Guide to Compressive

Membrane Action

7; Prepared for the - - -

Concrete Brldge Development Group
: . ' . . .. ,
A Guide to Compressive
Membrane Action in
Concrete Bridge
Notation
Summary
1.0 Introduction
2.0 Bridge deck slab design
2.1 T~ltroduction
2 2 Global and local behclr lour
3 3 British design codes
3 4 Other bridge design codes
2.5 Summan. of bridge design
3.0 Review sf literature
3.1 Introduction to compressi\-e membrane action
3 3 Early research
3 3 Research b!- Park
3 4 Research at Queen's Unir crsity Belfast
35 Research at U n ~ ~ e r s ~College.
t\.. London
36 Research at Cambridge Unn ersit\
37 Other research 111 the UK
3 S Research 111 Canada

3 9 Other research centres

3 10 Re\.ie~vof d!~namic/fatigue load tests
3 1 1 Summan. of methods
4.0 Sarnrmary of methods for predkctirmg the strength of lateraIly restrained
bridge deck slabs
4 1 Introduction
4 2 Comparison to the existing code (BS5100)
43 Queen's Unixws~t! of Belfast approach
44 Park's method
5.0 ConcIudlng remarks
Tables
Appendix
Bibliography
 cross sectional area width of section
equivalent area of support beam effective width of loaded slab
area steel reinforcement critical perimeter
compressive arching force width of square patch load
con~pressivebending force n-idth of patch load parallel to slab
compressive force in concrete span
compressive force in reinforcement width of patch load perpendicular to
concrete elastic modulus slab span
second moment of area of support effective depth of slab
beam about the vertical axis half the arching depth
axial stiffness concrete cylinder compressive
eqivalent stiffiless of support beam strength
stiffness of diaphragm and slab concrete compressive strength
combined stiffness of restraint concrete indirect tensile strength
stiffness of slab within effective width reinforcement yield strength
half the span of the arch length depth of slab
half the span of the rigidly restrained arch height
arch arch height
arching monlent of resistance Kirkpatrick et al's arching moment
M.,(,,,,,, maximum possible arching moment coefficient
of resistance shape factor
arching moment of resistance of h.IcDowell's maximum compressive
rigidly restrained slab strip stress
flexural moment of resistance at McDoweIl's non-dimensional
principal section parameter (deflection)
balanced moment of resistance deflection under the load point
moment ratio (non-dimensional) depth of concrete compression zone
applied load proportion of d l in contact with the
predicted ultimate arching capacity support
predicted ultimate flexural capacity proportional depth of stress block
predicted ultimate flexxral capacity (=0.9 in BS)
from BS5400 deflection under the load point
predicted ultimate capacity under strain
Kirkpatrick et al's method concrete compressive plastic strain
predicted ultimate capacity under value
proposed method concrete maximum compressive
predicted ultimate capacity under strain
Park's method width of circular patch load
measured ultimate load of test model reinforcement ratio at principal
predicted ultimate punching shear section
capacity from BS5400 effective reinforcement ratio at
flexural punching strength principal section
shear punching strength effective archiig reinforcement ratio
McDowell's non-dimensional at principal section
parameter (elastic deformation) stress
tensile force in reinforcement
Summary
Slabs in beam-and-slab bridge decks are nornlally designed for global bending effects and local
bending and shear stresses These are resisted by conventional reinforcement in the top and the
bottom of the slab. However it is widely accepted that, in addition to the strength provided by
the reinforcement, bridge deck slabs have an inherent strength due to the in-plane arching forces
set up as a result of restraint provided by the slab boundary conditions This has been attributed
to compressive membrane action (CMA) and by adopting arching theory it is possible to
produce bridge deck slabs which have lower quantities of reinforcement and are, therefore, both
more economical and more durable.

