadrants (1) to (4) in one
operation. This is a faster method of placement
Methods and sequence of concrete placement than that described in section 3.5 of [8], provided
that the concrete placers have sufficient
While some of the contractors present agreed with experience in free screeding to level between the
the method of concrete placement described in reference points. In terms of setting these
section 3.5 of [8], representatives from two major reference points, their placement and allowance
contractors advised that they would not use that for the expected wet concrete deflection, the
method, as it is too slow. Instead they would flood procedure in section 3.4.2 of [8] is still directly
a slab panel region (ie. the region of floor slab applicable.
between adjacent vertical supports), making
allowance for the expected wet concrete However, in setting the area to be flooded with
deflection at specific points during this concrete concrete, this would need to extend beyond the

HERA Steel Design & Construction Bulletin Page 17 No. 67, April/May 2002
edges bounded by the columns, so as to allow for from the seminars, these errors are easily spotted
the deflection of the interior support beams at the and have no technical impact.
panel edges due to the wet concrete loading from
the adjacent panel. This is shown in Fig. 67.19 Horizontal shear splitting is covered in session
and requires different overruns for the interior 4.3.
secondary beams than for the interior primary
beams. It is an approach that requires the The last notes page in session 2.2 of [7] presents
concrete placers to have an understanding of how rules of thumb for preliminary design and
unpropped composite floor systems will deflect evaluation. The rules for vibration assessment
under the wet concrete loading. are given in item 4. The deflection limits of 0.5
mm and 0.7 mm are expected to be satisfactory,
It is intended to include both methods in the final in terms of giving a satisfactory preliminary design
version of [8]; this involves also modifying Table solution that will also pass the final design check
3.2 on page 12. to [9-12], for floor systems with damping ratios of
0.03 or greater and acceleration limits of 0.5%g or
The contractors also advised that the area of greater. (See Table 4 of [9] for these quantities).
concrete per concrete placing gang recommended
from [8] which can be placed in a normal
For floor systems with acceleration limits ≤ 0.5%g
working day is too optimistic. In a normal
and with a damping ratio of 0.025, the preliminary
working day, a concrete placement gang can
design deflection limits may not be low enough.
lay 500 m2 to 600 m2 of concrete. However, on a
When the damping ratio is 0.02, these limits will
hot summer’s day, there is a risk of the delay
definitely be too lenient (ie. too high).
between placing and finishing becoming too long,
resulting in a poorly finished slab. This may
The final point made in this session was reference
require a reduction in the area placed.
to Table 5.1: Table of Equivalent Uniform Design
Testing for flatness: rapid determination of Loads, from the AISC Design Capacity Tables
critical areas to test with the freestanding [13]. This very useful table allows the equivalent
straightedge UDL for strength (moment or shear) or for
serviceability evaluation to be rapidly determined.
Section 2.3.5 of [8] specifies the use of the This will allow software which is based on UDLs,
freestanding straightedge to measure flatness. As such as the HiBond Design Wizard [12], to also be
illustrated in Fig. 2.8 of [8], this is applied over the used for point loading situations.
rows and columns of the grid of measurement
points established for levelness testing, as well as Session 3.1: Deflection
in between those rows and columns.
Note that Fig. 1.4 from [8] has an error in it; the
However, this method requires a lot of positions in second fraction in Fig. 1.4(b) should read 5/9 on
which to place the straightedge and undertake the the denominator, not 5/6.
recordings. During the seminars, advice was
sought on ways to cut down on the number of One of the slides in this session addressed the
testing locations required for flatness. issue of locations requiring particular care in
deflection considerations. It highlighted the case
One suggestion which has merit is to flood the of secondary beams running parallel to adjacent
floor with water, then to identify locations where vertically stiff elements, such as shear walls,
the depth of ponding water exceeds the flatness where a high local slope can develop between the
80% gap limits of section 2.3.5.6 of [8]. Such floor at the midspan of the secondary beam and
locations would then be subjected to the flatness the adjacent point at the shear wall. This situation
testing regime specified in section 2.3.5, while is also covered in section 1.9 (2) of [8]. A solution
other regions would not need specific testing for can be to place another secondary beam of the
flatness. The practicality of this will need to be same size midway between the shear wall and the
trialed; the HERA Structural Engineer is seeking adjacent secondary beam.
feedback on this from the industry.
The same solution can be employed where the
Deflection issues
lateral or vertical deflection of an unpropped
Deflections were covered in several sessions, but secondary edge beam needs to be minimised.
principally in session 3.1. Matters arising This is not a common requirement, but can occur
regarding deflections are all covered there. where a secondary edge beam is located hard
against lift shaft guide rails or similar. Edge
Session 2.2: Design Principles beams can “roll” slightly at midspan as the
concrete is placed, due to the slight torsional
Some minor editorial errors were noted in the eccentricity and pull-in from the decking wet
seminar notes. These have been corrected in the concrete loading. From a strength or a vertical
master document. For those with the copies of [7] deflection point of view, this is not critical.

