Structures 20 (2019) 365–373

Contents lists available at ScienceDirect

Structures
journal homepage: www.elsevier.com/locate/structures

Historical review of prescriptive design rules for robustness after the T

collapse of Ronan Point
⁎
J.M. Russella, J. Sagasetab, , D. Cormiec, A.E.K. Jonesd
a
School of Engineering, University of Warwick, Coventry, UK1
b
Department of Civil and Environmental Engineering, University of Surrey, UK
c
Ove Arup and Partners, UK
d
MPA The Concrete Centre, UK

A R T I C LE I N FO A B S T R A C T

Keywords: After the collapse of Ronan Point tower building in 1968 there was an unprecedented discussion about the issue
Progressive collapse of progressive collapse in structural design. In particular, recommendations were published that precast panel
Design codes structures should include tying elements to hold a structure together after an element loss. The initial in-
Robustness vestigation into the causes of the collapse and the majority of the subsequent discussion was focused on ensuring
Structural integrity
precast structures had the same monolithic behaviour as conventional forms. However, the prescriptive re-
commendations were then applied to all structural forms without amendment to account for the different me-
chanical behaviour. This paper presents the findings from a novel bibliographic study of historical documents
published soon after Ronan Point collapse which influenced the development of relevant design guidelines. The
technical information was analysed chronologically to determine the intended purpose of such requirements and
the assumptions they were based on. It then traces the development of progressive collapse design requirements
to the current Eurocodes to consider if they are being applied as intended. This critical review is timely since
robustness considerations in Eurocodes and other international codes are currently being reviewed and general
misconceptions regarding existing prescriptive rules have been identified among practitioners in the UK and
internationally.

1. Introduction meetings, for example in 2005 the Joint Committee on Structural Safety
and the International Association of Bridge and Structural Engineering,
It is well established that structures must be designed to ensure that concluded that current codes fail to ensure sufficient structural ro-
they are suitably robust and that they should not fail disproportionately bustness which is especially concerning for high-risk buildings. This
to a damaging event. Therefore modern design codes provide guidelines concern led to the development of a European COST project initiative in
to help designers with this issue. Many people are aware that the un- 2007–2011 (COST Action TU0601: Robustness of Structures). Around a
expected partial collapse of Ronan Point tower building in 1968 due to similar time in the US, the ASCE SEI Committee on Disproportionate
a relatively small gas explosion was the starting point for such con- Collapse Standards and Evidence was formed in order to develop con-
siderations [1]. However, it is important to understand how the code sensus-based design approaches from existing US guidelines. Research
requirements have changed and developed since then. In particular this on structural robustness has experienced a significant peak of interest as
matter should be considered carefully at this time as the new generation reflected by the number of publications in this topic since 2000, as
of Eurocodes 2020 is in development and robustness is in the agenda for reported by Adam et al. [2]. The Institution of Structural Engineers
EN 1990 and EN 1991-1-7. Additionally, other international codes are (IStructE) have also sought to address this issue with the publication of
being redrafted and consideration taken for their robustness require- design guides [3,4] which include an illustration that shows the loca-
ments. tions of the different types of ties (Fig. 1) that are generally re-
Furthermore, the issue of providing suitable protection against commended in design guidelines. This approach, consisting of a series
disproportionate collapse has been raised by a number of international of prescriptive details rules, is generally adopted by international codes

⁎
Corresponding author.
E-mail address: j.sagaseta@surrey.ac.uk (J. Sagaseta).
1
Research carried out during postdoctoral stay at University of Surrey.

https://doi.org/10.1016/j.istruc.2019.04.011
Received 7 June 2018; Received in revised form 11 April 2019; Accepted 13 April 2019
Available online 08 May 2019
2352-0124/ © 2019 Institution of Structural Engineers. Published by Elsevier Ltd. All rights reserved.
J.M. Russell, et al. Structures 20 (2019) 365–373