In the past 30 years it has become increasingly evident that corrosion of reinforcement due to
the effects of de-icing salts has been one of the major factors in the deterioration of reinforced
concrete bridge decks. Detailing to reduce the risk of corrosion is simpler if the percentage of
conventional reinforcement is low. It has been recognised for some time that restrained slabs
and beams exhibit strengths far in excess of those predicted by bending theory (Turner, 1909).
However, the phenomenon has remained largely esoteric for the practising engineer.
If the edges of the slab are restrained against lateral movement by a stiff boundary, an internal
arching mechanism or Compressive Membrane Action is induced as the slab deflects. This
enhances the flexural load capacity of the slab. The arching phenomenon occurs in concrete due
to the significant difference betw-een its tensile and compressive strengths. The weak strength in
tension causes cracking due to the application of load. This shifts the neutral axis towards the
compression face. If the edges of the slab are restrained by a stiff boundary, internal arching
action is induced (Fig 1).
The enhancement in slab strength, due to arching or Compressive Membrane Action (CMA),
has been incorporated into some design standards, including the Department of Regional
Development (NI) 'Design Specification for Bridge Decks' (1990) and more recently the draft
design guidance, BDS 110 1, 'Use of Con~pressiveMembrane Action in Bridge Deck Slabs'.
This highlights the practical application of a design theory which allows for compressive
membrane action in such decks

Fig. 1: Compressive membrme action in laterally restrained

reinforced concrete siab
It has been well documented that the degree of lateral restraint has a beneficial effect on the
strength of restrained slabs and bridge decks have inherent restraint due to the presence of
diaphragms, supportin,a beams and the surrounding area of deck slab Similarly, concrete
strength has a significant effect on the arching capacity. Therefore, by utilising the advantages
of high strength concrete and lateral restraint it should be possible to produce deck slabs with
very low percentages of reinforcement
This report presents a review of the literature in regard to CMA in bridge deck slabs. The
design procedures developed at Queen's University of Belfast and the University of Canterbury
in New Zealand are considered in detail together with a typical example of the Queen's
University procedure for predicting the strength of laterally restrained slabs.

2.0 Bridge deck slab design

2.1 Introduction

New bridges are usually designed to a Code of Practice, in the UK this is BS5400, Part 4 (1990)
and the associated Highways Agency Standards (BD's). Similar standards or codes exist for the
assessment of existing structures. These codes have been developed either from an accepted
rational theory or from empirical formulae based upon experimental evidence and engineering
experience. Therefore, any new design or re-discovered design concept, such as compressive
membrane action, is unlikely to be accepted in modem design unless it is included in the
relevant design codes Compressive membrane action has only recently been included in the
draft BD81101 and, to date, has remained largely unused in design This guide aims to address
one obstacle to the use of BD8 1/01 by setting out in detail the method by which the capacity of a
laterally restrained slab can be calculated.

2.2 Global and local behaviour

The global behaviour of a bridge structure is determined by the overall distribution of forces and
deflection both longitudinally and transversely. This analysis is normally carried out for a
complex envelope of loads I-epresenting the many different load combinations which are
achievable. The local behaviour is the effect on a particular element of the structure caused by
individual loads, for example, a wheel load on a section of the deck slab In the design, and the
assessment, of bridge structures it is difficult to assess the combined effects of both the local
and the global load Therefore, in most cases, the two effects are treated independently. In the
UK, bridge designers combine the worst coexistent local and global load effects and the detailed
analysis subsequently examines the stresses at particular elements in the structure In the design
of the deck slab of a beam-and-slab type bridge it is generally the local effects which dominate
Therefore, the focus of this guide is on the design of the deck slab in relation to the local effects
2.3 British design codes

