Stability of Buildings Part 4 Moment Frames

Stability of buildings
Part 4:
Moment frames
Author
A Gardner MEng(Hons) MA(Cantab) CEng MIStructE (The Institution of Structural Engineers)
Consultees
P Perry BSc(Hons) CEng MIStructE MICE MHKIE (CH2M Hill) Chairman of the Reviewing Panel
O Brooker BEng CEng MIStructE MICE MCS (Modulus)
N D Eckford CEng FIStructE MICE (Consultant to David French Partnership)
Dr G J Evans BSc PhD CEng FIStructE FICE MBCS (Seawork Marine Services Ltd)
J Guneratne BSc(Hons) CEng MIStructE (CH2M Hill)
G Mantzouridis MSc CEng MIStructE MICE (AMEC FW)
Dr Y Pan MSc PhD CEng MIStructE (Sir Robert McAlpine Design Group)
A Yates BSc(Hons) BEng(Hons) MBA CEng MIStructE FGS (Webb Yates Engineers)
Secretary to author and consultees
Y Da Silva (The Institution of Structural Engineers)
Acknowledgements
Photographs and other digital imagery have been supplied courtesy of (and are
published with the permission of) the following organisations and individuals:
Figure 3.10: British Standards Institution (BSI)
Figures 7.13, 7.26, 7.30, 7.35: Arup
Figure 7.17: Structural Timber Association and B&K
Figures 7.18, 7.20: Timberbuilt Solutions Pty Ltd
Figure 7.21: prefabricadosagustin.com
Figure 7.27: Concast Precast
Figure 7.31: Kirk Roberts
Figure 7.34: Rob Young
Figure 7.36: Carol M. Highsmith Archive, Library of Congress
Box 2.1: Kirk Roberts
Box 2.2: CH2M Hill
Box 6.1: Steel Designers’ Manual (7th ed.) # 2012 Steel Construction Institute/Wiley-
Blackwell
Box 6.3 (charts): Steel Construction Institute (SCI)
Box 7.3: Architecture:WK/Paul Smith
Hand-drawn illustrations: Y Da Silva, A Hammad and A Gardner (The Institution of
Structural Engineers)
Permission to reproduce extracts from British Standards is granted by the British

Standards Institution (BSI). No other use of this material is permitted. British
Standards can be obtained in PDF or hard copy formats from the BSI online shop:
www.bsigroup.com
Published by The Institution of Structural Engineers,

47–58 Bastwick Street, London EC1V 3PS, United Kingdom
Telephone: þ44(0)20 7235 4535 Fax: þ44(0)20 7235 4294
Email: [email protected] Website: www.istructe.org
First published: November 2015

This version (1.0) published: November 2015
ISBN 978-1-906335-31-1
# 2015 The Institution of Structural Engineers
The Institution of Structural Engineers and those individuals who contributed to this Guide have
endeavored to ensure the accuracy of its contents. However, the guidance and recommendations
given in the Guide should always be reviewed by those using the Guide in the light of the facts of
their particular case and specialist advice obtained as necessary. No liability for negligence or
otherwise in relation to this Guide and its contents is accepted by the Institution, the author, the
consultees, their servants or agents. Any person using this Guide should pay particular
attention to the provisions of this Condition.
No part of this publication may be reproduced, stored in a retrieval system or transmitted in any
form or by any means without prior permission of The Institution of Structural Engineers, who may
be contacted at 47–58 Bastwick Street, London EC1V 3PS, United Kingdom
Glossary
The following definitions are provided to explain how the listed terms are used specifically in this Guide. They
may differ to definitions found in other documents.
Term Definition
Battery A configuration of similar sub-frames arranged in an array; each parallel and at regular centres.
Boundary The stiffnesses applied to the degrees of freedom at the support points (extents) of a structural analytical model.
conditions These are typically at the ground interface but may be to any sources of stiffness (e.g. structural elements)
omitted from the particular model.
Elastic analysis Analysis in which stresses are linearly proportional to strains. In particular, analysis in which curvature is
proportional to the bending moment.
Elastic-plastic Analysis which is elastic until a plastic yield – at which point it becomes plastic.
analysis
In-plane The plane that the sub-frame of interest is arranged within. Where used to describe a characteristic (be this a
deformation, failure mechanism or resistance of either a sub-frame, element or joint), it is the characteristic that
is aligned to the plane of the sub-frame being considered.
Out-of-plane is the orthogonal plane/characteristic.
Note that these definitions are not to be confused with those for the major and minor axes of elements.
Instantaneous A point in space, defined for a specific geometry of a mechanism, about which a part of the mechanism will
centres instantaneously rotate.
Load reversal A load combination that causes stresses within an element to reverse relative to a norm or baseline case. The
norm is often the case with maximum gravity loads.
Major axis The axis of maximum flexural stiffness of an element in a plane that is orthogonal to the longitudinal axis of the
element. It is defined uniquely for each element of a frame.
Minor axis The axis orthogonal to both the major axis and the longitudinal axis of an element.
Plastic analysis Analysis in which energy dissipation is proportional to ductile deformation following yield, but deformation is not
proportional to stress.
Portal frame Herein, the term portal frame refers specifically to a single-storey moment frame used as the primary structure to
a pitched-roof ‘shed’.
Rigid-plastic Plastic analysis in which zero deformation (and zero energy dissipation) is assumed prior to yield.
analysis
Slender An element with slenderness greater than a critical limit and which is likely to buckle under an axial compression
less than the material capacity.
Slenderness A measure of the likelihood of a section, element or frame to buckle when experiencing a compressive stress
less than the yield stress. An element’s slenderness is traditionally measured as the effective length divided by
the radius of gyration of the section about the axis under consideration.
Statically A structure in which the static equilibrium equations (i.e. those defined by Newton’s first law of motion) are
indeterminate insufficient for determining the internal stress resultants and reactions.
Stocky An element with slenderness less than the critical limit and which is not expected to buckle when subject to a
pure axial compression.
Stress resultant A term used to reference bending moments, shear forces and axial forces collectively as they act across a
section of an element. Each of these can be evaluated as the integral of a stress component.
Sub-frame A 2-dimensional, planar system of structural elements that is a subset of a more extensive system. A sub-frame
will usually be expected to be stable in-plane.
Note that elements within one sub-frame will often be considered within an orthogonal sub-frame where the two
intersect.
iv The Institution of Structural Engineers Stability of buildings Part 4

Notation
The following notation is used for hand-drawn figures:
An applied force
A transverse uniform load acting along an element
A uniform axial force acting along an element (shown here to represent compression)
An axial force profile acting along an element
A transverse shear force profile acting along an element
A bending moment profile acting along an element
A movement/displacement
An idealised simple joint or boundary condition (i.e. a pinned or nominally pinned joint)
A moment-resisting joint of non-disclosed stiffness (i.e. a semi-continuous or continuous joint)
An idealised moment-resisting joint of infinite stiffness (i.e. a continuous joint)
A moment-resisting boundary condition of non-disclosed stiffness (i.e. a semi-continuous or

continuous condition)
A moment-resisting joint or boundary condition of finite stiffness (i.e. a semi-continuous joint)
The following notation is used in equations. Further notation is defined in the body text and within figures where
used.
A Cross-section area
E Young’s modulus of elasticity
I Section second moment of area
L Length
Le Effective length for Euler buckling
LLTB Effective length for lateral torsional buckling
My Major axis bending moment
Mz Minor axis bending moment
N Axial force
V Shear force
i Radius of gyration
k Stiffness coefficient
w Applied load
acr Critical load amplifier necessary to cause buckling
D Global displacement at node within frame
d Local displacement at point along element
1 Strain
l Slenderness
s Stress
f Angular rotation (deformation)
The Institution of Structural Engineers Stability of buildings Part 4 v

Contents
Glossary iv 7 Moment frames in application 36

7.1 Single-storey pitched roof ‘sheds’ 36
Notation v 7.1.1 Introduction 36
7.1.2 Overview 36
Foreword vi 7.1.3 Load paths and global stability 39
7.1.4 Structural hierarchy and local stability 39
Part 4: Moment frames 1 7.1.5 Plastic analysis 41
7.1.6 Frame materials 42
1 Introduction 1 7.1.6.1 Steel 42
1.1 References 1 7.1.6.2 Timber 43
7.1.6.3 Precast concrete 44
2 Moment frame overview 3 7.1.7 Further considerations 45
2.1 Form 3 7.1.7.1 Snap-through buckling 45
2.2 Failure mechanisms 4 7.1.7.2 Building length and thermal
2.3 Configuration and application 5 movement 45
2.4 Superstructure materials and construction 6 7.1.7.3 Serviceability deflection limits 46
2.5 Substructure materials and construction 8 7.1.7.4 Internal mezzanine floors 46
2.6 Robustness 10 7.2 Multi-storey frames 46
2.7 References 11 7.3 Further scenarios 47
7.4 References 51
3 Buckling stability – analysis and design 12
3.1 Introduction 12
3.2 Statics and elastic theory 12
3.3 Second-order PD effects 12
3.4 Second-order Pd effects 13
3.5 Designing using utilisation equations 14
3.6 Defining Ncapacity 15
3.6.1 Introduction 15
3.6.2 Effective lengths 16
3.6.3 Calculating the effective length of an
element 16
3.7 Out-of-plane failure mechanisms 18
3.8 Second-order frame buckling analysis 20
3.9 Buckling restraint 21
3.10 References 21
4 Plastic analysis 22
4.1 Introduction 22
4.2 Plastic hinges 22
4.3 Rigid-plastic and elastic-plastic analysis 23
4.4 Collapse mechanisms 23
4.5 Material suitability 24
4.6 References 25
5 Analytical modelling of joints 26

5.1 Introduction 26
5.2 Models 26
5.3 Classification 26
5.4 Modelling practice 28
5.5 Linking analysis to design 28
5.6 References 28
6 Simplified analysis methods 30

6.1 Introduction 30
6.2 Formulaic elastic analysis of standardised frames 30
6.3 Graph-based analysis of standardised frames 30
6.4 Wind-moment frame ultimate limit state analysis 30
6.5 Serviceability sway deflection checks 33
6.6 References 35
The Institution of Structural Engineers Stability of buildings Part 4 iii

Part 4: Moment frames
1 Introduction
This Guide is the fourth and final part of a series reading it will not in itself equate to ability. Rather, this
concerning lateral load resisting ‘stability’ structures. Guide aims to supplement supervised learning by
Its focus is the use, analysis and design of unbraced increasing awareness and promoting thoughtfulness
moment frames. in design.
Moment frames are deceptively simple. To the In line with Part 1 of this series1.1, earthquake design
untrained eye they play down stability; they have is omitted from this Guide. So too are the advanced
fewer elements than triangulated systems and topics of non-linear and dynamic analysis. Although
nothing as bold as a wall to form a focal point for introduced in Part 1, these are considered too
lateral resistance. Indeed, the regular arrangement of advanced for these introductory texts. This is not to
members can have the effect of being little more than say these matters are not important, or that moment
an efficient system to support gravity loads. This frames are not suited to earthquake-prone locations.
perception is possibly not helped by the widespread Indeed the opposite is often true and moment
use of moment frames in lightly utilised everyday frames, when suitably detailed to ensure ductility, can
objects such as tables and chairs. These structures behave favourably under seismic loads.
are everywhere and generally show little evidence of
careful engineering.
However, it is perhaps the immediate simplicity of 1.1 References

moment frames that masks their design complexity.
These are undoubtedly complicated systems and
their design requires a sound understanding of 1.1 Institution of Structural Engineers. Stability of buildings.
structural mechanics. This Guide describes common Parts 1 and 2: General philosophy and framed bracing.
characteristics of frames (Chapter 2) with a bias London: IStructE Ltd, 2014
towards fundamental concepts of frame behaviour,
analysis theorems, second-order effects and buckling
(Chapter 3).
Chapter 4 is dedicated to plastic frame analysis, while

Chapter 5 discusses the behaviour of joints. These
topics are not specific to moment frames but they are
of heightened relevance. A final analytical chapter
(Chapter 6) considers simplified methods that can be
used for manual calculations.
The Guide ends with a discussion on the application

of moment frames (Chapter 7). Throughout, the
Guide is illustrated where appropriate with project
examples.
Relevance and limitations

This Guide does not cover general considerations of
horizontal load resisting ‘stability’ systems or the
broader topics of actions and load paths. These are
discussed in Stability of buildings Parts 1 and 2:
General philosophy and framed bracing1.1. It is
recommended that designers familiarise themselves
with this before embarking on a stability design.
Figure 1.1 on page 2 reproduces the designers’
checklist from the previous parts of this series for
quick reference.
Note that the checklist given in Fig. 1.1 is intended to

serve as a prompt for designers. It only concerns the
design of lateral load-resisting systems and is
generic, to be considered with the project context in
mind.
This Guide is an introduction written primarily for

graduate design engineers, particularly those
approaching a professional review. It has not been
written to provide detailed ‘how to’ instruction, and
The Institution of Structural Engineers Stability of buildings Part 4 1

1.1 Introduction
Designers’ checklist
Actions – applied:
Are minimum and maximum gravity load cases considered?
Are wind, soil, ground surcharge and hydrostatic lateral forces considered?
Are accidental and extreme actions including impact, fire and earthquakes considered?
Actions – induced:
Will actions result from the restraint of arches, domes, catenaries, nets?
Will actions result from initial imperfections?
Will actions result from inelastic strains?
Will actions result from the restraint of post-tensioning and other elastic strains?
Second-order PD effects:
Is the structure sway sensitive/do PD effects need to be considered?
Combinations of actions:
Are all critical combinations for all elements/failure mechanisms evaluated?
Accommodating movement:
Are viable movements understood and quantified?
Are any movement joints necessary and/or incorporated?
Are these accurately portrayed in the analysis?
Are significant movements resisted by the structure?
Are corresponding forces (actions and reactions) allowed for throughout the load path?
Does the design take due account of force redistribution resulting from creep or ground movement?
Are all parts of the structure adequately served by load paths to ensure stability, noting load paths and movement joints are irreconcilable?
How many independent structures exist; is each one stable?
Load paths:
How do forces acting on the façade transfer to the horizontal stability systems? Where the façade spans onto beams, are they
restrained or bending in their minor axis?
How do forces acting on the horizontal stability structures transfer to the vertical stability structures?
How stiff are the horizontal stability structures and the connections from the horizontal to vertical stability structures?
How do forces transfer through the vertical stability structures?
How are forces transferred from the superstructure into the substructure?
How are forces transferred from the substructure into the soil?
Are the interfaces of the above six line items each adequate?
Are there any aspects of the structure, small or large, that do not follow the normal pattern? Do these have suitable load paths of
resistance?
Are all eccentricities accounted for in the analysis?
Braced or unbraced:
Is the structure braced, unbraced, or a hybrid?
Are effective heights correctly determined, taking account of relative stiffnesses and joint rotations where necessary?
Design – stability, strength, service and robustness:
Is the structure in static equilibrium: rotational and linear?
Are all elements and connections adequate to transfer the design actions?
Are deflections, rotations and the natural frequency each within permissible bounds?
Is the structure deemed robust in the event of failure to any of the stability structures? Does the design safeguard against
progressive collapse?
Construction:
Is the disposition of the stability system, and are all design assumptions communicated to the contractor?
Are all parties clear and in agreement on their responsibility?
Is the transfer of information understood by and compatible to all parties, e.g. are actions characteristic or factored values?
Where existing structures are involved, is the stability of these understood before demolition works start?
Are new and existing parts to be connected or isolated from one another?
Alterations and maintenance:
Will new structure provide support to, or act on existing structures?
Are ‘as built’ records available for the existing structure?
Are these accurate to the structure and inclusive of any previous modifications?
Can elements within the completed structure be maintained?
Figure 1.1 Designers’ checklist for design of lateral load resisting stability systems
2 The Institution of Structural Engineers Stability of buildings Part 4