HERA Steel Design & Construction Bulletin Page 18 No. 67, April/May 2002
However, it may cause the beam bottom flange to Top flanges of steel beams in carpark
more outwards at midspan sufficient to obstruct buildings
services etc. placed very close to the undeformed
edge of the beam. Placement of an additional Also, in this session, the issue of durability of steel
secondary beam between the edge beam and the beams in car parking buildings, which are
adjacent interior beam will eliminate this problem. receiving steel decking and through deck welded
This solution would not be necessary in typical shear studs, was discussed. HERA’s
building secondary edge beam applications, only recommendations in this instance is to leave the
in some special circumstances, such as adjacent top flange unpainted so that the shear studs can
to close tolerance lift guide rails, etc. be successfully welded through deck. DCB Nos.
49 and 20 (especially DCB No. 20, pp. 2-5)
Session 3.3: Control of Cracking and provides the design guidance on assessing the
Suppression of Leaks corrosion allowance to make on the top flange in
this case.
In the slide on performance of existing car park
suspended concrete slabs, it was stated that However, there are instances where this approach
decking does not corrode from the inside. This is is not viable. For example, on edge beams
taken from DCB No. 49 and is based on exposed to the weather or for car parks open to
observation of existing car park decks, plus prevailing wind off the sea and situated close to
consideration of the suppression of corrosion the sea – see note 2 to Table 49.3 of DCB No. 49.
obtained by the concrete in contact with the There are also instances where this approach is
decking. As stated in item (i) on page 9 of DCB not acceptable to the client.
No. 49, this protection will remain in place until the
concrete depassivates at that interface. Given In these instances, if decking is to be placed over
that depassivation has to work down from the top the beams and shear studs welded on-site, then
surface, the typical slab thickness is expected to the decking must be locally holed to accept the
provide 50+ years protection against the slab studs (50 mm square opening), the paint removed
rusting out from the inside. from the beam top flange where the studs are to
be placed, the decking fixed in place and the
However, an example was given in the studs welded directly to the beam. This sounds
Christchurch seminar of a case where the deck like a slow and expensive operation, but in
had started to rust through, from the inside, in only practice it isn’t. One deck layer/shear stud welder
a few years. In that case, there was visible offers this for an extra $1.50/stud, which typically
cracking on the slab top surface, but this was not amounts to no more than an additional $3.00/m2
considered excessive. However, there was also a of floor area. The stud welders undertake the
concrete additive used for rapid setting that would paint removal, rather than the main contractor,
have had the effect of causing rapid which is desirable as they know from experience
depassivation, leading to a severe corrosion the cleanliness of steel surface finish required to
environment being generated at the steel/concrete achieve a reliable stud weld.
interface. Thus the advice given in session 3.3 or
in DCB No. 49 on the protection against decking For more details on this option, contact
top down corrosion afforded by the concrete Ian Welch at Nelson Stud Welding Ltd, at phone
needs to be tempered with the cautionary note 04-233 9421.
that no concrete additives (or aggregates)
containing chlorides are to be used, unless the Discolouration of paint from stud welding
decking is considered as permanent formwork
only. When studs are site welded onto already painted
beams, the shear studs should be, if possible,
The HERA Structural Engineer also advised that placed on the beam centreline, so as to minimise
increasing the slab reinforcement content from the potential effect of discolouration, from the heat
minimum requirements to “strong crack control” of welding, of the paint on the lower surface
requirements can mean increasing reinforcement underneath where the shear studs are located.
areas (mm2/m width) by up to 4 times, with some The susceptibility of paint systems to this heat-
associated significant cost increase implications. induced discolouration depends on the paint
However, subsequent discussions with system used, with inorganic zinc silicates being
contractors shows that the cost increase is not not susceptible and acrylic or epoxy systems the
great, being in the order of $6/m2 floor area. This most susceptible. If the paint surface is
is typically cheaper than controlling cracking discoloured but not blistered, then its corrosion
through reducing the shrinkage by means of performance is not adversely effected. Thus it is
concrete additives. more of a visual issue and the visual impact can
be reduced by using a darker coloured paint and a
matt finish. As a last resort, on-site touchup of
locations of unacceptable discolouration can be