Currently the British Standard for reinforced concrete bridge design (BS5400 Parts 2 and -I), in
conjunction with the Departmental Standards BD15192, BD24192, BD37101 and BD44195
reconmends the use of Westergaard's (1930) or Pucher's (1964) charts to assess the local
effects of wheel load This generally gives a level of transverse reinforcement of between 1 2%
and 1.7% in the deck slab However, these charts have been developed from plate theory
assuming a linear elastic material and are not representative of the behaviour of a restrained
reinforced concrete slab. In other words, the effect of membrane action has been neglected.
A detailed program of field tests in Northern Ireland showed the deck slabs of RI-beam type
bridges with reinforcement ratios of less than one third of those calculated using the design
charts performed satisfactorily. In addition several laboratory models tested with reinforcement
ratios as low as 0 25% behaved adequately. Subsequently the Northern Ireland Standard
(DRD, 1990) was amended to reduce the amount of reinforcement The code suggests the use
of 0 6%, top and bottom, transverse reinforcement in the deck slab of M-beam type bridges
nith a main beam spacing of 2111 or less This includes a fairly high safety margin based upon
the results of field tests but halves the amount of reinforcement compared to that predicted by
the use of Pucher's or Westergaard's methods.
There are, however, a number of limitations in the use of this standard. The span to depth ratio is
restricted to less than 15 and the tests (Kirkpatrick, 1984) were based on a 160mm-deck slab
The code recognises the lack of lateral restraint in edge locations of the deck, compared to the
interior sections Therefore, the edge cantilevers are designed by the conventional methods and
this generally requires greater percentages of reinforcement.
Following the work by Kirkpatrick and Rankin the Highways Agency has now produced a
draft design guidance BD8 1/01, 'Use of Compressive Membrane Action in Bridge Deck Slabs'
This draft is currently being finalised but outlines the Queen's University of Belfast method
developed by Rankin and Kirkpatrick However, the same limitations apply as to the NI design
code and for slabs outside this criteria it is not clear how to assess the d e ~ r e eof external lateral
restraint

2.4 Other bridge design codes

There are hndamental differences in the methods adopted for the analysis of bridges throughout
the world and t h s is reflected in the relevant design codes For example, bridge engineers in the
LK combine the worst effects of the local and global loads In USA the deck slab is generally
designed for the local effects; the section is then deemed adequate for the effects of global
loading This is valid for the majority of bridge structures used; namely, steel-concrete
composite beam-and-slab bridges ~vithinternal diaphragms

A small number of countries have adapted their codes to recognise the presence of compressive
membrane action In addition to Northern Ireland, which has already been mentioned (DOE,
1990), the Canadian Code (OHBDC, 1992) has reduced the reinforcement requirements These
incorporate n~ostlyempirical design rules based upon the results of field and laboratory tests
The Canadian code (OHBDC, 1992) recommends 0.3% isotropic reinforcement in a deck slab
with a minimum thickness of 225mm. The span to depth ratio is limited to less than 15 which
corresponds to a maximum slab span of 3.7m. The code requires additional reinforcement in
certain circumstances, such as at the ends of highly skewed decks where the amount of
reinforcement is doubled; though recent research (Bakht and Aganval, 1995) has suggested that
this is too onerous Following a series of fatigue tests on similar skewed deck slabs (Bakht et al,
1996a), an addendum to the OHBDC code has been made. This removes the restriction for
additional reinforcement provided the explicit in-plane restraint has been provided, most
commonly by steel channel diaphragms placed with the web in the horizontal plane.
2.5 Summary of bridge design

There are substantial differences in the methods of design and the forms of construction
adopted for bridge structures throughout the world. The most common form of bridge
construction in the latter half of the twentieth century has been the beam-and-slab type
structure In North America this tends to be a steel/composite slab form but in the UK, pre-cast
concrete beams are also common Thus, theory based on a reinforced concrete deck slab is
relevant to both forms of bridge structure
As already stated, the global loading has an effect on the transverse moments but it is the local
loading which has the dominant effect. Traffic loading creates an infinite variety of load
combinations but for the deck slab the niorst position is a concentrated load at the midspan of
the slab With the exception of a few codes (OHBDC, 1992 and DRD, NI 1990) the effects of
compressive membrane action have been ignored although the slabs in a typical beam-and-slab
bridge deck have inherent restraint and represent a structure ideal for the development of
compressive membrane action

3.0 Review of literature

3.1 Introduction to compressive membrane actfoil
If the edges of a loaded slab are restrained against lateral movement by a stiff boundary, an
internal arching reaction or compressive membrane force is induced as the slab deflects (Fig. I ,
p. 1) The behaviour at low loading is elastic, region A to B in Fig.2, but reaches an elastic-
plastic phase, B to C, prior to the peak load at C. This peak load corresponds to the maximum
arching effect and, in under reinforced slabs, the maxinium bending strength. At increased
deflection, the subsequent load capacity reduces.