2 Moment frame overview
2.1 Form using a moment frame in a hybrid structure together

with framed bracing and/or walls. The different
systems may be used at different floors (common in
Moment frames (also known more accurately as mixed-use developments), in different axes, or
‘moment-resisting frames’ or ‘vierendeel frames’) are working together at a given floor in a single axis
non-triangulated framed structures that resist both (Figure 2.3). Note that the latter arrangement is not
vertical and lateral forces by a combination of flexural, uncommon – especially where bracing is eccentric
shear and axial stiffness of the constituent elements but leads to a complicated stress state in which the
and of their joints. relative stiffness of each system must be understood.
A number of possible frames are shown in Figure 2.1. Deflected shape and force transfer
Each is a system that has in-plane stability and will Unbraced moment frames can be identified by their
transfer forces acting on the structure in any in-plane characteristic ‘shear deflection’ when subject to
direction to the supports. lateral force (Figure 2.4). This is a global behaviour
A common characteristic of all the frames shown is

that they are statically indeterminate; i.e. the stresses
and reactions cannot be determined via the
equations of static equilibrium i.e. SF ¼ 0 and
SM ¼ 0 alone. Instead, the compatibility of strains is
critical to the force distribution.
A further common characteristic is that the frames

are planar. This is not an absolute rule but is a norm
that is widely adhered to. As such, moment frames
can often be considered as sub-frames within a
structure; each being a 2-dimensional system that
achieves stiffness in a vertical plane against in-plane
forces. In this respect a moment frame is not
dissimilar to a planar wall or a triangulated
arrangement of framed bracing.
Another characteristic of moment frames is that they

are relatively unobtrusive on the floor plan of the
accommodation when compared to walls or bracing
(Figure 2.2); this characteristic is discussed further in
Section 2.3.
Whether a moment frame is more or less of an

obstruction for service distributions will depend on the
coordinated (cross-discipline) design philosophy and
building requirements. Moment frames tend to have
beams that are deeper than simply-supported beams
near the columns, but these may well be shallower
than simply-supported beams at their mid-span.
Note that the stiffnesses of these frames may vary

significantly.
Fig. 2.1 shows various examples of moment frames

acting in isolation. This is common, but so too is Figure 2.1 Generic moment frames/sub-frames
Moment frame Braced frame Wall (often with penetrations)

(a) (b) (c)
Note that the stiffnesses of these frames may vary significantly.
Figure 2.2 Side-by-side comparison showing elevations for each of (a) moment frame (b) braced frame (c) shear wall

2.2 Moment frame overview
Braced at 1st floor
Figure 2.4 Deflected shapes of multi-storey moment frame
resulting from elastic flexural deformations of

Moment frame at ground floor individual elements and joints.
The deflected shape gives significant insight into the

force transfer. Both lateral and vertical actions transfer
through moment frames via a stress distribution that
combines coincident bending moments, shear forces
and axial forces (Figure 2.5). These components are
in equilibrium with the applied forces (and the
reactions) and are compatible with the deformations.
The rules that relate stress resultants within elastic
analysis are discussed further in Chapter 3.
Moment
frame
Moment-resisting joints are essential to maintain the
Braced frame static load path and prevent mechanisms forming.
These are a common feature of all the frames shown
Eccentric in Fig. 2.1 and are critical to the frame analysis.
braced Indeed each frame shown in Figure 2.6 is a stable
frame moment frame of similar geometry, yet each will have
a different stress distribution for any given load
because of the joints.
Ultimately it is the relative stiffnesses of the various

elements and joints that will dictate the magnitude of
the stress resultants from a given load (Figure 2.7).
Moment Flexural deformations of these components tend to
frame overshadow those resulting from shear and axial
forces. The bending stiffness will therefore tend to
Figure 2.3 Hybrid systems govern the global frame behaviour.
It should be noted that the stiffness of the element is

usually in the hands of the structural engineer who
will specify the element. The joint may, however, be
sub-contractor designed and it is imperative that both
the stiffness and strength requirements are
communicated between parties when this is the
case.
Actions and reactions
2.2 Failure mechanisms
In line with the stress resultants shown in Fig. 2.5,

moment frames are susceptible to a plethora of
failure mechanisms concerning each of the elements
and their joints. ‘Failure’ in this case, can relate to
both strength and ductility (at the ultimate limit state)
Bending moments Shear forces Axial forces or stiffness (at the serviceability limit state). Stiffness
will also influence the ultimate limit state in the form of
second-order effects.
However, of particular concern to moment frames is

the global resistance to in-plane buckling. Modes
include:
– Buckling of elements subject to combined
Curvature Deflected shape compression and bending
– Snap-through buckling of pitched rafters in multi-
Figure 2.5 Elastic analysis of simple frame bay portal frames

Moment frame overview 2.3
Figure 2.6 Joint configurations for four similar frames that will influence internal stress distribution and reactions
Frame disposition Bending moments Deflected shape
Note that a similar set of figures could be produced for frames of varying joint stiffness.
Figure 2.7 Bending moments and deflected shapes for similar frames with elements of varying stiffness
These are shown in Figure 2.8. Critical to each is the (for imperfections), indirect actions (e.g. those
analysis of second-order effects. Chapter 3 discusses resulting from thermal effects) and second-order
this topic in more detail. effects. Portal frames with pitched or curved roofs
need to be analysed for both symmetrical and
Continuity between elements means that pattern asymmetrical load configurations. Other symmetrical
loading should be considered. Furthermore, direct frames with horizontal floor and roof systems are
actions must be combined with equivalent actions generally governed by asymmetric load combinations
only (noting that all analytical results for symmetrical
frames must be reflected about the axis of
symmetry).
2.3 Configuration and application

Sway instability
Moment frames are almost always less stiff than
framed bracing or shear walls of similar overall plan
dimensions. This generally limits their application to
relatively low-rise buildings in which lateral shear forces,
sway and second-order global PD effects are modest.
Snap-through buckling
On account of their limited stiffness, moment frames
Figure 2.8 In-plane buckling of multi-bay portal frame will often extend across multiple column bays. This is

Bracing Moment frame
Plans
Sections
Two-way braced Orthogonal braced/moment frame Two-way moment frame
Figure 2.9 Example configurations of vertical stability systems
only possible because of their relatively non-invasive – Particular temporary conditions during construction,
form with only columns impacting on the e.g. where site constraints necessitate that an
accommodation at each floor (Fig. 2.2). access route is maintained across the ground floor
of the building footprint, or ahead of an infill panel
The most common configuration is to use a battery system being installed
arrangement of similar frames; each parallel and
ideally, spaced at a regular pitch. This has a number By far the most common application of 1-way
of advantages – one being to limit the load acting on moment frames is in single storey ‘sheds’
any single frame so that column and beam sizes can (Section 7.1). Contributing to their efficiency, and thus
be kept to a minimum. Further advantages include: widespread use, is a freedom to utilise an optimal
regularity in the design and fabrication of the frames; frame geometry with each of the pitch of the roof, the
greater potential to omit or vastly reduce horizontal size of the joints (including haunches) and the height
plan bracing in the floor and roof planes; greater of the eaves and/or ridge only loosely constrained by
stability during erection and greater overall operational requirements. Multi-storey buildings do
robustness. not often afford these luxuries. Storey heights,
structural beam depths and clearances under beams
Figure 2.9 shows (on plan and in elevation) how (with or without haunches) are often governing
battery arrangements can be used for 1-way and criteria.
2-way moment frames. A 2-way braced frame is
shown for comparison. Collectively the three images
highlight the openness of moment frame systems but Further reading: hybrid systems and arrangements of
also that the column sections are generally larger stability structures
than in braced frames.
– Institution of Structural Engineers. Stability of buildings.
Buildings are often anisotropic in their layout, Parts 1 and 2: General philosophy and framed bracing.
functionality and/or site constraints. It is therefore London: IStructE Ltd, 2014
often practical to adopt a hybrid system with
orthogonal moment-resisting and braced frames. This
will usually aid the construction allowing 1-way planar
moment frames to be designed, fabricated and
erected largely independently of one another.
2.4 Superstructure materials and
construction
Relative to 2-way moment frames, 1-way moment
frames (with orthogonal bracing) usually lead to less
cumbersome and costly joints. 1-way systems will Reinforced concrete, steel and timber are common
also often promote a natural hierarchy within floor or materials for moment frame construction. Each can
roof frames that is efficient for the floor and roof be designed to exhibit the combined shear, flexure
constructions and equally for services distribution. and axial compression resistance that is necessary
Hence, use of 2-way moment frames is generally for the force transfer.
restricted to a ‘needs-must’ basis. Common
scenarios where needs are such include: Material availability, local construction preferences and
– Sheds with loading bays along the entirety of the local environmental conditions together with regulated
eaves elevations performance requirements can influence the
– Open canopy roofs devoid of a solid envelope prevalence of material by region, while site constraints
– Low-rise and large footprint reinforced concrete can often rule out or favour options for a specific site
frames in seismic areas (Boxes 2.1 and 2.2).

Box 2.1 Multi-storey engineered timber moment frame

The Merritt Building, Christchurch, New Zealand, has a first-of-kind post-tensioned glulam moment frame, utilising a technique
pioneered locally at the University of Canterbury2.1. The system has since been replicated on a larger scale at the Trimble
Navigation building (also Christchurch).
Both buildings contain multi-storey accommodation stabilised in one axis by a battery-arrangement of post-tensioned timber
moment frames that have been shown via testing to provide good earthquake ductility. Timber is competitively priced in
New Zealand and the earthquake resilience suits the islands’ geology.
The monolithic nature of in situ reinforced concrete assembled on site. Limiting dimensions for
lends itself to moment frames. However, it is the transportation generally mean that joints are at or very
arrangement of reinforcement that is ultimately critical near to cranks (e.g. column-beam junctions) where
to the behaviour, not least to the overall ductility. bending moments tend to be high (Fig. 2.5). Hence,
site joints will often need to have comparable
Timber, steel and precast concrete each tend to be resistance (stiffness and strength) to that of the
transported and erected as a kit of parts to be elements (Chapter 5).
Box 2.2 Hybrid concrete frame construction at Bond Street Station upgrade, London
The over-site development above the new satellite ticket hall to the Bond Street Station upgrade has a structural system that was
dictated by the temporary use as a Tunnelling Crane Hall, prior to completion as a retail and residential development. Here, precast
concrete columns with moment connections to steel floor beams were adopted, providing a moment frame that works in unison
with an offset shear wall2.2. This allowed the first and second floor to be omitted in the temporary configuration.
Each precast column has cast-in steel splice details and connectors to which the floor beams fix.
Precast concrete
column
Steel floor
beams
Cast-in
steel splice

Moment frame to raised atrium roof



 Braced structure



Figure 2.10 Atrium roof acting as moment frame
Many local construction industries recognise the although this does depend on the nature of the joints
dangers of site welding and prefer bolted site through the moment frame (the frame in Fig. 2.11(c)
connections for steel. Timber is similarly bolted for example, demands that the substructure is
either with or without steel inserts. Thus, rotationally stiff).
connections in both steel and timber tend to
necessitate haunches to accommodate the bolt The degree of base fixity will always be critical to the
group. Each of reinforcement area, bar spacing, reactions and internal stress distribution. It must
bending radii and laps have similar influence on the therefore be considered in a similar manner to other
overall dimensions of reinforced concrete and it is joints throughout a frame (refer to Section 4). Each of
not uncommon to detail haunches within concrete the horizontal, vertical (Box 2.3) and rotational
profiles. stiffnesses are variables to be considered. As with all
variables, the assumptions adopted within analysis
must be suitably in-keeping with the behaviour of the
realised structure to avoid significant analytical errors
2.5 Substructure materials and misleading the design.
construction
Irrespective of the various joint stiffness (but not
independent of them), a notable characteristic of a
Moment frames can use traditional foundation moment frame’s reactions are the horizontal forces
systems (e.g. piles, pads, rafts, strips etc.) where that result from vertical loads. These thrusts are not
extending to ground. Alternatively they may found dissimilar to those acting on an arch. Different
onto a braced structural system (e.g. an atrium roof approaches to resist these forces in foundations are
that is stooled up above the surrounding structure shown in Figure 2.12. Strategies either contain the
(Figure 2.10)). Either way, the reactions at the base of force within the structure or exert it onto the
the moment frame will be dependent on the linear surrounding soil.
and rotational stiffnesses of the connections and
‘substructure’ elements below. It should be noted that examples shown in
Fig. 2.12(b–e) are specifically for self-opposing thrusts
Reactions on the substructure for three moment and these systems will not work for external
frames are shown in Figure 2.11. The figure illustrates horizontal actions (such as wind) that cause a net
how the frames may be stable (in plane) with either horizontal shear force. Net forces must be resisted by
pinned or moment-resisting column base details, the soil.
Load case 1
Load case 2
(a) (b) (c)
Figure 2.11 First-order reactions for moment frames

Box 2.3 Vertical foundation settlement and soil stiffness

Settlement comes about from the vertical stiffness of the soil upon which a foundation bears when subject to the load of a building.
How this influences the stress distribution in the superstructure can vary significantly depending on the frame geometry, its joint
types and their stiffness. The following illustrations show this for a braced frame with pinned joints and a moment frame with rigid
joints.
Deformation is local Potential stresses in façade

to the settlement elements and slab if continuous
Settlement (a) Braced structure with settlement away from bracing
Increased tension in bracing
Horizontal
deformation
extends through
structure
Increased load on foundation as forces redistribute

Settlement
(b) Braced structure with settlement to braced bay
Deformation
extends
through
structure
Increased load on foundation as forces redistribute
Settlement
(c) Moment frame structure with settlement
Differential settlement is the critical parameter here and would typically depend on the foundation system and site geology.
In addition to variation across a site, soil stiffness can also be time-dependent. In particular, pore-water pressures and permeability
mean that a soil’s stiffness can vary depending on the load duration. In practice, this means it may be possible and/or necessary to
assume different soil stiffnesses for short and long term actions (short term possibly affording a higher soil stiffness).
Fig. 2.12 makes no reference to base friction. Friction evaluate whether it is safe to utilise the passive
should only be assumed for the minimum reaction pressure, considering the specific circumstances of
pressure, taking account of any uplift action on the a given project (e.g. the site, building class, design
superstructure and hydrostatic action on the life and proximity to other structures, highways,
foundations. It is, however, common practice to hard standings, buried services, soakaways etc.).
ignore friction particularly for lightweight sheds.
Even where passive ground pressure can be relied
Unlike friction, the horizontal passive ground upon without unacceptable risk, the displacement
resistance is independent of the vertical reaction. necessary to mobilise a passive force may be
However, the loaded ground is often outside the too large to be within tolerances for the
building footprint and may be outside the site superstructure. The ground stiffness and its
boundary. As such, it can be uncontrollable and variability are critical parameters in this regard and
vulnerable to future or immediate excavation that may structural engineers should consult the project
influence both short and long term performance. geotechnical engineer to ascertain bounds on the
Engineers should carry out a risk assessment to reaction stiffness.

(a) Passive soil pressure
(b) Ground beams (c) Ties below slab
(d) Tied suspended slab (e) Tied slab
Ties and ground beams connect to opposing column and footing
Figure 2.12 Horizontal restraint of columns within pitched roof portal frames due to vertical actions
Ground beams and ties are the most direct and of battery configurations (as introduced in
certain means of providing resistance but are Section 2.3) which means that stability of the
expensive and need to be designed to withstand global structure is often shared across a number
long-term corrosion. They can also impede mobile of sub-frames.
plant and influence the installation of buried services
during construction. However, continuity imparts requirements that are not
applicable to simple-jointed braced frames.
Ties into a floor slab can provide an efficient solution
but only where the slab is free of free-movement It is generally favourable to design the elements of
joints in the direction of the thrust. Tying the slab to a moment frame so that failure will occur in beams
the frame can also make detailing for vertical in advance of the columns. This is commonly
movement more complicated (e.g. where the slab is referred to as a ‘weak beam-strong column’
ground bearing). Temporary restraint may also be approach (although this phrase is not universal and
needed in advance of the slab being cast. does not, for example, appear in BS EN 1992
Part 1-1 Clause 5.6.2(3)2.3). The logic is that there
Finally, it should be noted that it can be hard to is usually less redundancy in the vertical elements
determine whether passive pressure, ground beams, and therefore it is most important to maintain the
ties or ties into the slab are acting within existing integrity of these. This is illustrated in Figure 2.13
structures. Engineers working on new developments for a local extreme action.
should document clearly what systems are being
employed so as to provide a clear record. Meanwhile, For global actions (e.g. seismic actions), a weak
those engineers working on existing structures beam-strong column approach will usually
without thorough documentation should arrange for maximise the number of plastic hinges necessary
exploratory studies to be conducted to ascertain the to cause lateral instability. For the undamaged
tie system before any significant demolition is frame shown in Fig. 2.13, 16 hinges (two hinges
conducted. The same obviously applies for moment per beam) are needed in the beams to achieve a
frames atop braced structures, although these are mechanism, while only six are needed in the
usually more easily inspected. columns (two hinges per column). Meanwhile, if the
column bases were pinned, only three hinges
would be needed in the columns (one hinge per
column).
2.6 Robustness
In line with this, splice joints between lengths of
columns should have strength commensurate to the
Moment frames are often inherently resilient to strength of the ‘strong column’ and not the ‘weak
disproportionate collapse. This is in part a result beam’.