outlining the Institution's response and opinion to these matters. The

first report [12] opened by emphasising their recommendations were
focused on prefabricated concrete panels used in residential buildings
and added that for conventional construction (i.e. not precast concrete)
“…experience has shown that the structures are capable of safely sus-
taining abnormal condition of loading and remaining stable after the
removal of primary structural members.” Some of this experience was
based on field observations after World War Two that had demon-
strated that monolithic RC structures showed good robustness to local
damage [7,8,30]. Fig. 3 shows some concrete structures that experi-
enced local damage due to explosives but maintained integrity, it
should be noted that Baker [7] does talk about the role of structural ties
in his description of building failures.
In commenting on the application of the government's circular the
IStructE's report adds the comment that if the wall panels are not
Fig. 1. Example of the location of ties in a flat slab structure (IStructE 2010 connected to the floor slabs then “… adjacent floor panels may act as a
[3]). catenary over a twin span of a floor” [13], although this is limited to a
sag of half the storey height, and it was acknowledged this is dependent
for low risk structures. on the connections.
With this in mind it is vital that any misconceptions about the use An investigation into the Ronan Point collapse event was commis-
and purpose of the current prescriptive rules for providing tying ele- sioned and its findings presented in the ‘Report of the Inquiry into the
ments throughout a structure are addressed and understood correctly. Collapse of Flats at Ronan Point, Canning Town’ published in the same
Therefore, this paper describes considerations for robustness that were year [10]. This report was thorough in its consideration of the failures
produced in the UK after the collapse of Ronan Point and traces the involved ranging from flawed structural design to poor workmanship
various iterations during the development of design codes. This allows and out of date codes of practice. However, it is clear that its conclu-
the intended purpose of the rules to be understood and will help de- sions are intended to be very narrow in application. This arises from the
signers and future code writers appreciate the limitations of such ap- fact that the report limited itself to precast system buildings, and indeed
proaches. This analysis is timely in the current context of the devel- only those of the construction used for Ronan Point. The main re-
opment of different guidelines for robustness and it is particularly commendations from the report are that consideration should be given
useful to an international audience who might not be familiar with the to the use of town gas in high rise buildings, existing, tall ‘system-built
background of UK building regulations. Some of this background is blocks’ should be appraised and that codes of practice should be
distributed between different historical documents which are difficult brought up to date. This last point focused on wind loading on high
to access and trace. buildings and for large concrete panel construction.
There is however a simply worded recommendation presented that
“the Building Regulations should include provisions dealing with pro-
2. Ronan point collapse and the immediate discussion
gressive collapse.” The report commented that the existing steel or re-
inforced concrete frame tall buildings “…are not liable to progressive
Much in depth discussion has been made about the collapse of
collapse and accordingly nobody turned their minds to this specific
Ronan Point tower building [1,5], and this paper does not seek to
question” [10], and so, according to report, it was only due to the use of
readdress this analysis or provide additional information on the event
new construction techniques that resulted in this becoming an issue.
itself. Rather the focus is on the immediate discussion and reports that
The report also identified that large concrete panel system buildings
occurred (1968–1972) and their consequences for the structural en-
requires “continuity at the joints of a kind strong and tough enough to
gineering community. Fig. 2 provides a summary of some of the most
stand both the initial shock of local damage and the abnormal and, in
significant events, covering a time scale from 1905 with an early ex-
detail, unforeseeable loads they may subsequently have to bear” [10].
ample of a progressive collapse described in [6] with large social impact
In fact, the aim was to achieve a non-brittle monolithic structure. The
but negligible impact on codes and robustness considerations at that
authors state “Reinforced concrete buildings constructed of in-situ
time, until 2025 with the expected end of the work on the second
concrete have most of the properties required” [10].
generation of Eurocodes.
The report authors stress the fact that it does not consider it ap-
propriate that they should attempt to deal in detail with the measures
2.1. Initial reports needed to strengthen the joints. However, they comment that a “gen-
erous and general distribution” of mild steel between panels would
Shortly after the collapse the UK Ministry of Housing and Local have improved the design.
Government released Circular 62/68, ‘Flats constructed with pre-cast In summary, any comments from the Ronan Point inquiry are very
concrete panels. Appraisal and strengthening of existing high blocks: narrow in terms of application, dealing exclusively with large precast
Design of new blocks’ [11], which required an investigation into the concrete panels. The recommendations presented are mainly concerned
susceptibility to progressive collapse for all existing pre-cast load with the out of date Code of Practice for wind loading and the use of
bearing buildings over 6 storeys and a check for fire and wind re- town gas in tall structures, as well as the introduction of a code of
quirements. It described a method for resisting progressive collapse by practice for panel buildings. It is also understood that the authors be-
providing an alternative load path for the loadings above a removed lieved that progressive collapse is not a major concern for frame
wall, by utilising “arching, beam, or cantilever action by the provision structures or in-situ concrete. Finally, while broadly stating the re-
of suitable reinforcement” [11]. It alternatively specifies that elements quirement for a structure to be tied together the report makes no de-
could have a local resistance of 5 psi (34kN/m2), described as the tailed comment on how this should be achieved, or the design loads that
maximum likely pressure from an explosion in a block of flats, although would be appropriate. Additionally, common later terms such as can-
the scientific justification of this value was debated [27]. tilever, catenary, ties, robustness or sudden column loss are not used at all.
Over the next few months the Institution of Structural Engineers
(IStructE) released 5 reports (RP-68-01 to RP-68-05) [12,13,15,28,29]

366
J.M. Russell, et al. Structures 20 (2019) 365–373

Fig. 2. Time line of major events and documents.

2.2. Early discussions on code amendments internal and peripheral ties “are intended to impart a quasi-monolithic
character to the floor” [31] which could allow membranes to form.
A Short and J R Miles' 1969 [31] paper on the new draft require- Within their conclusion they state that due to the lack of reliable sci-
ments for large panel structures [14] describe the purpose of tying entific evidence the “requirements were based on the intuitive judge-
elements as to “ensure the integrity of the construction, to prevent ment of experienced structural engineers backed by a design study.”
structures from falling apart.” This is contrasted to monolithic struc- Finally, although they recognised that the addendum was only in-
tures where “these ties forces are generally present owing to the in- tended for large-panel structures they argued that the principles should
herent nature of the construction” [31]. be “considered for other types of construction that are sensitive to ex-
Short and Miles also expand further on the role of ties in different ceptional and unforeseen loadings” [31]. This went beyond the report
locations (see Fig. 1 for examples), stating the reinforcement along the into the collapse of Ronan point which was limited to precast con-
peripheral will strengthen the corner and “produces the necessary struction, and recommended only investigation and new regulation for
bridging reinforcement to take care of the danger of progressive col- large panel structures. However, the government decided the fifth
lapse caused by the elimination of an external wall” [31]. Whereas amendment to the building regulations should apply to all structural

367
J.M. Russell, et al. Structures 20 (2019) 365–373

(a) Local damage leading to the loss of one (b) Local damage leading to frame damage
column and slab in a flat slab building and partial loss in a frame building
Fig. 3. Arrest of progressive collapse and damage of monolithic RC buildings after internal explosions during WWII (Baker et al. [7], reproduced with permission of
ICE Publishing).