Midspan deflection +
Fig. 2 : Typicid load vs. deflection for restrained reinforced
concrete s h b

The following sumiiiarises the wealth of literature covering compressive membrane action with
specific attention to bridge deck slabs It has been shown that the arching effect is relatively
greater in slabs with lower reinforcement percentase, low span to depth ratio and a high degree
of lateral restraint That is, in comparison to the flexvral or yield line predicted ultimate
strengths which do not consider membrane effects.
3.2 Early research
The phenomenon of 'arching' has been recognised by engineers for many decades Turner i n
1909 was one of the first to acknowledge arching effects when he wrote 'such a slab will at first
act somewhat like a flat dome and slab combined' In 1921 Westergaard and Slater described
the effects of arching but in the ensuing years most research appears to have followed an
alternative path The first serious endeavour to rationalise ChfA in analysis was in the So\.iet
Union when Gvodzev (1936) published a paper referring to the Russian code of practice for
reinforced concrete, which took into account arching action. However, it was probably due to
the pioneering work of Johansen, on yield lines, that most other research in this field followed
another route and, combined with the effect of the Second World War, the arching phenomenon
did not regain interest until the 1950's.
It was brought to prominence in 1956 when Ockleston published the full-scale loading test
results from the Old Dental Hospital in Johannesburg. The measured failure loads were
considerably higher than those predicted by yield line theory, which had become globally
accepted at the time. Some years later in South Africa, Liebenberg (1966) carried out field tests
on an eight-storey reinforced concrete building in Cape Town which clearly illustrated the
nature of CMA. Around the same time that Ockleston was carrying out his work, arch action in
masonry panels subjected to transverse loading was observed by hllcDowell et a1 (1956) They
proposed a theory for arching action which represented a radical deviation from current thinking
at that time.
Brotchie (1963) used classical elastic theory to provide a solution for the capacity of restrained
slabs but did not consider an elastic lateral restraint. Also in 1963, Christiansen developed a
theory for one-way spanning slabs restrained by a flexible boundary. Nearly two decades later
Christiansen (1983) published two papers setting out a simplified approach t o assess the
strength of laterally restrained slabs based upon the following:

Brn = Ptest - Pj WWI

\%-here P,,, = load due to con~pressivemembrane action
P,,,, = maximum total load on the slab
P, = Johansen's loads (i e flexural capacity using yield line analysis)

Although it was appreciated that the flexural and arching effects were not separate mechanisms
in reality, the consideration as two independent effects was put fonvard as a rational approach
for predicting the peak load. Taylor and Hayes (1965) carried out tests on 22 unreinforced and
reinforced square slabs in pairs. They concluded that the enhancement in strength was greatest
in the pairs where the simply supported model had been close to flexural failure prior to
punching, that is, in the slabs with a lower percentage of reinforcement Roberts (1969)
conveyed his theory for slabs with varying restraint conditions although the tests carried out
included of a relatively stiff surround compared to slabs in practice. Compressive nlembrane
action received significant attention in 1971. The American Concrete Institute held a seminar,
which was aimed at bringing together researchers in the field of concrete slab systems The
special publication (SP30-14, 1971) contains several papers devoted to compressive membrane
action.
-4 gzm& 10 can~pressi\vmembrmre crcirorl in bridge deck slabs 6