Failed column
Failure propagates
as columns fail
Action
(e.g. explosion)
Weak column
Failure contained
‘Strong’ column designed for locally
lateral tie force of catenary floor
Failed beam. Catenary action
prevents propagation to floor below
Strong column
Note that in the absence of more accurate data, the accidental loading on key elements should be taken as 34kN/m2 when
designng in accordance with EN1991 Part 1-72.4.
Figure 2.13 Disproportionate collapse: strong vs. weak columns
Further reading: robustness and disproportionate

collapse
– Institution of Structural Engineers. Practical guide to
structural robustness and disproportionate collapse in
buildings. London: IStructE Ltd, 2010
2.7 References
2.1 Buchanan, A. et al. ‘Post-tensioned timber frame

buildings’. The Structural Engineer, 89(17),
6 September 2011, pp24-30
2.2 Institution of Structural Engineers. Stability of buildings.

Part 3: Shear walls. London: IStructE Ltd, 2015
2.3 BS EN 1992-1-1: 2004: Eurocode 2: Design of

concrete structures – Part 1-1: General rules and rules
for buildings. London: BSI, 2004
2.4 BS EN 1991-1-7: 2006 þ A1: 2014: Eurocode 1:

Actions on structures. Part 1-7: General actions –
Accidental actions. London: BSI, 2014

3 Buckling stability – analysis and design
3.1 Introduction to a distance along the longitudinal axis of the

element.
This chapter provides an overview of the behaviour of The rules apply irrespective of whether the frame is
unbraced, statically indeterminate moment frames. statically determinate or indeterminate.
The focus is on in-plane behaviour as this is the most
influential characteristic in differentiating unbraced The first and second rules ensure overall equilibrium
moment-resisting frames from braced counterparts. while the third and fourth rules define relationships
Elastic analysis is considered as this goes hand-in- outlined in Box 3.2.
hand with buckling stability. However, parts of this
chapter apply equally to plastic analysis which is Collectively the four rules allow the approximate
discussed separately in Chapter 4. shape of the first-order bending moment, shear force
and axial force profiles to be determined manually for
Both local Pd and global PD second-order effects are regular sub-frames undergoing elastic deformation.
discussed herein, with both being critical to moment Figure 3.1 shows an example of this, displaying the
frame design. The difference between these is profiles for a simple rectangular ‘goalpost’ frame
defined in Box 3.1. subject to a uniform load; it reveals how elements
within a loaded moment frame are generally subject
to each of axial forces, shear and bending moments
concurrently.
3.2 Statics and elastic theory
For moment frames in particular it should be noted
that it is impossible to quantify salient values within
Four non-material rules apply in the static analysis of these profiles without considering component
a sub-frame: stiffnesses (because moment frames are statically
– Forces must be in equilibrium at all points S(F ) ¼ 0 indeterminate).
– Moments must be in equilibrium at all points
S(M) ¼ 0
– The differential of moment is transverse shear
f 0 (M ) ¼ V 3.3 Second-order PD effects
– The differential of a transverse shear is a transverse
force f 0 (V ) ¼ w
Global PD effects are introduced in Stability of
In these rules, ‘transverse’ describes a direction that buildings Part 1 and 2 3.1, where a derivation of the
is orthogonal to the longitudinal axis of the element sway modification technique is provided. The
and the differentials (denoted f 0 ( _ )) are with respect discussion therein holds true for braced and
Box 3.1 Second-order Pd and PD effects

It is important to distinguish between local and global second-order effects. Herein, these are labelled Pd and PD respectively:
– A Pd effect is the local second-order effect on an element due to a localised deformation of that element; the most usual being
a compression bow that results in an additional (or amplified) bending moment due to the eccentricity of the axial force.
– A PD effect is the global second-order effect on a frame or sub-frame; the most usual mode shape is horizontal sway.
Second-order Pd and PD effects must be combined as illustrated:
∆
2nd order ∆
Initial state 1st order Total 2nd order

Pδ + P∆
δ δ
2nd order δ
The extent to which a frame is affected by global PD effects is dependent on the sway stiffness which, for a moment frame,
is dependent on the local Pd effects on the elements within that frame.

Buckling stability – analysis and design 3.4
Box 3.2 Elastic theory: relationships between actions, shear forces and bending moments
The following relationships can be determined from the differential equations that relate actions, shear forces and bending
moments:
– A free body that is not subject to a transverse action throughout its length will exhibit constant shear and a linear bending
moment profile:
Load Shear force Bending moment
In the particular case where the shear force is zero, the bending moment will be constant and may be zero.
– A free body subject to a constant (or uniform) transverse action will exhibit a linear shear profile and a quadratic bending
moment profile:
– A free body subject to a concentrated transverse action will exhibit a stepped shear profile and an abrupt change in gradient of
the bending moment profile:
unbraced structures alike; anywhere that a vertical with the bending moment, stiffness and support
action acts on an initially vertical or near vertical conditions. This deflection can be determined from
element, it will induce a horizontal component force first-order analysis and can be estimated accurately
or bending moment that is dependent on the sway using elastic theory (Figure 3.3). It can also be
deformation of the element. proven experimentally using a simple physical
model.
Elements with moment-resisting joints can
themselves offer resistance against PD effects. An axial compression can cause a similar element to
Meanwhile orthogonal pinned-frame elements, in the deflect or bow. Like bending, this deformed shape
absence of bracing, cannot resist PD effects and will can be proven easily with a physical model.
always cause a force to act on an adjoining stability However, less obvious, the behaviour is not apparent
frame. Thus, any pin-ended prop within a moment from a first-order mathematical analysis. Indeed
frame will cause a force on the moment-frame that bowing is dependent on second-order Pd effects
will increase the PD effects, increasing the overall (Figure 3.4).
sway-sensitivity. This is shown in Figure 3.2.
Euler accurately predicted that, in the absence of a
Moment-resisting frames are often particularly prone transverse action, failure of a pin-ended initially
to sway instability because they deform in a shear straight element (or ‘strut’) will occur at a critical axial
mode with maximum inter-storey sway coinciding force Ncr where:
with maximum column axial force at or close to the
ground floor. Ncr ¼ p2EI/L2 . . . Eqn 3.1
where:
E is the Young’s modulus of the element
3.4 Second-order Pd effects I is the second moment of area of the element
L is the length of the element between pinned
supports (the ‘strut length’)
A transverse force will cause an element to deflect
or bend with rate of change of curvature consistent This is an elastic buckling failure.
Actions and reactions Axial forces Shear forces Bedding moments
Note that the elastic deflected shape can be approximated from the bending moment profile, knowing that the elastic bending
moment is proportional to curvature.
Figure 3.1 Example elastic distribution of stress resultants throughout a sub-frame

3.5 Buckling stability – analysis and design
Destabilising force
onto bracing structure
δB δB
Self stable Pin-connected strut
(a) With pinned joints (b) With moment-resisting joints (c) Combined
Figure 3.2 Sway-induced PD effects for elements with and without moment-resisting joints
When an axial force acts concurrently with a actions, and that second-order analysis is needed to
transverse load the axial force at which buckling evaluate local Pd effects accurately.
occurs reduces as the transverse load increases
(Figure 3.5)3.2. Reversing this statement, the bending Significant Pd effects can prohibit the use of the sway
moment capacity of a section decreases as the modification method (as introduced in Parts 1 and 2
coincident axial compression increases. An axial of this series3.1) for global PD analysis. BS EN 1993
compression thus has a destabilising effect on Part 1-1 Clause 5.2.13.3 defines a limit on what it
flexural elements. considers ‘significant’ for steel moment frames.
Bringing Figs 3.1 and 3.5 together, it should be BS EN 1992 Part 1-1 Clauses 5.8.7 and 5.8.83.4 deal
apparent that moment frames are susceptible to the with Pd effects by stipulating that either an additional
destabilising effects of combined axial and flexural moment should be applied or the applied moment
amplified (both to take account of the increased
eccentricity).
w 3.5 Designing using utilisation equations
δ Bending moments and axial compression each cause

longitudinal normal stresses on sections of an
5wL4 element. These must be accommodated locally (to
δ= avoid cross-section failure) and globally (to avoid
384EI element buckling).
Figure 3.3 Flexural deflection, d of a pin-ended beam subject A typical resistance curve for compression elements
to transverse force w is shown in Figure 3.6. The graph plots the capacities
of an element of constant cross section when varying
the length between pinned supports. It exhibits two
dominant component lines: a non-linear buckling limit
N N as defined by Equation 3.1, and an upper bound
plateau defined by the material strength.
1st order: N N N
Ncrit 1
2nd order: N δ Increasing transverse load

N
M = δN
The moment M results in a curvature δ
that causes a compatible deflection δ Note that initial imperfections have a similar effect on the
compression resistance as a transverse load.
Figure 3.4 Analysis of a pin-ended strut subject to an axial
force N Figure 3.5 Variation of axial load with deflection

Compressive stress Strength failure N
Buckling failure
Stocky Slender
Strut length
Figure 3.6 Idealised resistance curve for axial compression, Reinforced concrete M
N capacity
N
Fig. 3.6 defines the horizontal axis as the ‘strut
length’ consistent with Eqn 3.1. However, it is typical
to assign ‘slenderness’ to this axis (Box 3.3). It
should be noted that element slenderness, as would
be used in this instance, is different to section
slenderness used primarily to determine the plastic
bending moment capacity of a section. The two must
not be confused.
Box 3.3 Slenderness

Slenderness l is a non-dimensional measurement that takes
account of the section stiffness I, allowing a buckling curve to
be defined that is independent of element cross-section M
Steel
geometry. The relationship between slenderness and buckling
can be derived directly from Eqn 3.1:
Figure 3.7 Compressive failure envelopes for elements
Critical axial strut force restrained out-of-plane. The hatched areas satisfy Eqn 3.2
Ncr ¼ p2EI /L2 . . . Eqn 3.1
Ncapacity and Mi.capacity can be defined for each of the
Corresponding critical stress strength and buckling limits and both must be
considered in turn. Definitions for Ncapacity and
I Mi.capacity differ by material as a function of the
scr ¼ p2 E =L2 ¼ p2 E ½i 2 =L2 ¼ p2 EI =l2
A dominant characteristics and are not necessarily
independent of one another (e.g. Mi.capacity may be a
Where: function of Ncapacity). This impacts the relationship
l is slenderness between limiting Ndesign and Mi.design which can be
i is the radius of gyration non-linear. Figure 3.7 illustrates this, plotting typical
A is the cross section area envelopes for reinforced concrete and steel. In both
In this relationship, scr =sy (where sy is the material cases, the envelopes are for elements of a fixed
strength) will vary as a function of E leading to unique length and cross section (i.e. at a fixed point on the
resistance curves for any combination of E and sy . This Ncapacity curve shown in Fig. 3.6) and subject to a
can be avoided by defining a normalised slenderness l , uni-axial bending moment.

typically with l ¼ 1 for scr =sy ¼ 1 (i.e. when
Ncr ¼ A sy ). In practice, an equation equivalent to Eqn 3.2 is not
given in EN1992 and is not easily applied in practice
Note that BS EN 1993 confuses this matter by describing l
to reinforced concrete because of the intricate way in
as the ‘‘non-dimensional slenderness’’ and not a normalised
which Mi and N interact.
slenderness. All measures of slenderness are non-
dimensional.
Most codes of practice consider the utilisation ratios 3.6 Defining Ncapacity
of both an element (against buckling) and its cross-
section (against crushing/yielding) by means of two or
more utilisation expressions. A generalised form of 3.6.1 Introduction
which is listed as Equation 3.2:
Fig. 3.6 shows that two dominant lines define
Ndesign My:design Mz:design Ncapacity : a straight line for the strength limit and a
þ þ 1:0 . . . Eqn 3:2 curved line for the buckling limit. These are material
Ncapacity My:capacity Mz:capacity
and system characteristics and apply irrespective of
the analysis technique.
where:
N is an axial force The strength limit relates to the net cross-section area
Mi is a bending moment in the i’th axis and the material strength. It is independent of the

Note that the action must

align with the axis of the
equivalent strut
Equivalent pin Equivalent pin

(M = 0) (M = 0)
Le/2
Equivalent point of
Le/2
max curvature
Point of
Le
max curvature
Le
<Le/2
Le/2 Equivalent pin
(M = 0)
Axis of equivalent strut

(a) With pinned bases (b) With moment-resisting bases
Figure 3.8 Effective lengths of columns within single storey unbraced moment frames
element length, the position of the element within a two example frames. The effective lengths for the
frame, and of the behaviour of the frame. It is also beams will be similar in both frames and the form is
independent of second-order effects (although these shown in Figure 3.9.
may influence Ndesign ). This capacity applies only
where the element is ‘stocky’. 3.6.3 Calculating the effective length of an element
To the contrary, the buckling limit is intrinsically linked The effective length of an element is dependent on
to the strut length and, hence, to the frame geometry. the rotational stiffnesses of the two end fixings, on
Furthermore, in moment frames where elements the position of any points of inflection along the
cannot be considered to act as pin-ended Euler element, and on the restraint against lateral sway.
struts, it is also linked to the joint/frame stiffness. In Figure 3.10 (a reproduction of BS EN 1992 Part 1-1
this scenario, Eqn 3.1 needs to be modified to Figure 5.73.4) summarises the different scenarios.
account for this stiffness, usually via the adoption of Columns b, e and g relate to unbraced sub-frames
an ‘effective length’. where the element is not restrained against lateral
sway. Of these, column e is an idealisation that
3.6.2 Effective lengths assumes infinitely rigid end conditions.
An effective length Le (otherwise known as an All real-world moment frames have non-ideal end
equivalent buckling length3.5 ) is the analytical length conditions with finite rotational stiffness provided by
used to account for the buckling behaviour of an the stiffnesses of the joint and the connecting
element. It equates the real element, as it exists elements. These need to be considered when
within a frame, to the strut length (that of an evaluating effective lengths.
equivalent pin-ended element considered in isolation
that conforms to the Euler buckling limit). As such, it Finally, the effective length is dependent on the
can be substituted for the strut length L in Eqn 3.1. accuracy of the analysis; in particular, how accurately
Figure 3.8 defines effective lengths for columns within it evaluates Ndesign . Eqn 3.2 requires that the force
vectors Ncapacity and Ndesign share a common axis to
be comparable. Two analytical methods are
Dimension varies as a function introduced that deal with this requirement quite
of the column stiffness differently:
Le Global second-order analysis

Le/2 Global second-order analysis takes account of initial
Point of max frame imperfections (usually by using equivalent
curvature horizontal forces) as well as the globally applied
actions, boundary conditions, and Pd and PD effects
to determine maximum magnitude force vectors for
Equivalent pin Ndesign . These vectors are in keeping with the
(M = 0) Equivalent pin deformed geometry of the frame and remain
(M = 0) compatible with the effective lengths that can be
derived directly from the frame curvature (without
Figure 3.9 Effective lengths of a beam within a single storey simplifying assumptions or boundary conditions
moment frame (Figure 3.11a)).