forms. This was a matter of some controversy and was among the issues long as the floors and columns are tied together. For box-frame struc-
debated on this subject in the UK House of Commons, as Nicholas tures they also state that for an effective tie “The magnitude is ob-
Ridley MP (himself a qualified civil engineer) stated, “Instead of ac- viously related to the span of the floor and beam components and
cepting the recommendations of the report, [the government] launched should be capable of sustaining practical catenary effects” [34].
out on their own with a series of building regulations which are far too Research into the issue of progressive collapse continued into the
technical and far too expensive and which will do great harm to the 1970's and 80's with particular interest in concrete structures and ca-
advancement of building techniques” [27]. tenary action [32,35–38]. However, there was no major changes to
However, the Minister of State for Housing at the time, Mr. Reginald design approaches as a result of these later studies, except for some
Freeson MP, stated that “The amendment proposals are primarily aimed rules introduced for detailing specific structural elements.
at protection against the effects of damage, however it is caused, or The original report into the Ronan Point collapse made clear that its
against progressive collapse experience, however it might be caused, in recommendations were intended only for precast, indeed it states that
the future” [27] and then explained why the new provisions were ap- RC frames are unlikely to undergo progressive collapse. Its proposed
plied broader than just precast structures. He viewed that the “new method of preventing progressive collapse was to attempt to give pre-
mandatory requirement should not be selective in its application and so cast structures the same monolithic behaviour that occurs naturally
not be thought to penalise a particular form of construction” [27]. He with in-situ construction. Although it is now known that RC frames and
also argued that some framed and brick buildings were still susceptible slab structures may still be susceptible to progressive collapse, (for
to progressive collapse, but the new requirements would not be un- example the Skyline Plaza in 1973 and Sampoong Department Store in
reasonable for structures that were not susceptible and that “the new 1995 [39]) this is due to their different mechanical behaviour (e.g.
requirements have been deliberately drawn up in fairly functional brittle connection failures, long spans, less optimal construction control
terms…to facilitate their application to varying forms of construction compared to prefabricated structures) rather than their loss of con-
and to permit of the flexibility of design” [27]. tinuity.
By the time the code of practice for the structural use of precast
concrete, CP116 [14], was updated to include the Ronan Point Inquiry 3. Review of code tying force requirements
Report's recommendations, some aspects took the lead from a pre-
viously published CEB recommendations for large panel precast struc- 3.1. CEB bulletin 60
tures [9]. This document suggested that the panels should be tied to-
gether to maintain continuity. However, the purpose of those ties was In 1967 CEB released a report entitled ‘International recommenda-
to resist horizontal loadings (e.g. wind, seismic, eccentricities) and tions for the design and construction of large-panel structures’ [9]. This
forces due to differential settlement and no considerations was given to document included some broad guidance for ensuring the continuity of
element loss. Similarly, as result of Ronan Point and the subsequent precast panels such as “Within the thickness of each floor, or close to
updating of regulations, Fintel and Schultz conducted investigations the floor, mechanically continuous steel “ties” should be provided in
during the 1970's into the integrity of large panel buildings in America both directions.” Both peripheral and internal ties (or chains to use the
[32]. Their recommendations of including ties between elements to original French term) were required, although these are specified be-
maintain integrity and bridge over potential damage were developed as tween panels and not flooring elements. The purposes of these are de-
a basis for the minimum tie force requirements still in use in the ACI scribed as:
Standard 318 [33].
After building regulations and some codes of practice were updated • resist the forces acting on the external panels (in their load-bearing
to include the Ronan Point Inquiry Report's recommendations, the role) owing to inaccuracies of installation;
Institution of Structural Engineers published the report “Stability of • resist the horizontal reactions directly exerted by the panels (wind
modern buildings” [34] which notes on the topic of providing alter- and earthquake);
native load paths that “…there is no fully detailed design procedure by • resist the tensile forces developed in the floors performing their
which to implement this philosophy. There is, however, plenty of wind-bracing function;
practical experience to demonstrate that a damaged structure will re- • resist where necessary the horizontal components of the diagonal
main stable in favourable circumstances after an incident that has se- forces in the wall panels subjected to tensile load by the effect of
verely damaged and possibly removed a section or component of the lateral forces, and transmit the horizontal floor reactions to the
framework” [34]. This report also viewed that the formation of a ca- windward edges of the wind-bracing cantilevers;
tenary or membrane could maintain stability after a damaging event, as • resist the tensile forces developed in the walls if differences of level

368
J.M. Russell, et al. Structures 20 (2019) 365–373

develop in the supports can be seen. Firstly that if a structure loses its primary load path, it is
assumed secondary mechanisms could be utilised and the structure can
From a review of these clauses it is clear that although most of the develop the required tie forces, without additional checks. The tensile
terminology used (e.g. peripheral and internal ties, continuity, canti- force within these ties can usually be determined based on the overall
levers etc.) is similar to later progressive collapse design guidelines, structural geometry, i.e. number of stories, storey height, span lengths.
these requirements were not intended to address the sorts of extreme Only internal or vertical ties require consideration of the direct loading
scenarios that occur after a sudden column loss. Although, it has been on that floor. Additionally, as the consequences of failure of a taller
noted that if Ronan Point had been designed and constructed to comply structures may be more severe, the secondary mechanisms must resist
with CEB bulletin 60 the consequences of the event would have been higher forces to reduce the chance of progressive collapse. Finally, by
significantly lower [40]. providing vertical ties, the probability of a column failure is reduced
and there is potential for load sharing between floors, although this
3.2. CP 110: Part 1: 1972 mechanism is not explicitly checked.