3.3 Research by Park

Park has been one of the major contributors to research into compressive membrane action and
actually devotes a whole chapter to the subject in his book (Park and Gamble, 1980) Park
(1961a) presented a theory where the ultimate moment of a rigid-plastic strip was determined,
based upon a yield line pattern and using horizontal equilibrium combined with geometric
compatibility. Park concluded that, due t o the sensitivity to the concrete strength, the strength of
restrained slabs is highly dependent upon the stress diagram employed in the calculations He
used Hognestad's relationship, which assumed an elasto-plastic material property.
Later, he presented a refined theory which included an adjustment for the lateral restraint and
axial strain, caused by shrinkage and creep in the concrete (Park, 1961b). The following year,
Park (1965) presented the results of tests on twenty small-scale mortar models where the span
to depth ratio varied between 18 and 30. Park refined his theory and summarised the many years
of work in his book (Park and Gamble, 1980) and his method is discussed later.

3.4 Research at Queen's University Belfast

In the late 1970's the effects of compressive membrane action were being investigated at
Queen's University, Belfast and have continued until the present day. Masterton and Long
(1 974) first suggested that the punching strengh of slab-column structures would be enhanced
by the development of CLMA. Later, Rankin (1982) developed a rational approach for the
strength of laterally restrained reinforced concrete slabs. Rankin allowed for a less than rigid
restraint by using the analogy of a three-hinged arch (Fig.3). The analysis was exxended to
predict the strength of two-way spanning slabs, under a concentrated !oad, by u i n g a two-phase
approach, similar in concept t o that proposed by Long (1975). It separately predicted the
enhanced flexural and shear punching strengths of the slab and the ultimate strensh was the
lesser of the two values.

a) elastically restrained arch

b) equivalent rigidly restrained arch

Fig. 3 : Analogy of three-lringed arch (Rankin, 1982)

In 1987 Rankin published his refined theory including the effects of compressi~.emembrane
action Rankin attempted to p r o ~ i d ea relatilrely simple method for the prediction of flexural and
shear punching strengths of interior slab colurnn specimens and he showed that the theorjc
provided more realistic values for the actual strength compared to the design codes While there
is general agreement for the bending strength of slabs there are major discrepancies betneen the
design codes for the prediction of shear stren,nth The British Standard equations for shear have
been developed from n o r k by Regan whereas the American codes ha1.e progressed from the
theories of Moe
ParaIlel to the work by Rankin at Queen's University, Kirkpatrick (1982) was investigating
arching in the deck slabs of M-beam bridge deck slabs. This included both field and laboratory
tests and is summarised in several papers. The first paper (Kirkpatrick et al, 1982) described
tests carried out on four bridges around Northern Ireland and a subsequent paper (Kirkpatrick et
al, 1984a) presents the punching strength test results on a third scale model of Clinghan's
bridge The analysis of punching was developed by modi@ing Rankin's model and the arching
capacity was given by the equation.

M,, = k. f c.h2 where k = u . M , ( h41 = 4 for rigid plastic material [Eqn.2]

4 < I for elastic-plastic material)