(a) Le = L (b) Le = 2L (c) Le = 0.7L (d) Le = L/2 (e) Le = L (f) L/2 < Le < L (g) Le > 2L
Figure 3.10 Effective lengths of elements
N=V
V H
N= + N=V
cos θ sin θ
Le > 2L
High stiffness beam

V V V
Le = 2L
Le = 2.2L
H H H
Negligible
θ
L
joint rotation
H
H H
V V V
(a) (b) (c)
Figure 3.11 Equivalent struts for elements with one pin-end in unbraced frames
The method is highly accurate and also suitable for in-plane stiffnesses of connected elements to
irregular frame geometries that do not fit standard determine the approximate end rotation and
simplifying assumptions. It is however almost curvature of a subject element. Every element that is
impossible to implement without the use of a suitable loaded in compression should be evaluated as the
computer analysis package. subject element in turn, considering the influence of
sufficient connecting elements so that the end
Sub-frame analysis conditions can be accurately assessed. As a
Effective lengths can be approximated by first-order minimum, all elements connecting directly to the
sub-frame analysis. This method uses the relative subject element are usually considered (Figure 3.12).
Simplified boundary
conditions assumed
Subject Subject element

element
Frame Sub-frame (for nominated subject element)
Figure 3.12 Sub-frame model for determining the effective length of a subject element

BS EN 1992 Part 1-1 provides simple formulae

from which the end rotational stiffnesses can be
calculated for elements that fit columns f and g of
Fig. 3.10. Commentary on this is provided in
Clause 2.11 of PD 6687-1: 20103.6. Other codes,
including the Australian Standard AS 41003.7, use a z
charted approach to achieve a similar outcome.
n
Lex-colum
However, equivalent guidance is omitted from all α
parts of BS EN 1993. It is instead included within
Le
UK Non-Contradictory Complementary x-b
ea
Information3.5. θ m
Being a first-order approximation, the axial design θ

force Ndesign will be aligned with the initial state and Out-of plane
not the deformed state (this is irrespective of whether axis (y) α
a horizontal sway modification factor is applied). The
effective length must take account of this, ensuring
In p
that Ncapacity shares the axis of Ndesign . This generally lan
ea
means an effective length for each element that is xis
(x)
longer than that derived from global second-order
analysis (Figure 3.11b). θ and α are rotations in the z–x plane
In line with this last concept, AS 4100 Figure 4.6.3.2

Figure 3.13 In-plane effective lengths for compression
suggests an effective length of Le ¼ 2.2L is used for
buckling
the case where rotation at the continuous end of a
cantilever is negligible (Figure 3.11c). This is obviously
more conservative than Fig. 3.10 column b, but
makes allowance for the omission of second-order may ultimately be the same (this is common in
effects. In the absence of a more accurate ‘simple’ construction) but will often differ in moment
assessment, we recommend this is used where the frames due to the dependency on the frame stiffness.
flexural stiffness of the subject element is less than
half the summed stiffness of the connecting Elements may also have differing stiffnesses in two
elements. In this calculation the stiffness can be orthogonal directions, often having ‘major’ (more stiff)
determined for each element as E,I/L where E, I and and ‘minor’ (less stiff) axes.
L are the Young’s modulus, second moment of area
and length of the element respectively. Combining the different stiffnesses with the effective
lengths, it is typical that elements are of different
The primary advantage of the sub-frame method over slenderness in two orthogonal axes and therefore
a global second-order analysis is that it can be have different critical buckling loads in the two axes.
completed without the aid of a computer. Hence it Ncapacity , as contained within Eqn 3.2, should always
remains a convenient method for preliminary design be taken as the lesser of these critical buckling loads.
and detailed design validation (checking). The key
simplifications are: Box 3.4 Orthogonal axes for compression
– That boundary conditions are assumed at the buckling
periphery to the sub-frame elements
– That it does not take explicit account of the When evaluating an element, two orthogonal axes are needed
deformation mode shape nor PD effects because these consider orthogonal stiffness vectors; i.e.
(assumptions are instead applied via the effective component vectors that are independent of one another and
length charts/formulae) collectively define the element’s stiffness characteristics.
Whether these axes are defined in- or out-of-plane (as
defined for the global sub-frame analysis) is arbitrary,
however it is convenient and convention to align at least one
3.7 Out-of-plane failure mechanisms of these axes to either a dominant axis of a structure, or a
dominant axis of the element (preferably both simultaneously).
Figure 3.13 replicates parts of Figs 3.8 and 3.9,

showing the in-plane effective lengths for Major axis bending moments My.capacity
compression of both the column and beam. Any element that has ‘major’ and ‘minor’ axes of
However, depending on the arrangement of structure differing stiffnesses can experience lateral torsional
that is projecting out of the plane of the frame, buckling when subject to a major axis bending
various additional ‘out-of-plane’ failure mechanisms moment and is free to deform in the minor axis. The
may be at play. Examples are shown for similar mechanism is most prevalent in sections that have a
frames in Figure 3.14. high major axis to minor axis stiffness ratio (as these
can be loaded most highly in the major axis), and in
The out-of-plane failure mechanisms can influence open sections that exhibit low torsional resistance.
both Ncapacity , My:capacity and Mz:capacity from Eqn 3.2 Steel universal beams, channels and angles are highly
for the design of elements, as well as influencing the sensitive, but none of steel rectangular hollow
stability of the joints. Collectively with in-plane sections, timber joists, glulam timber beams or
mechanisms, they are discussed as follows: reinforced concrete sections of large depth to
breadth ratios is immune from lateral torsional
Compression Ncapacity buckling as a potential failure mechanism.
All elements subject to compression must be
considered to have two effective strut lengths: one in Similar to compression buckling, consideration of
each of two orthogonal axes (Box 3.4). These lengths lateral torsional buckling will lead to an element’s

Columns are unbraced and part of

a moment frame out-of plane
z
φ
z
Ley
-be
am
φ
Frame is ‘simple’ and

braced out-of plane Ley-column
e
lan
t - of p (y) ne
Ou axis o f pla
t- )
Ou axis (y
Ley-column
In p
lan In p
ea lan
xis ea
(x) xis
(x)
φ is a rotation in the z–y plane
(a) (b)
Le-
LTB
.be
am
Beam and columns unrestrained

along their lengths
ne
o f pla
t- )
Ou axis (y
Le-LTB column
In p
lan
ea
xis
(x)
(c)
(a) Compression buckling of a sub-frame that is ‘simple’ and braced out-of-plane
(b) Compression buckling of a sub-frame that is unbraced with columns part of orthogonal moment frames
(c) Lateral torsional buckling of unrestrained elements subject to in-plane major-axis bending moments
Figure 3.14 Out-of-plane failure mechanisms (in-plane movements omitted for clarity)
capacity being defined by two component curves: a the capacities of an element of constant cross
section strength limit (determined by either elastic or section when varying the length between points of
plastic methods as appropriate) and an element restraint.
buckling limit (Figure 3.15).
An effective length LLTB for lateral torsional buckling is
Note that the intersection of the two curves is itself a typically used to determine the slenderness and
curve for real sections owing to element hence My:capacity . It must be recognised that this
imperfections. This is not shown in Fig. 3.15 but is effective length is independent of (but may be similar
included in Figure 3.16. Both these figures plot to) the compression effective lengths Le .
My.capacity
Section strength limit

My.capacity
Lateral torsional
buckling failure
Slenderness
Slenderness
Figure 3.16 Resistance curves for major axis bending
Figure 3.15 Idealised resistance curve for major axis bending moments M y:capacity , showing curves for different bending
moments M y:capacity moment profiles

n mn
olum fc olu
of c x is o
axis or a
M ajo
r Min
Min Ma
or a jor
xis axis
of c of c
olu olu
mn mn
This column
will not undergo
lateral torsional
buckling due to
This column may in-plane actions
be susceptible to
lateral torsional
buckling due to
in-plane actions
ane pla
ne
o f -pl f -
t- s t-o s
Ou axi Ou axi
In-p In-p
lan lan
ea ea
xis xis
(a) (b)
Notes
Requirements of the joint such as stiffeners or haunches are omitted.
Case (b) is typically preferred only where an orthogonal sub-frame dictates the column to be in this orientation.
Figure 3.17 Cases where lateral torsional buckling is either possible or impossible for columns subject to in-plane bending moments
Lateral torsional buckling can only occur in the minor 3.8 Second-order frame buckling analysis
axis due to a major axis bending moment. Although it
is usual for elements to have their major axis in the
plane of the dominant moment-frame, this will not As an alternative to completing an element utilisation
always be the case. Lateral torsional buckling can calculation (as introduced in Section 3.5), a second-
therefore occur either in the in- or out-of-plane axis order frame buckling analysis can be completed. This
with respect to a moment frame (Figure 3.17). is largely similar to, and an extension of, the global
second-order analysis introduced in Section 3.6.3.
As an aside to the effective length, the shape of the
bending moment profile can be critical to the major Analogous to an element having a critical buckling
axis capacity My:capacity, potentially influencing this for force Ncr, a frame (whether a 2-dimensional sub-
a particular slenderness. This is shown indicatively in frame, or a complete 3-dimensional frame) subject to
Fig. 3.16. As the figure illustrates, a bending moment a specific load combination will have a critical
profile that causes contraflexure within a segment buckling load. This can be defined as the load which
(as is common in continuous beams and columns of causes the first instance of buckling failure within the
multi-bay/multi-storey moment frames) is far less frame (Figure 3.18).
influential on lateral torsional buckling than a uniform
bending moment (as is common in single-storey/ A separate frame buckling analysis is needed for
single-bay moment frames). each combination of loads. Provided that the smallest
acr (as defined in Fig. 3.18) from each of these
Further reading: lateral torsional buckling
– Gardner, L. Stability of steel beams and columns: in w αcrw
accordance with Eurocodes and the UK National Annexes. N αcrN
SCI Publication P360. Ascot: SCI, 2011
– Access Steel. NCCI: Effective lengths and destabilizing load
parameters for beams and cantilevers – common cases.
SN009a-EN-EU
– Access Steel. NCCI: Elastic critical moment for lateral
Applied load Critical buckling load
torsional buckling. SN003b-EN-EU
where:
Minor axis bending moments Mz:capacity αcr is the minimum load amplifier for the applied loads
The minor axis bending moment Mz:capacity of an to reach the critical buckling load
element will always be the strength capacity of the N, w are applied loads
section, determined via either elastic or plastic
methods as appropriate. No buckling modes affect Figure 3.18 Elastic critical buckling of a 2-dimensional sub-
this. frame

analyses is larger than unity, the structure is

theoretically adequate to prevent buckling. However,
codes apply a factor of safety and define a more
conservative lower bound on acr. BS EN 1993
Part 1-1, for example, states that acr must be greater
than or equal to 3.0 unless a particularly accurate
method of second order analysis is adopted.
It should be noted that a frame buckling analysis

does not negate the need for section utilisation
checks to be performed: i.e. checks using Eqn 3.2
where Ncapacity and/or Mi:capacity are limited by the
section strength and not the buckling limit. Some
software will complete this check, though not all; for
the software to do so, the modelled material
properties must acknowledge the material strength
limit (i.e. the material properties must be modelled as
non-linear).
3.9 Buckling restraint
Moment frames are most efficient when restrained

such that out-of-plane buckling failure is prevented.
This is usually via considered use (position and
detailing) of ‘secondary’ structure.
It should be noted:
– It is usually impossible to restrain against in-plane
buckling without adding bracing elements that
fundamentally change the structure.
– Parallel out-of-plane elements do not provide any
restraint in themselves unless they are either part of
an out-of-plane moment frame (i.e. have moment-
resisting connections) or link to a braced system
(a diaphragm or triangulated frames).
3.10 References

Parts 1 and 2: General philosophy and framed bracing.
London: IStructE Ltd, 2014
3.2 Lim, J.B.P. et al. ‘Eurocode 3 and the in-plane stability

of portal frames’. The Structural Engineer, 83(21),
1 November 2005, pp43-49
3.3 BS EN 1993-1-1: 2005: Eurocode 3: Design of steel

structures. Part 1-1: General rules and rules for
buildings. London: BSI, 2005

concrete structures – Part 1-1: General rules and rules
3.5 Access Steel. NCCI: Buckling lengths of columns:

rigorous approach. SN008a-EN-EU. Available at: http://
www.steel-ncci.co.uk/Clauses/BS-EN-1993-1-1-2005
[Accessed: 3 August 2015]
3.6 PD 6687-1: 2010: Background paper to the National

Annexes to BS EN 1992-1 and BS EN 1992-3.
London: BSI, 2010
3.7 AS 4100-1998: Steel structures. Homebush, NSW:

Standards Australia, 1998

4 Plastic analysis
4.1 Introduction determine a system of viable internal stresses. Such

analysis leads to the evaluation of conservative
failure loads and thus an upper lower-bound can be
Plastic analysis is a method used at the ultimate limit sought
state to determine the load-carrying capacity of
ductile frames. It was pioneered in the early to mid- Despite being the less conservative of the two
20th Century and, today, is commonly used with approaches, the mechanism method is the most
moment frames to achieve maximum efficiency. The appropriate to frame analysis. Engineers must
technique considers the dissipation of energy via recognise this as an upper-bound method – and
ductile deformation at points in the system commonly there may be a need to consider multiple failure
referred to as ‘plastic hinges’. This deformation allows mechanisms in order to obtain the lowest failure
bending moments to be redistributed and, through load. Box 4.1 illustrates this point, considering a
this redistribution, more sections of a system can fairly trivial but immediately obvious scenario of a
become more highly stressed at the point of ultimate uniform beam with two point loads of differing
failure. magnitudes.
Plastic analysis is not, however, universally applicable.

Rather its application is dependent on a number of
prerequisites. These include:
– The material of the structure must have the
Box 4.1 An illustration of the need to correctly
capacity to undergo considerable plastic
identify the critical collapse
deformation without risk of fracture or local buckling
mechanism for use within the
of the cross-section (i.e. the designer must avoid
mechanism method
non-compact element cross-sections)
– The material must have an approximately stable Scenario
yield plateau that extends over a significant
L/3 L/3 L/3
strain
– No in- or out-of-plane instability (element buckling)
can occur prior to yield
w 3w
Plastic analysis is widely used in the UK for steel
frame design. Its use internationally, and with other
materials in the UK, is much less common.
Mechanism 1
Note that a plastic frame analysis is different to a
plastic section design. The latter is used to justify 3δ 3δ
θ= φ=
resistance at the ultimate limit state and is applicable 2L L
in conjunction with either elastic or plastic frame
θ φ
analysis. It is widely used in both steel and reinforced
concrete design. δ
L L
Work done = w θ + 3w φ
4.2 Plastic hinges 3 3
Energy dissipation = Mp(θ + φ)

Plastic analysis concerns the formation of plastic
hinges: points of concentrated deformation resulting 9Mp
where the applied bending moment equals the plastic w=
7L
capacity of the element. In this way it differs
significantly to elastic analysis that requires element Mechanism 2
curvature to remain proportional to the bending 3δ 3δ
moment. θ= φ=
L 2L
There are two principal methods for plastic θ φ
analysis4.1. These are: δ
– The ‘mechanism method’: an upper-bound
procedure that determines the maximum load
corresponding to a collapse mechanism,
considering both equilibrium and compatibility. Equating work done and energy dissipation:
Depending on the mechanism, this maximum load 9M p
will match or exceed the true failure load, and w=
therefore analysis should always look to determine 5L
the lowest upper-bound
Mechanism 1 is proven here to be critical; predicting failure
– The ‘statical method’: a lower-bound method that
at a lower load.
considers only equilibrium and yield limits to

Plastic analysis 4.3
4.3 Rigid-plastic and elastic-plastic

analysis Plastic yield
σ
Plastic analysis can be either ‘rigid-plastic’ or ‘elastic-

plastic’. Both are implemented in an idealised form
but rigid-plastic is a more significant idealisation that
discounts any elastic deformation in advance of
plastic yield. The material model is shown in Elastic stiffness
Figure 4.1.
Elastic-plastic analysis requires non-linear material ε

properties and an analysis algorithm that is iterative.
The common idealisation is to use a bi-linear material Figure 4.2 Non-linear ’elastic perfectly-plastic’ material model
model (Figure 4.2), often labelled an ‘elastic-perfectly
plastic’ model (ignoring such effects as strain
hardening and cracking). Such an analysis can take Then for each mechanism in turn:
account of second-order effects and also determine – Establish work equations in accordance with the
the critical collapse mechanism. mechanism and determine collapse loads from the
plastic moment capacities.
Due to the iterative requirements of a bi-linear model, – Select the collapse mechanism that gives the
elastic-plastic analysis is generally not suitable for lowest collapse load.
hand calculations except for the simplest of elements. – Complete the bending moment diagram (using the
Rather, it is commonly carried out using analytical elastic rules listed in Section 3.2); check that the
software programs. system can achieve equilibrium and that there are
no points between plastic hinges where the
In the absence of an iterative analysis, rigid-plastic bending moment exceeds the plastic section
analysis provides a convenient approximation that capacity.
can be carried out using linear-analysis software
programs or by hand calculations. However, the
material model fails to take account of any elastic 4.4 Collapse mechanisms
deformation prior to yield and therefore cannot
determine second-order effects. It also requires the
collapse mechanism to be pre-defined. Where a frame and/or the actions are more
complicated than those shown in Box 4.1, the critical
Stages to a rigid-plastic mechanism method can be collapse mechanism can be unclear. For example,
summarised as follows: Figure 4.3 shows options for a pitched roof portal
– Determine the format of the mechanisms that can frame subject to coincident vertical and horizontal
potentially form, identifying approximate positions forces. Here, the only mechanism that can be
for the plastic hinges within compatible systems. immediately discounted is mechanism (D), and even
this is dependent on the assumption that both
columns, and separately both rafters, are of matching
sections. All other mechanisms depend on the
relative capacities of the rafters and columns, and on
Plastic yield
σ the relative magnitude of the applied forces V and H.
Where the collapse mechanism is unclear, and/or the

relative rotations are not easily determined for the
work equations, the mechanical concept of
‘instantaneous centres’ is a useful technique to aid
manual rigid-plastic analysis (Figure 4.4).
It should be noted that each of the mechanisms

shown in Fig. 4.3 and 4.4 are what are known as
ε ‘correct’ and ‘complete’ mechanisms4.2. A ‘correct’
mechanism is one in which the rotational deformation
Figure 4.1 Rigid-plastic material model of a hinge is in the direction of the bending moment
(a) (b) (c) (d)
















