One of the earliest national codes to include consideration for pro- 3.3. BS 8110: Part 1: 1985 and 1997
gressive collapse was Code of Practice 110 in 1972 [16]. Here the re-
quirement is stated that, “The layout of the structure on plan, and the The British Standard, BS 8110 [17], for concrete was introduced in
interaction between the structural members, should be such as to en- 1985. Clause 3.12.3 of this document stated that “The necessary in-
sure a robust and stable design…there should be reasonable probability teraction between elements is obtained by tying the structure together
that it will not collapse catastrophically under the effect of misuse or using the following types of tie…” [17], and described the use of hor-
accident.” It is also clarified that the structure is not expected to resist izontal and vertical ties. It is noteworthy that reference to the me-
extreme loads of events, but not fail disproportionately. To meet this chanisms by which ties operate (e.g. cantilever, catenary) have been
requirement ties are recommended to maintain the structures robust- removed and replaced with a general statement regarding tying the
ness. Initially the purpose of ties in buildings was stated in the addi- structure, however the section continues in an almost identical wording
tional clause that “…the ties should be so placed as to provide the best and requirement to CP 110, indicating they are based on the same as-
assistance in resisting by cantilever, catenary or other actions the results sumption. The requirements for vertical ties is expanded on in this
of extreme damage by accidental causes.” Emphasis added to highlight standard and introduces the additional requirement that the ties must
the mechanisms by which the ties are utilised for, and therefore the able to carry in tension the vertical design loading on that column from
expected state of the structure for which they are relevant. one floor. The update to BS 8110 in 1997 [41] made no change to the
The required tensile capacity of these ties depends on their location requirements.
within the structure, however all are based on a function of Ft, in kN.
This value is a function of the number of stories, with an imposed 3.4. BS 5950: Part 1: 1985, 1990, 2000
minimum. From this it can be seen that it is assumed that a taller
structure requires higher tensile capacity to maintain integrity, al- For steel structures, the original BS 5950 in 1985 states in the in-
though this is capped at 10 floors. Therefore Ft has a range of between troduction that “The structure should behave as one three-dimensional
24 and 60 kN. entity. The layout of its constituent parts, such as foundations, steel-
work, connections and other structural components should constitute a
(20 + 4ns )
Ft = min ⎧ kN robust and stable structure under normal loading to ensure that in the
⎨
⎩ 60 (1)
event of misuse or accident, damage will not be disproportionate to the
Peripheral ties must resist Ft kN within 1.2 m of the edge of the cause” [18]. It also highlights that minor incidental loads should not
building. External columns and wall must also be capable of resisting a jeopardise the safety of other parts.
force which increases with the floor to ceiling hight, lo. This value varies The requirements of the tie include the following clauses: “All ties
between 0.8 and 2 times Ft for floor to ceiling heights of 2 to 5 m, and is and their end connections should be of a standard of robustness com-
capped at 2Ft. Additionally this value must be greater than 3% of the mensurate with the structure of which they form a part and should be
ultimate design vertical load. Ftie, col is given in Eq. (2). Note the symbol capable of carrying a factored tensile load of not less than 75kN at
Ftie, col is not used within CP 110. floors or 40kN at roof level.” This is clarified later with the requirement
that the ties should be, where practical, arranged in continuous lines
⎧ ⎧ 2Ft and at two directions. The required tensile strength of these ties, and
⎪ min lo ⎞
Ftie, col = max ⎨⎛ Ft kN their connections, is given in Eqs. (4) and (5), for internal and periphery
⎨ ⎩ ⎝ 2.5 ⎠ ties respectively.
⎪
⎩3%of the ultimate vertical load (2) Ti = 0.5wf st La (4)
Internal ties must resist at least Ft kN per meter width, with larger
Tp = 0.25wf st La (5)
values required for longer spans or larger characteristic loading.
Assuming the floor to ceiling height is less than 4 m and the span is where wf is the factored loading, per area, on the floor. st is the spacing
under 10 m, then for typical loading levels the tie force is under 2.67Ft between ties and La is the length of the tie under investigation.
kN/m.
3.5. Approved Document A
⎧ Ft
Ftie, int = max Ft (gk + qk ) l kN/m
⎨ Within the UK, the publication of Approved Document A [19]
⎩ 7.5 5 (3)
provided additional guidance for designing structures against dis-
gk and qk are the characteristic dead and imposed UDLs and l is the proportionate collapse. This document is equivalent in status to a code
span length, limited to 5 times the clear storey height. of practice, i.e. that compliance with the guidance does not absolve the
Finally vertical ties, acting across all levels should be provided. The designer of his legal responsibilities. The legal requirement is that
area of reinforcement specified should be at least the minimum re- contained in the Building Regulations 2000 (as amended), which state
quirement for normal conditions. that “the building shall be constructed so that in the event of an acci-
Based on these requirements a number of underlining assumptions dent the building will not suffer collapse to an extent disproportionate