It was postulated that the arching effect could be equated to an 'equivalent' reillforcement
percentage which had a similar effect on the depth of the neutral axis. By the substitution of the
equivalent reinforcement index into the equation for the punching shear strength the enhanced
punching strength was established The theory showed good agreement for thick slabs (span to
depth ratio less than 15) with near rigid restraint and showed that the then current design codes
were highly conse~ative.A second paper by Kirkpatrick et al (1986) presented the results of
serviceability tests on the actual bridge, Clinghan's bridge In this paper they reiterated the
conservative nature of the design codes. Westergaard's charts, used t o establish the effect of
local wheel loading, had required 1 776 transverse reinforcement However, the actual range was
detailed with between 0 25 and 1.7% transverse reinforcement and, under an abnormal wheel
load of 1 12 5kN, all the regions of the deck slab remained uncracked. They concluded that
co~npressivemembrane action controlled the onset of cracking and that the BS5400 semi-
empirical equation for the prediction of crack widths was not applicable to restrained slabs
In the same pear another researcher at Queen's University, Niblock (1986), proposed an
analjtical method for predicting the strength of restrained slabs under uniformly distributed
loading At about the same time Skates (1987) was investi,oating the feasibility of utilising
arching action in cellular structures. Ruddle (1989) considered the effect of compressive
membrane action in laterally restrained rib slab systems.
In 1988 Long et a1 presented an overview of the tests on h4-beam bridge decks. It was
concluded that the percentage of reinforcement in the deck slab could be reduced to 0.6% with a
beam spacing of between 1 5 and 2113 This represented a 35% saving in the cost of a typical
bridge deck. It was shortly after this that the Department of the Environment (Northern Ireland)
modified the bridge design code to allow for compressive membrane action if certain conditions
of restraint were present The following year Long and Rankin (1939) published an overvjew of
research where the vast majority of the tests show strengths far in excess of those predicted by
the current theory. This was mainly attributed to the presence of arching action. A subsequent
paper in 1995 (Long et al, 1995) highlighted the benefits of including compressive membrane
action in bridge assessment In 1997 Rankin and Long presented, in detail, Rankin's theory for
the predicting capacity of laterally restrained slab strips.
Some of the most recent research at Queen's University of Belfast into compressive membrane
action investigated its effect in continuous steellconcrete composite floor slabs (Peel-Cross et al.
2001) More recently Taylor et a1 (2001a) presented the effects of compressive membrane
action in high strength concrete bridge deck slabs It is known that the compressive strength of
concrete has a significant effect on the strength of laterally restrained slabs. This research
extended the existing knowledge of compressive membrane action for concrete ~vith
compressive drenghs up to 100~1mm' and by utilising the advantages of high performance
concrete it was possible to produce decks with very low percentages of reinforcement. Fifteen
one-way slabs typical of a section of bridge deck were tested. The variables included concrete
strength, degree of edge restraint and the percentage, position and type of reinforcement.
The extent of arching action is dependent upon the degree of lateral restraint and this has proved
difficult t o quantify Taylor (2000) provided a method for assessing the degree of lateral
restraint by using a restraint model and this is described more hlly in section 4.4. Taylor et a1
(2001b), developed a method for predicting the ultimate load carrying capacity of bridge deck
type slabs with a range of boundary conditions The proposed method was found to more
accurately predict the strength of these slabs compared to current methods. The research also
investigated novel ways in which the slab could be more effectively and efficiently reinforced
This included the use of polypropylene fibres within the concrete which reduce thermal and
shrinkage cracking to further enhance the long term durability.

3.5 Research at University College, London

In the 1980's and 1990's research into compressive membrane action in the LK was also being
carried out under Morley in Cambridge University and Keinp and Eyre in University College,
London. Eyre and Kemp (1984) postulated that the load enhancement versus deflection
relationship was a reflection of the yield curve for rigidly restrained slabs. In 1990 Eyre
highlighted the difference between flow and deformation rules and concluded that for a rigid-
plastic system the two rules give identical solutions for the membrane force at maximum
deflection Eyre's theory predicts a safe working load for a rigidly restrained slab from the
calculated total load versus deformation response.

Eyre and Kemp (1994) provided evidence that the axial stiffness of the slab was dependent upon
the stress state in the slab and much smaller than the stiffness given by the full depth of the slab
Most rigid plastic theories (e g , Christiansen, Roberts, and Braestrup) assume that the stiffness
is given by the following equation.
K= E.A [EW.31
L/2

Other researchers reduced the value to account for the cracked section properties. Eyre
suggested a value of stiffness incorporating only the area of concrete, which was hlly stressed,
similar to Rankin (1982).