Mechanisms
Figure 4.3 Potential collapse mechanisms for a pitched-roof moment frame

4.5 Plastic analysis
further analysis. These mechanisms are collectively

Instantaneous centre known as ‘false’ mechanisms (Figure 4.5).
In accordance with these comments on hinge

correctness, hinges (and indeed redistribution) can
θ only take effect once the plastic moment resistance
h1 of the section is reached. Subsequently, the plastic
moment resistance cannot be exceeded, nor can
redistribution reduce the bending moment back
below the plastic moment resistance.
δ = θh1
δ = θa2
h2 δ = θh1 4.5 Material suitability
h1 a2 h1
θ =θ θ Plastic hinges can only form where an element can
h2 a1 h2 undergo considerable rotation without local failure.
On the basis that global and element buckling are
restrained, this concerns the material ductility and
a1 a2 the element’s cross sectional (local) buckling
resistance.
Figure 4.4 Use of ‘instantaneous centres’ to determine a compatible collapse mechanism
Hinges suitable for plastic analysis must sustain a
from which the hinge forms. Conversely, an ‘incorrect’ reasonably constant rotational stiffness throughout
mechanism is one which needs to deform against the their ductile range. Both steel and reinforced concrete
bending moment; such a mechanism would never can be suitable.
form naturally as a result of component yield.
Standard grades of steel are generally ductile
Meanwhile, a ‘complete’ mechanism is one which when used within standard environmental
has exactly the number of hinges needed to provide conditions and without heightened risk of fatigue.
the necessary degrees of freedom for the mechanism However, cross sectional buckling can often be
to take hold (a number equal to the degrees of critical. BS EN 1993 Part 1-1 Clause 5.54.3
redundancy plus one). A system with an ‘incomplete’ classifies cross section geometries as Class 1, 2,
number of hinges cannot form a mechanism and a 3 or 4. Sections must be Class 1 within twice the
system that is ‘over-complete’ has superfluous section depth of a plastic hinge location in order to
hinges. While theoretically possible, over-complete permit plastic analysis.
mechanisms will never be realised in a real structure
with real-world imperfections. Note that the local buckling that may occur in a steel
section with geometry of Class 2, 3 or 4 is different to
Mechanisms that are either ‘incorrect’ and/or ‘over- the member and frame buckling discussed in
complete’ can be immediately disregarded from Chapter 3. It is an additional failure mechanism.
Frame and loading Elastic bending moment
(a) (b) (c) (d)

Incomplete Complete and correct Complete and incorrect Over-complete

















































Mechanisms
Note that many more hypothetical mechanisms exist for the frame shown. Which mechanisms are ‘false’ will depend on the loading.
Figure 4.5 Example hinge configurations; (a) is an incomplete mechanism, while (c) and (d) are ’false’

Plastic analysis 4.6
Cross section buckling is less of an issue for

reinforced concrete but rotational ductility must be
considered. BS EN 1992 Part 1-1 Clause 5.64.4
defines indirect criteria to confirm rotational capacity
that are dependent on both reinforcement ductility
and area.
Note that it is with rotational ductility in mind that

‘under-reinforced’ concrete sections are always
preferred; this ensures yield of the steel occurs in
advance of brittle crushing of the concrete. Care
must be taken wherever reinforcement is bent or
welded as this can reduce its ductility.
Global plastic analysis is not applicable to timber

structures.
4.6 References
4.1 Megson, T.H.G. Structural and stress analysis. 3rd ed.

Oxford: Butterworth-Heinemann, 2014
4.2 Davies, J.M. ‘False mechanisms in elastic-plastic

analysis’. The Structural Engineer, 66(16), 16 August
1988, p268


concrete structures. Part 1-1: General rules and rules

5 Analytical modelling of joints
5.1 Introduction simplification does not introduce errors that are

significant to the frame analysis. Simple joints are
those assumed to behave as an ideal pin; they
Joints (otherwise referred to as connections, nodes transfer only shear and axial forces. Conversely,
or junctions) play an important role in moment continuous joints are those that are infinitely stiff and/
frames, with both their stiffness and strength or strong. Both are idealisations that render the frame
characteristics influencing the overall frame behaviour dependent on the properties of the
performance. They are also critical elements in the elements only.
design, often being the most complicated and costly
components of the structure to fabricate and
sometimes the heaviest elements in the erection. This
is true in most forms of construction. Joints are 5.3 Classification
typically: regions of high congestion in reinforced
concrete; stiffened, welded and/or cast sections in
steel; and stiffened sections in timber. Classifications set limits on the applicability of ‘simple’
and ‘continuous’ joint models.
Joints are situated between elements. Analytically,
they are the zone within which the behavioural BS EN 19935.1 classifies steel joints based on their
characteristics differ from those of the connecting stiffness and strength. The same classifications are
elements. Their presence is often most pronounced not replicated in BS EN 19925.2 or BS EN 19955.3 for
in assemblies but they can also be of influence in concrete and timber respectively. However, the
monolithic construction. principles of the classifications translate across
materials. Indeed BS EN 1995 Part 1-1 Clause 5.1
Critical to moment frames is the rotational does instruct designers to consider joint slip within
characteristic of the joint. This is the focus of this global analysis, acknowledging that most timber-
chapter. joints are semi-continuous. Classifications are also
not unique to Eurocodes.
It is worth emphasising that the discussion herein and
design-ramifications thereof are as applicable to the Whether stiffness or strength governs the
appraisal and engineering necessary when working classification is a function of the frame analysis
on existing structures as they are in the design of model: (whether it is elastic, rigid-plastic or elastic-
new structures. plastic). Details are listed in Table 5.1, consistent with
BS EN 1993 Part 1-85.4.
Elastic analysis
5.2 Models Elastic analysis concerns elastic material behaviours
and the compatibility of deformations throughout a
system. Thus, a stiffness classification is necessary.
Joints may be modelled as either: This is based on the rotational stiffness of the joint,
– simple usually measured relative to the in-plane flexural
– semi-continuous stiffnesses of the connecting elements. It concerns
– continuous the influence of the joint on the wider stress
distribution. Stiffnesses are classified as either
The behaviours are shown in Figure 5.1. ‘rigid’, ‘semi-rigid’ or ‘nominally-pinned’ (Figure 5.2).
BS EN 1993-1-85.4 defines the cut-off limits
In reality all joints (even pins experiencing between zones 1 and 2, and 2 and 3. Joints that
imperfections and friction) are semi-continuous with can be classed as either ‘rigid’ or ‘nominally-pinned’
an elastic stiffness and plastic yield and/or rupture can be modelled as idealised continuous or simple
capacity. However, in the interest of simplicity, simple components (of infinite or zero stiffness
and continuous idealisations are permitted where the respectively).
(a) (b) (c)

Simple Semi-continuous Continuous





































Unstable Stable and statically indeterminate
Figure 5.1 Joint types (shown with simple base plate joints)

Analytical modelling of joints 5.3
Table 5.1 Joint classifications

Method of global analysis Classification of joint
Elastic Nominally-pinned Rigid Semi-rigid
Rigid-plastic Nominally-pinned Full-strength Partial-strength
Elastic-plastic Nominally-pinned Rigid þ Full-strength Semi-rigid þ Partial-strength
Semi-rigid þ Full-strength
Rigid þ Partial-strength
Type of joint model Simple Continuous Semi-continuous
Note that separate to the use of classifications and models, it is essential that all joints have sufficient strength to transmit the
forces and moments they attract by virtue of their stiffness. Conversely, if a joint is to be designed as anything other than rigid, it is
important that it is modelled as such. This is to ensure that stresses are accurately distributed and catered for through other parts
of the structure.
have a minimum of four bolts to provide a convenient

M level of stability during their erection. The SCI
recommends that nominally pinned bases of this
Rigid nature may be modelled with stiffness up to 20% of
Continuous the column stiffness for serviceability deflection
Semi-rigid analysis and up to 10% of the column stiffness for
assessing second-order sway; both on the
assumption that the foundation has adequate
strength and stiffness to resist the bending moment
transferred5.5. These stiffnesses can be applied as
boundary conditions on nodes for a superstructure
φ
Nominally-pinned Simple analysis model. For timber frames with steel
baseplate assemblies, both the steel-substructure
and timber-steel joint stiffnesses should be
where:
considered.
M is the applied bending moment through the joint
φ is the angular rotation (deformation) of the joint
Rigid-plastic analysis
Rigid-plastic analysis concerns only the strength
Figure 5.2 Stiffness classification of joints classification that is derived from the plastic moment
capacity.
Where a joint is semi-rigid, it is usual to model the As with stiffness, three classifications relate to the
joint behaviour using a linear rotational stiffness model three modelling assumptions (Figure 5.4). ‘Nominally-
when completing an elastic analysis. With this model, pinned’ joints can transfer forces (shear and axial) but
the angular rotation within the joint is linearly can rotate without developing significant bending
proportional to the applied bending moment moments. Here, ‘significant’ is a relative measure
(Figure 5.3). relating to the plastic hinge capacity of the elements;
nominally-pinned joints should be sufficiently ductile
It is worth noting that detailed analysis of deflections to accommodate a rotational strain of 0.03 radians
at the serviceability limit state, as well as assessment (1.7 degrees)5.6.
of second-order sway, can assume some resistance
from nominally pinned connections. This is most In contrast, a ‘full-strength’ joint must have capacity
applicable to steel column baseplates which generally not less than that of the connecting elements to
ensure that plastic hinges form away from the joint.
Any joint that falls between these classifications is

‘partial-strength’ and should be modelled ‘semi-
M continuous’. A graphical representation of this,
analogous to Fig. 5.2, is shown in Fig. 5.4.
M Rd
M
Full-strength
M Rd element
Gradient defines
rotational stiffness Partial-strength
φ
φ Simple Nominally-pinned
Figure 5.3 Linear elastic stiffness model Figure 5.4 Rigid-plastic classification of joints

5.4 Analytical modelling of joints
M Semi-continuous Continuous
Moment resistance of beam Full-strength joint column
Moment resistance of connection
Partial-
Rigid strength
Semi-rigid Non-ductile Ductile
Nominally-pinned Rotation, φ Rotational spring Rigid node
Figure 5.5 Strength and stiffness classification of joints
Elastic-plastic analysis
An elastic-plastic analysis needs to take account of
both stiffness and strength characteristics. As a Figure 5.6 Joint modelling
minimum, this requires a bi-linear moment-stiffness
model for the joint material.
arrangement of the bolts are each critical. The
Figure 5.5 amalgamates Figs 5.2 and 5.4. It also detailing of reinforcement is equally influential for
plots typical true and idealised characteristics for a reinforced concrete.
semi-rigid partial-strength steel connection.
It is not always possible to provide a joint that is as
Further reading: analytical modelling of joints strong or stiff as the members, and analytical models
must reflect this. It is particularly common in light
– Access Steel. NCCI: Column base stiffness for global gauge steel and timber construction.
analysis. SN045a-EN-EU
Further reading: joint design

5.4 Modelling practice Steel
– Steel Construction Institute and British Constructional
Steelwork Association. Joints in steel construction:
Simple and/or continuous joints are almost always moment-resisting joints to Eurocode 3. SCI Publication
the default settings in structural analysis software. P398. Ascot: SCI, 2013
Traditional analysis packages which attribute section – Steel Construction Institute and British Constructional
properties to 1-dimensional elements tend to have Steelwork Association. Joints in steel construction: simple
joints continuous by default, while some of the more joints to Eurocode 3. SCI Publication P358. Ascot: SCI,
modern ‘BIM’ (Building Information Modelling) 2011 [Reprinted 2014]
packages identify components by type (e.g. beams,
Timber
columns, etc.) and set different defaults to different
– Ross, P. et al. Concise illustrated guide to timber
elements. Irrespective of the software, it is important
connections. High Wycombe: TRADA Technology, 2012
that the engineer understands the way in which
element joints are modelled and corrects the defaults Concrete
as necessary. – Institution of Structural Engineers and The Concrete
Society. Standard method of detailing structural concrete:
Where joints of three or more branching elements are A manual for best practice. 3rd ed. London: IStructE Ltd,
not fully continuous for all elements, it is important 2006
that they are represented in a manner that does not
oversimplify the load path. In this regard, it is
important that the cases shown in Figure 5.6 by way
of example, can be differentiated within the analysis.
BS EN 1993-1-8 Clause 5.3 gives further guidance 5.6 References
on this.

5.5 Linking analysis to design buildings. London: BSI, 2005

In practical terms, joint classification is dependent on concrete structures. Part 1-1: General rules and rules
the nature of the joint, together with the connecting for buildings. London: BSI, 2004
elements. However, more often than not, it is
governed by the detailing at the connection. 5.3 BS EN 1995-1-1: 2004 þ A1: 2008: Eurocode 5:
Design of timber structures – Part 1-1: General –
For steelwork (and similarly for timber) the use of Common rules and rules for buildings. London: BSI,
stiffeners, choice of plate thicknesses or the 2009

Analytical modelling of joints 5.6

structures. Design of joints. London: BSI, 2005
5.5 Koschmidder, D.M. and Brown, D.G. Elastic design of

single-span steel portal frame buildings to Eurocode 3.
5.6 Salter, P.R., Couchman G.H. and Anderson D. Wind-

moment Design of Low Rise Frames. SCI Publication
P263. Ascot: SCI, 1999

6 Simplified analysis methods
6.1 Introduction of that book have been translated into English6.1 and
direct extracts of the content have been adopted by
the SCI6.2,6.3. Similar data is contained in Reynolds’s
It is widely recognised that the analysis and design of Reinforced Concrete Designer’s Handbook6.4.
frames is increasingly being completed by purpose-
written software. Indeed computer-aided design has Kleinlogel set out algebraic equations for salient
made manual calculations almost redundant for values within statically indeterminate frames; an
detailed design. Manual methods do, however, remain example of which is contained in Box 6.1. Critical
valued for initial sizing of members and model underlying assumptions include:
verification. This chapter looks at common techniques – Linear elastic behaviour of all elements
for both the ultimate and serviceability limit states. – Infinite stiffness (rigid) moment-resisting joints
– Ideal (rigid/free) boundary conditions with regard to
both rotational and translational stiffness
– Axial and shear deformations are neglected
6.2 Formulaic elastic analysis of – Constant element properties along the full length of
each element
standardised frames
Global second-order PD effects are neglected, and
Kleinlogel first published Rahmenformeln in 1913, the analysis makes no attempt to consider section-
listing design formulae for standardised rigid, single- or buckling-capacity limits. However, being linear,
span frames subject to regular actions. Later editions superposition of results is possible, allowing the
effects of different loads to be combined.
Box 6.1 Extract from Kleinlogel’s Rahmenformeln Another publication, Multibay frames: ready-to-use
formulas...6.5, provides similar, although less
simplistic, formulae for multi-span frames.
6.3 Graph-based analysis of standardised

frames
A graphical method can be used for analysis of

standardised frames. This is largely the same as the
formulaic method outlined in Section 6.2.
The method is used primarily for initial trial sizing of

elements in steel single-storey sheds for which simple
graphs can relate the primary variables: column
height, rafter span and rafter pitch. The Institution’s
Manual for the design of steelwork building structures
to Eurocode 36.6 provides a detailed procedure and
graphs that are relevant to symmetrical duo-pitch
single-span frames. These were derived via a plastic
frame analysis; the notes contained within
Section 4.5 concerning the suitability of plastic
analysis apply.
6.4 Wind-moment frame ultimate limit

state analysis
‘Wind-moment frame analysis’ is a phrase used

throughout the UK for a particular analytical
technique. It is applicable to both single- and multi-
storey buildings.
The key benefit of the technique is its simplicity. It

sets a procedure for analysing statically indeterminate
2-dimensional frames via statically determinate
Note that the notation here is copied directly from the referenced source and is not
equivalent models. Thus it allows analysis to be
necessarily consistent with the notation used elsewhere in this Guide.
completed manually without concern for elastic or

