Professional Documents
Culture Documents
assessment of high-risk
structures against
disproportionate collapse
September 2013
Technical author
D Cormie MEng(Hons) CEng CEnv FIStructE FICE M.ASCE (Arup)
Membership of the steering group
O Brooker BEng CEng MIStructE MICE MCS (Modulus) Chairman
J Carpenter BSc(Eng) CEng FIStructE FICE CFIOSH (Consultant)
G Harding OBE DIC CEng FIStructE MICE (Consultant)
J Lane MSc CEng FICE (Rail Safety and Standards Board)
Dr A Mann PhD FREng CEng FIStructE MICE (Consultant)
Dr S Matthews PhD CEng FIStructE FICE CSci CWEM MCIWEM (BRE)
S McKechnie BSc(Hons) CEng MIStructE (Arup)
A T Moolan BSc(Hons) CEng MICE FRICS (LABC)
C O’Regan BEng(Hons) CEng MIStructE (The Institution of Structural Engineers)
P Watt BSc(Eng) (Department for Communities and Local Government)
Consultees
Dr J Agarwal PhD (University of Bristol)
W Baker MSc SE PE FIStructE F.ASCE NAE (Skidmore, Owings & Merrill LLP)
M Banfi MA CEng FIStructE FICE (Arup)
J Burridge MA(Cantab) CEng MIStructE MICE (MPA The Concrete Centre)
Dr M Byfield PhD CEng MIStructE MICE (University of Southampton)
Professor M Chryssanthopoulos PhD DIC CEng FIStructE FICE (University of Surrey)
K Claber MSc CEng MICE
Dr S Dogruel PhD (Arup)
D Dusenberry PE F.SEI F.ASCE (Simpson Gumpertz & Heger Inc.)
S Inglis MA (Arup)
Professor B A Izzuddin PhD CEng FIStructE M.ASCE (Imperial College London)
R Johnson MArch SE SECB ASCE (Skidmore, Owings & Merrill LLP)
Dr H S Lew PhD PE Hon M.ACI F.SEI F.ASCE (National Institute of Standards and Technology)
Dr S Marjanishvili PhD SE PE M.ASCE (Hinman Consulting Engineers Inc.)
K Moazami PE F.ASCE (WSP)
Dr M O’Connor PhD MBA (WSP)
L Robertson SE PE CEng FIStructE FIEI Dist.M.ASCE M.NAE (Leslie E. Robertson Associates RLLP)
Dr R Smilowitz PhD PE M.ASCE (Weidlinger Associates)
Dr D Stevens PhD PE F.ACI M.ASCE (Protection Engineering Consultants LLC)
Professor A Vrouwenvelder MSc (TNO/Delft University)
M Waggoner SE PE PEng (Walter P Moore and Associates Inc.)
A Way MEng CEng MICE (The Steel Construction Institute)
Building Standards Division, Scottish Government
Director: Engineering and Technical Services
S Fray BSc(Hons) CEng MIStructE M.SEI MICE (The Institution of Structural Engineers)
Secretary to the steering group
L Allen MEng(Hons) (The Institution of Structural Engineers) (until July 2012)
L Kirk MEng(Hons) (The Institution of Structural Engineers) (from September 2012)
Acknowledgements
This publication was supported through grant funding from the Department for
Communities and Local Government and received funding from Arup’s Design and
Technology Fund.
Image copyright:
Figure 2.1: Contains public sector information published by the Health and Safety
Executive and licensed under the Open Government Licence v1.0
Box 3.1: Huddersfield Examiner
Box 3.3: AP/Press Association Images
Box 3.9: Institution of Civil Engineers (Professor John Burland)
Box 3.10: Jonathan G M Wood Structural Studies & Design
Box 4.1: Dr Man-Seop Lee and Prof. Uwe Starossek
Figures 4.3, 4.5 and 4.6: Arup
Figure 4.7: Cox Rayner Architects
Box 4.2: Victoria Police
Box 4.3: AP/Press Association Images
Box 4.5: AP/Press Association Images
Box 5.1: Colin G Bailey
Published by The Institution of Structural Engineers
International HQ, 11 Upper Belgrave Street, London SW1X 8BH
Telephone: +44 (0)20 7235 4535 Fax: +44 (0)20 7235 4294
Email: mail@istructe.org Website: www.istructe.org
First published 2013
ISBN 978-1-906335-24-3
# 2013 The Institution of Structural Engineers
The Institution of Structural Engineers and those individuals who contributed to this Manual have
endeavored to ensure the accuracy of its contents. However, the guidance and recommendations
given in the Manual should always be reviewed by those using the Manual 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 Manual and its contents is accepted by the Institution, the technical
author, the members of the steering group, the consultees, their servants or agents. Any person
using this Manual 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 11 Upper Belgrave Street, London SW1X 8BH.
Tables Boxes
iv The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Foreword
Robustness is an indisputably desirable quality of a the risk, so that resources can be judiciously targeted
structural design. Designing a structure to be less to give the most effective reduction in risk.
sensitive to damage renders it safer and reduces
risks. A small perturbation which produces a gross Risks may be derived not only from accidental or
structural response is not automatically unsafe, but malicious actions on the structure but also from
usually has the potential to be. Designing for so-called ‘soft’, or procedural, hazards such as
robustness aims to eliminate such characteristics design errors, the failure to adequately
such that the structure is less sensitive to communicate design information leading to errors
perturbations which have the potential to cause in its construction, a lack of requirements for the
harm. In design there is much that we do not know: inspection and maintenance of the structure, or
robust design is much more than following a set of the failure to adhere to them. Rarely is there a
rules and designing for a certain set of scenarios, it is single cause leading to an accident and usually
about ensuring our design is less sensitive to these human error are a significant contributing factor,
unknowns. It is famously stated in The Structural the risk of which the designer would do well to
Engineer, May 1978: reduce. Indeed, ‘soft’ hazards may be so
significant that they will be the governing factor in
‘‘Structural engineering is the art of moulding the overall level of risk associated with the design.
materials we do not wholly understand into shapes
that we cannot precisely analyse, so as to withstand This Manual was commissioned by the Institution
forces we cannot really assess, in such a way that with financial support from the Department for
the community at large has no reason to suspect the Communities and Local Government, and is written
extent of our ignorance.’’ primarily for the design of those structures classified
– Dr A. Robert Dykes, as Class 3 in EN 1991-1-7. The principles can also
Institution of Structural Engineers be readily applied to all types of structures
Scottish Branch, 1978 irrespective of the design standards being used, and
it is recommended that they are adopted wherever a
Robustness is part of our protection against this structure might be considered high-risk. The Manual
uncertainty and in the UK has been recognised since provides guidance to the practising structural
the collapse of Ronan Point in 1968 as an essential engineer on the design of such structures against
aspect of good structural design practice. disproportionate collapse by systematic risk
assessment, as well as providing useful guidance for
Since 2004 there has been a particular requirement client bodies and building control authorities on the
to specifically consider the risk associated with a use of such design approaches.
high-risk, or ‘Class 3’ structure through the use of a
systematic risk assessment. In high-risk structures
the potential to cause harm is heightened compared
with other structures due either to the number of
occupants, a characteristic of those occupants, or
some feature of the structure’s function, form and
construction, location or purpose. Consequently there
is a need to take greater care to assess the David Cormie,
robustness of the structure and to take greater steps Technical Author
to reduce the risks to which it is subjected.
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse vii
Contents
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse iii
1 Introduction
1.1 Scope of this Manual The reader is referred to three related Institution
publications:
– Safety in tall buildings and other buildings of large
This Manual proposes a systematic risk assessment occupancy1.3 provides many recommendations for
framework for the design of high-risk structures designing tall buildings to be intrinsically safe and is
against disproportionate collapse. Risk assessment is a direct complement to this Manual.
a requirement for all structures, but for Class 1 and 2 – Risk in structural engineering1.4 provides a
buildings (BS EN 1991-1-7 Annex A1.1) is usually foundation in the principles of risk management and
addressed through good quality management and how to manage risk in design.
through adherence to recognised codes of practice. – Appraisal of existing structures1.5 has an equally
important relationship to this Manual, which in
This Manual is concerned with systematic risk Chapter 6 outlines how systematic risk assessment
assessment as defined in BS EN 1991-1-7 can be applied to designing for robustness in the
Annex A1.1, principally with regard to the design of retrofit of existing structures, where the existing
Class 3 buildings. It complements the Institution of nature of the building is likely to place constraints
Structural Engineers’ publication Practical guide to on the design feasibility.
structural robustness and disproportionate collapse in
buildings1.2, which provides a useful primer on the
concepts of robustness and the (largely) qualitative
methods used in the design of lower-risk buildings 1.2 Purpose
against disproportionate collapse. The Practical guide
does not, however, give guidance on the design of
high-risk (Class 3) buildings, which is the primary The risk-based approach outlined in this Manual may
focus of this Manual. be used to consider the full spectrum of natural,
accidental and malicious hazards. It is intended to
The approach set out for systematic risk assessment provide an approach through which the Eurocode
in this Manual is designed to be equally applicable to requirement (BS EN 1991-1-7 Annex A1.1) to
any high-risk structure. This may include structures undertake a systematic risk assessment for Class 3
and buildings that do not fit into the building risk buildings can be met, together with the
classification system in BS EN 1991-1-71.1, recommendations of the Building Regulations
structures where there are particular risks associated Approved Document A1.6 and the equivalent
with the design itself – perhaps by virtue of it being documents in the devolved administrations of the
novel or unusual – or where there are particular risks United Kingdom1.7-1.9 for Class 3 buildings.
to which the structure may be exposed, such as
malicious risks including terrorism. The Manual is written for the professionally qualified
structural engineer who has a certain level of
It may also include structures where there is experience in safety engineering and the principles of
exposure or vulnerability in a temporary state, and the risk assessment. It is intended to aid the practising
design of the extension, alteration or change of use of structural engineer in the design of robustness in a
existing structures where design constraints may limit high-risk building, though the structural engineer
the level of robustness than can practically be tasked with preparing the robustness design has a
achieved. In addition to consideration of risks to life responsibility to ensure that they are suitably qualified
safety, the same methodology may equally well be and experienced (SQEP) to do so. No formal register
applied for the consideration of risk unrelated to exists designating individuals with suitable
safety, such as where a higher level of performance is qualification and experience for the systematic risk
necessary in the structure to support some critical assessment of high-risk buildings, but it will normally
function, purpose or use. be expected that the systematic risk assessment will
be authored by a senior structural engineer with a
Boxed examples are given throughout the text to track record of designing similar buildings, while past
highlight either lessons from past failures, examples experience of designing against extreme hazards will
that would constitute a poor design, or examples of be advantageous.
good design practice. This chapter includes the legal
background, while in Chapter 2 some of the key While intended to aid the structural engineer in the
concepts are introduced together with a discussion process, the Manual is not intended to be the sole
the Eurocode requirements. Chapter 3 presents the means by which the engineer will gain the skills and
proposed framework for a systematic risk knowledge required and hence become competent in
assessment and is the core of the document, and is the preparation of a systematic risk assessment.
followed by discussion of each part of the risk
assessment process and the factors that should be In addition to providing guidance to the practising
considered in its development. Chapter 4 presents a structural engineer, the Manual highlights the duty of
commentary on the spectrum of risk reduction care owed by the structural engineer to deliver a
measures that could be employed, both ‘hard’ robust design, and the responsibilities held by the
measures (changes to the design) and ‘soft’ client in the decisions made on the basis of the
measures (quality assurance, use of peer review and assessment findings. It is also written to provide
so on). Chapter 5 gives a worked example for a guidance to the building control body in evaluating a
typical Class 3 building. building control submission for a Class 3 building.
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 1
1.3 Introduction
The designer has a statutory duty to, so far as is 1.1 British Standards Institution. BS EN 1991-1-7:2006.
reasonably practicable: Eurocode 1: Actions on structures. Part 1-7: General
– eliminate hazards which may give rise to risks, and actions: Accidental actions. London: BSI, September
– reduce the risks associated with any remaining 2006.
hazards (Regulation 11(4)).
1.2 Institution of Structural Engineers. Practical guide to
If any significant risks remain, the designer must take structural robustness and disproportionate collapse in
all reasonable steps to provide adequate information buildings. London: IStructE, October 2010.
to those who are likely to need to identify and
manage the remaining risks (Regulation 11(6)). 1.3 Institution of Structural Engineers. Safety in tall
buildings and other buildings of large occupancy.
Legislation requires duty holders to consider all London: IStructE, October 2010.
reasonably foreseeable hazards which will, depending
on the structure, include natural hazards such as 1.4 Institution of Structural Engineers. Risk in structural
windstorm and flood, accidental hazards such as fire engineering. [due for publication 2013]
and vehicle impact, and malicious hazards such as
malevolent vandalism and terrorism. A hazard must 1.5 Institution of Structural Engineers. Appraisal of existing
be eliminated or the risk arising reduced unless, structures. 3rd ed. London: IStructE, November 2010.
compared to the risk, it can be demonstrated that it
is grossly disproportionate in terms of time, cost and 1.6 Office of the Deputy Prime Minister. The Building
effort to do so. Regulations 2000 — Approved Document A:
Structure. A3 — Disproportionate Collapse. 2004
Some guidance on determining whether something is edition with 2004 amendments. NBS, 2004.
in gross disproportion to the risk and what
constitutes ‘so far as reasonably practicable’ is given 1.7 Scottish Building Standards Agency. Technical
by the Health and Safety Executive1.13 and the handbook – non-domestic. Section 1 – Structure.
Institution of Civil Engineers1.14. Further information Scottish Building Standards Agency, 1 October 2010.
about the duties of a designer is found in the CDM
Regulations Approved Code of Practice1.15 and 1.8 Scottish Building Standards Agency. Technical
Industry Guidance for Designers1.16. handbook – domestic. Section 1 – Structure. Scottish
Building Standards Agency, 1 October 2010.
The structural engineer will hold a duty of care to
develop a safe structural design. However, duties will 1.9 Building Control Northern Ireland. Technical Document
not fall solely upon the structural engineer: any D – Structure. Department of Finance and Personnel
individual or organization who makes a project- Northern Ireland (DFPNI), 2009.
related decision that might significantly impact the
safety of others may be deemed a designer under the 1.10 The Health and Safety at Work etc. Act 1974 (as
Regulations. amended). 1974.
In designing against disproportionate collapse, risk 1.11 The Construction (Design and Management)
reduction may impact upon the aesthetics of the Regulations 2007. 2007 (SI 2007/320).
design or may increase the cost of the structural
package or the project as a whole. Other parties 1.12 The Building Regulations 2010. 2010 (SI 2010/2214).
such as clients and contractors may be implicated as
designers by virtue of making decisions about 1.13 Health and Safety Executive. ALARP at a glance: How
particular risk reduction measures in the context of to tell if a risk is ALARP: Deciding by good practice.
aesthetics, amenity, function, cost, programme and http://www.hse.gov.uk/risk/theory/alarpglance.htm.
so on that have a bearing on the safety of the design. [Accessed 2012-11-28].