369
J.M. Russell, et al. Structures 20 (2019) 365–373

to the cause.” The document is based on the principle of categorising structure shall be designed and executed in such a way that it will be
buildings into consequence classes based on their size and use and not be damaged by events such as: explosion, impact, and the con-
describes the provision of ties and the conducting of alternate load path sequences of human errors, to an extent disproportionate to the original
analysis. The latter approach is necessary if more than 15% or 70 m2 of cause.” On inspection, this is just a restatement of the requirement that
floor area would collapse after an element removal. The 70 m2 is existed in CP 110. Under the list of options for avoiding or limiting the
equivalent to two perimeter bays on a 6 m structural grid. However, potential damage within EC0's basic requirements, three are applicable
Eurocode 1991-1-7 [42] increases this limit to 100 m2 in recognition of for this study. These are “selecting a structural form and design that can
increasing structural spans (equivalent to two perimeter bays on a 7.5 m survive adequately the accidental removal of an individual member or a
grid) and later revisions of Approved Document A incorporate the same limited part of the structure, or the occurrence of acceptable localised
change. If such an alternate load path cannot be achieved, the element damage; avoiding as far as possible structural systems that can collapse
is defined as ‘key’ and must be able to withstand a notional accidental without warning; and tying the structure together” [21].
load of 34kN/m2.
The 1992 edition also made an attempt to allow designers to avoid 3.6.2. EN 1991-1-7
or reduce the hazards to which the building may be exposed (for ex- EN 1991-1-7 [42] provides in the information for tie strengths, and
ample removing potentially explosive sources or keeping vehicles from for the first time, explicitly separates frame and load-bearing wall
approaching a building), as an alternative to any of the structural construction, although its requirements cover all structural material
measures described. While a conceptually valid approach, it is unlikely forms.
this would be sufficient to justify the omission of any tying or robust- For framed structures, the ties must be able to carry the required
ness requirements from a design. However it may be a consideration tensile load, according to Eqs. (6) and (7) for internal and perimeter ties
when dealing with existing buildings, particularly those designed pre- respectively which given as a function of the loading and the span
Ronan Point. This approach, as an alternative, has been deleted from length. These equations and methodology are similar to the previous
subsequent editions of the approved document. British Standard for steel structures (BS 5950 [18]), see Section 3.4. The
In 2004 the document was revised to extend the previous require- separate requirement for external columns or walls to be tied in has
ments, mandating horizontal tying for buildings below five stories for been moved to the different material sections.
the first time (except single-occupancy residential dwellings). Above
Ti = 0.8(gk + ψqk ) sL or 75kN, whichever is the greater. (6)
five stories, the previously-existing general requirements continued to
apply (horizontal and vertical tying, alternate load path analysis and Tp = 0.4(gk + ψqk ) sL or 75kN, whichever is the greater. (7)
key element design). Basements became excluded from the storey
count, provided they satisfy Class 2B, potentially to reduce the re- For load bearing wall cases, a structure in Class 2b risk group, ty-
quirements for horizontal tying. A significant difference with the 2004 pically buildings between 4 and 15 storeys, requires horizontal ties in
edition was the subtle change to the layout and punctuation that re- the floors. The value of Ft is similar in practice for internal and per-
sulted in horizontal tying no longer being required with notional ele- ipheral ties as previous codes, although the load used in the equivalent
ment removal. However, in previous editions of Approved Document A, equation is that for the accidental situation rather than the ultimate
the intent is quite clear: horizontal ties are to be provided regardless of limit state.
whether vertical tying or alternate load path analysis is chosen.
This edition also includes the option of “effective anchorage of 3.6.3. EN 1992-1
suspended slabs to walls” given as an alternative to horizontal tying. For concrete structures EN 1992-1 [44] provides further instruction.
Although similar to the provision of horizontal ties, reference to ma- Firstly, the purpose of tying systems is stated clearly: “Structures that
sonry code reveals that it is intended to describe deemed-to-satisfy are not designed to withstand accidental actions shall have a suitable
details in masonry construction of joist hangers for timber floor joists or tying system, to prevent progressive collapse by providing alternative
built-in ends of precast concrete slabs which satisfy the Class 2A re- load paths after local damage” [44]. Then follows the list of four types
quirement without tying reinforcement. The Code gives no detail about of ties, that have been mentioned in previous codes above. i.e. per-
the resilience of the details proposed, but as they are only valid for Class ipheral, internal, horizontal column/wall ties and vertical ties. Section
2A buildings, it can be inferred that they provide less resilience than full 9.10.2.2 in EN 1992-1, provides recommendations for these tie forces,
horizontal tying. Additionally, note that as prior to 2004, horizontal however, the UK National Annex differs in calculating these values. For
tying did not apply in buildings less than 5 stories, this potentially al- peripheral ties, EC states Ftie, per from Eq. (8).
lowed a reduced requirement for the masonry and timber industries for Ftie, per = li ·q1 ≤ Q2 (8)
this building class.
Finally, in 2004 the Class 3 category, was introduced covering the where li is the length of the end-span and EC recommends q1 = 10kN/m
highest risk buildings, for which systematic risk assessment is required. and Q2 = 70kN. However, the UK decision in their National Annex is to
Intended for buildings exceeding 15 stories or 5000 m2 per storey, those maintain the previous equations from the CP110 and BS 8110, with
containing hazardous substances or processes, stadia and grandstands, Q2 = 60kN and defines q1 with Eq. (9).
or buildings into which the public are admitted in significant numbers. q1 = (20 + 4no ) li (9)
The IStructE (practical guide to robustness) also recommend that cer-
tain buildings merit treatment as Class 3 buildings because of their This can be simplified down to the system used for CP 110 and BS
value, vulnerability or the consequences of their failure [4,43]. 8110, i.e. Ftie, per = Ft.
For internal ties, EC simply recommends a value of 20 kN/m while
3.6. Eurocode the UK again sticks with existing equation. For most applications the UK
requirements of tying forces are higher than the EC suggested values.
With the introduction of the Eurocodes, a number of changes were For floors without screeds and where ties are grouped at beam lines
made regarding robustness considerations in design. However, the then EC suggests Eq. (10) while the UK uses the previous internal tie
majority of these refer to a reorganisation of design requirements rather equation.
than a substantial change in methodologies.
Ftie = q3·(l1 + l2)/2 ≤ q4 (10)
3.6.1. EN 1990 with q3 = 20kN/m and q4 = 70kN. l1 and l2 are the span lengths either
Eurocode EN 1990 [21] states as a basic requirement that “A side of the beam.