Simplified analysis methods 6.4
plastic theory, and without needing to estimate

member stiffnesses in advance. The technique has Assumed pins
been well documented by the SCI with a series of
publications, the latest of which is Wind-moment
Design of Low Rise Frames published in 19996.7.
This references BS 59506.8.
Vertical
The method remains useful for computer model actions
verification and scheme design. It generally produces
conservative designs for buildings up to four storeys.
It is also appropriate for use with all materials.
Two simplifying assumptions that define the

technique are now considered. The suitability and
limitations of the method are then discussed.
Assumption 1: Statically determinate ‘equivalent’

structures Horizontal
The wind-moment frame method requires that two actions
models are analysed, with each introducing
simplifications into the frames. One model is for
vertical actions and the other for horizontal actions.
Each model places pins in the frame, introducing
degrees of freedom that render the frame statically
determinate (Figure 6.1). Actions Bending moment
For vertical loads, the frame is assumed to be
Figure 6.1 Wind-moment equivalent frame models for vertical actions (top) and horizontal
‘simple’ and the beams can be sized for a design
actions (bottom)
bending moment of wL2/8 and a deflection of
5wL4/384EI. At the ultimate limit state, 10% moment
redistribution is typically assumed to reduce peak shown for a typical storey within a generic (but
sagging bending moments (to 0.9wL2/8)6.7. The regular) frame in Box 6.2. Figure 6.3 shows how
redistributed hogging bending moment (0.1wL2/8), column base fixity can affect the ground floor
together with any moment resulting from eccentricity columns.
in the joint, must be transferred by the joints and
accommodated by the connecting columns It should be noted that the level of the foundations,
(Figure 6.2). as well as the base joint fixity, must be taken into
account when evaluating the ground storey column
For horizontal loads, pins would ideally be positioned bending moments. The level of the column base
at points of contraflexure. However, in practice it is connection, which is usually in excess of 200mm
standard to assume they are at mid-spans of beams below the slab finish level, will define the column
and mid-storey of continuous columns. Columns length denoted ‘h’ in Fig. 6.3.
pinned at ground are exceptions that do not require
additional pins. Assumption 2: Second-order PD effects
The sub-frame analysis method (Section 3.6.3) is
Internal forces and bending moments can be used to evaluate in-plane effective lengths. This
determined for each load condition by considering makes allowance for the sway instability but falls
free-body parts to the frames, knowing that the short of the accuracy of an explicit second-order
bending moment at a pin must be zero. This is analysis.
0.1wL2
w
8
0.9wL2
2
wL 8
8
Idealisation Redistributed moment

L
Figure 6.2 Redistribution of bending moment from vertical loads

6.4 Simplified analysis methods
Box 6.2 Derivation of internal bending moments and shear forces resulting from horizontal forces
‘n ’ columns























L L L L L
F1
h
F2
.
.
. h
Fi
h
Fi + 1
External column (E)

Internal column (I)
L/2 L
(i –1) (i –1)
VCI(i – 1) = ∑1 F ik I VCE(i – 1) = ∑1 Fi k E
•
h/2 h h M Bi = M CEi – M CE(i – 1)

M CI(i – 1) = VCI(i – 1) M CE(i – 1) = VCE(i – 1)
2 2
i th floor
M
VBi = Bi M CEi = VCEi
h
h (L /2) 2
M CIi = VCIi
2
h i i
VCI(i) = ∑1 F ikI VCE(i) = ∑1 Fi k E •
M B(i +1) = MCE(i +1) – M CEi
(i +1 floor)
M
VB(i +1) = B(i +1) M CE(i +1) = VCE(i +1)
h
h/2 (L /2) 2
h
M CI(i +1) = VCI(i +1)
2
(i +1) (i +1)
VCI(i +1) = ∑1 Fi k I VCE(i +1) = ∑1 Fi k E
where:
kI 1/(n 1) when columns and beams are of similar stiffness
1/n when beam stiffness column stiffness
kE 1/[2(n 1)] when columns and beams are of similar stiffness
1/n when beam stiffness column stiffness
n is the number of columns
Procedure:
1. Determine shear forces in the columns Vc
2. Determine the bending moments in the columns Mc ¼ Vc h/2
3. Determine the bending moment necessary in the beams to maintain moment-equilibrium at the beam-column joints MB
4. Determine the shear forces in the beams V B ¼ M B /(L/2)
Note that the expressions listed for k I and k E assume all columns (and separately all beams for a given storey or floor) are of equal
stiffness. For frames where the element stiffnesses vary (either by virtue of their cross section or length), the column shear forces
must be apportioned in accordance with this stiffness.

MCE . VCE . h/2 MCE = VCE . h

MCI . VCI . h/2 MCI = VCI . h
h h
VCI VCE VCI VCE

Fixed bases Pinned bases
Figure 6.3 Column base fixity and ground floor column bending moments
An amplification (sway modification) of the first-order For a design to comply with the first assumption, it is
horizontal actions can be applied. This can be in line important that the joints are:
with both the guidance given in Stability of buildings – Sufficiently strong and stiff to resist moments and
Parts 1 and 26.9, and by using the deflection deflections owing to lateral actions
calculation outlined in Section 6.5 of this publication. – Sufficiently ductile in rotation to accommodate
However, in the absence of any insight into the sway curvature in the beams under vertical actions
sensitivity, it is recommended that a 15% without attracting bending moments that exceed
amplification is applied to horizontal actions for a first- those derived for lateral actions
pass analysis.
Satisfactory connections are generally semi-
The amplification will increase VC and therefore continuous (semi-rigid þ partial-strength) as
increase MC , MB and VB , (each as defined in described in Section 5.3.
Box 6.2). However, it will have no impact on the
design axial force in the columns.
Suitability and limitations 6.5 Serviceability sway deflection checks

The SCI advises that the wind frame technique is only
suitable for low rise buildings up to four storeys6.7.
This is, in part, owing to the limitations of the second Box 6.3 sets out a method for calculating the
assumption. approximate sway deflection of a rigid jointed frame.
Box 6.3 Sway deflection of rigid jointed frame

The following guidelines set out one method for approximating the sway deflection of a rigid jointed frame. Results can be amplified
for semi-rigid joints. The method makes no allowance for second-order PD effects, although force amplification can be applied
where applicable.
The first step is to convert the actual frame to an equivalent ‘summed’ frame and determine equivalent beam and column
stiffnesses kB and kC respectively:
Li Lii Liii
F3 F3 kB3
kC3 h3
F2 F2 kB2
kC2 h2
F1 F1 kB1
kC1 h1

I I
Beams: kB ¼ 3S B:i Columns: kC ¼ S C:i
LB:i h C:i
where:
I B:i is the in-plane flexural stiffness of the i’th beam
LB:i is the length of the i’th beam
I C:i is the in-plane flexural stiffness of the i’th column
h C:i is the height of the i’th column

6.5 Simplified analysis methods
Box 6.3 Continued

Joint stiffness coefficients can subsequently be determined for each end (top and bottom) of each column in the equivalent frame
(i.e. at each storey). These are coefficients used to approximate the distribution of forces through the frame; they are not to be
confused with the rigidity of the joint (which is assumed infinite).
kCt
k Bt
kt
kC
kb
k Bb
kCb
kC þ kCt kC þ kCb
Top: kt ¼ Bottom: k b ¼
kC þ kCt þ kBt kC þ kCb þ kBb
where:
kC is the stiffness of the equivalent column at a given storey
kCt is the stiffness of the equivalent column above the given storey
kCb is the stiffness of the equivalent column below the given storey
kBt is the stiffness of the equivalent beam above the given storey
kBb is the stiffness of the equivalent beam below the given storey
Note that kCt ¼ 0 for the upper-most storey, kb ¼ 0 for a fixed base and kCb ¼ kBb ¼ 0 for a pinned base.
Using kt and kb , the sway index f can be read from the following graph for each storey:
Finally, the inter-storey sway deflection D can be calculated for each storey:
Fj h 2j fj
Dj ¼
12EkCj
where:
Fj is the total (cumulative) horizontal force acting on the equivalent column at the j’th storey
hj is the storey height
E is the Young’s elastic modulus

A crude approach to combat the shortfall in stiffness

owing to semi-rigid joints is to derive the sway
assuming a rigid joint system and subsequently apply
an amplification factor to the deflection. The SCI6.7
recommends this amplification is:
– 1.5 for column bays larger than or equal to 6.0m
– 2.0 for column bays less than or equal to 4.5m
Linear interpolation should be used for column bays

of between 4.5–6.0m.
Note that these factors are derived for steel frames

with standard joint types. They also assume a storey
height in the range 3.5–5.0m and a bay width to
storey height ratio in the range 0.9–3.0.
6.6 References
6.1 Kleinlogel, A. Rigid-frame formulas. 2nd ed. New York:

Ungar, 1958 [Translated from the German
Rahmenformeln, 12th ed]
6.2 Davison, B. and Owens, G.W. Eds. Steel Designers’

Manual. 7th ed. Chichester: Wiley-Blackwell, 2012

6.4 Reynolds, C.E. et al. Reynolds’s Reinforced Concrete

Designer’s Handbook. 11th ed. Abingdon: Taylor &
Francis, 2008
6.5 Kleinlogel, A. and Haselbach, A. Multibay frames:

Ready-to-use formulas for single- and two-storey
continuous frames comprising any number of bays,
elastically restrained at the supports and with or
without sidesway. London: Crosby Lockwood, 1963
[Translated from the German Mehrfeldrahmen, 7th ed]
6.6 Institution of Structural Engineers. Manual for the

design of steelwork building structures to Eurocode 3.
6.7 Steel Construction Institute. Wind-moment Design of

Low Rise Frames. SCI Publication P263. Ascot: SCI,
1999
6.8 BS 5950: Part 1: 1990: Structural use of steelwork in

building. Part 1: Code of practice for design in simple
and continuous construction: hot rolled sections.
London: BSI, 1990


7 Moment frames in application
7.1 Single-storey pitched roof ‘sheds’ axis aligned to the gable ends and a longitudinal axis
aligned to the eaves.
7.1.1 Introduction Standard form

The standard form is a battery of planar portal frames,
Single-storey pitched roof moment frames are very each identical and spanning in the transverse axis.
economical structures used widely for agricultural, These frames comprise pitched rafters and vertical
industrial, retail and leisure facilities. They are quick to columns (also known as stanchions) connected with
erect, efficient in their material embodiment, cost moment-resisting joints at each of the eaves and ridge.
effective and ultimately provide large, unobstructed
accommodation that is flexible in use. As shown in Figure 7.2, column bases can be pinned
or nominally pinned (creating ‘2-pin’ portals), and
The buildings are often referred to as ‘sheds’ semi-continuous or continuous (creating ‘fixed-base’
because they can be devoid of internal partitions. portals). Nominally pinned connections are usually
Although the phrase ‘portal frame’ can be used for preferred as these reduce the extent to which the
any moment frame structure, it is most synonymous frame is dependent on the ground stiffness, while still
with this form (possibly because of the sheer providing some stability in construction. Semi-
abundance of these unmistakeable structures). continuous and continuous connections can however
increase overall frame stiffness and lead to a more
These structures tend to be highly refined, lean efficient superstructure.
systems that target material efficiency above all else.
This section (7.1) discusses the salient features and Portal frames are usually spaced at a regular interval
considerations. (‘bay’ or ‘grid’ spacing) in the longitudinal direction. Roof
purlins span parallel to one another between frames
7.1.2 Overview and support the roof covering which is usually one-way
spanning. Similar cladding rails (sometimes referred to
Figure 7.1 illustrates a basic single-span duo-pitch as ‘girts’) span horizontally between columns on the
shed annotated with two dominant axes: a transverse eaves elevations and carry the wall cladding.
Bay


Purlin

Ridge


Portal frame rafter

Eaves
Roof-plane bracing
Eaves elevation
Gable end
Cladding rails
Wind posts
Longitudinal axis
Transverse axis
Portal frame column
Vertical bracing Haunch
Figure 7.1 Single-span duo-pitch shed
(a) 2-pin frame (b) Fixed-base frame
Figure 7.2 2-pin and fixed-base portal frames

Moment frames in application 7.1
Efficiency
Simply supported The span, column height and roof pitch collectively
rafters influence the efficiency of the portal frames and more
broadly the building as a whole. A well-conditioned
structure is one that:
– Gains benefit from thrust within the pitched rafters
without excessively penalising the columns
(Box 7.1)
– Has elements restrained against buckling
– Is not unduly sensitive to either sway or splay
Braced gable bay deformation
– Is not unduly tall and susceptible to excessive wind
loads
Load-bearing columns – Is not unduly voluminous and inefficient to heat or
cool
– Does not have an excessively large envelope (either
Figure 7.3 Single-span duo-pitch shed with a braced gable
due to eaves height or roof pitch) that is inefficient
end
to clad
– Meets all internal and external operational
Transverse gable end bays generally also have girts requirements including minimum clear headroom,
spanning horizontally. These end bays may contain column-free clear spans etc.
portal frames that are similar or identical to the – Achieves necessary roof drainage
intermediate ones; in which case vertical elements
can be non-loadbearing windposts (Fig. 7.1). The portal frame spacing (bay size) should be
Alternatively, gable ends can often be braced in plane decided by considering:
with diagonal bracing and load-bearing intermediate – The span capacity and size of the cladding rails/
columns (Figure 7.3). Despite the additional elements, purlins
this is often a more cost-effective solution that allows – The overall structural depth of the rafters (including
end rafters to be a series of shorter simply supported haunches) together with the purlins and the
beams. The bracing also adds stiffness to the façade influence this has on the building height
and can be an economical solution when the – The size and quantity of the foundations
cladding system demands a deflection limit that is – The operational requirements of the facility and the
more onerous than that adopted for the internal bays. restrictions these impose on column centres
Box 7.1 Pitched-roof frame efficiency

The efficiency of a pitched roof portal frame can be illustrated by considering the
(A) sequence A to G:
– A continuous beam (B) subject to a middle-span uniform load is more efficient
(B) than an equivalent simply supported frame (A).
(C) – Converting the continuous beam into a goal-post frame (C), bending moment
continuity applies around the corners and the bending moment profile remains
unchanged. However, to achieve force equilibrium, shear and axial forces
develop at the eaves (D) that are compatible with the action and the bending
moment (Section 3.2).
(D)
(E) – The shear and axial force acting on the rafter at the eaves create a diagonal
resultant that is most efficiently resisted when aligned as a pure axial force in
the rafter (E).
(F.1) – However, the horizontal thrust that develops from this pitch and the bending
moment at the eaves are collectively dependent on the overall frame geometry
and stiffness. Thus there is an optimal balance (G), considering each of the
(F.2) rafters, the columns, and the joints, that lies between the four idealisations
(F1 to F4).
(F.3)
(F.4) Notes
– With the loading shown, the only stress resultant that is independent of the
frame stiffness is the axial force in the columns which is a function of the span
and load magnitude only.
(G)
– The example illustrates a 2-pin portal frame. Similar can be presented for
fixed-base and 3-pin frames alike.