The Building Regulations1.12 require the building to be 1.14 Institution of Civil Engineers. A review of, and
constructed so that ‘‘in the event of an accident [it] will commentary on, the legal requirement to exercise a
2 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Introduction 1.4
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 3
2 Background and design objectives
4 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Background and design objectives 2.4
structure to structure. Quite clearly, designing a threshold generally fall into the ALARP region and
building for all possible accidental and malicious must be driven down to a level which is as low as
actions to which any structure might be subjected is reasonably practicable (ALARP). Below this level,
simply neither practical nor proportionate, nor good available resources could be used more effectively on
engineering design. other risks. ALARP can, for practical purposes, be
taken as being synonymous with the statutory duty to
The purpose of a risk-based approach to design reduce risks so far as is reasonably practicable
against disproportionate collapse is to allow a proper (SFARP)2.3, 2.4 and forms the basis of the approach
examination of the risks to which the structure is used in this Manual.
subjected; and in so doing to produce a
proportionate, economic design. Good risk As highlighted by the statutory duties of a designer
management does not seek to minimise all risks, but outlined in Section 1.3, it is better to eliminate a
ensures that the costs of risk reduction measures hazard in the first place than to protect against the
implemented in the design remain proportionate to risk which arises from it. Similarly, it is better to
the magnitude of the risk. implement measures to reduce the risk than to simply
communicate the risk such that it has to be
A risk management approach also has an important controlled through operational measures.
secondary benefit in that the remaining risks have
been identified and assessed, enabling those affected One model for expressing this which is suited to
to take action to control the risks to which they are structural engineering design is the ERIC risk
exposed. If, for example, a building houses a data reduction model2.5 (Figure 2.2). This states that
centre, actions may be taken to ensure appropriate hazards should be eliminated unless the cost of
fail-over and back-up facilities are designed for the doing so is disproportionate when weighed in relation
appropriate actions, such that the tenant’s business to the risk. The designer has an almost unique
is able continue uninterrupted and their consequential opportunity to eliminate hazards from a design, and
losses (and hence the building owner’s potential every opportunity should be taken to do so in order
liability) are reduced. to avoid the need to mitigate the associated risks
later on. If, for example, a building can be designed
such that an errant vehicle cannot impact the building
columns, it will always be preferable to do so than to
2.4 Reduction of risk reduce the risks associated with the hazard. By the
same logic, it will always be preferable to design a
building so that it is able to withstand the loss of a
Two basic approaches exist with regard to reduction column than to protect against its loss, because the
of risk. The first is to define a threshold above which inherent vulnerability associated with the column is
a risk is considered intolerable, and to reduce the removed and the risks derived from the hazards to
risks to this threshold. The second is to reduce all which the column may be subjected become less
risks so far as is reasonably practicable, irrespective significant (Box 2.1).
of whether they already fall below the tolerable
threshold. Any risk which remains above the tolerable If the hazard cannot reasonably be eliminated, the
threshold is unacceptable in all circumstances. This risk arising should be reduced to a level which is
second approach is the one required by England and ALARP (Box 2.2). Deciding whether a risk is ALARP
Wales statute for safety-related risks. can be challenging because it requires exercise of
judgement. The designer must reach decisions about
The first approach ensures that intolerable risks are whether a risk reduction measure is proportionate or
addressed, but fails to consider whether a significant disproportionate, and therefore whether a risk has
further benefit could be gained beyond this threshold been reduced so far as reasonably practicable. In the
for a small additional investment. It also fails to majority of cases the decision can be reached
consider the risks which initially lie below the tolerable through engineering judgement and with reference to
threshold. contemporary good practice, examples of which are
given throughout this Manual. Guidance on deciding
The second approach goes much further: it defines a whether a risk is ALARP is given by the Institution of
threshold beyond which risks must be mitigated, but Civil Engineers2.3 and the Health and Safety
also requires the designer to consider all risks which Executive2.4.
fall within a range beneath this threshold and mitigate
these so far as is reasonably practicable (SFARP).
This forces the designer to consider the potential to Unacceptable
eliminate hazards or reduce risks for a small, region
sometimes negligible, design change. Such changes
can have a potentially significant impact on the level
of risk and will often be missed by the first approach,
and it is for this reason that England and Wales
statute adopts this approach focussed on driving ALARP region
down all risks.
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 5
2.4 Background and design objectives
6 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Background and design objectives 2.5
Some risks will remain in the design: these will relate approaching the robustness of the structure and
to hazards for which a reduction in the risk has not perhaps the structural design as a whole in a different
been possible for some reason or where the risk has way.
been reduced as far as possible but is still
significant. These risks should be communicated
to those affected (‘inform’) such that they may
take steps to control the risk through working 2.5 Building risk classification
practices or other measures (Boxes 2.2 and 2.3).
It is often here that things are most likely to go
wrong – from the study of past failures it is readily A structure designed to the Eurocodes is required by
apparent that poor communication is often a BS EN 1990 to ‘‘be designed, and executed in such
significant causal factor2.6-2.11. Safeguarding the a way that it will, during its intended life, with
reduction in risk achieved through good quality appropriate degrees of reliability and in an economical
management and reducing any significant remaining way sustain all actions and influences likely to occur
risks are therefore both fundamental to a successful during execution and use, and meet the specified
design. serviceability requirements for a structure or a
structural element’’2.17. This overarching requirement
If the cost of a proposed measure is shown to be describing the philosophy of the structural design is
disproportionate when weighed in relation to the supplemented by more specific requirements
magnitude of the risk, it is unnecessary to implement contained in Annex A of BS EN 1991-1-72.18 (refer
the risk reduction measure in the design (Box 2.4). also to the background document by
Consequently the assessment of risks using Equation Vrouwenvelder2.19). These vary according to the risk
2.1 forms one input to a cost-benefit assessment, classification ascribed to the building, see Table 2.1.
upon the basis of which decisions will be taken about
the reduction of these risks. The building risk classification system and the
associated design requirements given in Annex A of
If, after going through these stages, it has not been BS EN 1991-1-72.18 represent contemporary good
possible to reduce the risk to a tolerably low level, the practice which the engineer would be expected to
designer should consider whether the proposed follow. Annex A has been accepted in the UK
design is worth the risk. In some cases this will mean National Annex2.20. Where the status of the Annex is
going back to the conceptual stage of design and informative, as a minimum the building risk
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 7
2.5 Background and design objectives
classification system and the performance of a circumstances surrounding the design that warrant
systematic risk assessment as outlined in this Manual the categorisation of a building nominally designated
should still be adopted in the design of high-risk as a lower risk class as a Class 3 building. Such
buildings. The engineer may also decide to adopt the situations in which the designation of a building as
Annex in the design of lower risk buildings. Class 3 might be appropriate include:
– Buildings that do not fall into the building risk
Class 3 framed structures designed and detailed in classification system given in Table 2.1.
accordance with the Eurocodes with the UK National – Buildings housing hazardous operations such as
Annexes and accompanying non-contradictory laboratories, or hazardous plant or materials such as
complementary information (NCCI) will normally oil-filled electrical transformers and bottled gases.
incorporate the requirements of the risk classes up to – Structures featuring innovative, complex or unusual
and including Class 2B. Nevertheless, as part of their structural framing, stability arrangements, materials
design duties the structural engineer has a or construction techniques2.21.
responsibility to review the codes of practice – Structures with a high degree of modularisation and
proposed for design and to ensure they are repetition where there is the potential for a systemic
appropriate and contain all necessary design error.
provisions. This includes the robustness – Structures designed at the margins of the codes of
requirements. practice and the underlying assumptions upon
which they are based.
The building risk classification system in Table 2.1 is – Structures designed ‘down to the bone’ or
merely given as guidance, and the structural engineer optimised ‘to within an inch of their lives’, and
should determine the most appropriate risk therefore very highly stressed and with very little
classification based on the risks associated with the redundancy.
design. Rarely will it be detrimental to classify a lower – Buildings exposed to abnormal, significant or
risk class building as Class 3 and undertake a extreme risks, including those to which there is a
systematic risk assessment. At the very least it results credible risk from terrorist attack.
in a better understanding of the risks associated with – Structures having exposure or vulnerability in a
the structural design, and will often be beneficial in temporary state.
terms of reducing the level of risk in the design or – Existing buildings undergoing extension, alteration
identifying that an alternative design approach is or change of use, particularly where there are
more suitable. unknown structural characteristics and/or the
potential for hidden defects.
The structural engineer should therefore consider – Structures required to exhibit a high level of
whether there are unusual features, actions or other reliability or a greater level of performance than
8 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Background and design objectives 2.6
required solely for life safety, owing to some critical equal: there may be nothing inherent in the design of
function, purpose or use they are required to a four-storey hospital that justifies a different
support. approach being used from a building with one fewer
storey. However, for a more significant structure,
There will also often be components of Class 3 some level of structural analysis of the performance
buildings in a building which would otherwise be of the structure under some of the hazards identified
designated as Class 2A or Class 2B, in which case would be expected, rather than merely relying on the
the entire building should be considered as a Class 3 notional level of robustness imparted by prescriptive
building. methods.
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 9
2.7 Background and design objectives
duty ‘so far as is reasonably practicable’ with specific structures – Part 1-7: Accidental actions. London: BSI,
regard to designers in the construction industry. December 2008.
London: ICE, January 2010.
2.21 Pugsley AG. The safety of structures. London: Edward
2.4 Health and Safety Executive. ALARP at a glance: How Arnold (Publishers) Ltd, 1966.
to tell if a risk is ALARP: Deciding by good practice.
http://www.hse.gov.uk/risk/theory/alarpglance.htm. 2.22 Department for Communities and Local Government.
[Accessed 2012-11-28]. Guide to the use of EN 1991-1-7 – Accidental
Actions. DCLG Publications, September 2006.
2.5 Health and Safety Executive/Construction Skills. The
Construction (Design and Management) Regulations 2.23 American Society of Civil Engineers. ASCE/SEI 7-05:
2007 – Industry Guidance for Designers. CDM07/4. Minimum design loads of buildings and other
ConstructionSkills, 2007. structures. ASCE, 2005.
2.6 Agarwal J, Haberland M, Holickẏ M et al. Robustness 2.24 American Society of Civil Engineers. ASCE/SEI 7-10:
of structures: lessons from failures.. Structural Minimum design loads of buildings and other
Engineering International 2012;1/2012:105-111. structures. ASCE, 2010.
2.7 Rihani S. Structural failures – a case study: Hartford 2.25 International Code Council. International Building Code.
Coliseum. ASCE Structural Engineering Institute (SEI), ICC, 2012.
September 2005.
2.26 United States Department of Defense. Unified Facilities
2.8 Delatte NJ. Forensic case studies for civil engineers. Criteria UFC 4-023-03: Design of Buildings to Resist
ASCE Press, 2009. Progressive Collapse. Washington, D.C., 14 July 2009
(including Change 1 – 27 January 2010).
2.9 Levy M, Salvadori M. Why buildings fall down: how
structures fail. New York, NY: W.W. Norton, 1992.
10 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
3 Development of a systematic risk assessment
3.1 Framework for a systematic risk element rather than being hardened against a given
load, it no longer matters whether this load is slightly
assessment larger than expected or the strength slightly lower.
The cliff edge defined by the ultimate capacity of the
This chapter forms the core of this Manual: it outlines column has been eliminated; consequently the global
the systematic risk assessment process which is structural response has been rendered insensitive to
proposed, together with some commentary on each the ultimate capacity of the column and therefore to
step of the process. This process is supplemented by the risks derived from the hazards to which it could
the discussion in Chapter 4 on the design of ‘hard’ be subjected.
measures (i.e. changes to the design) and ‘soft’
measures (i.e. procedural measures) that may be
adopted as risk reduction measures, and an example
systematic risk assessment built on this methodology
is given in Chapter 5.
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 11
3.1 Development of a systematic risk assessment
Step 1:
Identify the hazards
Figure 3.1 Framework for a systematic risk assessment for design against disproportionate collapse
12 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Development of a systematic risk assessment 3.1
– Step 1: Identify the hazards (Section 3.2). for the controlling the residual risks, both the
Identify the foreseeable and unforeseeable client and other risk stakeholders must agree
hazards to which the building might reasonably with and accept the level of residual risk inherent
be subjected. in the design.
– Step 2: Eliminate hazards where feasible to do – Step 9: Check the sensitivity of the risk
so (Section 3.3). Consider whether it is possible assessment (Section 3.10). Check the sensitivity
to eliminate some of these hazards through an of the findings of the risk assessment in the
amendment to the design. For example, the following three respects:
design of a grade change between an adjacent 9(a) Check cliff edge effects. For each hazard,
roadway and a building may eliminate the risk of check whether small changes in the
vehicle impact, or simple anti-vandal measures assumptions will lead to a gross change in
may eliminate the risk of malevolent vandalism to the consequences of the hazard. If so,
critical structural elements. consider the beyond-design basis event in
which the action is slightly worse than
– Step 3: Determine the level of tolerable risk assumed in the risk assessment and either
(Section 3.4). Define scales of likelihood and eliminate the cliff edge or move it to be
consequence appropriate for the evaluation of sufficiently distant from the design basis so
risk for the particular project being undertaken, as to reduce the sensitivity in the design.
and determine the extent of damage that would 9(b) Review low likelihood/high consequence
be considered intolerable. Plot this threshold on hazards. Due to the sensitivity of such
the risk matrix. hazards to small changes in the underlying
assumptions, use a sensitivity study to
– Step 4: Evaluate the risks (Section 3.5). For determine whether the risk to occupants
each hazard that remains: will be significantly worse if the hazard is
4(a) Determine the likelihood of the hazard slightly larger than is assumed, or the
materialising. structural response slightly worse than
4(b) Determine the consequences of the hazard calculated.
should it materialise, taking into account 9(c) Review combined hazards. Combined
the existing design provisions. hazards may not have been considered
4(c) Quantify the risk associated with each directly in the preceding risk assessment.
hazard according to the likelihood and Consider combined hazards in their own
consequence evaluated in 4(a) and 4(b). right on the basis of the previous decision.
Consider the sensitivity of the design to the Consider the hazard from two angles when
underlying assumptions and, where determining whether the level of damage is
necessary, perform sensitivity studies using intolerable:
engineering analysis to identify how the – the likelihood of each event separately and
evaluated risk might vary with minor changes the level of damage that is considered
in the input parameters. intolerable
– given that the first hazard may occur, how
– Step 5: Identify risk reduction measures (Section much damage is considered proportionate
3.6). For each risk in turn, irrespective of to the risk of the second hazard?
whether the risks lie below the tolerable risk
threshold, identify scenario-independent and – Step 10: Review the overall level of risk (Section
scenario-dependent measures that could be 3.11). Review the overall level of both the
implemented to reduce the risk. unmitigated and the mitigated risk. Unmitigated
risks may be high but highly-effective mitigation
– Step 6: Perform a cost-benefit assessment for may have been identified and incorporated.
each potential risk reduction measure (Section Equally, the overall level of risk may be high but
3.7). Weigh the reduction in risk in relation to the it may not have been possible to make much
cost of the measure. Usually, one mitigation impact through the mitigation measures.
measure will give a reduction in the risks across
several hazards. The total risk reduction across In either case, the design of the building against
all impacted hazards must be used as the basis disproportionate collapse has been found to
for the cost-benefit assessment. have particular importance, and an independent
peer review of the systematic risk assessment
– Step 7: Implement risk reduction measures by a suitably qualified and experienced person is
(Section 3.8). Identify risk reduction measures to warranted.
be implemented, and incorporate them in the
design. Risk reduction measures which are – Step 11: Effectively communicate information
cheap but have a significant impact on the risk about any risks that remain (Section 3.12). For
(a high cost-benefit ratio), or that represent those risks that remain in the design, provide
contemporary good practice, should always be adequate information to those who might be
implemented regardless of whether the risk is affected by the risks such as clients, other
already below the tolerable threshold. designers, contractors and sub-contractors. The
duty is incumbent on the designer to take all
– Step 8: Review the residual risk (Section 3.9). reasonable steps to ensure the information is
Review the residual risks that remain in the communicated effectively: it is not sufficient to
design following the application of the risk merely prove that information has been issued.
reduction measures with the client and any other
risk stakeholders who will become the risk
owners for the residual risks in the design. As
the eventual risk owners who will be responsible
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 13
3.1 Development of a systematic risk assessment
Equation 3.1
Annual probability of exceedance
¼ 1 ½1 required probability of exceedance in N years]1=N
This is equal to the inverse of the return period.
Example: for an event with a 10% probability of exceedence in 50 years, the annual probability of exceedence is
1 ½1 0:11=50 ¼ 0:2% per annum, and the corresponding return period is 1=0:2% ¼ 475 years, or a 1 in 475 year event.