HERA Steel Design & Construction Bulletin Page 19 No. 67, April/May 2002
used. Leaving the top coat to be applied on site a repair situation. As a general alternative,
after the studs are placed overcomes this look at using channel connectors to
discolouration issue, but requires masking of the NZS 3404 Clause 13.3.2.1 (b).
decking and on-site painting of significant areas of
steelwork. Testing of shear stud weld quality

A specification is given in this session for testing

Session 4.2: Shear Stud Design, Including
of shear stud weld quality. There are two parts to
Testing, and Other Floor this specification. The first part is for shear studs
System welded to the beams with a machine that does not
Good practice provisions record weld cycle parameters. The second part is
for a machine that records weld cycle parameters.
In this session, there is a note slide covering the
topic good practice provisions in regards to the In both parts, but explicitly stated under item 6 of
welding of shear studs. However, a number of the first part, there are requirements to select a
important points were missed out of these notes. random number of extra studs for testing. The
These are: levels of studs to be selected are stated at 5% for
15o bend testing or 100% for ring testing.
• There are special Ramset pins and Ramset
guns for fixing decking to steel that are These requirements are appropriate for studs in
available. The notes correctly state that floor systems subject to fatigue loading, such as
Ramset pins (and guns) used for fastening bridge decks. However, for floor systems subject
timber to steel are not suitable, however to static loading, they are excessive and should
they do not mention that specialist Ramset be reduced to 1% for 15 o bend testing or 20% for
deck pins and guns are available. ring testing.

Performance of floor systems incorporating

• The ratio of (stud diameter / parent metal
Hollowcore units (HCUs)
thickness) ≤ 2.5 is required from NZS 3404
Clause 13.3.2.3 (c). This means that Three issues were covered in this session,
19 mm diameter studs cannot be used on namely:
steel flanges less than 7.6 mm thick. See
also the other provisions of that clause from (1) Problems due to HCU failure under severe
[1] for stud spacing, etc. earthquake action
(2) Determination of shear stud design shear
• Normally, the largest diameter shear stud capacity for composite beams supporting
possible for a given application should be HCUs
used. Also, different shear stud diameters (3) Problems due to HCU failure in severe fire
should not be used on the same floor conditions.
system.
The question was asked after one seminar as to
• When, as is typical, shear studs are to be whether these problems are particular to HCUs on
welded on site, and especially through steel beams and whether the same problems
deck, a good power supply must be would be expected for HCUs on a concrete frame.
available. Minimum recommended is 450
amps/phase, 3 phase, with a 125 amp The answer to that is as follows:
motor rated fuse.
(1) The severe earthquake action problem is
common to all framed building systems, in
• Welded headed studs are designed to be
either steel or concrete. This is well
welded with specialist stud welding
illustrated by the identical failure modes
equipment. The alternative, of welding with
being seen in a steel framed building at
a fillet weld around the base, requires the
Northridge (1994 Northridge earthquake)
use of a suitable prequalified weld detail.
and in a concrete framed building at the
Such a detail would be AF-1 from the draft
University of Canterbury research testing.
standard AS/NZS 1554.3, Structural Steel
In terms of the transverse splitting of the
Welding: Welding of Reinforcing Steel [14].
ends of the units, this is likely to be a
The choice of electrode would be as for
greater problem when the unit is supported
grade 300 rebar from that standard.
on a torsionally stiff concrete beam than
Testing of such welds would be to the 15o
when the unit is supported on a torsionally
bend test, with 1 in 4 studs so tested. This
flexible open steel section.
process is not economic compared with use
of specialist stud welding equipment and
(2) The determination of shear stud design
should be considered only for placement of
shear capacity is particular to the use of
limited numbers of studs, as for example in
HCUs on steel beams.