7.1 Moment frames in application
Tied frame These frames include a tie usually positioned at eaves level to limit splaying of the
eaves. This can be particularly favourable for structures with travelling cranes.
The tie will reduce the bending moments in the columns and rafters but can add
significantly to the compressive force in the rafters. It adds a degree of
indeterminacy to the analysis, and each of the relative stiffnesses of the elements
and joints, the construction sequence and any locked-in stresses (including any
introduced by post-tensioning) should be considered.
Lateral forces can be far more significant on the design of tied systems because
the tie locks up the pitched roof.
Propped frames Props can be used to reduce the bending moment resulting from vertical loads,
allowing reduced element sizes. They are suited to large spans, where the ridge
height is not critical and the internal columns can be accommodated.
Props are usually simple pin-ended struts that do not contribute to the lateral
resistance. Any sway results in a destabilising second-order effect on the portal
frame.
Mansard and curved frames Mansard (faceted) and curved frames are used primarily for aesthetic reasons.
They can also reduce overall height and internal volume; the latter offering a
potential saving for heating and cooling mechanical services.
It should be noted that ribbon-cut steel cellular beams can often be curved at
little/no additional cost. These sections cannot develop plastic hinges, preventing
the use of plastic analysis.
Mono-pitched frames Mono-pitched frames are used primarily for short spans or where site constraints
prevent drainage from a duo-pitch roof.
These frames can be particularly sensitive to asymmetric wind loading.
Note
– Each statement of a relative nature is a comparison to the basic frame shown in Fig. 7.1.
– In each instance where the bending moments are said to be reduced, column base shear will also be reduced (Section 3.2).
– Normal span ranges are qualified for various materials in Section 7.1.6.
Figure 7.4 Variations on principal frame
Variations on the standard form only the transverse moment frames and are
Variations on the standard 2-pin and fixed-base commented on in turn in Figure 7.4.
portals include tied, propped, mansard, curved and
mono-pitched frames. These variants each concern Two further variations on the transverse portal frame
are:
– A multi-span frame (Figure 7.5)
– A 3-pin frame (Figure 7.6)
These can be used in tandem with all variations, bar

the mono-pitched frame (listed in Fig. 7.4).
Valley
Ridge
Eaves
Figure 7.5 Multi-span frame Figure 7.6 3-pin portal frame

3-pin frames are rare in steel framed structures but

common in both precast concrete and timber.
7.1.3 Load paths and global stability
Vertical load path

Vertical actions transfer from the cladding to the
purlins, main rafters and columns, and into the
foundations.
Net uplift due to wind is commonplace given the

lightweight nature of these structures together with
heightened internal wind pressures that can occur
when the envelope has dominant openings. Global
Eaves elevation
net uplift is normally resisted by the self-weight of the
moment frames
foundations (discounting hydrostatic buoyancy where
the ground may become saturated). Locally it must Transverse axis
Longitudinal axis
be considered within the element and joint design
and ‘load reversal’ is commonplace.
Figure 7.7 Longitudinally unbraced shed
It should be noted that the hydrostatic buoyancy
force is dependent on the permeability of the soil and
it may not be necessary to consider this with short longitudinal steelwork in the roof. These elements act
duration wind loads. Geotechnical engineers should as ties, fixing intermediate frames to the bracing at
be consulted in this regard to provide site specific each end of the building; they must be designed for
advice. the cumulative axial force acting through successive
bays.
Transverse load path
Lateral actions acting in the transverse axis are Vertical stability provided in the eave elevations can
resisted by each of the individual portal frames, take either a braced or unbraced form. Bracing
together with any bracing provided in the gable ends. (framed bracing or shear walls) is usually preferred as
a more efficient and often cheaper option.
With each frame self-stable (as either a portal or with Consequently, unbraced (moment frame) systems
bracing in the gables), there is no need for a roof tend to be used only where bracing cannot be
diaphragm or for a longitudinal arrangement of roof accommodated. Where an eaves moment frame is
bracing. Furthermore, the overall longitudinal length of needed, this tends to run the full length of the
the building (defined by the number of bays) has very building and generally invokes minor-axis bending in
little impact on the transverse stability; only the the columns (Figures 7.7 and 3.17b).
spacing of the portal frames is significant as this
defines the tributary load per frame. The local Where bracing is provided in the eaves, the
magnitude of the lateral force is, however, critical. foundations and column sizes immediately flanking
Peak wind pressures tend to occur at a building’s the braced bays need to be sufficient to take the
corners and act most heavily on the final and concentration of load that the bracing creates.
penultimate transverse frames. Other loads (e.g. the
notional forces from a mezzanine) may lead to 7.1.4 Structural hierarchy and local stability
another frame being more critically loaded.
Hierarchy
Longitudinal load path As with other framed structures, it is tradition to label
Lateral actions acting in the longitudinal axis act the elements within a shed with hierarchical titles.
primarily on the face of the gable ends, typically ‘Primary’ elements are generally those making up the
transferring vertically to each of the ground and roof transverse portal frames together with roof plane and
planes. This transfer is often via vertical-spanning vertical bracing elements and the substructure. Other
cladding panels, horizontal-spanning cladding rails elements are traditionally labelled as being
and vertical-spanning wind-posts or columns. ‘secondary’.
Friction on the eaves and roof surfaces can add a This labelling is in keeping with global load paths; it is
further significant component of lateral action and also usually in keeping with the scale of the
is critical for very large buildings with large surface components and with the erection sequence.
area. However, it downplays the interdependency of the
elements with regard to out-of-plane buckling
Actions on the roof plane must transfer laterally to the restraint.
eave elevations. This can be via diaphragm action of
the roof sheeting but is usually via in-plane roof Careful coordination of primary, secondary and
bracing (a ‘wind girder’) spanning from eave to eave. cladding systems can lead to frames being designed
This is commonly positioned in the end bay to most efficiently. An approach working ‘from the
provide immediate support to the gable elevation outside in’ is often optimal for the derivation of
without inducing unnecessary longitudinal resistance and derivation of actions simultaneously
compression in the roof plane. (Figure 7.8). Indeed this workflow could be
advantageous for many types of building. However,
Additional lateral actions caused by frame the simplicity of sheds means that they are often
imperfections (equivalent horizontal forces) acting on commoditised systems. This allows a highly refined,
the intermediate portal frames may be resisted by integrated design to be developed without
tension in the purlins and/or tension in any additional unreconciled design assumptions.

Operational performance requirements

Envelope requirements (insulation, Self-weight of suspended services
Critical dimensions
permeability dominant openings etc.) and/or plant
Environmental actions
Wind, snow etc.
Cladding design
Span limit defines the spacing of the secondary elements. Self-weight established
Secondary out-of-plane structure design
Span limit defines the spacing of the portal frames. Self-weight established
Primary portal frame superstructure and foundation design
Figure 7.8 Shed design workflow
Local stability of the primary moment frame The rails provide direct restraint to the outer
The primary moment frame will be most efficient with extremities of the section to which they are fixed.
elements that are: However, the bending moment envelope (Figure 7.10)
– Orientated to bend about their major axis dictates that additional restraint is often needed to
– Restrained against out-of-plane Euler and lateral the inner extremities along most of the frame. This is
torsional buckling traditionally provided by ‘fly bracing’: steel ties that
are fixed at 458 and locally brace the inner extremities
Cladding rails and purlins complete with bracing can of the portal frame elements to the cladding rails
provide a component of this restraint. These fix to the (Figure 7.11).
outer extremities of the elements (Figure 7.9).
It should be noted that parallel purlins and cladding
rails only provide restraint where they are braced
against racking (Figure 7.12). The bracing that is used
is commonly the same wind girder system used to
Purlins brace the gable elevations. However, care is needed
where this does not node out with each purlin.
It can be impossible or highly impractical to brace

columns when large openings (e.g. loading bays)
Rafter Negative pressure Positive pressure

bending moment bending moment
envelope envelope
Cladding
rails Haunch
Column
Figure 7.10 Bending moment envelope (shown for

Figure 7.9 Arrangement of cladding rails and purlins, showing cold formed thin-gauge symmetrical positive and negative vertical actions only and
steel profiles within a 2-pin frame)

Cladding rail or purlin Bracing provides critical buckling

restraint to primary frame and
prevents purlins from racking
Purlin
Fly bracing
Primary frame rafter or column Primary frame

Figure 7.11 Fly bracing (shown for a steel moment frame)
Buckling of
parallel rafters
interrupt cladding rails in the elevation (Figure 7.13).
where unbraced
The same is true for internal columns to propped or
multi-bay frames. In these scenarios the columns
must be considered unrestrained in both their major
and minor axes.
7.1.5 Plastic analysis
Plastic analysis is widely regarded as the most

efficient means by which to design portal frames
where applicable. For a standard frame with pinned
base-plates, a minimum of two hinges are needed to
achieve a complete mechanism. A typical mechanism Purlins do not node out
is shown (Figure 7.14(a)). with bracing and are
therefore free to rack
Note that the symmetrical mechanism shown in (b) –
and often published in design literature – is an over- Figure 7.12 Adequate (top) and inadequate (bottom) buckling
complete mechanism (Section 3.3). restraint of parallel frames
Discontinuous side rail

twists with the column
Unrestrained columns
with discontinuous Column
side rails
Section through column and
discontinuous façade frame
Figure 7.13 Unrestrained columns

Instantaneous centre
Plastic hinges
below haunches
(a) (b)
With combined vertical With vertical load only
and horizontal loading
Figure 7.14 Plastic hinge mechanisms for standard portal frame sheds
(Classes 1 and 2 being defined in BS EN 1993

Further reading: plastic analysis of portal frame sheds
Part 1-17.5).
– King, C.M. Plastic design of single-storey pitched-roof
portal frames to Eurocode 3. SCI Publication P147. Ascot: Fabricated sections may be used where needed to
SCI, 1995 achieve section properties and/or meet geometric
requirements that are not possible with standard
rolled sections. These may include built up sections
7.1.6 Frame materials
often made from two or more standard rolled
sections (e.g. nested columns that are ideal for
Frames in the UK are predominantly steel. Outside
unrestrained elements – particularly columns that
the UK, reinforced concrete and various engineered
contribute to moment frames in two axes
timber products are competitive and command a
(Figure 7.15)).
significant market share.
Unless there is good reason not to (e.g. aesthetic
7.1.6.1 Steel
considerations), columns and rafters should be
Steel frames are efficient for spans 15–50m, with
independently sized to maximise efficiency. The
columns 5–15m tall and with rafters set 5–108 to the
stiffness of the column is usually in the region of 50%
horizontal7.1.
greater than that of the rafters, although the optimum
will vary with the height to span ratio.
The portal frames are usually spaced at 5–8m
centres; this is a range well catered for with light-
Haunches are standard at each of the eaves and
gauge cold-formed purlins and cladding rails.
ridge joints. These cater for the peak bending
moments and ensure joints can be ‘full strength’
Material grade
(Section 5.3) without reducing the efficiency of less
Where the local steel industry produces more than
highly-stressed sections. Eaves haunches typically
one ‘standard’ grade (e.g. S275 and S355 in the UK),
extend 10% of the total span for a duo-pitch frame,
the lower of the two has traditionally been considered
at which length the hogging moment at the haunch
more economical for shed structures7.2. This is mainly
owing to the fact that deflection criteria tend to
govern even when quite large deflections are
accepted. However, more recent guidance7.3
suggests higher grades are increasingly more
economical as the cost differential between grades
has diminished.
Steel sub-grades should be selected to suit the

exposure conditions. Most sheds in the UK should
use category ‘welded generally’ as defined by
PD 6695-1-107.4, though this may not be suitable for
cold stores, un-conditioned (external), or other
onerous environmental exposure conditions.
Sections
Primary portal frames will often use standard rolled
sections, usually I-section (universal beams)
orientated with major axis orthogonal to the in-plane
bending moments. Where plastic design is adopted,
sections must be Class 1 ‘plastic’ at the positions of Figure 7.15 Nested steel column, used to enhance minor axis
plastic hinges, and Class 2 ‘compact’ elsewhere buckling resistance and stiffness

Haunch typically cut

Figure 7.17 Notre Dame Catholic College, Liverpool
from same section
as rafter
Figure 7.16 Steel frame eaves haunch depth profiled to include general tapers and/or local
haunches.
nib is approximately equal to the sagging moment Standard glulam sections are limited to a maximum
closer to the rafter midspan. Eaves haunches are breadth of 185mm (the largest standard cross-
usually cut from the same standard section as that of section dimension for planed timber). Although the
the rafter (Figure 7.16). This is an industry norm. depth is largely unconstrained, lateral instability can
make very deep sections inefficient.
Apex haunches tend to be fabricated from plate.
Their size tends to be dominated by the requirements Plywood web beams are a far less aesthetically
of the bolted rafter splice, acknowledging the access pleasing solution than either LVL or glulam but are
requirements to fit and tighten bolts (in addition to the most economical. Larger spans generally use a
strength and stiffness requirements). box construction with glulam, LVL or sawn timber
flanges, connected with face-fixed plywood webs
Connections (Figure 7.18). Elements are suitable for span of the
Each of the eaves and ridge connections tend to be order of 40m and can be site-spliced. For optimal
bolted end-plate connections with all bolts contained efficiency, the depth of the web should be
within the depth of the rafter and haunch combined. receptive to the standard dimensions of locally
available plywood sheets (2400 1200mm in the
Baseplates are usually nominally pinned with bolts set UK) – i.e. a section that is 2600mm deep could be
between the column flanges. significantly more expensive than one that is
2400mm deep.
7.1.6.2 Timber
Sawn timber, glue laminated timber (glulam), Purlins tend to be sawn timber or plywood webbed
laminated veneer lumber (LVL) and plywood webbed beams and are economical for spans in the range
beams are all common for portal frames. 3.6–6m. They tend to be simply supported and sit
within the depth of the rafters. This generally means
Load-duration factors mean that the dead plus long- their depth is inconsequential (neither adding anything
term imposed loads often govern. Timber is therefore to the overall building height nor infringing on the
often most competitive in regions with relatively high clear headroom). Where sufficiently deep, the purlins
short-duration wind and/or seismic loads.
Sawn timber and engineered products

Sawn timber is practical for duo-pitch spans less
than 12m (i.e. rafters up to 6m); beyond this limit the
raw material is simply not available in sufficient
quantity to be economical. Engineered products are
far more versatile – being available with larger all-
round dimensions (length, breadth and depth).
Further advantages of engineered wood products
include:
– Reduction in imperfections/greater performance
uniformity
– Opportunity to tailor strength grades throughout the
section to minimise the use of high strength (high
cost) timber
– Greater geometrical stability
Both glulam and LVL are solid products that are well
suited to moderately-loaded portal frames, especially
those where the appearance of the structure is critical
(Figure 7.17). Both are suitable for spans up to
approximately 35m and can have sections with their Figure 7.18 Plywood web beam of a box-format (with cavity end blanking piece)

Glulam rafter Threaded rod anchored Steel or timber gusset or

into column with resin concealed flitch plates
Eaves
Glulam column
(a) (b)
Through-bolts
Ridge
(c) (d)
Figure 7.19 Moment-resisting timber connections with steel inserts
can provide torsional restraint to the rafters without Bracing

need for fly bracing. Both vertical and in-plane roof bracing elements are
usually steel rod or cable, crossed in a Modified Fink
Connections truss format. Further information can be found in
While factory-made moment-resisting glued ‘finger’ Stability of buildings Parts 1 and 27.6.
joints, tight-radius glulam beams, and site-glued
‘butt’ joints are possible, most portal frames adopt Erection
site-bolted joints with steel inserts. These generally It is standard practice to assemble pairs of timber
take of one of two forms: rafters, together with purlins, at ground level and lift
– Timber on timber bearing with tensile threaded rod as an assembly into place (Figure 7.20). This can lead
set in the plane of the frame to fast, safe construction with minimal work at height.
– Timber or steel gusset plates with screws or It also minimises the risk of unfavourable out-of-plane
through-bolts acting in shear buckling of the rafters during lifting.
These are shown for each of moment-resisting eaves 7.1.6.3 Precast concrete
and ridge joints in Figure 7.19. Here, face-mounted Precast, reinforced or pre-tensioned concrete can be
gusset plates are shown which are economical but used efficiently for portal frames.
often aesthetically unattractive; centrally located ‘flitch’
plates can be used where the appearance is critical. Site splices are inefficient and thus economic spans
are limited by the length of rafter that can be
Moment-resisting ridge joints can be particularly transported.
onerous within timber portals. Where adopted, it is
best to have rafters pitched above 108 to the Sections and connections
horizontal. Frames typically use non industry-standard
proprietary sections. Rafters are mostly of an I- or
Alternative 3-pinned timber portals are common T-profile with a general taper or haunch increasing the
(Fig. 7.6). These frames are efficient with the roof section depth towards the moment-resisting column
pitched 20–508 to the horizontal. joint (which typically comprises a ‘turn-key’ or similar
Figure 7.20 Timber portal frame erection

locking mechanism for easy assembly). Rafters are

generally at their shallowest towards the ridge, where
a pinned connection is standard. This arrangement
minimises the combined weight of the rafters and
mid-span connection towards the centre of the span;
it also simplifies the reinforcement and pre-tensioning
by eliminating contraflexure towards the apex.
Precast columns will generally be nominally pinned at

their bases for ease of detailing, noting that they
should have sufficient moment resistance to be self-
stable during erection. Full moment-resisting
connections are a viable alternative and can offer
greater stiffness.
Secondary purlins can be precast concrete; however,

light-gauge, cold-formed steel sections are more
common as these are much lighter than a concrete
alternative.
Bracing
In-plane roof bracing will usually comprise steel rods
or cables, crossed in a Modified Fink truss format. Figure 7.21 Precast concrete portal frame with pinned apex connection
Framed bracing may also be used in the eaves walls.