14 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Development of a systematic risk assessment 3.1
Consequence
Catastrophic
Substantial
Significant
Minimal
Serious
Severe
Minor
Likelihood
Frequent/common
Likely
Unlikely
Rare
Improbable
Negligible
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 15
3.1 Development of a systematic risk assessment
The systematic risk assessment process proposed in – Involve other members of the design team or
this Manual assumes the involvement of several specialists e.g. geotechnical engineers, fire
parties, the suggested roles of whom are described engineers, blast engineers, transport engineers and
below. external stakeholders such as emergency
responders, whose input is needed to properly
Client/building owner inform the systematic risk assessment. Consider
– Ensure that the designer is competent and is involving an independent chair with particular
adequately resourced to carry out the work. expertise in risk management to facilitate a hazard
– Define the brief for the structural design, including identification and risk assessment workshop: the
any specific design requirements which may CDM coordinator or another risk specialist such as
indicate an enhanced level of performance is a HAZOP chairman, who is experienced in risk
necessary than that required solely for life safety. assessment but not directly involved in the design,
– Where the level of residual risk affects the client’s will often be a useful chair for a risk assessment
interests (e.g. because of an enhanced level of workshop.
performance, or because of the maintenance and – Develop the systematic risk assessment outlined in
inspection requirements associated with a this Manual, fulfilling the statutory duty to, so far as
particular design solution), agree this with the reasonably practicable, eliminate hazards from the
design team. In all cases the structural engineer, as design and reduce the risks associated with any
a designer, will have a duty to provide a minimum hazards that cannot be eliminated.
level of safety in the design, regardless of any – Develop the design of any risk reduction measures
client opinion. and coordinate the design with other members of
– Take due regard of safety such that the safety of the design team.
the design is not adversely impacted by any client – Provide the client, other designers (e.g. temporary
decisions. Design decisions reconciling cost, works designers, contractor’s designers),
programme and conflicting design requirements will contractors and sub-contractors with adequate
often lead to the client being defined as a designer information about any significant residual risks that
under health and safety legislation, then having the remain in the design.
same duty of care as the design team regarding the
elimination of hazards and reduction of risks Other designers (e.g. temporary works designers,
(see Section 1.3). contractors’ designers)
– Put in place risk management procedures – Provide input to the systematic risk assessment on
recommended by the structural engineer (e.g. any risks within each individual’s area of expertise,
fatigue/corrosion inspection and maintenance in particular on any risks associated with each
regimes, systems to limit the imposed loads applied design that could have an adverse impact on the
to weak floors, etc.) to control any risks that remain robustness of the main structure.
in the design. – Provide advice on any measures that may be used
– Communicate the residual risks to others (e.g. to reduce any risks found to be intolerable and the
building occupants) who may be affected by the implications of any proposed risk reduction
risks that remain in the design, so that they may measures upon each design.
identify and implement procedures to control their
risks which are appropriate to their operations. Other design team members/external stakeholders
– Provide input to the systematic risk assessment on
Structural engineer (designer) any risks within each individual’s area of expertise,
– Ensure the author of the systematic risk and on any measures that may be used to reduce
assessment possesses suitable qualification and any risks found to be intolerable.
experience (SQEP). Where this is not the case,
engage individuals who do have suitable Building control authority
qualification and experience. A single structural – Provide input to the structural engineer on the level
engineer should take overall responsibility for the of risk that the building control authority would
systematic risk assessment and will be the consider tolerable in the design, and any specific
named author of it, but the risk assessment will hazards about which the building control authority
have contributions from several individuals to has particular concern or specific expectations.
provide a wide range of expertise. SQEPness – Once prepared by the structural engineer, evaluate
should normally be demonstrated through the systematic risk assessment and its findings in
substantial previous experience in the design of terms of compliance of the design with the relevant
similar high-risk buildings, with experience in building regulations.
design against extreme hazards being
advantageous. CDM coordinator/risk management specialist
– Develop the structural design philosophy and the – May act as the chair for the hazard identification
approach to design against disproportionate and risk assessment workshop.
collapse to be adopted in the design of the – As an external and independent specialist
structure, and agree this with the client. experienced in risk assessment and not directly
– Review the normal design criteria and ensure their involved in the design, provide input into the
appropriateness for developing a safe structural identification of hazards and assessment of risks as
design. required by the structural engineer.
– Evaluate the tolerable risk threshold, based on the – Review coordination and information flow between
minimum level of risk reduction necessary to designers and ensure that no gaps exist between
achieve a safe design and any client requirement different designers’ responsibilities. This can be
for additional risk reduction to achieve an enhanced particularly relevant when some designers (e.g.
level of performance. temporary works designers or contractors’
– Seek advice from the relevant building control designers) are not appointed at the early stages of
authority on their expectations for the systematic a project when the risk assessment is first
risk assessment. prepared.
16 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Development of a systematic risk assessment 3.2
– CDM coordinator: as the duty holder responsible Some of the hazards that a systematic risk
for the preparation of the health and safety file, assessment might typically need to consider in the
ensure information is adequately provided by all design of a building against disproportionate collapse
designers on residual risks in the design, including are listed in Table 3.3. This is best used as a list of
properly developed proposals that allow the risk prompts to think about the hazards to which a
owner to control the residual risks, for example building might reasonably be subjected. It is
through inspection and maintenance regimes. intentionally incomplete: the structural engineer is
responsible for ensuring that all reasonably
foreseeable hazards with the potential to cause harm
are identified, and must ensure suitable qualification
3.2 Step 1: Identify the hazards and experience is available to do so. The
identification of hazards will be a team effort,
coordinated by the structural engineer. In some cases
3.2.1 Identification of hazards it may involve specialists being brought into the
design team to provide advice in specialist design
At the outset of the design, it is recommended that fields such as fire engineering and blast engineering
the structural engineer should document the or pre-construction advice.
proposed structural design philosophy and the
approach to design against disproportionate collapse Hazards are best identified by approaching the
and agree this with the client. This will be a live design from several different angles (Figure 3.3):
document that will be updated throughout the – by the nature of the design (e.g. whether innovative,
project. The purpose of preparing this prior to starting complex or unusual) and by reference to what has
the systematic risk assessment is to establish some gone before (including past failures and the lessons
principles and consensus about items such as: learned)
– the basis of design – by the type of action (e.g. environmental, imposed)
– the assumptions that will be made in the analysis – by the nature of the action (whether normal or
and design, particularly any simplifying abnormal; or whether natural, accidental or
assumptions malicious)
– the basic approach envisaged for design against – by the origin of the hazard (from the structure itself,
disproportionate collapse. external to/independent of the structure, deriving
from the use/maintenance of the structure or owing
The identification of the hazards is often best done to procedures surrounding the design, construction
through a hazard identification and risk assessment or use of the building)
workshop. A workshop, which might usefully be – by the state of construction: during construction
chaired by a person with specialist expertise in risk (including specific states at the different stages of
identification such as a CDM coordinator or a HAZOP construction e.g. due to lack of stability,
chairman, will often lead to a greater flow of ideas overloading, lack of fit, the occurrence of specific
than a desktop exercise. In some cases, particularly significant hazards or governing loads, or
in buildings where hazards are less readily identified, unforeseen loading conditions), during the building’s
a formal HAZOP will be prudent (refer to guidance operation and maintenance, due to deterioration or
from the Chemical Industries Association3.1, the lack of maintenance, or during demolition
Health and Safety Laboratory3.2 and Kletz3.3). – by part or section of the structure (substructure,
braced frames, sway frames, stability system;
The structural engineer should give careful original construction, extension, alteration; transfer
consideration to who should be involved in the structures, long spans, connection details)
hazard identification and risk assessment workshop. – by what has the potential to go wrong, be ignored,
As well as considering whether to have a CDM overlooked or misinterpreted (e.g. due to lack of
coordinator or HAZOP chairman leading the knowledge, failings in communication, the cutting of
workshop, ensure that different designers (e.g. corners).
temporary works designers, contractors’ designers)
are represented. The client should be involved. The
architect, geotechnical engineers, fire engineers, blast
engineers, transport engineers and external
stakeholders (such as emergency responders and
counter-terrorist security advisers) should be involved Nature of Type of
where relevant. the design action
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 17
3.2 Development of a systematic risk assessment
18 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Development of a systematic risk assessment 3.2
Note
a Risks during construction, alteration or demolition must be considered under the CDM Regulations but do not fall under the
Building Regulations. Determining the need for robustness in a temporary condition would normally fall under the duties of the
relevant contractor except if some sort of accidental damage would have a disproportionate effect, such as on an urban site or
an air rights building where the consequences might spread beyond the site itself.
Approaching the design from different angles is Complacency, lack of knowledge, programme and
advantageous on the basis that several attempts to commercial pressures all play their part in
identify the hazards affecting the design are better generating hazards.
than one. It will lead to repetition, but will also help
identify some hazards that had been missed through Advanced structural analysis tends to persuade us
the earlier approaches. Designers tend to think first of that we understand the structural behaviour of a
the technical things that could go wrong – a car system: a BRE/CIRIA survey of 120 structural failures
crashing into a column, a sustained fire, a fatigue- found that by far the biggest single reason for failure
induced failure. However, this is often at the expense is a grossly inadequate understanding of real loadings
of the procedural failings which are often much more and behaviour of the structure3.5.
important factors in the actual level of risk to which
the design is exposed. Study of past failures is simultaneously illuminating and
sobering, and is an essential but under-utilised tool in
It must be recognised that hazards are not merely the structural engineer’s armoury. Agarwal3.6, Delatte3.7
technical but also procedural. An error in design is rarely Levy3.8, Feld3.9, Kaminetzky3.10, Campbell3.11, Kletz3.12
the sole cause of the collapse. Almost without and Mann3.13 contain extensive discussions on lessons
exception it will be accompanied by errors in the learned from failures.
communication of the design, errors in the construction,
failures in the quality assurance and quality control The importance of identifying and considering all such
procedures for the design and construction, failure to procedural (‘soft’) hazards alongside the technical
stay within the limits of the defined usage for which the hazards in a systematic risk assessment cannot be
building has been designed, malfunction of mechanical emphasised too strongly. In some cases procedural
plant or machinery, unauthorised alteration of the hazards will be considered individually, and in other
building, and so on. cases they will be considered in relation to the effect
they may have upon other hazards that could occur.
Human error is persistent and invidious. People
make mistakes, take short-cuts, are put in a role 3.2.2 Foreseeable and unforeseeable hazards
they are not qualified to undertake, fail to carry out
effective checks, fail to ensure they are understood, It is a well-established principle of our codes of
fail to understand what others have done, and fail to practice that in the design of a structure the designer
ensure design interfaces are defined and controlled must consider all reasonably foreseeable actions
(‘I thought someone else was doing that’). Checking (hazards). However, a detailed consideration of a
procedures break down, audits fail to detect errors, particular load action is often deliberately neglected in
safeguards fail to operate, quality management conventional design because it is deemed ‘outside
systems and procedures are circumvented. the normal design envelope’.
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 19
3.2 Development of a systematic risk assessment
It is, however, foreseeable that loads could be answer lies in Vrouwenvelder3.16, who comments that
higher than designed, or that actions for which the hazards may be unforeseeable due to a lack of
structure is not specifically designed could professional knowledge. Mistakes will always be
materialise, or that the assumptions upon which the made, but the risk of mistakes will be significantly
structure is designed could be invalidated. It is also higher when the body of past experience available to
foreseeable that the structure may not be designers is smaller (Box 3.1). These are the known
constructed according to the design intent, or may unknowns made famous by Plato and latterly by
not be used as designed. It is not only foreseeable Rumsfeld3.18, and discussed in the Blackett review of
that mistakes could be made in the construction high impact low probability risks for UK
but is also foreseeable that the structural engineer Government3.19.
could themselves have made an error which goes
undetected3.14, 3.15. Could the problems have been foreseen, for example,
with high-alumina cement (HAC) that led to the
In addition to foreseeable hazards, BS EN 1991-1-73.4 collapses at Camden School for Girls and the Sir
explicitly requires the consideration of unforeseeable John Cass Secondary School in the 1970s (Box 3.2)?
hazards. At first this appears paradoxical – if a hazard Probably – certainly what was widely known was that
is unforeseeable, how can it be considered? The the body of knowledge related to high-alumina
Box 3.2 Camden School for Girls assembly hall, 1973 and Sir John Cass Secondary School swimming pool, 1974
Roofs constructed and designed with prestressed concrete beams using high-alumina cement (HAC) collapsed at Camden School
for Girls and the Sir John Cass Secondary School in Stepney in 1973 and 1974 respectively. High-alumina cement had enjoyed a
rapid rise in popularity due to its rapid hydration and early strength compared with Portland cement, giving faster and more
economic construction. Unfortunately it was not widely recognised that the calcium aluminates in the cement were unstable and
undergo a spontaneous chemical conversion, which is accompanied by a decrease in strength, an increase in porosity and therefore
an increase in susceptibility to sulphate and alkali attack. The chemical conversion is accelerated by increased temperature and
humidity, and is highly sensitive to the water/cement ratio of the concrete. Both roof structures failed spontaneously, in-service – in
Camden the roof was 18 years old at the time of the collapse which happened to occur after school hours, while at the Sir John
Cass school the roof was eight years old and a swimming class was using the pool until moments before the collapse.
20 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Development of a systematic risk assessment 3.2
cement was considerably smaller than that for properly identify them. In law this is no defence: the
Portland cement. Earlier failures had occurred: one in structural engineer holds a responsibility to ensure
1962 at Sheephouses reservoir3.20, 3.21, while several that they are suitably equipped to advise the client on
collapses had occurred abroad – several agricultural the hazards which are reasonably foreseeable.
sheds, each less than eight years old, suddenly Evaluation of potential actions is not merely based on
collapsed in Bavaria in 1961 in similar climatic the systematic modelling of a defined set of actions:
conditions to those in the UK. The chemical the structural engineer is expected to remain abreast
conversion and associated progressive loss of of the state of practice and must ensure that
strength was discussed as early as 19493.22-3.24 and modelling, analysis and design is consistent with
the occurrence of failures was predicted, while a contemporary good practice. This may be through
substantial increase in porosity of HAC was reported being conversant with the causes of recent near-
by the Institution of Structural Engineers in 19643.25. misses and structural failures from which much can
Greater caution could have and arguably should have be learned. These are reported by bodies such as
been applied noting the relatively unknown nature of Structural-Safety3.28 (incorporating the Standing
the material. Committee on Structural Safety (SCOSS) and the
system of Confidential Reporting on Structural Safety
Greater caution should be taken through the (CROSS)), but there is much that can be learned from
reduction in the tolerable risk threshold or judicious failures beyond the immediately applicable3.13.
increase in partial factors when working with new Learning can also come through being conversant
materials, manufacturing processes, unique or with research literature on advances in the
unusual structural schemes. This reflects the understanding of structural phenomena, material
increased risk of unwittingly overstepping some behaviour and mechanisms of failure. It is expected
bound in the design approach or the manifestation of that the structural engineer designing a high-risk
previously unrecognised phenomena3.26. building will be conversant with mechanisms and
behaviours that might not have previously been
The Royal Commission into the collapse of the West considered in an otherwise similar design.
Gate Bridge3.27 found that ‘‘. . . when leading
designers are working as pioneers, and only just 3.2.3 Cliff edge effects
within the bounds of the engineer’s knowledge, some
slight misjudgement, or failure to appreciate every In the Gloucestershire floods in 2007, river flows on
aspect of a new problem may prove disastrous and the Teme and Avon were the highest ever recorded;
bring tragic and fatal results. Under these conditions, river levels in the city reaching 4.92m. This was only
it is more than ever essential to employ really 1cm lower than the highest recorded level in the
adequate margins of safety and to ensure that they 1947 floods due to flood protection works that had
are not eroded by various unexpected and accidental been carried out on the Gloucester floodplain.
factors, including of course, imponderables and
human fallibility.’’ The probability of occurrence of the flood was
calculated as 0.5% per year, i.e. a 1 in 200 year
A key plank in protection against unforeseeable event. Overtopping of the perimeter wall of the main
events is a scenario-independent design approach electrical switching station was only prevented due
(Section 3.6), in which the building is designed to to the installation of temporary flood barriers by the
withstand the loss of any given loadbearing armed forces and emergency pumping, and the
element. If the building can sustain a degree of floodwaters came within 5cm of overtopping the
damage irrespective of the cause, it is immaterial wall. So while the consequences were already
that a particular hazard is unforeseeable because extremely serious, had the flood event been just
the structure is less sensitive to damage. It is slightly rarer or the flood waters slightly higher, the
primarily for this reason that the scenario- consequences would have been much worse had
independent design approach is so important in electrical power been lost that the event dominates
designing for robustness, because it ensures focus the overall risk profile, and all because of the simple
is placed on designing for a safe mode of failure, choice about the height of construction of a
rather than merely considering the behaviour under reinforced concrete wall.
a given set of loads.