3.6 Research at Cambridge University

Recent work was highlighted in a paper by Kuang and Morley (1992) Tests were carried out on
twelve restrained slab specimens at 115 scale. Contrary to the effect of the reinforcement found
in the tests at Queen's (Kirkpatrick, 1982), they found a 51% increase in the non-dimensional
strength of the 40mm thick slab when the reinforcement was increased fiom 0 3% to 1.0%.
However, this particular slab had a high slenderness value and was, perhaps, not typical of many
d gzu~leto con~pre.s.si~v
--
n~emhrzrr2eacfiorr it? 1T1ricfgedeck sInhs 9

structures in practice, particularly bridge decks Kuang and Morley \yere in ageement lvitli
most other researchers in concluding that the degree of edge restraint was significant to the
ultimate strength of the restrained slabs They also recognised that the presence of arching
played an important role in controlling cracking. The ultimate strenghs were much greater than
the punching strengths predicted by both the British Standard (BSSI 10) and the American code
(ACE 18-89) which highlights the conservative nature of the design codes A paper the
following year (Kuang and Morley, 1993) described a plastic theory using hlohr's failure
criterion for the punching shear. The most recent n o r k at Cambridge in this field was published
the following year (Morley et al, 1995). This described the 'Galerkin' approach to computing
the capacity of restrained slabs. A stiffness ratio was adapted to model the edge restraint and the
procedure considered both a rotational and lateral restraint

3.7 Other research in the UK

One of the first attempts to incorporate finite element analysis into arching theory was by Cope
and Rao (1977) and a subsequent paper (Cope et al, 1979) used a non-linear finite element
analysis (NLFEA) for determining the strength of bridge deck type slabs. Jackson, 1989,
presented an elastic method for predicting the strength of a restrained slab strip but he only
considered an unreinforced strip. He went on to outline a relatively sin~plenon-linear finite
element analysis which accounted for membrane action in bridge deck slabs.
Jackson validated the analysis by testing two half scale bridge models under full HI3 loading He
concluded that diaphragms were not required to enable the development of compressive
membrane action but restraint was inherent in the surrounding under-stressed material This
followed similar conclusions by Rankin (1982). Jackson's laboratory tests were some of only a
few which considered the effects of compressive membrane action on the global analysis
Jackson determined that the failure mechanism was 'influenced by the global behaviour'. In
1990 Jackson and Cope published several papers outlining the advantages of compressive
membrane action in the design and assessment of bridges particularly where costly
strengthening programmes would be the alternative
Research at Bristol University into compressive membrane action has been described by Lahlou
and Waldron (1992) They tested three slab strips v,ith varying end wall details and the results
were compared to a NLFEA package called ANSYS, vvhich consisted of an 8-node solid
element with two degrees of freedom at each node and showed good agreement The most
recently reported research in the UK into compressive membrane action has been at Aberdeen
University. Famiyesin and Hossain (1998) made use of a FEA package to develop a series of
design charts. The FEA package had been verified using the results of several test models. The
charts showed good agreement with 36 test results reported in literature. However, the charts are
only valid for rigidly restrained slabs and therefore have little bearing on the strength of many
bridge deck slabs in practice

3.8 Research in Canada

I-Iewitt and Batchelor (1975) presented a rational theory for membrane action and the following
year Batchelor and Tissington (1976) advanced the research They implemented tests on bridge
models and suggested a theory, based upon Christiansen's concept of a combined flexural and
arching moment Batchelor et a1 (1978) set out a detailed test progranme to assess the
endurance limit of slabs with a various amounts and arrangements of reinforcement Five
models of a steeVconcrete composite type bridge, at '18 scale, were tested The tests at Queen's
University, Ontario, led to hrther tests on actual bridge structures in a comprehensive study
A grtidt? to con7pressi~vntembrcrne actro??itr bridge deck slcrhs 10