370
J.M. Russell, et al. Structures 20 (2019) 365–373

Edge columns and walls should be tied horizontally into the struc- an economic disadvantage. Related to this, the initial purpose of the ties
ture. EC recommends Ftie, fac = 20kN/m and Ftie, col = 150kN. Again, the was to help precast structures behave as monolithic insitu concrete
UK decision is to maintain the historical approach, with the same value ones, with the view that these forms were able to resist disproportionate
for columns and walls, although the units differ. collapse. It is now known that such structures may fail despite con-
tinuity, therefore, the same methodology may not be appropriate for all
3.7. US provisions structural forms.
The majority of the knowledge about collapse of structures in 1968
Although this paper is focused on the direct development of UK was based on observations during World War Two and specialised
progressive collapse guidelines as a result of the Ronan Point collapse, knowledge for the precast industry. There is very little experimental
other counties have naturally developed their own recommendations and theoretical research from the time period to back up the pre-
[2]. Many of these international codes show the evolution of some of scriptive rules proposed to prevent progressive collapse, and even less
the early concepts introduced soon after Ronan Point such as the con- to provide evidence for the tying forces and values prescribed.
cept of robustness itself or the fundamentals behind the notional ele- It appears the early codes intended tie requirements to be a broad,
ment removal approach. This evolution was particularly relevant in the prescriptive method of ensuring that a structure does not fail dis-
United States of America, especially after events such as the Alfred P. proportionately to a local damaging event. However, the actual method
Murrah Federal Building collapse in Oklahoma in 1995 and the World by which this is accomplished, is general at best and vague at worst
Trade Center in New York in 2001 leading to a number of well estab- with different mechanisms described at different times (e.g. catenary,
lished codes such as ACI 318 [33], ASCE 7-98 [20], ASCE 7-16 [26], IBC membrane, beam action, arching, cantilever). In addition, due to the
2009 [23], UFC 4-023-03 [24] and GSA [22–25]. prescriptive nature of the tying requirements it is not possible to
ASCE 7-16 [26] describes a requirement of general structural in- quantify their effectiveness or what minimum level of robustness they
tegrity with similar language to Eurocode 1990 stating that a structure provide. However, despite its limitations, these and similar rules are
must be able “to sustain local damage with the structural system as a used in many codes in addition to Eurocode (e.g. ASCE 7-16 [26], IBC
whole remaining stable and not being damaged to an extent dis- 2009 [23], UFC 4-023-03 [24] and also the Chinese code CECS
proportionate to the original local damage”. It lists continuity, re- 392:2014 [47] and Canadian NBCC 1995 [48]). However, such an ap-
dundancy and ductility as methods to achieve this. UFC 4-023-03 De- proach is not universal as other international codes, such as GSA 2013
sign of Buildings to Resist Progressive Collapse [24] describes a tie force [25], NYC BC 2014 [49] (New York) and NCC 2016 [50] (Australia),
method very similar to European approaches stating that “the building exclude this method. See Adam et al. for more information [2].
is mechanically tied together, enhancing continuity, ductility, and de- The traditional prescriptive based codes do not give any con-
velopment of alternate load paths” by use of vertical, longitudinal, sideration to the fact that a structure may not be able to utilise these ties
transverse, and peripheral ties. Significantly, the existing structural without deforming to an extent which would cause brittle failure. For
members can only be used to provide the tie force if they are cable of example, in concrete flat slab construction it is known that punching
undergoing a 0.20 rad (11°) rotation. The latest General Services Ad- around the column will occur before the slab can reach the catenary
ministration guidelines (GSA) [25] move away from tie force require- deformations needed to resist gravity loads by means of tensile mem-
ment to focus on alternative load path methods requiring full analysis of brane action alone [51,52]. However some codes (e.g UFC 4-023-03 in
the structure after the loss of a load bearing element. Stevens et al. [45] the US and also the Chinese code) are starting to introduce rotation
provides an overview of Department of Defense (DoD) requirements for limits for the ties and linking the tying force to physical concepts [53],
high risk structures, typically government or military. which is a step forward according to the authors. See Fig. 4 from Chi-
The majority of approaches mentioned above remain threat in- nese code CECS 392 that includes ductility and consideration of the
dependent, although in cases where a known threat exists more detailed beam end moments while utilising the tie.
design is provided for. More usual designs include tie force checks for
low risk locations and alternative load path analysis in more critical 5. Conclusions
scenarios. A general classification of design methods against dis-
proportionate collapse was put forward by the ASCE SEI Sub-committee After the progressive collapse of Ronan Point tower building in 1968
on Terminology and Procedures [46], dividing methods into non- there was an unprecedented discussion about how buildings were de-
structural and structural. The non-structural methods focuses on event signed and constructed and whether existing or future structures were
control whereas the structural focuses on the collapse resistance (ro- vulnerable to such events. The investigating report into the incident
bustness and vulnerability), e.g. alternative load path method, seg- concerned itself exclusively with large panel precast construction used
mentation, protection, increased local resistance (key element design). for Ronan Point and along with concerns with the use of gas in tall
A similar design strategy is proposed in Eurocodes [42] for accidental buildings it recommended that design codes be brought up to date,
design situations although in this case the same methods are classified especially for wind loading. Soon after the accident the British
according to whether the accidental action is identified or the method Government released amendments to the building regulations requiring
focuses on limiting the extent of localised failure. In both classifications, considering of progressive collapse, and applied these specifications to
it is understood that the tying approach would contribute towards all structural forms. From the bibliographic study carried out in this
limiting the extent of the localised failure and provide integrity. work it was concluded that:

4. Discussion of code development at an international context 1) The new requirements for ties are based on a methodology proposed
in CEB bulletin 60 [9], this document predated the collapse of
From considering all the development of the requirements, culmi- Ronan Point and was concerned with maintaining integrity across
nating in the current Eurocode guidelines, a number of points can be precast joints due to minor eccentricities and so they were not in-
made. Firstly, it is clear that although the collapse of Ronan Point, tended for use with other structural forms or for dealing with ex-
which triggered the introduction of robustness considerations in design treme events such as element removal. The literature shows dif-
guidelines, only occurred due to a combination of the nature of the ferent interpretations of the purpose of the tying approach,
structural form used and deficiencies in design, the solutions were including preventing local failure and maintaining integrity after
equally applied to all forms of construction. This was not due to tech- member loss.
nical considerations, but to make the new codes of practice more gen- 2) While the original investigators were no doubt aware of the lim-
eral in their application and to prevent the precast industry becoming at itations and the focus of their recommendations, the current form of