The underlying flaw in the design of the substation is
The continuous advancement of the state of that it might have been designed for a 1 in ‘n’ year
knowledge is one pitfall for the unwary structural event, but that at a 1 in ‘n þ 1’ year event the
engineer working on the design of a high-risk likelihood was correspondingly marginally lower but
building; a second is working outside one’s area of the consequences substantially greater, meaning the
competence. In preparing the systematic risk overall level of risk was also substantially greater. This
assessment, it is to be expected that competence is symptomatic of low likelihood/high consequence
does not reside wholly within one person. In risks – because the consequences can exhibit this
particular, in identifying the hazards and cliff edge, such risks can dominate the overall risk
understanding the state of practice in response to profile.
particular hazards, recourse will often need to be
made to other specialists within the design team for The parallels in structural design are clear: in the
geotechnical engineering, fire engineering, evacuation design of a structural element which is critical to the
modelling, traffic modelling and blast engineering stability of the structure lacking redundancy through
advice. Recourse may also be needed to external alternate load paths, any cliff edges in the structural
advisers for advice on the likelihood and nature of behaviour must not be sufficiently close to the
malicious threats and the performance criteria for expected structural response to compromise the
which a particular asset should be designed. basis of the robustness design (Box 3.3).
Unforeseeable hazards are not the same as Further examples of cliff edges are given in
unforeseen hazards, which implies the failure to Section 3.10.2.
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 21
3.2 Development of a systematic risk assessment
Box 3.3 Cliff edge effects – Paris Charles de Gaulle Terminal 2E collapse, 2004
On 23 May 2004, a 30m section of the 650m long elliptical vaulted roof in Terminal 2E at the Charles de Gaulle Airport collapsed,
killing four people. The collapse occurred less than two and a half years since construction was completed, and less than a year
since the terminal opened. The roof structure consisted of external elliptical steel trusses acting compositely with a 300mm precast
concrete shell on the interior.
The official investigation3.29 found that the steel dowels supporting the concrete shells were too deeply embedded into the shells,
which caused cracking in the concrete that in turn led to a weakening of the roof, decreasing the stability of the structure. This also
increased the spreading forces at the springing points of the arch. The concrete blocks were also found to have been poorly
reinforced, meaning that the stresses flowed to the weakest points of the elliptical structure. The arches were supported on parallel
reinforced concrete longitudinal edge beams, supported by slim rectangular columns. The tendency for the superstructure to spread
under load was contained by steel stirrups tied around the column head and bolted to the longitudinal beam, though it is thought
that these were added after the construction had started on site. During construction, repairs were made to the supporting columns
after cracks appeared – a forewarning of the collapse.
The cause of the collapse was found to be long-term concrete creep, exacerbated by the rapid thermal expansion/contraction of the
structure. The temperature dropped from 258C a week before the collapse to 48C, on the morning of the collapse. The differential
thermal movement between the steel truss and the concrete panel acted to exacerbate the cracking due to concrete creep, critically
weakening the highly-stressed concrete.
The report concluded that the edge beam supporting the arches fractured, causing the loss of the springing point of the arch and
the collapse of the roof. In this the design was literally a cliff edge: the lack of redundancy in the arch meant that collapse across
the full width of the terminal was inevitable, and this led to the collapse of the entire length of the structure between expansion
joints. If the structure had been continuous across the expansion joints, it seems unlikely that the forces that occurred during
collapse could have been supported by the adjacent sections and a progressive collapse through the length of the structure could
have been possible.
The cliff edge, a shear-critical design, violates the safe design theorem (see Hambly’s paradox in Box 3.9). It fails to consider whether
the structural component has the strength to transmit its calculated force, the ductility to retain its strength while deforming, and
sufficient stiffness to keep deflections small and to avoid the onset of a local instability before design loads are reached.
The report of the investigation also questions the existence of proper design and construction processes and technical controls on a
project where owner, project manager and architect were essentially the same. The state-owned airport company not only designed
the building but managed the construction, virtually eliminating the distinction between design team and client. This again
emphasises the need for independent peer review, robust analysis and greater redundancy, particularly in unique structures: Leslie
Robertson, the structural engineer responsible for the design of the World Trade Center, commented in an interview to the New York
Times that ‘‘when problems occur, it is usually in the interface’’ between architects, engineers and contractors – ‘‘seldom can one
say with any certainty, ‘That’s it, that right there is where the trouble happened.’’’3.30
22 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Development of a systematic risk assessment 3.2
× =
Likelihood Consequence Risk Abnormal action but within the design basis
× =
× =
Box 3.4 Low likelihood/high consequence hazards – Eschede train disaster, 1998
On 3 June 1998, the growth of a fatigue crack in the steel tyre of an intercity express train from Munich to Hamburg caused the
sudden failure of the tyre, which embedded in the floor of the train car3.31. As the train travelled over two sets of points outside
Eschede, the embedded tyre caught in the guide rail of the points, switching them and causing the derailment of the rear ten
coaches of the train. The derailed coaches impacted the central support of a reinforced concrete overbridge, causing the
catastrophic collapse of the bridge onto the train and 101 fatalities.
Shortcomings were found in the design of the wheel, which was a dual bloc design to minimise noise and vibration, comprising a
wheel body surrounded by a rubber damper and a relatively thin steel tyre. Specifically, the following were not considered in the
design:
– the flexure of the tyres on the soft rubber damper with each revolution of the wheel and the corresponding fatigue effects due to
the oscillating load
– the potential for crack growth to initiate on the inside of the tyre
– as the tyre wears and the thickness reduces, the increase in dynamic forces and their exacerbation of crack growth
– the exacerbation of dynamic forces and consequential acceleration of wear due to low or high spots on the tyre.
Despite these failings and the failure of the train operator to take heed of warning signs prior to the accident, the likelihood of the
impact of the train onto the bridge would have been calculated as exceptionally low, because of the condition for the failure of the
tyre and the derailment to occur precisely when they did. Nevertheless the consequences, given that the hazard materialised, were
catastrophic, and the intrinsic risk therefore substantial.
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 23
3.2 Development of a systematic risk assessment
There are, however, a number of adverse events easily identified, including where human error has the
which have a high likelihood of causing a secondary potential to be a significant factor in the generation of
event that is not statistically independent, and should hazards or the control of risks through procedural
therefore be considered in combination measures.
(consequential hazards). Many of these consequential
hazards involve fire (Box 3.5), for example an 3.2.7 Malicious hazards
incendiary fire following an earthquake or explosion
due to the fracture of a gas main, leak of fuel or the Malicious hazards should be considered in the design
ignition of combustible materials. It is necessary to of a building when the hazard is reasonably
assess the potential consequences of an adverse foreseeable. The judgement about what to consider
event and properly consider where combined is necessarily subjective, but when the consequences
hazards need to be considered. Again, a HAZOP is of an attack could be disproportionate – for example
often the best vehicle for properly considering the theft of copper cables or lead roof cladding, the
consequential hazards. malevolent vandalism of a railway bridge or the
cutting of the stay cables on a cable-stayed bridge –
it is clear that they should be considered (Box 3.6).
Box 3.5 World Trade Center, 11 September 2001
The Twin Towers were designed in the mid-1960s, and were Box 3.6 Mersey Valley Footbridge, 4 January
designed to withstand the accidental impact of a Boeing 707 2007
aircraft. In 1945, a B-25 bomber had crashed into the
79th floor of the Empire State Building after becoming lost in The Mersey Valley and Sale Water Park footbridges are two
fog while trying to land. The design of One and Two World cable-stayed footbridges over the M60 in Manchester, each
Trade Center assumed an impact would be during the aircraft with a solid abutment and A-frame supported by a 28m back
descent to land, and therefore at relatively low speed and span on one side of the motorway, with stay cables
relatively low on fuel. It did not consider a higher-velocity or supporting the 74m main span. Each bridge has 20 stay
more massive impact, or the effect of a subsequent fuel cables, of which half anchor the A-frame to the back span
fire3.32, 3.33. and the remaining ten (five on each side) support the main
span. On 4 January 2007, a member of the public alerted
Nevertheless, the substantial structural robustness in the police to damage of the stay cables supporting the
design of the Twin Towers meant that when deliberately Manchester Valley footbridge. Four stay cables had been
impacted by a much larger aircraft, a Boeing 767 (flying at found to have been severely damaged with high tensile wire
similar speed to that considered in design but a heavier strands being severed, and broken hacksaw blades were
aircraft, in addition to the mass of a full fuel load comprising found nearby. Subsequently protective sleeves were fitted
nearly an additional one-third of the aircraft’s weight), the over the cables on both bridges to prevent direct access to
design of the towers exhibited a remarkable degree of the cables.
robustness. They were able to sustain the loss of a significant
number of perimeter columns over several floors of the
building, yet redistribute these loads and remain standing
under the ensuing fire for nearly 60 minutes after the attack. Similarly, when the likelihood of a terrorist attack is
non-negligible, consideration should be given to
While the consideration of fire following impact was not
terrorist threats. At a high level, the judgement about
common practice in the 1960s or for several decades
what to design for is most readily determined by
thereafter, the combined hazard of a sustained fuel fire
considering the criticality and vulnerability of the
following impact is now only too foreseeable in the design of
asset. The asset may be critical either to the asset
such buildings.
owner or to a third party. Criticality may be
considered in terms of consequences such as the
economic cost of downtime or asset reinstatement,
3.2.6 Human error the loss of human life in the event of failure, or the
iconic importance of an asset. Vulnerability expresses
Human error is perhaps the most difficult type of the fragility of the asset, i.e. how close it is to failing. If
hazard to spot. Kletz has written extensively on the an asset is both critical and vulnerable, this is a
subject of human error3.12 and outlines five types of strong indicator that the designer should consider
human error which are a useful aid in identifying the terrorism in the risk assessment. A specialist threat,
types of errors that can arise in each phase – design, risk and vulnerability assessment (TRVA) is usually
construction, operation and maintenance: necessary to evaluate the risk due to terrorism.
– errors due to a slip or momentary lapse of attention Guidance on malicious hazards and vulnerability of
– errors due to poor training or instructions, i.e. structures to such hazards should be sought from
through ignorance or for wont of knowing any the UK Centre for the Protection of National
better Infrastructure3.34 or from a member of the Register of
– errors due to a lack of physical or mental ability Security Engineers and Specialists3.35, or the
– errors due to a lack of motivation, or to a equivalent resources in the project’s local jurisdiction.
deliberate decision not to follow instructions or
expectations 3.2.8 What to include
– errors made by managers, often due to a lack of
appreciation of the part they should or could play. A systematic risk assessment should include all
reasonably foreseeable hazards (notwithstanding that
Kletz developed the technique known as HAZard and Eurocodes refer to unforeseeable hazards). The
OPerability studies (HAZOP) aimed at identifying emphasis is on reasonably foreseeable, i.e. hazards
hazards associated with both technical and human for which the risk is not so vanishingly small that there
fallibility. While developed for the process industry3.3, would be no design impact.
it is equally applicable across the engineering
profession. It is recommended in this Manual for the Consequently the systematic risk assessment should
design of those structures for which hazards are less include natural ‘perils’ such as earthquake,
24 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Development of a systematic risk assessment 3.3
windstorm, fire, flood, rainwater ponding, adverse by equally high-rise or higher-rise buildings. However,
snow loading, and ice accretion, where relevant. Risk for buildings on the margins of a high-rise area or for
assessments have been prepared which include buildings which exceed the height of the surrounding
‘volcano eruption’ amongst the structural hazards, for buildings, past events show that aircraft impact is a
buildings in no proximity even to a dormant volcano. credible risk, particularly of light aircraft due to pilot
Similarly, tsunami would only be relevant in the error (Box 3.7). Most observers would accept that
coastal region of a tsunami-prone area such as the collapse resulting from the impact of a heavy
Pacific Rim or the Indian Ocean. However, sea level passenger or cargo aircraft would not be
rise due to climate change, in combination with the disproportionate, but collapse would not be expected
storm surge due to the coincidence of a deep to result from the impact of a light aircraft (including
meteorological depression, spring tides and adverse post-impact fire).
winds, should be considered where the overtopping
of flood protection would have structural
consequences. Similarly, it might be valuable to have Box 3.7 Light aircraft crash, New York City,
determined that the extreme wind load is not that 11 October 2006
much larger than the 1 in 50 year load, and that the
structure is insensitive to this change. Hazards from On 11 October 2006, a two-seater fixed-wing single-engine
adjacent or nearby sites (e.g. industrial hazards) are aircraft crashed into the 30th floor of an apartment block on
more difficult to evaluate but should be included. the Upper East Side of Manhattan while attempting to
Information should be available from the site owner or execute a 180 degree banked turn above the East River.
operator. The National Transportation Safety Board determined that
pilot error was the cause of the crash, the bank angle being
The systematic risk assessment should also include insufficient to execute a 180 degree turn within the
man-made accidental and malicious hazards. This will available width; this being further reduced due to 13 knot
include hazards such as accidental impact and crosswinds. The crash killed both occupants of the aircraft
malicious hazards such malevolent vandalism and and caused a post-impact fire and 21 injuries,
terrorism. predominantly burns.
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 25
3.4 Development of a systematic risk assessment
out the hazard – eliminating the hazard from the further than the immediately adjacent storeys3.4.
design – results in a both a safer and a more Taking into account the accidental nature of the
cost-effective solution than having to reduce the risk events to which this relates implies that this might be
arising from it through design measures or through deemed tolerable for events with up to, say, a 2%
operational maintenance and inspection. probability of occurring once during the lifetime of the
building. This guidance is given in the context of
typical forms of construction and accidental
loadcases, and BS EN 1991-1-7 Annex A notes that
the extents of damage considered to be
3.4 Step 3: Determine the tolerable risk disproportionate may vary between different types of
threshold building.
For the reasons described in relation to Figure 3.2, The inclination to point the finger of blame applies
area of collapse is commonly used as a measure of particularly to man-made hazards, but also holds for
consequence, and therefore of proportion or environmental hazards because perception tends to
disproportion. In determining the tolerable threshold it be that the hazard should have been foreseen.
may be useful to consider fatalities, economic loss Perception is often markedly different with the benefit
and asset disruption, but to facilitate the workings of of hindsight, and it is strongly recommended that
the risk assessment area of collapse is used as a decisions about the amount of tolerable damage
measure of consequence. A collapse may be said to should be made by looking at them through a lens of
be proportionate if the area of collapse is sufficiently how the decisions would be judged if the event were
limited when compared with the likelihood of the to actually occur.
hazard that brings about the collapse. For typical
accidental events, the area of collapse in BS EN
1991-1-73.4 and typical occupancy densities
suggests that fewer than, say, five fatalities might be 3.5 Step 4: Evaluate the risk
deemed tolerable.
Practically, the determination of the tolerable risk is As defined in Equation 2.1 (Section 2.2), risk is
usually best undertaken using likelihood as a frame of evaluated as likelihood consequence. Both likelihood
reference: for a hazard which is expected to have and consequence, and the resulting evaluation of the
greater than a 2% chance of occurring once during magnitude of risk, are usually best described using
the lifetime of the building, how significant an extent qualitative scales. Quantitative risk analysis can be
of collapse would be considered intolerable? How undertaken in some circumstances and, while the
would this vary for more frequent or more infrequent principles remain the same, the details are beyond the
hazards? scope of this Manual; however, information is available
from Vrouwenvelder3.16, Faber3.43, Sørensen3.44,
In current codes of practice, a collapse is typically Izzuddin3.45 and Janssens3.46.
judged as being proportionate provided the area of
floor at risk of collapse does not exceed 15% or In evaluating the consequences of the hazard,
100m2 (whichever is the lesser) of the area of the consider the sensitivity of the design to the underlying
floor area of the affected storey and does not extend assumptions and undertake sensitivity studies where
26 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Development of a systematic risk assessment 3.5
Catastrophic
Substantial
With reference to the risk matrix shown in Figure 3.2
Significant
Minimal
or a suitable alternative developed for the project, plot
Serious
Severe
Minor
the risks arising from each of the hazards identified
(example in Figure 3.5). These are the ‘raw’ or
unmitigated risks. Some could fall above the tolerable
Likelihood
Frequent/common
risk threshold; typically the majority will fall below it
but as discussed in Section 2.4 will still warrant
Likely
further reduction such that the residual risk is as low
as is reasonably practicable.