However, it is far more common to stabilise the by contributing a restoring force. However, simple
structure in the vertical longitudinal direction using prop columns will always exert a destabilising force
precast shear wall panels that double up to act as on the rafters as they move away from plumb and
part of the façade (Figure 7.21). These may be of generate PD effects.
‘tilt-up’ construction where too large to be
transported (tilt-up construction being particularly Further reading: snap-through buckling
viable in sheds where the large floor-plan area can
facilitate the flat-bed casting). A major benefit of using – Access Steel. NCCI: Simple methods for second order
precast shear walls rather than steel bracing effects in portal frames. SN033a-EN-EU
elements is that they can be readily procured and
erected, together with the portal frames, by a single 7.1.7.2 Building length and thermal movement
subcontractor. Thermal effects start to become significant for
buildings with linear plan dimension greater than
7.1.7 Further considerations approx. 60m. For such buildings, consideration must
be given to thermal movement joints and/or thermal
7.1.7.1 Snap-through buckling locked-in stresses. The maximum permissible
Snap-through bucking is an in-plane failure spacing of joints will be dependent on the material,
mechanism found primarily in multi-span frames and construction and environmental conditions. In the
only with pitched or arching (‘banana’) rafters. It extreme, steel framed buildings subject to a
occurs when the thrust within the rafters of one span temperate climate, and without onerous performance
is inadequately opposed by that of the rafters making requirements, can extend up to 150m between
up the adjacent spans. The imbalance causes the dedicated movement joints when detailed carefully7.7.
valley to displace laterally allowing the rafter with the This is not to say, however, that this will be the most
greater thrust to splay and flatten (Figure 7.22). If not cost-effective or pragmatic solution. Factors that may
resisted, the thrust will increase as the rafter flattens dictate the maximum permissible movement include:
(i.e. it is a destabilising effect) and the flattened rafter – The capacity of the structure to withstand thermal
will invert (‘snap-through’). stresses
– The integrity and air-tightness of the envelope
Any lateral stiffness provided by the element – The integrity of roof drains
supporting the valley, and/or by a moment-resisting – The integrity of internal suspended services
ridge connection, can help overcome the imbalance (particularly gas and water distribution)
w2
w1 w1 w2 > w1
Centre span splays and

Outward
flattens until eventual
splay
‘snap-through’
Figure 7.22 Snap-through buckling of a multi-span frame

Table 7.1 Lateral deflection limits for sheds

(a) Mezzanine (subject to local
Cladding material Deflection at eaves
lateral load, applies
Profiled metal sheeting h/100 load to primary frame)
Fibre reinforced sheeting h/150
Precast concrete h/200
Concrete blockwork h/200
Clay brickwork h/300
Note that limiting deflections can be measured relative to the
initial frame geometry at the time the cladding is installed.
(b)
7.1.7.3 Serviceability deflection limits Externally-

Standard lateral deflection limits (e.g. those found in applied
the Institution’s Manual for the design of building load
structures to Eurocode 17.8 ) are largely inappropriate
for sheds. Instead, one or more specific criteria will
often apply. These can include: Mezzanine bracing
– Operational limits for large doors loaded by
– Operational limits for travelling cranes and other externally-applied load
suspended or braced plant
– Operational limits for internal services
Figure 7.23 Horizontal stability of mezzanine floors where
– Internal brittle finishes and partitions
fixed to portal frame, (a) without bracing and (b) with bracing.
Both (a) and (b) show major axis bending moments in frame due
In the absence of any such criterion, the deflection
to actions shown
limit is generally governed by the cladding material
(Table 7.17.3).
7.1.7.4 Internal mezzanine floors 7.2 Multi-storey frames

‘Single storey’ portal frame sheds frequently include
internal mezzanine floors. These typically utilise the
primary portal frame columns for vertical support and Multi-storey frames tend to have horizontal floor and
may or may not utilise the frame for stability. roof planes. Consequently, there are few features of
these structures that have not been introduced
Where a mezzanine is stabilised by the moment previously within this Guide.
frame, the mezzanine will exert a horizontal force on
the frame (Figure 7.23a). This force is largely Notably, the frame layout is usually of a form
governed by the equivalent horizontal force owing to resembling the battery arrangements shown in
initial sway, amplified to take account of PD global Fig. 2.9, and the deformed shape is of a form
second-order effects. resembling that shown in Fig. 2.4. Critical load
combinations are usually asymmetric, with equivalent
Conversely, where a mezzanine is braced but linked horizontal forces and PD effects (as a minimum)
to the moment frame, the effect of this bracing on the acting laterally on the structure.
global frame should be considered. The bracing may
provide a point of stiffness attracting global actions Accentuated in multi-storey buildings, it is critical that
into the mezzanine’s bracing (Figure 7.23b). To beam-column joints can transfer bending moments
prevent this from happening, the braced mezzanine from the beams into the column and vice versa. This
can be horizontally isolated from the portal frame. primarily requires careful detailing to ensure that the
Full or partial
strength end plate
joint to beams
Section
Deformed
column web
Continuous column
Elevation
Figure 7.24 Inadequate beam-column joints resulting in local distortion of column web

design reciprocates the analytical model. Figure 7.24 7.3 Further scenarios
illustrates a steel example whereby moment-resisting
joints maintain moment-continuity through the beams
but the unstiffened web fails to engage the columns. Sections 7.1 and 7.2 each consider moment frames
Such a detail would be adequate in a braced frame as primary systems on a building-wide scale.
to achieve continuity of the beams, but is critically Moment frames are, however, used far more
inadequate in an unbraced moment frame where commonly for both primary and secondary systems
continuity between the beams and columns is of stability. This section introduces a few common but
essential (Box 7.2). Similar can be true in reinforced less obvious scenarios.
concrete where either the area, effective depth and/or
curtailment to the reinforcement (in either the column Cantilevering columns
or beam) is inadequate to resolve the bending Arguably the most simple ‘moment frame’, tied
moment around the corner. cantilevering columns of dissimilar inertias are
statically indeterminate and adhere to all the
discussion herein. They are effectively inverted
Box 7.2 Design communication moment frames and share the same characteristics
as regularly orientated counterparts (Figure 7.25).
Where joints are contractor-detailed based on performance
They rely on moment continuity at their bases,
data, it is critical that the required resistance and load path is
provided via connecting superstructure or
clear without ambiguity. Notation such as the following is
substructure/soil-structure interaction.
widely adopted on general arrangement (GA) drawings:
Cantilevering columns are often used for canopy roof
Simple, nominally pinned connection
structures (Figure 7.26). They are also used for sheds
as an alternative to the portal frame structures
Moment resisting (continuous or discussed in Section 7.1. This variation on the portal
semi-continuous) connection frame form is particularly common with precast
However, this notation can be ambiguous. When shown on
plan, as per the following figure, it is unclear whether
bending moment-resistance at the beam ends is to be
transferred from beam to beam only (as it would in a
continuous floor frame supported on simple columns) or into
the column (as is essential for a moment frame). The
engineer must provide further clarity, generally via detailed
drawings (elevations and or sections) for the connection that
clearly shows the arrangement of stiffeners (for steel
sections) or reinforcement (for reinforced concrete).
V = 150 kN
M = ±75 kNm M = ±75 kNm

V = 100 kN V = 100 kN
V = 150 kN
Note that this example intentionally shows moment continuity

in the minor axis of the H-section column. Although this is
less likely, it is most problematic (Fig. 7.24). Figure 7.26 Canopy roof supported on cantilevering columns
Apportioned load to
single cantilever column
Semi-rigid or idealised
rigid connection
Figure 7.25 Cantilevering columns

– Clearance under a bridge prevents downstand

beams
– The bridge traffic prevents bracing to the top flange
In buildings, the form is adopted most widely for

staircases and link bridges that use the full depth of
the balustrade as the structural stringer (Figure 7.31).
At a smaller scale, stiffeners to open sections can
also behave in a similar manner.
The most detailed guidance for the design of

through-girders is contained within codes of practice
Figure 7.27 Precast concrete shed with cantilevering columns and simply supported relating to the design of bridge structures (e.g. BS EN
rafters (shown here with a central, internal line of structure to support gantry crane 1993 Part 27.11).
independent of roof rafters)
Post-fixing goalpost frames and box frames within
concrete construction, to simplify the beam-column existing structures
connections at the eaves (Figure 7.27). Opening up existing structures to achieve more
convenient ‘open-plan’ accommodation is a regular
Cranked beams undertaking that will often impact both vertical and
Cranked beams tend to exhibit all the characteristics lateral load-resisting structures. Goalpost and box
of a more pronounced moment frame, transferring frames are both moment frames that may be used as
actions via combined bending moments, shear and replacement to structural walls (Figure 7.32 and
axial forces. Such beams are often used in mansard Box 7.3).
and dormer roofs (Figure 7.28)7.9, on stair flights and
for ramps. Box 7.3 Goalpost moment frame to a domestic
extension
Drawing parallels between the structure shown in
Fig. 7.28 and those shown in Fig. 7.2 and Box 6.1, An example of a goalpost frame, replacing the previous back
engineers must recognise that horizontal reactions wall of a domestic house. Accepting that the moment frame
can develop even when such a structure is loaded by would require substantial elements, the design team has
vertical actions only. Support conditions are therefore chosen to make an architectural feature of the frame. Timber
critical and the support stiffnesses will impact the is adopted with elements glued in situ at the corner joints.
force distribution when statically indeterminate.
Out-of-plane restraint to unrestrained compression

flanges and chords
Figure 7.29 shows how a beam or truss can be
modelled as a moment frame – not dissimilar to the
cantilevering column systems discussed previously.
As drawn, the vertical force is the shear in the global

system and the horizontal force is the equivalent
restoring force necessary to prevent lateral buckling of
the compression chord. This is usually taken as 2.5%
of the maximum compression force in the flange7.10.
‘Through-girders’ (Figure 7.30) are a common

example used widely by bridge engineers where: Box frames are generally stiffer than goalpost frames.
They also tend to distribute forces more evenly at the
floor level, better replicating the vertical reactions of a
wall (Figure 7.33). They are not, however, always
Cranked moment-resisting
suitable; with the added complication of a beam at
beam-column frame
floor level.

When designing these frames, care is needed so as



 not to locally overload any part of the adjoining load

 path. It is also important that the frame provides
Mansard  Haunch adequate stiffness; to not only satisfy serviceability
roof requirements but to also ensure that any retained non-



 ductile systems are not at risk of being overloaded.


 Highrise buildings
There are occasions where engineers and architects
will break away from the norm and develop systems
that prioritise quite unique objectives. Moment frames
have played a part in this spectacle, dominating the
appearance and functionality of highrise buildings
around the world.
The HSBC Building in Hong Kong is an example at

Figure 7.28 Cranked moment-frame structure to a two-storey mansard roof at one extreme, using mega-frame elements to achieve
One Grafton Street, Mayfair, London a hierarchical system. Eight vertical moment frame

Simply supported through-girder bridge

subject to uniform load
Bending moment
Shear force
Section forces on a free body
Horizontal restraint force

necessary to prevent buckling
of the compression chord
Bending moment associated

with the restraint force
Figure 7.29 Moment frame behaviour within a ‘through-girder’ bridge
Figure 7.31 ‘Through-girder’ staircase with unrestrained

Figure 7.30 ‘Through-girder’ bridge structural balustrades

(a) Wall (b) Box frame (c) Goalpost frame

(with lower beam within the floor)
Figure 7.32 Box and goalpost frame options
• Relatively uniform shear • Relatively uniform shear • Concentrated shear

• Linear distribution of • Non-linear distribution of • Concentrated normal
normal forces normal forces forces
(a) Wall (b) Box frame (c) Goalpost frame
Figure 7.33 Relative reactions from walls, box and goalpost frames
clusters, each one a compact core-like arrangement It should be noted that the global moment frame
of four columns linked with beam elements at each resulting from the mega-bracing is not dissimilar to
storey, work locally as a series of vierendeel trusses. the statically indeterminate system created when
These are braced via mega-bracing at intervals outriggers are used in combination with a central
through their height and work collectively in pairs to core, even where the outrigger to perimeter column
form a global moment frame system that is devoid of connection is pinned. This is also the system adopted
diagonal bracing through much of the building’s for the International Finance Centre, Hong Kong
height (Figure 7.34). (Figure 7.35).
Local behaviour
Global behaviour (due to local
horizontal floor
loading)
Figure 7.34 HSBC Building, Hong Kong

Global behaviour
Figure 7.35 International Finance Centre, Hong Kong
By contrast, Towers 1 and 2 of the World Trade engineering principals set out in this Guide.
Center, New York, were at the opposite extreme to Collectively, the buildings show how engineering is
the HSBC Building. These towers employed a both a science and an art. Take these, and other
hierarchically flat approach, with a uniform moment precedent projects as inspiration: hone your skills and
frame system wrapping around the entire façade of practice ingenuity.
each tower. The moment frames formed tube-like
systems, with global behaviour for each tower akin to
that of single-element statically determinate vertical
cantilevers (Figure 7.36). 7.4 References
Perhaps the beauty of structural engineering is that
these three structures – the HSBC Building, the 7.1 Salter, P.R. et al. Design of single-span steel portal
International Finance Centre and the World Trade frames to BS 5950-1: 2000. SCI Publication P252.
Center – are almost as visually different as they are Ascot: SCI, 2004
systematically similar. Each was conceived by
engineers who overcame the project-specific 7.2 Steel Construction Institute. Wind-moment Design of Low
challenges by applying a sound understanding of the Rise Frames. SCI Publication P263. Ascot: SCI, 1999
Local
behaviour
Global
behaviour
Figure 7.36 World Trade Center Towers 1 and 2, New York


7.4 PD 6695-1-10: 2009: Recommendations for the

design of structures to BS EN 1993-1-10. London:
BSI, 2009


7.7 Alexander, S.J. Design for movement in buildings.

CIRIA C734. London: CIRIA, 2014

design of building structures to Eurocode 1 and Basis
of Structural Design. London: IStructE Ltd, 2010
7.9 Perry, P. ‘Development over London Underground

tunnels: No. 1 Grafton Street’. ICE Proceedings,
Structures and Buildings, 167(12), December 2014,
pp728-742

design of steelwork building structures to Eurocode 3.
7.11 BS EN 1993-2: 2006: Eurocode 3 – Design of steel

structures – Part 2: Steel bridges. London: BSI, 2006

Stability of Buildings Part 4 Moment Frames

Uploaded by

Copyright:

Available Formats

Stability of Buildings Part 4 Moment Frames

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Stability of Buildings Part 4 Moment Frames

Uploaded by

Copyright:

Available Formats

Stability of buildings

Permission to reproduce extracts from British Standards is granted by the British

Published by The Institution of Structural Engineers,

First published: November 2015

# 2015 The Institution of Structural Engineers

iv The Institution of Structural Engineers Stability of buildings Part 4

The following notation is used for hand-drawn figures:

A transverse shear force profile acting along an element

A bending moment profile acting along an element

A moment-resisting joint of non-disclosed stiffness (i.e. a semi-continuous or continuous joint)

An idealised moment-resisting joint of infinite stiffness (i.e. a continuous joint)

A moment-resisting boundary condition of non-disclosed stiffness (i.e. a semi-continuous or

The Institution of Structural Engineers Stability of buildings Part 4 v

Glossary iv 7 Moment frames in application 36

5 Analytical modelling of joints 26

6 Simplified analysis methods 30

The Institution of Structural Engineers Stability of buildings Part 4 iii

However, it is perhaps the immediate simplicity of 1.1 References

Chapter 4 is dedicated to plastic frame analysis, while

The Guide ends with a discussion on the application

Relevance and limitations

Note that the checklist given in Fig. 1.1 is intended to

This Guide is an introduction written primarily for

The Institution of Structural Engineers Stability of buildings Part 4 1

2 The Institution of Structural Engineers Stability of buildings Part 4

2.1 Form using a moment frame in a hybrid structure together

A common characteristic of all the frames shown is

A further common characteristic is that the frames

Another characteristic of moment frames is that they

Whether a moment frame is more or less of an

Note that the stiffnesses of these frames may vary

Fig. 2.1 shows various examples of moment frames

Moment frame Braced frame Wall (often with penetrations)

Note that the stiffnesses of these frames may vary significantly.

The Institution of Structural Engineers Stability of buildings Part 4 3

Braced at 1st floor

Figure 2.4 Deflected shapes of multi-storey moment frame

resulting from elastic flexural deformations of

The deflected shape gives significant insight into the

Ultimately it is the relative stiffnesses of the various

It should be noted that the stiffness of the element is

Actions and reactions

2.2 Failure mechanisms

In line with the stress resultants shown in Fig. 2.5,

However, of particular concern to moment frames is

4 The Institution of Structural Engineers Stability of buildings Part 4

Frame disposition Bending moments Deflected shape

2.3 Configuration and application

The Institution of Structural Engineers Stability of buildings Part 4 5

Bracing Moment frame

Two-way braced Orthogonal braced/moment frame Two-way moment frame

Figure 2.9 Example configurations of vertical stability systems

6 The Institution of Structural Engineers Stability of buildings Part 4

Box 2.1 Multi-storey engineered timber moment frame

The Institution of Structural Engineers Stability of buildings Part 4 7

Moment frame to raised atrium roof

Figure 2.10 Atrium roof acting as moment frame

(a) (b) (c)

Figure 2.11 First-order reactions for moment frames