371
J.M. Russell, et al. Structures 20 (2019) 365–373

Fig. 4. Tie mechanism for a) internal, b) border and c) corner column failure locations from Chinese code CECS 392:2014 [47].

the guidelines, developed through the British Standards and now the References
Eurocodes, is not directly based on their scenarios. Although the
provisions were based on the best information and experience [1] Pearson C, Delatte N. Ronan point apartment tower collapse and its effect on
available at that time, further experimental testing and solid re- building codes. J Perform Constr Facil 2005;19(2):172–7.
[2] Adam Jose M, Parisi Fulvio, Sagaseta Juan, Lu Xinzheng. Research and practice on
search is needed to support these prescriptive rules. progressive collapse and robustness of building structures in the 21st century. Eng
3) There are significant mechanical differences in the behaviour of Struct 2018;173:122–49.
large precast panels and modern construction methods and there- [3] Institution of Structural Engineers. Practical guide to structural robustness and
disproportionate collapse in buildings. London, UK: IStructE; 2010.
fore the same approach may not be appropriate. This causes a [4] Cormie D. (Arup). Manual for the systematic risk assessment of high-risk structures
problem for drafting general rules for structures (such as for the against disproportionate collapse. IStructE Ltd. 2013.
Eurocodes) as the vulnerability and behaviour could be very dif- [5] Bussell MN, Jones AEK. Robustness and the relevance of Ronan Point today. The
Structural Engineer 2010;88(23):20–5.
ferent for different structural forms. [6] Díaz-Pavón Cuaresma Eduardo, León González Javier, Ley Urzáiz Jorge.
4) Although the original intention of the ties was to prevent brittle Robustness: the quality Ribera missed in 1905. Hormigón y Acero
failure, this is not explicitly checked and therefore it is not certain 2017;68(283):e23–34.
[7] Baker JF, Williams Edward Leader, Lax D. The design of framed buildings against
that the tie, or secondary mechanism can form without further
high-explosive bombs. The civil engineer in war: a symposium of papers on war-
failures occurring. This issue is being addressed in some codes and is time engineering problems. Thomas Telford Ltd; 1948. p. 3–80.
an important area of future development. [8] F Walley. The design of bomb-resisting structures against H.E. attack. 1954.
5) Future revisions for international design codes such the Eurocodes Available in the Walley Collection at the (UK) Institution of Civil Engineers.
[9] C V [tr] Amerongen. International recommendations for the design and construction
should be orientated towards improving confidence in prescriptive of large-panel structures. Technical report, C. & C. A. A translation of the report in
rules and their range of applicability. Such rules should be reviewed French ‘Recommandations internationales unifiees pour le calcul et l'execution des
based on the variability of the risk of progressive collapse, form of constructions en panneaux assembles de grand format’. Paris, Comite Europeen du
Beton, April 1967. pp. 198. Information bulletin no 1967:60.
construction, hazards identified and robustness considerations [10] H. Griffiths, A. Pugsley, and O. Saunders. Report of the inquiry into the collapse of
during the design, construction and operation phases. flats at Ronan Point, Canning Town. H. M. S. O, [S.l.], 1968. [presented by H.
Griffiths, Sir A. Pugsley and Sir O. Saunders.]
[11] Ministry of Housing Government and Local. Flats constructed with pre-cast concrete
There have been very few cases of complete progressive collapse in panels - appraisal and strengthening of existing high blocks: design of new blocks.
the last few decades and so, while useful for practical cases of low risk, Technical report, HMSO Ministry of Housing and Local Government; 1968.
it is not possible to quantify the efficiency of implementing ties in [Circular 62/68].
[12] IStructE. Ronan Point - the aftermath - RP/68/01: structural stability and the pre-
building structures. In many cases, designers are unaware of the in- vention of progressive collapse, and covering letter titled ‘Ronan Point collapse’
tended purpose and assumptions behind these guidelines. In some cases Technical report IStructE; 1968.
applying these guidelines may lead to the false sense of security and in [13] IStructE. Ronan Point - the aftermath - RP/68/02: notes for guidance which may
assist in the interpretation of appendix 1 to Ministry of Housing and Local
some other cases designers might lean towards alternative load path
Government Circular 62/68. Technical report. IStructE; 1968.
approaches which are quantitative. With this in mind it is therefore [14] BSI. CP 116: addendum no. 1:1970 - the structural use of precast concrete. Large-
recommended that designers using the current tie guidelines should panel structures and structural connections in precast concrete. Addendum to CP
also take into account the required mechanisms for tie to be utilised 116:1965 and CP 116–2:1969. 1970.
[15] IStructE. Ronan Point - the aftermath - RP/68/05: the resistance of buildings to
after an extreme event. In particular joint ductility for large deforma- accidental damage. Technical report. IStructE; 1971.
tions and the prevention of brittle failures (e.g. shear failures) is vital [16] BSI. CP 110-1:1972 - code of practice for the structural use of concrete. Design,
for such an approach to be effective. materials and workmanship. 1972.
[17] BSI. BS8110: Part 1: 1985: structural use of concrete: part 1: code of practice for
design and construction. 1985.
[18] BSI. BS5950: Part 1: 1985: structural use of steelwork in building. Code of practice
Acknowledgements for design. Rolled and welded sections. 1985.
[19] HM Government. Building regulations 2000. Approved document a. 2010.
This work is part of a research project financially supported by the [20] American Society of Civil Engineers. Minimum design loads and associated criteria
for buildings and other structures. 1998. p. 7–16.
EPSRC Impact Acceleration Account held by the University of Surrey [21] BSI. BS EN1990:2002: Eurocode 0 - basis of structural design. 2002.
(grant ref.: EP/K503939); linked with a previous project funded by the [22] General Services Administration (GSA). Alternate path analysis & design guidelines
Engineering and Physical Sciences Research Council of the U.K. (grant for progressive collapse resistance. 2003.
[23] ICC IBC. International building code. 2009.
ref.: EP/K008153/1). The authors would also like to thank the IStructE [24] US DoD. UFC 4-023-03: design of buildings to resist progressive collapse. 2009.
librarians for their assistance in obtaining some of the documents dis- [25] General Services Administration (GSA). Alternate path analysis & design guidelines
cussed in this paper. for progressive collapse resistance. 2016.