Unlikely
Sensitivity studies must be undertaken using well-
Rare
founded engineering analysis to evaluate the
sensitivity of the design to the underlying
Improbable
assumptions and the effect that small changes will
have on the consequences calculated for the given
hazard (Box 3.9). Note: the likelihood of an event also
Negligible
potentially exhibits sensitivity but is generally less Note
likely to do so. Each dot plots the risk arising from a particular hazard.
Figure 3.5 Example risk matrix showing the ‘raw’/unmitigated risks arising from the
hazards considered in the design
60kg 60kg
A three-legged stool, one of the few situations that engineers can analyse with certainty;
and a four-legged stool, one of the innumerable that we cannot (after Burland3.49)
A robust, near-rigid stool, standing on a firm, nearly-rigid floor, will rock: three of the legs will be in contact, supporting the weight
of the man, but the fourth will be clear of the floor. Even if this fourth leg is clear by only a fraction of a millimetre, then it is certain
that the force it is carrying is zero. By simple statics, the force in the leg diagonally opposite will also be zero, even if this leg
appears to be in contact with the floor. The weight of the man is in fact supported symmetrically by the other two legs of the stool,
and each must therefore be designed to carry a force of 300N.
Now the stool may be imagined to be placed arbitrarily on a randomly rough floor and, as there is no way of deciding which leg will
be off the floor, all legs must therefore be designed to carry a force of 300N. There is thus a paradox apparent in the problem –
the addition of a fourth leg to a three-legged stool increases, rather than decreases, the force for which each leg must be
designed. This would be missed entirely assuming perfect geometry and ideal boundary conditions in simple analysis. The boundary
conditions are apparently so simple that no thought is given to their specification.
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 27
3.6 Development of a systematic risk assessment
Hambly recognised that current design methods work for ductile structures because of the safe design theorem, which derives from
the lower bound theory of plasticity3.50-3.52 and states that a structure can carry its design loads if, and only if:
– the calculated system of forces is in equilibrium
– each component has the strength to transmit its calculated force and the ductility to retain its strength while deforming
– the structure has sufficient stiffness to keep deflections small and to avoid the onset of a local instability before design loads are
reached.
Thus, even if the real structure deforms under load with a different flow of forces from that calculated, it will still be safe as long
as:
– the failure modes are adequately ductile
– there is no risk of local instability.
Good robustness and design against disproportionate collapse is little more than the application of this principle, albeit with
the significant question for the engineer of confirming that all the failure modes of the structure as designed are adequately
ductile3.53.
3.6 Step 5: Identify risk reduction result from the implementation of each (Figure 3.6).
measures Necessarily, any evaluation of benefit will usually
involve a degree of subjectivity, either because the
measures are ‘soft’ measures for which the benefit
For each risk in turn, irrespective of where the risk lies cannot be quantified with precision, or simply because
in relation to the tolerable risk threshold, consider the benefit is determined through analysis which
measures that could be implemented to reduce the contains inherent uncertainties and approximations.
risk. For any given hazard, several potential risk
reduction options will typically exist, which will vary in Some measures will have a stronger effect upon the
terms of cost, effectiveness and utility/design impact. likelihood than upon the consequence and vice versa.
The structural engineer must, of course, check that Some, as with one of the risk reduction measures
any risk reduction measures do not cause an proposed in Figure 3.6, might decrease the likelihood
increase in the risk owing to another hazard, for but result in an increase in the consequence, though
example by causing a change in stiffness which still be effective in reducing the risk overall. An
affects the seismic demand on the structure. example might be spacing the expansion joints
further apart in the structure, improving the tying in
For all hazards, risk reduction measures must at least the structure such that a horizontally progressive
reduce the risk below the tolerable threshold. The collapse is less likely to occur. If it does occur,
engineer is duty-bound to reduce risks to a level however, it will automatically be greater in its extents.
which is as low as reasonably practicable, and should
therefore consider both measures that reduce the risk In the same way as sensitivity studies were carried
to just below the threshold as well as more powerful out in Step 4 to evaluate the sensitivity of the
risk mitigation measures. Determination as to which unmitigated risk to changes in the underlying
of the potential risk reduction measures is most assumptions, it is essential that as part of this step a
appropriate happens through a cost-benefit sensitivity study is undertaken on the mitigated risk.
assessment in Step 6. This sensitivity study should examine whether, when
each potential risk reduction measure is
To support the subsequent cost-benefit assessment, implemented, the resulting design remains sensitive
the risk reduction measures associated with each or is made insensitive to changes in the initiating
hazard should be plotted on the risk matrix by event. This study can usually be undertaken by
evaluating the residual or mitigated risk that would considering the effect of each potential risk reduction
28 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Development of a systematic risk assessment 3.7
Catastrophic
Substantial
the structure to withstand the sudden loss of the
Significant
Minimal
Serious
column will automatically make the building able to
Severe
Minor
withstand the impact of a larger vehicle (Figure 3.7).
Thus under larger initiating events, measures that
eliminate or reduce the sensitivity will show a greater
Likelihood
Frequent/common
reduction in the consequences of the hazard than
those that do not. Likely
Designing the building using a scenario-independent
Unlikely
approach such that it is capable of withstanding the
sudden loss of a single element forming part of the
Rare
vertical load path or the lateral stability system, will
always be preferable to hardening the structure in a
Improbable
scenario-dependent approach. This is simply
because it renders the structure less sensitive to the
Negligible
magnitude of the hazard and is effective across
multiple hazards. The structural design will be
insensitive to assumptions about, for example, the
Figure 3.6 Residual risks owing to various different risk reduction measures for a given
velocity of a vehicle impacting the column, the
hazard
magnitude of an explosion or the quality of welding
compromising the ductility of the connections under
stress. As a result, the structure is better able to
resist events of any cause, irrespective of whether
they have been explicitly considered, and a better Consequence
level of robustness is achieved in the design.
Catastrophic
Substantial
Significant
Consequently a scenario-independent approach is
Minimal
Serious
Severe
Minor
preferred over those approaches in which sensitivities
remain. It is therefore recommended that a structure
is, wherever possible, designed to withstand the
Likelihood
Serious
Severe
Minor
supply chain
– an external and independent check of the design
(known in the UK as a Category 3 check3.54)
Likelihood
Rare ii)
Such procedural risk reduction measures must be
much more than just an after-thought, and are some
of the most important risk reduction measures that Improbable
can be implemented in a design.
Negligible
(b) scenario-independent design
Note
3.7 Step 6: Cost-benefit assessment The risk reduction associated with option (b) is greater than option (a).
In this step, the benefit resulting from the risk Figure 3.7 Raw and mitigated risks associated with the accidental impact from i) a 3.5T
reduction measure determined in Step 5 is weighed van and ii) a 40T truck
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 29
3.8 Development of a systematic risk assessment
against the cost of implementing the measure. ‘Cost’ The statutory duty to reduce risks to a level which is
may be measured in financial cost, programme delay, as low as reasonably practicable was emphasised in
impact on design aspirations such as aesthetics, loss Step 5, and in all cases the mitigated risk must fall
in amenity, utility or function, and so on. Benefit below the tolerable threshold for that risk. A
should not be underestimated: while costs are by responsible structural engineer experienced in risk
comparison easily quantified, benefits are not. The assessment will be familiar with these concepts and
engineer must remember that the real benefit is a will be able to guide the decision-making process
reduction in the fatalities and injuries that will result, such that the overall level of risk is appropriately
should an accidental or malicious hazard develop. It managed. If it is concluded from the systematic risk
is often more illuminating to consider whether the assessment that one or more risk reduction
benefits are ‘worth having’ through a conditional measures are necessary, the cost of which is
assessment: given that the event may occur, is the significant, the basis for these requirements has been
reduction in risk worth the cost? defined in a reasoned, justified manner and the
commercial aspects of the project must be adjusted
The cost-benefit assessment will often be qualitative accordingly. It should be remembered however that in
and based upon good engineering judgement. Often the risk management process, the action taken (and
a relative cost-benefit assessment of whether the cost expended) should at all times remain
Option A has greater potency than Option B is more proportionate to the risk. If this has been shown to be
useful than an absolute assessment, particularly for the case, it indicates a necessary risk reduction
‘soft’ measures where the cost cannot be easily measure and the viability of the project must be
determined. Contemporary good practice will often reviewed: cost alone is not a reason for not
be the ultimate measure against which different risk implementing a particular measure.
reduction measures are evaluated.
Equally and as highlighted in Section 2.4, if after
In the cost-benefit assessment, each of the risk going through the previous steps it has not been
reduction measures is considered in turn, rather possible to reduce the risk to a tolerably low level, it
than each of the hazards in turn as above. The total must be considered whether the proposed design is
benefit of a given risk reduction measure is merely worth the risk. This may mean going back to the
the total risk reduction determined in the previous conceptual stage of design and approaching the
step, aggregated across all of the hazards impacted robustness of the structure and perhaps the
by the particular risk reduction measure being structural design as a whole in a different way.
considered.
The cost-benefit assessment permits the total Any risks which fall above the tolerable risk threshold
reduction in risk from each of a range of mitigation must be mitigated, as a minimum such that they are
measures to be evaluated, and importantly, to be brought below the threshold; all raw risks falling
compared in deciding where to target finite resources below the threshold should also be reduced so far as
to derive the maximum value. Some notes are helpful it is reasonably practicable to do so. As highlighted in
in differentiating between different risk reduction Section 1.3, efforts to reduce the risks that remain in
measures where the reduction in risk is similar: the design should continue until the point of gross
– Where one measure is scenario-independent and disproportion is reached. It is through the cost-benefit
the other is scenario-dependent, for the reasons assessment that the location of what represents
described earlier the scenario-independent ALARP and thus the point of gross disproportion is
measure should be prioritised. judged. This is likely to be realised through
– Where the above distinction does not apply, the contemporary good practice and professional
measure impacting the larger raw risks is the more judgement. Guidance on deciding whether a risk is
important of the two. ALARP is given by the Institution of Civil Engineers3.56
– Where one measure mitigates a greater number of and the Health and Safety Executive3.57.
risks than the other but the total level of risk
reduction between the two is broadly similar, the It is by this point that both design changes and ‘soft’
one impacting the larger number of risks is procedural risk mitigation measures will have been
generally the more important. identified. The structural engineer should now be able
to articulate the findings of the systematic risk
The total cost of all risk reduction measures and the assessment in terms of modifications to the structural
aggregated risk reduction can be a useful, though design. They should, however, also be able to
ancillary, measure for demonstrating the value given articulate the necessary precautions to be taken in
by the risk assessment. the procurement process to ensure quality, the effect
30 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Development of a systematic risk assessment 3.9
upon the construction sequence, controls to ensure ultimate force to which it might be subjected, and yet
adequate quality management through the supply still be governed by a brittle failure mode such as
chain and during construction, inspection and shear but be designed ‘in accordance with the codes
maintenance regimes to be implemented during the of practice’. This would completely overlook the
operation of the building, and so on. The financial importance of the element to the robustness of the
cost associated with the risk mitigation measures structure. If an element is shown to be critical to the
should also now be available. What follows in the design of the structure, a sensitivity study should be
remaining steps is largely to do with testing the undertaken to ensure that there is only a gradual
sensitivity of the risk assessment to the underlying change in the response of the structure at a load of
assumptions, which may modify the findings but is up to, say, 40% above the design (i.e. factored) load.
unlikely to cause wholesale changes, and the review
and acceptance of the residual risks by others.
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 31
3.11 Development of a systematic risk assessment
Serious
Severe
Unlikely
Serious
Severe
Frequent/common
instance the mitigation measures are not particularly
Likely important but there remains a significant risk to
occupants (Figure 3.8b), and the design is sensitive to
Unlikely the assumptions made in the risk assessment.
32 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Development of a systematic risk assessment 3.12
not be a mechanistic review of the content of the risk disproportionate collapse have been documented,
assessment and the analysis undertaken but an one vehicle for the communication of this information
independent review of the principles adopted, the is through an update of this document. Often it will be
sensitivity of the design and the assumptions made. preferable to also provide information to each
One potential approach is given by SCOSS3.63. affected party that just details the risks and
precaution measures about which they need to be
aware.
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 33
3.13 Development of a systematic risk assessment
3.17 Bunni NG. Risk and insurance in construction. 2nd ed. report of the National Construction Safety Team on the
Taylor & Francis, 2005. collapses of the World Trade Center tower. September
2005. Available from: http://wtc.nist.gov.
3.18 Rumsfeld D. US Department of Defense news briefing.
12 February 2002 Available from: www.defenselink.mil/ 3.34 Centre for the Protection of National Infrastructure.
Transcripts/Transcript.aspx?TranscriptID=2636. www.cpni.gov.uk.
[Accessed 2012-11-28].
3.35 Register of Security Engineers and Specialists.
3.19 Government Office for Science. Blackett review of high www.ice.org.uk/rses.
impact low probability risks. Department for Business,
Innovation and Skills, 2011. 3.36 Department for Transport. Managing the accidental
obstruction of the railway by road vehicles. London:
3.20 Neville AJM. History of high-alumina cement. Part 1: Department for Transport, February 2003.
Problems and the Stone report. Proceedings of the
Institution of Civil Engineers – Engineering History and 3.37 Rail Accident Investigation Branch. Rail accident report
Heritage 2009;162(EH2):81-91. – road vehicle incursion and subsequent collision with
a train at Stowmarket Road, 30 November 2011.
3.21 Neville AJM. History of high-alumina cement. Part 2: Report 25/2012. Rail Accident Investigation Branch,
Background to issues. Proceedings of the Institution of November 2012.
Civil Engineers – Engineering History and Heritage
2009;162(EH2):93-102. 3.38 Ellingwood BR. Acceptable risk bases for design of
structures. Progress in Structural Engineering and
3.22 Neville AJM. A study of deterioration of structural Materials 2001;3(170-179).
concrete made with high-alumina cement. Proceedings
of the Institution of Civil Engineers 3.39 Hambly EC, Hambly EA. Risk evaluation and realism.
1963;25(3):287-324. Proceedings of the Institution of Civil Engineers
1994;102(5):64-71.
3.23 Neville AJM. A study of deterioration of structural
concrete made with high-alumina cement – 3.40 Ale BJM. The use of risk information in the
discussion. Proceedings of the Institution of Civil Netherlands. 7th Annual European Summer School on
Engineers 1964;28(1):57-84. Major Hazards. Christ College Cambridge: 6-10 July
1992.
3.24 Lea FM, Davey N. The deterioration of concrete in
structures. Journal of the Institution of Civil Engineers 3.41 Ball DJ, Floyd PJ. Societal risks. Report prepared for
1949;32(7):248-295. the risk assessment policy unit, Health and Safety
Executive. Risk & Policy Analysts Ltd, 1998.
3.25 Institution of Structural Engineers. Report on the use of
high-alumina cement in structural engineering. London: 3.42 Health and Safety Executive. Societal risk technical
IStructE, 1964. seminar progress – Summary Note 04/09. Available
from: http://www.hse.gov.uk/societalrisk/seminar0409/
3.26 Pugsley AG. The safety of structures. London: Edward summary-note.htm. [Accessed 2012-09-25].
Arnold (Publishers) Ltd, 1966.
3.43 Faber MH, Narasimhan H. Robustness of structures:
3.27 Report of the Royal Commission into the failure of the final report of COST Action TU0601. September
West Gate Bridge. 2.-7037/71. Melbourne, Victoria: 2011. Available from: Available from: www.cost-
1971. tu0601.ethz.ch.