HERA Steel Design & Construction Bulletin Page 20 No. 67, April/May 2002
(3) The potential poor performance in fire Connections; HERA, Manukau City, New
comprises two aspects. First is poor Zealand, 2002, HERA Report R4-114.
performance of the units due to deformation
of their supporting beams in fire. This is a 5. Clifton, GC; Structural Steelwork Limit
potential problem only for steel supporting State Design Guides Volume 1; HERA,
beams and is overcome by insulation, Manukau City, 1994, HERA Report R4-80.
where required, as specified in [7]. Second
is poor performance due to the unit 6. ABAQUS/Standard; Finite Element Analysis
delamination under fire-induced stresses. Program; HKS Inc, Pawtucket RI, USA,
This is independent of the supporting 2001.
structural system.
7. Clifton, GC (Editor); Notes Prepared for a
At the time of writing this part of the Seminar on Composite Steel Design and
seminar notes [7], the HERA Structural Construction; HERA Manukau City, New
Engineer was not aware of any test or Zealand, 2002, HERA Report R4-113.
actual fire examples of HCU delamination in
severe fires. During the seminar series, he 8. Clifton, GC; Draft for Comment: Guide to
was advised of an example of standard fire the Practical Aspects of Composite Floor
tests from Denmark in which the units System Design and Construction, Including
suffered the postulated delamination. Concrete Placement; HERA, Manukau City,
However, these results have not been New Zealand, 2002, HERA Report R4-
officially published. 107DD.

Acknowledgments 9. Khwaounjoo, YR; Report and User’s

Manual for NZF1_Vib 1 Program (Program
The HERA Structural Engineer would like to for the Analysis of Floor Vibration); HERA,
acknowledge the contribution of all persons and Manukau City, New Zealand, 2002, HERA
organisations involved in these composite floor Report R4-112.
system design and construction seminars, with
special mention of: 10. Murray, TM et. al.; Floor Vibration due to
Human Activity; American Institute of Steel
• Geoff Bird, New Zealand Steel, and Michael Construction, 1997, Steel Design Guide
Klemick, Dimond, for their contribution to Series 11.
the seminar notes, seminar organisation
and presentations. The seminar series 11. Allen, DE et. al.;Minimising Floor Vibration;
would not have been possible without their Applied Technology Council, Allen, DE et.
inputs al.;Minimising Floor Vibration; Applied
Technology Council, Redwood City, USA,
• Dimond and New Zealand Steel, for 1999, ATC Design Guide; 1. Redwood City,
sponsorship and support of their USA, 1999, ATC Design Guide; 1.
presenters’ costs
12. Bird, GD and Klemick, MP; HiBond Design
• Speedfloor Ltd for sponsorship. Wizard for Composite Design of the Hi-
Bond Flooring System, Version 1.0;
Dimond, Auckland, 2002.
References
13. Design Capacity Tables for Structural Steel,
1. NZS 3404: 1997, plus Amendment No. 1: Third Edition, Volume 1: Open Sections;
2001, Steel Structures Standard; Standards Australian Institute of Steel Construction,
New Zealand, Wellington, New Zealand. Sydney, Australia, 2000.
2. Joints in Steel Construction: Moment 14. Draft Standard AS/NZS 1554.3:200X,
Connections; The Steel Construction Structural Steel Welding: Welding of
Institute, Ascot, England, 1997 (with Reinforcing Steel; Standards New Zealand,
amendments), SCI Publication No: 207/95. Wellington.
3. AS/NZS 1252:1996, High Strength Bolts
With Associated Nuts and Washers for
Structural Engineering; Standards New
Zealand, Wellington.

4. Mago, N and Clifton, GC; Finite Element

Analysis of Circular Bolted Flange Annulus

HERA Steel Design & Construction Bulletin Page 21 No. 67, April/May 2002