372
J.M. Russell, et al. Structures 20 (2019) 365–373

[26] American Society of Civil Engineers. Minimum design loads and associated criteria Nicholas J. Best practices for reducing the potential for progressive collapse in
for buildings and other structures. 2016. p. 7–16. buildings. NIST Interagency/Internal Report (NISTIR)-7396. 2007.
[27] UK House of Commons debate. HC Deb 09 March 1970 vol 797 cc1059–1080. [41] BSI. BS8110: Part 1: 1997: structural use of concrete: part 1: code of practice for
[28] IStructE. Ronan Point - the aftermath - RP/68/03: guidance on the design of do- design and construction. 1997.
mestic accommodation in loadbearing brickwork and blockwork to avoid collapse [42] BSI. BS EN1991-1-7:2006: Eurocode 1 - actions on structures - part 1–7: general
following an internal explosion, and covering letters titled ‘Ronan Point collapse’ actions - accidental actions. 2006.
and ‘loadbearing brickwork and blockwork: RP/68/02 Technical report IStructE; [43] Harding G, Carpenter J. Disproportionate collapse of ‘class 3’ buildings: the use of
1969. risk assessment. Struct Eng 2009;87:15–6.
[29] IStructE. Ronan Point - the aftermath - RP/68/04: the building (fifth amendment) [44] BSI. NA to BS EN 1992-1-1:2004 - UK National Annex to Eurocode 2. Design of
regulations 1970: notes for discussion at a meeting to be held IStructE 30 June concrete structures. General rules and rules for buildings. 2004.
1970. Technical report. IStructE; 1970. [45] Stevens David, Crowder Brian, Sunshine Doug, Marchand Kirk, Smilowitz Robert,
[30] Christopherson DG. Structural defence. British Ministry of Home Security; 1945. Williamson Eric, et al. Dod research and criteria for the design of buildings to resist
[31] Short A, Miles JR. Large panel structures: notes on draft addendum 1 to CP 116 progressive collapse. J Struct Eng 2011;137(9):870–80.
(1965) Building Research Station, [S.l.]. Building Research station current [46] Starossek U, Smilowitz R, Waggoner M, Rubenacker KJ, Haberland Marco. Report
papersGreat Britain: BRS CP 30/69 Building Research Station; 1969. of the terminology and procedures sub-committee (SC1): recommendations for
[32] Fintel Mark, Schultz Donald M. A philosophy for structural integrity of large panel design against disproportionate collapse of structures. Proceedings of the Structures
buildings. PCI J 1976;21(3):46–69. Congress. 2011. p. 2090–103.
[33] Standard, ACI. ACI 318-14 building code requirements for structural concrete. [47] China Association for Engineering Construction Standardization (CECS). Cecs
2014. 292:2014 code for anti-collapse design of building structures. 2014.
[34] Engineers Institution of Structural. Stability of modern buildings. London (11 Upper [48] National Research Council of Canada. Canadian commission on building and fire
Belgrave St., SW1X 8BH): Institution of Structural Engineers; 1971. [GB7206735 codes. 1995.
bnb The Institution of Structural Engineers]. [49] NYC Department of Buildings. Building code (chapter 16 structural design). 2014.
[35] Lewicki B, Olesen SO. Limiting the possibility of progressive collapse. 1974. [50] Australian Building Codes Board (ABCB). National Construction Code (NCC): vo-
[36] Regan PE. Catenary action in damage concrete structures. Special Publication lumes 1 & 2. Council of Australian Governments; 2016.
1975;48:191–224. [51] Regan PE. Report 89: behaviour of reinforced concrete flat slabs. Technical report.
[37] Ellingwood Bruce R, Leyendecker EV. Approaches for design against progressive Construction Industry Research and Information Association (CIRIA); 1981.
collapse. J Struct Div 1978;104(3):413–23. [52] Sagaseta J, Ulaeto N, Russell J. Structural robustness of concrete flat slab structures.
[38] ZA Zielinski OA Pekau, Lee AWK, Hum D. Dynamic effects of panel failure in Special publication honoring Neil M. Hawkins, ACI SP-315. 315. Fédération inter-
precast concrete shear walls. Struct J 1988;85(3):277–85. nationale du bton (fib) and American Concrete Institute (ACI); 2017. p. 273–98.
[39] Gardner NJ, Huh J, Chung L. Lessons from the sampoong department store collapse. [53] Li Y, Lu XZ, Guan H, Ye LP. An improved tie force method for progressive collapse
Cem Concr Compos 2002;24(6):523–9. resistance design of reinforced concrete frame structures. Eng Struct
[40] Ellingwood BR, Smilowitz R, Dusenberry DO, Duthinh D, Lew Hai S, Carino 2011;33(10):2931–42.

373