3.28 Structural-Safety website. Available at http://www. 3.44 Sørensen JD. Theoretical framework on structural
structural-safety.org [Accessed: 27 February 2013]. robustness. June 2011. Available from: Available from:
www.cost-tu0601.ethz.ch.
3.29 Conseil National des Ingenieurs et Scientifiques de
France. Synthese des travaux de la commission 3.45 Izzuddin BA, Pereira MF, Kuhlmann U et al. Application
administrative sur les causes de l’effondrement d’une of probabilistic robustness framework: risk
partie du terminal 2E de Paris-Charles de Gaulle. assessment of multi-storey buildings under extreme
Paris: Ministry of Transportation, Urban Design, loading Structural Engineering International
Tourism, and Sea, 2005. 2012;22(1):79-85.
3.30 Hawthorne C. Design notebook: the architectural 3.46 Janssens V, Sørensen JD, O’Dwyer DW et al.
blame game. The New York Times, 2004; Assessing the consequences of building failures.
27 May 2004. Structural Engineering International
2012;22(1):99-104.
3.31 Danger ahead – historic railway disasters. Special
feature: Eschede, Germany: ICE high-speed train 3.47 Hambly EC. Oil rigs dance to Newton’s tune.
disaster. http://danger-ahead.railfan.net/features/ Proceedings of the Royal Institution of Great Britain
eschede.htm. [Accessed 2012-11-17]. 1985;57:79-104.
3.32 Leslie E. Robertson Associates. WTC Sept 11, 2001. 3.48 Heyman J, Hambly EC. Hambly’s paradox: why design
http://www.lera.com/sep11.htm. [Accessed calculations do not reflect real behaviour. Proceedings
2012-11-28]. of the Institution of Civil Engineers: Civil Engineering
1996;114(4):161-166.
3.33 National Institute of Science and Technology. NIST
NCSTAR 1: Federal building and fire safety 3.49 Burland J. Interaction between structural and
investigation of the World Trade Center disaster: Final geotechnical engineers. 6th Athenian Geotechnical
34 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Development of a systematic risk assessment 3.13
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 35
4 Approaches to design against disproportionate
collapse
36 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Approaches to design against disproportionate collapse 4.5
Eurocode 0 (BS EN 1990:20024.11) defines the Alternative load path analysis Compartmentalisation
combination of actions for accidental loading. A
typical building with imposed office floor loading and
wind as accompanying variable actions gives the Load redistribution (LR) methods Local protection (LP) methods
two loadcases shown in Equations 4.1 and 4.2.
The partial (g-) factors and combination (c-) factors
are taken from the UK National Annex4.12. Figure 4.1 Categorisation of approaches to design against disproportionate collapse
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 37
4.7 Approaches to design against disproportionate collapse
Cormie4.14, 4.22.
(b) Effect of load redistribution (LR) approaches on the structural response 4.7 Scenario-independent design
Element failure and
onset of collapse Scenario-independent design, whereby the structure
Capacity
38 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Approaches to design against disproportionate collapse 4.8
The potential mechanisms of resistance against 4.3c). Depending on the type of construction, it may
progressive collapse and the analytical approaches also be possible to mobilise resistance in shear walls,
for undertaking a scenario-independent design are internal partitions, and in secondary elements such as
discussed in Sections 4.8 and 4.9 respectively. through racking in the façade system, though these
ancillary mechanisms will often be difficult to quantify.
Additional
reaction
Section forces
(a) Catenary action in structural beam/column frame of an internal column after removal of a
supporting column
Compression ring
Tension
membrane
Plan
(b) Tensile membrane action developed in a flat slab after the removal of the central column
C C
T T
Section
(c) Compressive arching action between composite metal deck slab and steel floor beams
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 39
4.9 Approaches to design against disproportionate collapse
4.9.2 Linear static procedures response of the system and therefore unknown, must
be assumed as an input variable. Assuming a
Linear static procedures are the simplest analytical dynamic load factor of 2.0 for elastic behaviour will
procedure, but contain significant assumptions and generally lead to a grossly conservative design, so for
are suited only to very limited circumstances. most practical purposes design based on linear static
Dynamic effects are modelled through the use of a analysis is not feasible or economic.
dynamic load factor, for which the maximum value is
usually 2.0 if the building is to remain elastic. 4.9.3 Nonlinear static analysis based on dynamic
energy balance
In a linear static analysis it is assumed that the
system is both geometrically and materially linear, i.e. A nonlinear static analysis which accounts for both
that large displacements do not occur and that the geometric and material nonlinearities in the
material remains elastic. Neither of these are usually structural response can provide a relatively
true in a robustness problem, and these assumptions straightforward analytical solution without the need
are therefore approximations which significantly limit for recourse to more advanced nonlinear dynamic
the extent of the application of this type of analysis. analysis. Pushover analysis is a technique
‘Overstress’ can be used as a predictor of damage originating from the seismic field, whereby the
but only when the levels of plasticity are low such that ultimate static capacity of the structure is derived
load redistribution due to yielding is insignificant. through a statically-applied lateral (‘pushover’) load.
Consequently, linear static analysis should only be Pushover analysis can be undertaken to evaluate
used when the structure remains broadly elastic. For the ultimate capacity of the structure following the
linear static analysis to be valid, none of the following loss of a column/structural element. The element is
should apply: notionally removed from the model and resistance
– irregular geometry mechanisms such as slab flexure, compressive
– irregular distribution of stiffness arching, tensile membrane action and catenary
– non-orthogonal lateral stability system action are modelled. A notional load is applied in the
– asymmetric lateral stability system direction of the removed force (Figure 4.4) and its
– irregular transfer structures. magnitude gradually increased, so that plasticity
successively develops in the structure as load is
In addition, and as a general rule of thumb, ductility redistributed until a collapse mechanism forms. The
ratios (for which ‘overstress’ factors, or force-based ultimate load capacity of the structure is hence
demand/capacity ratios, are often used as a proxy) determined as a function of the resistance
must be less than or equal to approximately 2.0 for a mechanisms modelled and the nonlinear resistance
linear static analytical procedure to be valid4.15 if the function derived. A pushover analysis does depend
failure mode is ductile. If the failure mode is not on correctly identifying the degree of freedom in
known to be ductile, more stringent overstress which failure will occur: in Figure 4.4, it is assumed
factors should be used. that the response of the structure is characterised
by the capacity in the vertical degree of freedom at
The key disadvantage of linear static analysis is that the location of the removed column.
the dynamic load factor, which is based on the
Dynamic energy balance is then undertaken by
comparing the internal strain energy, which is the
integral of this resistance function, with the dynamic
work done when the load carried in the column prior
to its removal is applied suddenly to the rest of the
structure (Figure 4.5). The force applied is of the
same magnitude as the static force; the only
difference is that it is applied suddenly. The
displacement increases until equilibrium is found
between the integrals of both curves, i.e. until the
work done equals the internal strain energy. If the
ultimate load capacity of the structure is reached
before this occurs, the system is not in equilibrium
and collapse results.
40 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Approaches to design against disproportionate collapse 4.9
Force,
Force,
Resistance
Resistance
R
F, R F
Figure 4.5 Work done vs. internal energy in a linear elastic system when the same magnitude of load is a) applied statically, and
b) suddenly applied dynamic load
Force,
Resistance
Fstat
u Displacement
over-estimated. The ductility supply is highly There are two major advantages of nonlinear
dependent upon detailing, and this must therefore pushover analysis over more complex dynamic time
be approached (and specified if undertaken by history analysis:
others) with care. The assumptions about allowable – The sensitivity of the structure to cliff edge effects
ductility should be based on dynamic parameters: can be observed directly from the plots of strain
under dynamic loading the basic yield strength of energy/dynamic work done.
the material is usually enhanced, but in both steel – The sensitivity of the structural response to small
and reinforced concrete an element can sometimes changes in the variables (material yield strength,
exhibit a more brittle response than when loaded input load and so on) can also be rapidly evaluated.
statically, or sometimes the failure mode can change In some cases these might be obvious from the
altogether4.25-4.29. results already computed, but because analyses
are computationally inexpensive, sensitivities can be
Nonlinear static pushover analysis based on dynamic thoroughly investigated with relative efficiency.
energy balance has a number of key advantages over
linear static analysis: The key limitations of this type of analysis are:
– It permits ductility to be accounted for in the – It implicitly assumes there is a single degree of
structural response. freedom which dominates the structural response,
– It permits load redistribution to be accurately and that this is correctly identified by the analyst.
modelled. – The results may be invalid where complex structural
– Through the energy-based nature of the analysis, effects during the shedding of load cause behaviour
assumptions about the dynamic load factor to change from one degree of freedom to another,
associated with the removal of an element are not or where contact will occur which significantly
required. changes the pattern of loading/mode of response.
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 41
4.10 Approaches to design against disproportionate collapse
Such structures will typically be unusual, complex materials for typical Class 3 buildings. Timber is
structures and include some cable-supported and advancing rapidly in design technique and usage,
other tensile structures, structures susceptible to and designers should consult with material bodies for
buckling, and structures sensitive to torsional advice on contemporary good practice.
integrity (Figure 4.7).
General
4.9.4 Nonlinear dynamic time history analysis – Provide ties in two approximately orthogonal
horizontal directions and continuously through the
Nonlinear dynamic analysis is the most analytically structure.
rigorous and complex form of analysis described – Other than perimeter columns which should be
here, and has the ability to model the robustness anchored with transverse perimeter ties, ensure
problem with the greatest accuracy. Analysis is that columns have a balanced (symmetric) tying
generally highly complex and requires expertise in arrangement.
computational structural dynamics, and extensive – Design the tying system so that no column is tied in
verification and validation of the findings is only one horizontal direction.
necessary. Unlike nonlinear static pushover analysis – Tie each column continuously from the lowest to
whereby sensitivities can be readily observed, in a the highest level of the structure.
nonlinear dynamic analysis the investigation of – Design each column for a tensile force at least
sensitivities requires re-analysis of multiple equal to the largest total ultimate vertical dead and
scenarios. Computational run times can be very imposed load applied to the column at a single
substantial. floor level (i.e. the sum of the reactions from all the
beams/slabs connected to a column at that floor
As with the nonlinear static approach, the results of level) to which the column may be subjected in
a nonlinear dynamic analysis will only be as good as tension if support is lost from below.
the underlying assumptions and the engineer – Ensure there is a load path to successfully transfer
should be wary of the ‘black box’ phenomenon. these forces through the structure to the
The solution is computed explicitly with few foundation.
simplifying assumptions being necessary: dynamic – Make peripheral ties to perimeter columns
load factors, strain-rate dependency in the material continuous with the internal ties such that peripheral
models, damping and inertial effects can all be ties are capable of reacting the internal tie forces.
modelled explicitly. While this is attractive in terms of – Ensure peripheral ties are continuous around the
reducing the assumptions and therefore perimeter of the building.
uncertainties upon which the analysis is based, the – Wherever practicable, distribute ties evenly
results of a single dynamic time history analysis for throughout each floor and roof level.
a given case should rarely be relied upon in isolation – If beams are to be used as ties, then the floor
and sensitivity studies are essential before the slabs should be effectively anchored in the
results of a nonlinear dynamic analysis should be direction of their span; either to each other
accepted. Because of the fact that the structural continuous over their supports, or directly to
behaviour is computed explicitly, sensitivity studies their supports.
are not necessarily straightforward to undertake, nor – Design ties to be ductile, considering the strains
quick to run. associated with the rotations that connections will
have to undergo to develop resistance against
progressive collapse.
– Provide direct and robust connection between
4.10 Tying horizontal ties and vertical elements.
– At re-entrant corners or at substantial changes in
construction, ensure that ties are adequately
Effective tying is of vital importance to designing anchored or otherwise made effective.
against disproportionate collapse. Listed below are – Where the building is divided by expansion joints
some recommended principles for the design and into structurally independent sections, ensure each
detailing of ties based on steel and reinforced section is designed with an independent tying
concrete design practices as the established system.
42 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Approaches to design against disproportionate collapse 4.11
forces.
Characteristic Mixed-bound
4.11 Local protection of critical structural strength strengths
elements
Figure 4.8 Illustration of how consideration of upper-bound and lower-bound assumptions
in the derivation of flexural and shear capacities can result in a change from a ductile to a
As discussed in Section 4.6, load redistribution (LR) brittle response, while at the same time increasing the calculated capacity of the element
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 43
4.12 Approaches to design against disproportionate collapse
44 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Approaches to design against disproportionate collapse 4.13
Box 4.2 Failures in quality control of design and construction, West Gate Bridge, Melbourne, October 1970
On 15 October 1970, a 367ft span of western approach of the West Gate Bridge over the River Yarra suddenly collapsed during
construction. The bridge was of a similar trapezoidal box girder design to the Milford Haven Bridge which collapsed a year earlier,
in response to which changes were made to strengthen the design. Each span of the bridge was constructed as two full-length
half-width girders which were bolted together along the longitudinal splice so created. The transverse beams which formed
simply supported beams in their final condition were cantilevered at this time, and the half-girders susceptible to buckling.
The specific action initiating the collapse was the removal of a number of bolts from a longitudinal splice in the upper flange plating
close to mid-span. The bolts were removed in an attempt to straighten out a buckle which had occurred in one of the panels
constituting the upper flange. The buckle had, in turn, been caused by the application of 80T kentledge as a point load at midspan,
in an attempt to overcome difficulties caused by errors in camber. A camber difference of approximately 4.5 inches had existed
between the two half-girders of the span.
The previous span on the eastern approach had suffered a similar but less severe camber difference in the height of the webs,
together with severe buckling of the projecting plate edge forming the upper flange. The camber difference was removed by
jacking up the lower span using four 200T hydraulic jacks and local hydraulic jacks to remove the buckle in the projecting plate
edge.
For the west span the contractor proposed to use kentledge on account of the fact that time might be saved, though no
supporting calculation was undertaken either by the contractor or the designer. During the operation the use of kentledge was
noted to have caused a buckle to develop and the ‘obvious’ overstress due to kentledge noted by the site engineers. The
unfastening of the bolts immediately prior to the collapse in an attempt to correct this buckle had caused it to spread to adjacent
panels.
The Royal Commission into the collapse4.35 attributed the failure of the collapse to two main causes. Primarily, the designers had
‘‘failed to give a proper and careful regard to the process of structural design and to the check of the safety of the erection
proposals put forward by the contractors’’. As a consequence, the factors of safety for the bridge were inadequate during erection
and would also have been inadequate in-service had the bridge been completed. The secondary cause leading to the disaster was
the unusual method proposed by the contractor to use kentledge to correct the camber difference in the erection of the span. The
erection, if to be successful, required more than usual care on the part of the contractor and a consequential responsibility on the
consultants to ensure that such care was indeed exercised, neither of which were met.
A total of 35 people died in the collapse. The Royal Commission found that ‘‘the disaster. . .was utterly unnecessary. The reasons for
the collapse are to be found in the acts and omissions of those entrusted with building a bridge of a new and highly sophisticated
design. [. ..] Among those engaged upon the design and construction of the steel spans there were mistakes, miscalculations,
errors of judgement, failure of communication and sheer inefficiency. [.. .] Error begat error. . .and the events which led to the
disaster moved with the inevitability of a Greek Tragedy’’.
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 45
4.13 Approaches to design against disproportionate collapse
Box 4.3 Gross design error and lack of single point of responsibility, Hartford Civic Center, Connecticut,
January 1978
The Hartford Civic Center Coliseum collapsed in the early hours of the morning of 18 January 19784.36, 4.37 under the weight of
snow and ice following a 10 day storm which, while significant, was not out of the ordinary. The construction of the building had
been completed in 1973 and had been used to house ice hockey games for five years, including one game the night immediately
before the collapse. The roof was designed as a pyramidal space frame, innovative at the time and noted for being one of the first
large-span roofs made possible by computer design and analysis.
The roof was designed to span 82 64m over the arena, which seated 10000 spectators. Horizontal chords were spaced 6.4m
apart, diagonally braced with an intermediate layer of horizontal members providing stability to the diagonal bracing members.
Back-to-back angle sections were used to form cruciform built-up sections for the main chord members, with offsets between the
node and the intersection of the bracing members due to the open nature of the chord sections. Computational analysis was used
to verify the design against elastic buckling.
During construction, the inspection agency notified the engineers that it had found excessive deflection in some of the nodes but no
action was taken. Once lifted into its final position, the measured deflection was found to be twice that predicted by the analysis,
but the engineers responded that such discrepancies between the actual and the theoretical should be expected4.38. After
completion of the roof, a member of the public expressed concern about the large downward deflection in the arena roof, but the
engineer and the contractor assured the authorities that the building was safe.
The investigation into the collapse found that three major design errors in the original design contributed to the collapse. The joints
of truss members were modelled as having zero eccentricity, an incorrect assumption which put additional stresses due to bending
moments into the members. Compression members in the top chord were under-designed and braced against buckling only in one
plane. The self-weight of the roof was under-estimated by 20% as 0.9kPa in the analysis, compared with an actual self-weight of
1.1kPa. In a separate investigation, torsional buckling capacity of members near the roof buttresses was also cited as a possible
reason for the collapse. On the night of the collapse the total load acting on the roof was 3.2-3.5kPa (of which the snow load was
0.7-0.9kPa), well below what the design capacity of the roof should have been without the errors, which was at least 6.7kPa.
The Hartford Arena contract was divided into multiple subcontracts, leading to a lack of clarity over who was responsible for the
project as a whole. Although the appointment of a qualified structural engineer to oversee the construction was recommended, the
construction manager opted to take on all the inspection responsibilities himself: as noted by the investigation, ‘‘inspection of his
own work by the construction manager is an awkward arrangement’’4.39. As a result, the system of checks and balances was
eliminated at a stroke and the design engineers were not on site during construction. The construction manager asserted his
responsibility to be only for the correct execution of the design and not for the performance of the project as a whole. Consequently
the design engineer did not have access to full information during the construction, and the significance as a precursor of collapse
of obvious bowing of structural members in the top layer of the space frame which occurred during construction was not
recognised.
The issues in management of the construction in the Hartford Arena collapse illustrates the importance of the responsibility for the
integrity of the design resting with one person. Further, the city of Hartford did not require the peer review of the arena design
usually required for projects of this magnitude, thus making it more likely that errors went undetected. Peer reviews for high-
capacity buildings and structures designed using new techniques are cited as essential by Feld4.40 and SCOSS4.41. The
consideration of the impact of a possible failure in determining suitable partial safety factors and the use of increased partial safety
factors in buildings with high occupancy is recommended by Kaminetzky4.42.
46 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Approaches to design against disproportionate collapse 4.13
Box 4.5 Bad Reichenhall Ice Rink roof collapse, 2 January 2006
On 2 January 2006, the roof of Bad Reichenhall Ice Rink collapsed due to a 30-40cm snow fall on New Year’s Day, killing 15 and
injuring 304.48. The snow load was well within the calculated design snow load for the structure. The structure was a 75 48m
timber roof construction with 2.87m high glulam box girder beams at 7.5m on centre. The investigation into the collapse found that
design errors and a limited understanding of durability issues were the main causes of the failure. In the structural design, the
weakening of the structure as a result of finger-joints in the timber was disregarded. The design also used urea-formaldehyde glue,
which is moisture-sensitive and is therefore unsuitable for high humidity conditions such as in an unheated, non-air conditioned ice
rink where the relative humidity is high and the thermal air movements lead to condensation on the underside of the roof structure.
Furthermore, there were repeated cases of water penetration as a result of leaks in both the roof membrane and the roof drainage
system, which exacerbated the water damage to the adhesive. The failure was triggered by the snow load in a box girder on the
east side of the building. Cross-beams transferred the load to the adjacent girders, and a horizontal progressive collapse
propagated through the roof structure.
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 47
4.14 Approaches to design against disproportionate collapse
4.1 Institution of Structural Engineers. Safety in tall 4.16 Faber MH, Narasimhan H. Robustness of structures:
buildings and other buildings of large occupancy. final report of COST Action TU0601. September
London: IStructE, October 2010. 2011. Available from: Available from: www.
cost-tu0601.ethz.ch.
4.2 Beeby AW. Safety of structures, and a new approach
to robustness. The Structural Engineer 4.17 Ellingwood BR, Dusenberry DO. Building design for
1999;77(4):16-21. abnormal loads and progressive collapse.
Computer-aided civil and infrastructure engineering
4.3 Starossek U. Avoiding disproportionate collapse of 2005;20(3).
major bridges. Structural Engineering International
2009;3/2009:289-297. 4.18 Ellingwood BR, Smilowitz R, Dusenberry DO et al. Best
practices for reducing the potential for progressive
4.4 Bussell MN, Jones AEK. Robustness and the relevance collapse in buildings. NISTIR 7396. National Institute of
of Ronan Point today. The Structural Engineer Science and Technology, U.S. Department of
2010;88(23/24):20-25. Commerce, February 2007.
4.5 Bussell MN, Jones AEK. Robustness and the relevance 4.19 Dusenberry DO, Hamburger RO. Practical means for
of Ronan Point today – discussion. The Structural energy-based analyses of disproportionate collapse
Engineer 2011;89(4):Online extra 1-2 et seq. potential. Journal of Performance of Constructed
Facilities 2006;20(4):336-348.
4.6 Standing Committee on Structural Safety (SCOSS).
15th biennial report. London: June 2005. 4.20 Izzuddin BA, Vlassis AG, Elghazouli AY et al.
Progressive collapse of multi-storey buildings due to
4.7 Carpenter J. The systematic risk assessment of UK sudden column loss – Part I: simplified assessment
Building Regulations class 3 structures. Proceedings of framework. Engineering Structures
the Institution of Civil Engineers – Structures and 2008;30(5):1308-1318.
Buildings 2012. DOI: http://dx.doi.org/10.1680/stbu.
12.00004. 4.21 Vlassis AG, Izzuddin BA, Elghazouli AY et al.
Progressive collapse of multi-storey buildings due to
4.8 Institution of Structural Engineers. Practical guide to sudden column loss – Part II: application. Engineering
structural robustness and disproportionate collapse in Structures 2008;30(5):1424-1438.
buildings. London: IStructE, October 2010.
4.22 Cormie D, Mays G, Smith P. Blast effects on buildings.
4.9 Institution of Civil Engineers, Institution of Structural 2nd edition ed. London: Thomas Telford, 2009.
Engineers. Meeting report: avoiding disproportionate
collapse – the use of ‘systematic risk assessment’. 4.23 Kunishima M. Collapse accident under New Haengju
Joint meeting between the Institution of Civil Engineers Bridge construction work [Seoul City, Korea. July 31st,
and the Institution of Structural Engineers held at 1992]. Failure Knowledge Database/100 Selected
11 Upper Belgrave Street on 13 February 2012. Cases. Available from: www.sozogaku.com/fkd.
London: ICE, February 2012. Available from: http://
www.ice.org.uk/topics/structuresandbuildings/ 4.24 Ruth P, Marchand KA, Williamson EB. Static
Avoiding-Disproportionate-Collapse [Accessed equivalency in progressive collapse alternate path
2012-12-28]. analysis: reducing conservatism while retaining
structural integrity. Journal of Performance of
4.10 Standing Committee on Structural Safety (SCOSS). Constructed Facilities 2006;20(4):349-363.
Guidance note: independent review through peer
assist. SCOSS topic paper SC/09/034. January 2009. 4.25 Ansell A. The dynamic shear capacity of concrete
structures. Structural Concrete 2007;8(3):147-154.
4.11 British Standards Institution. BS EN
1990:2002 þ A1:2005 incorporating corrigenda 4.26 Magnusson J, Hallgren M, Ansell A. Air-blast-loaded,
December 2008 and April 2010. Eurocode: Basis of high-strength concrete beams. Part I: Experimental
structural design. London: BSI, 27 July 2002 and 2010. investigation. Magazine of Concrete Research
2010;62(2):127-136.
4.12 British Standards Institution. NA to BS EN
1990:2002 þ A1:2005 incorporating National 4.27 Magnusson J, Hallgren M, Ansell A. Air-blast-loaded,
Amendment No. 1. UK National Annex for Eurocode: high-strength concrete beams. Part II: Numerical
Basis of structural design. London: BSI, 15 December investigation. Magazine of Concrete Research
2004 and 2009. 2010;62(4):235-242.
4.13 federation international du béton (fib). fib Bulletin 63: 4.28 Tyas A, Warren JA, Stoddart EP et al. A methodology
Design of precast concrete structures against for combined rotation-extension testing of simple steel
accidental actions. Guide to good practice. Lausanne: beam to column joints at high rates of loading.
2012. Experimental mechanics 2012;52:1097-1109.
DOI: 10.1007/s11340-011-9562-4.
4.14 Cormie D. Robustness. In: ICE manual of structural
design: buildings, Bull J, Editor. London: Thomas 4.29 Stoddart EP, Byfield MP, Davison JB et al. Strain rate
Telford, 2012. dependent component based connection modelling for
use in non-linear dynamic progressive collapse
4.15 United States Department of Defense. Unified Facilities analysis. Engineering Structures 2012. DOI: 10.1016/
Criteria UFC 4-023-03: Design of Buildings to Resist j.engstruct.2012.05.042.
48 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Approaches to design against disproportionate collapse 4.14
4.30 The Steel Construction Institute. Joints in steel span timber structures. Proceedings of the 10th world
construction. Simple joints to Eurocode 3 (P358) (the conference on timber engineering. Miyazaki, Japan:
‘‘Green Book’’). Ascot: The Steel Construction Institute/ June 2-5, 2008.
The British Constructional Steelwork Association, 2011.
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 49
5 Example systematic risk assessment
50 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Example systematic risk assessment 5.1
Likelihoods are ascribed to the hazards by reference structural collapse. The likelihood of such a
to the scale of likelihood in Table 3.1, based on how failure occurring somewhere in the design is
often each hazard might be expected to occur during again judged as ‘rare’ (2% probability of
the design life of the building. The consequences of occurrence).
each hazard are evaluated through analysis, the results
of which (with reference to Table 3.2) are as follows: Note that hazards 7 and 9 have been eliminated
1 Errant truck/bus impact at ground level: from the design and are not shown.
maximum impact velocities are calculated and
impact analysis shows that column will be The risks so evaluated are shown in Figure 5.1.
damaged and need repair but will continue to
carry self-weight. Step 5: Identify risk reduction measures
Scenario-independent design is undertaken using the
2 Fire during occupation: structural fire analysis nonlinear static approach based on dynamic energy
shows that complete collapse of the fire floor will balance outlined in Section 4.9.3. This finds that it is
occur, with consequent impairment of egress feasible to sustain the loss of any typical column on a
routes from floors above. 9m grid, but that the perimeter columns at ground
3 Fire during refurbishment/refit: structural fire level on an 18m grid cannot be lost. The loss of a
analysis shows that complete collapse of the fire column in the loading bay can also not be sustained.
floor will occur, with consequent impairment of Dealing with the hazards in order of their ‘raw’ risk:
egress routes from floors above.
2 Fire during occupation or refurbishment/refit: The
4 Vehicle-borne IED at ground level: advice is
3 nonlinear static analysis shows that substitution
sought from CPNI on suitable charge weights for of fin plate connections for full-depth end plates
design. Blast analysis is undertaken which shows on the secondary beams, together with a slight
that the column closest to the blast will be lost, increase in the section weight, avoids the need
but that loss of multiple columns will not occur. for fire-proofing of the sections. This provides
Loss of floor slabs will occur local to the device robustness by design and also eliminates any
over two storeys. additional through-life hazards associated with
5 Vehicle-borne IED in loading bay: advice is fire following damage to the fire protection during
sought from CPNI on suitable charge weights for the life of the building or a fire while the fire
design. Blast analysis is undertaken which shows protection is being replaced during maintenance
that the complete structural collapse will not or refurbishment (see Box 5.1).
occur, but that the escape routes will be Commentary: although shown on the general risk
impaired due to the internal propagation of blast matrix as being in the intolerable range, through
pressures. discussion with the Building Control Officer a
variation is agreed in the level of tolerable risk
6 Person-borne IED: internal structural columns are
and the appropriate performance criterion for this
analysed at the ground floor entrance and in the
hazard is agreed as being that collapse of the
exhibition space. Under the charge weight
complete fire floor will not occur. This is agreed
advised by CPNI, it is shown that structural
on the basis that the assumptions in evaluating a
columns in the exhibition space can resist the
fire during occupation (hazard 2 ) are extreme,
blast load without loss of their loadbearing
requiring the failure of the fire detection and
capacity under gravity load, but at ground floor
suppression systems.
level collapse occurs.
For fire during refurbishment/refit (hazard 3 ),
8 Failure of temporary works to provide necessary a supplementary risk assessment will be
interim stability during construction: top-down prepared at the time of the work, when any
construction will be used for the basement additional necessary measures can be identified.
and temporary works will be required to The requirement is therefore set as being that the
provide support to the superstructure frame passive fire resistance of the building is sufficient
during the excavation of the basement. The to contain the collapse to the fire floor. This
ground floor slab acts as a transfer slab and performance is achieved, and the magnitude of
its failure will cause a significant structural this risk is therefore accepted as tolerable for this
collapse. hazard.
10 Numerical or conceptual mistakes in the analysis The risks associated with a fire during the future
and design: the potential is recognised for retrofit of the building, when the fire detection and/
mistakes to be made by the design team. or sprinkler systems may be inactive, are identified
Mistakes may be a modelling error due to misuse using structural fire analysis. The risks cannot be
of the computational analysis software or may be evaluated in detail but a retrofit plan is prepared
in the design and detailing of the structure. which highlights the particular factors that will
Mistakes could occur through inadequate need further, more detailed consideration. This
experience or if there is inadequate supervision precludes the removal of firestopping and puts
during the design. If in the analysis or design of a limitations on the storage of combustible materials
critical structural element such as a transfer beam, to minimise the fire load when the fire detection
they may result in a severe structural collapse. It is and/or sprinkler systems are deactivated, and
not feasible to give a statistical probability of the includes the results of a structural fire analysis
occurrence of such an error, but considering the documenting the structural behaviour in this
design as a whole and the quality management scenario.
systems in place, the likelihood of an error 8 Failure of temporary works to provide necessary
sufficient to cause a severe structural collapse is interim stability during construction: the desired
judged as ‘rare’ (2% probability of occurrence). construction sequence requires that the
11 Failure of a critical structural element: the failure superstructure is constructed simultaneously with
of a critical structural element such as a the excavation of the basement, and therefore
transfer element could lead to a severe significant propping is required to support
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 51
5.1 Example systematic risk assessment
Consequence
Catastrophic
Substantial
Significant
Minimal
Serious
Severe
Minor
Likelihood
Frequent/common
Likely 1
2 3
Unlikely 8
6
Rare 10 11
Improbable 4 5
Negligible
Key
Raw risk Mitigated risk
52 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Example systematic risk assessment 5.1
basement slabs prior to the gain of full strength. PD 6705-2:20105.4) in the construction
The temporary works designer is not appointed, specification of the execution standards for
and therefore the risk cannot be immediately structural steel, and of execution class EXC3
addressed in the temporary works design; (BS EN 13670:20095.5) for in situ reinforced
however the PCSA contractor helps to develop concrete. For other parts of the structure,
outline method statements and a propping EXC3 and EXC2 are specified for structural
scheme, which together provide sufficient plant steel and reinforced concrete respectively.
access for excavation. The risks associated with Similar execution quality standards are
the impact of the temporary works design on the specified for other materials.
main structure are articulated and will be – Design of these elements using off-site
included in the temporary works designer’s brief. fabrication to permit higher quality control.
10 Numerical or conceptual mistakes in the analysis – Third-party inspection during off-site fabrication
and design: the firm’s normal (QA) procedures of these elements.
will apply and, as is normal design practice, hand – Supervision of the construction of these
calculations are used to check the results of elements on-site by the designer.
computer output. For this type of building the 5 Vehicle-borne IED in loading bay: a vehicle barrier
firm’s QA procedures include a Category 2 designed in accordance with PAS 685.6 and
check, which is undertaken by a suitably qualified PAS 695.7 to withstand a hostile impact can be
and experienced team that may be from the installed at the head of the ramp (option 5a). This
same organisation but is independent from the provides a technically functional solution but the
design team and using a different software structural engineer identifies the risk of human
package. error or the failure of the vehicle blocker which
would compromise the security. An alternative
In addition to these measures, it is also proposed scheme is proposed based on vehicle lifts
to: (option 5b), which has the advantage of reducing
– Provide a peer reviewer at commencement of the overall footprint of the basement. The vehicle
the design to examine the proposed structural lifts prevent inadvertent loss of security to the
design philosophy and approach to design basement due to human error or duress in the
against disproportionate collapse, and to follow operation of the vehicle blocker. In either
this as it evolves through the design process. scheme, a gatehouse will be incorporated at the
– For critical structural elements as identified in service vehicle entrance such that vehicles can
the nonlinear static analysis, increase the level be checked prior to being admitted to the
of checking by introducing ‘gateway’ checks basement.
where the designers must justify the design to Note that these measures reduce risk so far as is
the peer reviewer before proceeding. reasonably practicable, given the loading bay
11 Failure of a critical structural element: it is cannot be relocated from under the building. Of
proposed that enhanced levels of quality the two options, option 5b (vehicle lifts) gives the
management are implemented for critical greater reduction in risk.
structural elements for all participants in the 4 Vehicle-borne IED at ground floor: two schemes
supply chain, both off- and on-site. This will be are considered in the structural design (Figure 5.2).
implemented as follows: A transfer beam can be incorporated at first floor
– Specification of reliability class RC3 and level to transfer from the 9m to the 18m grid. The
inspection level IL3 (BS EN 1990:2002 transfer beam cannot sustain the loss of the
Annex B5.2 ) in the construction specification for intermediate column (i.e. a 36m span) and this
these elements for extended and third party column therefore is a critical structural element
inspection. For other parts of the structure, and requires local protection. This may be
RC2 and inspection level IL2 is specified. provided by strengthening the structural section,
– Specification of execution class EXC4 or by providing vehicle bollards at the kerbline to
(BS EN 1090-2:2008 Annex B5.3 and increase stand-off from the column (option 4a).
(a) Transfer beam at first floor level (b) Vee-columns from ground to third floor level
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 53
5.1 Example systematic risk assessment
Alternatively (option 4b), vee-columns are internal area. However, loss of a node at ground
proposed from ground to third floor level which level produces an unsupported span of only 18m.
involves the first and second floor levels also being Local protection as a critical structural element is
cut back and the consequential loss of this net therefore not required. While analysis shows that
the transfer beam is unlikely to fail, its critical
nature is noted and it being moved up by one
storey reduces the risk of failure of this element.
(a) 300mm (b) 1 Errant truck/bus at ground level: structural
measures proposed in response to the risk from
an explosion at the ground floor (refer to option
4b) provide the same reduction in consequences
resulting from loss of a column under a vehicle
impact. This therefore reduces the risk by
providing scenario-independence; however, an
analysis undertaken of the impact of the vehicle
onto the structural section shows that the
column will suffer moderate damage but will
continue to be able to carry gravity load.
6 Person-borne IED: a sensitivity study is
undertaken which establishes that 300mm
stand-off achieved through the architectural
cladding dimensions will be sufficient to prevent
loss of the column (Figure 5.3a). Alternatively,
Figure 5.3 Options for protection of columns against a person-borne IED tubular rather than open sections (Figure 5.3b)
Consequence
Catastrophic
Substantial
Significant
Minimal
Serious
Severe
Minor
Likelihood
Frequent/common
Likely 1 1
2,3
2 3 2 3
Unlikely 8
8
6 6
6
Rare 10 11
10 11
Improbable 4 4 4 5
4b 4a
5 5a
Negligible 8
10 11 5 5b
Key
Raw risk Mitigated risk
54 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Example systematic risk assessment 5.1
are less sensitive to the blast effects and will – independent internal (Category 2) check and
avoid collapse, but have a 20% cost premium on external peer review for critical structural elements
the cost of connections due to the additional – enhanced quality management of supply chain and
complexity of forming the beam/column construction for critical structural elements
connections. – vehicle lift.
The risk matrix showing both the raw and the Based on the cost-benefit assessment, the
mitigated risks is shown in Figure 5.4. implementation of these measures is judged to
reduce the risks in the design so far as is reasonably
practicable.
Step 6: Perform a cost-benefit assessment
Weigh the reduction in risk in relation to the cost Step 8: Review the residual risk
of the measure. A given mitigation measure will The residual risks are shown in the risk matrix in
sometimes reduce the risks associated with Figure 5.5. These are discussed with the client,
several hazards, and the total risk reduction together with the inspection and maintenance plans
across all impacted hazards must be used in the for the building which includes the fire retrofit plan
cost-benefit assessment of each measure. The developed in Step 5.
cost-benefit assessment is shown in Table 5.1.
‘Benefit’ is derived directly from the risk matrix in Step 9: Check the risk assessment
Figure 5.4. ‘Cost’ is all other metrics against which Sensitivity studies are performed to investigate the
this is weighed, i.e. financial cost, programme, following:
amenity and so on. Risk reduction measures are – Cliff edge effects: modes of failure are generally
ranked from top to bottom in terms of the risk found to be ductile. The only cliff edge effect
benefit. found is the risk of two vee-columns failing – this
is found to occur significantly above the
Step 7: Identify the risk reduction measures to maximum loads associated with the hazards
be implemented and incorporate them in the considered.
design – Low likelihood/high consequence hazards:
Based on the evaluation of the risks and the tolerable changes in the assumptions made about
risk threshold, it is agreed with the client that the risk likelihood and consequence of the VBIED events
reduction measures to be implemented are as are not found to change the decisions made
follows: about the measures to be implemented to
– stand-off around columns at ground floor level reduce risks.
– full-depth end plates to secondary beams – Combined hazards: combined hazards have not
– initial development of temporary works design been considered directly in the risk assessment. Fire
– vee-columns/move transfer beam to third floor level may follow an explosion. The lack of reliance on
Change of internal columns from open to tubular Cost: 20% premium on connections package 6 6
sections
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 55
5.1 Example systematic risk assessment
Consequence
Catastrophic
Substantial
Significant
Minimal
Serious
Severe
Minor
Likelihood
Frequent/common
Likely 1 1
2,3
2 3 2 3
Unlikely 8
8
6 6
6
Rare 10 11
10 11
Improbable 4 4 5
4b
Negligible 8
10 11 5 5b
Key
Raw risk Mitigated risk
passive fire protection renders the building better Consequently the structural engineer recommends
able to withstand ensuing fire because there is no that an independent review of the risk assessment
risk associated with fire protection being dislodged. be undertaken without liability following the
approach given in the SCOSS topic paper
Step 10: Review the overall level of risk SC/09/0345.8, and that reviews of the structural
The overall levels of both the unmitigated and the fire engineering design and nonlinear static
mitigated risk are reviewed to examine sensitivity to pushover analysis are undertaken by suitably
the assumptions made in the risk assessment, or qualified independent reviewers.
sensitivity to the risk reduction measures.
It is further recommended that the design of the vee-
The risk evaluation is generally insensitive to the column connections are included in the scope of the
underlying assumptions, and there is no ‘bunching’ external peer reviewer, and that the fabrication and
of risks around the threshold whereby small changes erection of the vee-columns, connections and
would make a significant difference in the transfer beam are subjected to an enhanced level of
assessment of risk (sensitivity to the assumptions quality management. This will include tighter
made). Some of the measures implemented in the specification, enhanced factory inspection and
design provide a significant degree of risk reduction witness acceptance testing, and the supervision of
(sensitivity to the risk reduction measures), foremost the construction of these elements on site by the
among which are: designer.
– Vee-columns/move transfer beam to third floor
level. Step 11: Provide adequate information about any
– Full-depth end plates to secondary beams. risks that remain
The robustness design is therefore sensitive to these The findings of the risk assessment are
two risk mitigation measures in particular. documented in an update of the structural design
56 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Example systematic risk assessment 5.2
5.2 References
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 57
6 Application of systematic risk assessment to existing
buildings
58 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Application of systematic risk assessment to existing buildings 6.4
The structural engineer dismisses option 2 in that, although the cost is limited in terms of the design impact, there is also limited
benefit. The cost-benefit assessment of this option is therefore not proven. Options 1 and 3 have significant benefit in terms of
improving the robustness of the structure, but each have design implications in terms of either an unacceptable impact on listed
status or the need for an increase of the project scope and cost. Option 4 will provide protection against this hazard but will
increase time on site and the consequent cost of construction while planning consent is sought. The risk in option 5 is shown to be
disproportionate.
The structural engineer tables options 1, 3 and 4 with the client, recommending negotiation with the listing authority on the basis of
precedent for externally-fixed ties in historic buildings. This recommendation is made on the basis that options 1 and 3 provide
scenario-independence in the design, noting that there could be other hazards such as explosion which causes the potential loss of
a pier. Implementation of a perimeter beam at eaves level as described in option 3 is recommended if the negotiation with the
listing bodies proves unsuccessful, with appropriate adjustment to the budget for the project for the additional design and
construction works. The design and installation of bollards to prevent the impact of a vehicle into the pier is recommended as the
option of last resort if neither of these options prove viable.
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 59
6.5 Application of systematic risk assessment to existing buildings
6.5 References
60 The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Tables Boxes
iv The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse
Appendix A Further reading
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse 61
Glossary and abbreviations
Term Definition
Accidental action An accidental action is an unintended action to which the structure might be subjected during its
design life.
Class 1 building A building defined as Consequence Class 1, 2A, 2B and 3 respectively according to Annex A of BS EN
Class 2A building 1991-1-7:2006 Eurocode 1: Actions on structures. Part 1-7: General actions: Accidental actions.
Class 2B building Definitions are generally similar to those given in Approved Document A to the Building Regulations and
Class 3 building in corresponding documents in the devolved administrations of the United Kingdom.
Cliff edge A cliff edge is an informal term which refers to the sudden loss of strength or stiffness in a structural
system. It is usually used on a system level rather than an element level, and is an undesirable
characteristic of a system when the cliff edge lies too close to the expected structural response. A cliff
edge cannot always be eliminated but can be moved further away from the expected structural
response or made less steep.
Consequence Consequence is the harm that will be caused should the specified hazard materialise. In structural
terms, it is an expression of the extents of damage considering the vulnerability of the structure to a
hazard of a given severity. Sometimes referred to in risk terminology as severity or impact, consequence
may be measured in terms of the extents of collapse, numbers of fatalities, injuries, economic loss,
downtime or some other metric.
Disproportionate collapse A disproportionate collapse is one which is greater than should be expected, given the magnitude of the
initiating event. Disproportion is a measure of the extents of the damage given the magnitude of the
initiating event (hazard), against some limit defined for the scenario.
Note: the threshold of disproportion will vary for events of different magnitude – BS EN 1991-1-7
recognises that complete collapse may sometimes be acceptable if the event is sufficiently severe. The
threshold of disproportion is informed by the tolerability of risk associated with a particular accidental or
malicious action. In turn this is informed by regulatory requirement, by client or professional perception,
or by societal expectation.
Ductility, ductility ratio Ductility describes a characteristic of a structural element which exhibits an ability to undergo inelastic
displacements beyond the point of first yield, while exhibiting only a gradual loss of strength as the load
is increased. This is in contrast to a brittle failure mode, which is characterised by a sudden loss of
strength for a small increase in load beyond the point of first yield. A ductility ratio is the displacement
of the system, nondimensionalised with respect to the limiting elastic displacement, and is used as a
measure of the amount of damage sustained by an element.
Hazard A hazard is something which has the potential to cause harm. It may be an accidental or a malicious
action, insufficient strength or resistance, or excessive deviation from intended limits.
Note: the term ‘accidental’ is generally used to encompass both natural and man-made hazards,
although sometimes it is useful to make a distinction between natural hazards and (man-made)
accidental hazards, when the separate terms will be used.
Likelihood Likelihood is an expression of how frequently the hazard of specified magnitude can be expected to
materialise.
Malicious action A malicious action is a deliberate action to which the structure might be subjected during its design life.
Mitigation Mitigation is the reduction in the severity (consequences) of the hazard, given that the hazard
materialises. Often also used to refer to reduction in likelihood. ‘Risk reduction’ is generally the
preferred term for measures that act on either the likelihood and/or the consequences of a hazard.
Progressive collapse A progressive collapse is a particular type of collapse which develops in a progressive manner akin to
the collapse of a row of dominos. A collapse may be progressive horizontally e.g. successively from one
structural bay to those adjacent to it and propagating through the structure in a framed structure, or
successively from one cable to the next in a cable-stayed structure. A collapse may also be progressive
vertically e.g. the collapse of the columns supporting a floor slab due to the dynamic shock load caused
by the collapse onto it of the storey above it, or the successive collapse of the columns supporting a
number of floors due to the dynamic shock load as the block of mass is brought to rest as it impacts
with more rigid structure below.
Note: disproportionate collapse and progressive collapse are often, mistakenly, taken to be synonymous.
A collapse may be disproportionate but not necessarily progressive if, for example, the collapse is
limited in its extents to a single structural bay but the structural bays are large. Vice versa, a collapse
may be progressive in nature but not necessarily disproportionate in its extents, for example if arrested
after it progresses through a number of structural bays and the bays are small such that the overall
extents of damage is limited. Progressive collapse refers to the mechanism of the collapse;
disproportionate collapse refers to its extents.
Reduction of risk Reduction of risk refers to the general concept of making changes that result in a reduction in the level
of risk owing to a particular hazard. Methods used to reduce risks are typically organised according to a
hierarchy of risk reduction which describes the order of preference in which the methods should be
applied, based on their effect on the level of risk.
The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse v
Term Definition
Residual risk The residual level of risk that exists after risk reduction measures have been put in place. The term
‘residual risk’ is used interchangeably with mitigated risk, though strictly risk reduction may act on both
likelihood and consequence while mitigation acts on consequence alone. This term also introduces the
concept of unmitigated, or ‘raw’, risk.
Risk The risk owing to a particular hazard is the combination of the likelihood that the hazard will be realised
and the consequence should it do so.
Note: in practice, one hazard may give rise to many risks, each of differing likelihood and consequence.
Severity The severity of a hazard defines the magnitude of the action – for example a 1 in 100 year wind load,
a vehicle impact of given mass and velocity and so on.
Structural robustness Structural robustness is a quality in a structural system that describes its ability to accept a certain
amount of damage without that structure failing to any great degree. That is, modest damage (whether
due to accidental or malicious action) causes only a similarly modest change in the structural behaviour:
this implies an insensitivity of the structure to local failure.
Robustness is a quality of the structural system alone, and is independent of the cause of the damage
and/or the local failure. A robust structure specifically is not one which is over-designed: the ability to
resist damage is achieved through consideration of the global structural behaviour and failure modes so
that the effects of a localised structural failure can be mitigated by the ability of the structure to
redistribute the load elsewhere. The effects of the initial failure are gradual in onset and structural
instability is avoided in the response of the structure.
Eurocode 1 (BS EN 1991-1-7) describes robustness as ‘‘the ability of a structure to withstand events
like fire, explosions, impact or the consequences of human error without being damaged to an extent
disproportionate to the original cause’’, thereby linking it explicitly to the concept of disproportionate
collapse while recognising that total collapse is an acceptable outcome from a gross hazard.
Vulnerability The vulnerability of the structure is used in evaluating the consequences of a given hazard, considering
the type of action and the structural response to that action. The vulnerability of a structure will vary
between different hazards – a structure may be sensitive to vehicle impact but not to seismic loading.
Abbreviation Definition
ALARP As Low As Reasonably Practicable.
ERIC Eliminate, Reduce, Inform, Control (see Figure 2.2).
HAZOP HAZard and OPerability review.
QRA Quantitative Risk Assessment.
SFARP So Far As Reasonably Practicable.
SQEP Suitably Qualified and Experienced (Person)/Suitable Qualification and Experience.
vi The Institution of Structural Engineers Manual for the systematic risk assessment of high-risk structures against disproportionate collapse