sponsored by the Ontario Ministry of Transport. Subsequently the design code (OHBDC, 1979)
was changed to allow a reduction in reinforcement to 0.3% in the deck slabs provided certain
boundary conditions existed
In the 1980's and 1990's research into the behaviour of bridge decks continued to be prevalent
in Canada Bakht, Mufti and Jaeger have published many papers on this subject In 1992,
Bakht and Jaeger summarked the results of tests on short span simply supported 'slab on girder'
bridges. The transverse distribution of loads appeared to improve at higher loads. The FEA
packages adopted were capable of modelling arching by incorporating sufficient degrees of
freedom to allow for the in-plane restraint. The following year, Mufti et a1 (1993) carried out
tests on M-scale bridge models containing no conventional reinforcement The third model
included external straps welded to the underside of the top flanges. This provided sufficient
lateral restraint to ensure a punching failure at a load far in excess of the required ultimate
bridge loading. However, the total area OF the steel straps was equivalent to a 1 4% area of
conventional reinforcement, which is over three times the current recommendations in the
Canadian bridge code (OHBDC, 1992). Another possible problem with external straps is the
introduction of increased maintenance in the life of such a deck. A second paper (Thorburn and
Mufti, 1995) details the second series of test using similar models to the above A range of
values for the external restraint was established and a preliminary method adopted for
optimising the value of the required restraint.
In 1995 Selvadurai and B a h t provided information on tests simulating a rolling wheel load on a
FRC deck. The tests were carried out on full-scale laboratory models and it was concluded that
the decks could adequately withstand four million passes at 98kN loading which was greater
than the equivalent working bridge load The same year Bakht and Aganval (1995) presented
the results of test on a series of third scale model skew bridges. These were aimed at testing the
ability of arching action in the end regions of slabs in highly skewed bridges They concluded
that it was not the geometry of the structure but the degree of lateral restraint which governed
the strengh of this region of the deck slab
Bakht et a1 (1996) sun~rnarisedthe recent work on steel-free bridse decks. They outlined the
current requirements in the OHBDC for a 0.3% isotropic mesh and emphasised that the main
cause for the deterioration of existing bridge decks was the corrosion of the reinforcement.
They emphasised that decks devoid of steel reinforcement have hisher durability and m-here
structural reinforcement is required, such as that catering for the hogging moment in the
cantilevers, plastic reinforcement was preferable. A rudimentary cost analysis on this type of
deck predicted an increase of only 8% in the cost of the deck.

The construction of five steel-fiee bridges in Canada (Bakht et al, 1998) enabled an accurate
cost analysis to be performed. The first bridge constructed included a 6% increase in the tender
price for the steel-free deck slab compared to the conventional reinforced concrete deck slab.
However, in a subsequent tender, the lowest bid quoted was for the FRC bridge deck. The
paper also described a NLFEA method developed by L/fufti and Wegner but concluded that
accurate comparisons with experimental work can only be achieved after 'tuning' the
parameters and this required prior knowledge of the experimental results

The ex7ensive research into compressive membrane action in bridge decks in Canada appears to
have yielded a beneficial solution. However, the differences in both the form, the type of
loading and the analysis used for bridge construction limit the use of such bridge decks in the
UK; particularly in Northern Ireland where the use of steel composite bridges is less prevalent.
3.9 Other research centres
A major contribution to work into the effect of the edge restraint in restrained slabs m-as made b\,
Guice and Rhornberg (1988) in the USA. These tests represented some in a limited number 0-f
tests which measured and varied the edge restraint In 1990 Fang et al presented the results of
field tests on steel con~positebeam-and-slab bridges The NLFEA package, SAPIV, modelled
the arching phenomenon in two layers of finite elements and the analysis shoued reasonable
correlation with the deck detailed under the Texas design code (an Ontario-type deck detail) It
was also noted that the behaviour of the deck was essentially linear up to three times the current
design wheel load A subsequent paper by Fang et a1 (1994) provided the results of tests on
eighteen slabs with two supporting edge beanis

Azad et a1 (1993) carried out field test on the punching failure of composite beam-and-slab
type decks in Saudi Arabia and found that the orientation of the crack width was critical to the
ultimate strength. A subsequent paper ( h a d et al, 1994) described the interaction of flexure and
shear in punching and compared the results of twelve '/2 scale bridge models with several
theories The ACI predictions of the strength were highly conservative but the ATFEA gave
good predictions Petrou and Perdikaris (1996) presented a theory similar to Rankin's approach
A three-hinged compressive stmt with finite edge restraint was used to model the arching
behaviour. A linear elastic FEA incorporating 8-noded brick element was used to establish the
value of restraint.

3.10 Review of dynamiclfatigue load tests

The majority of research carried out and sumniarised in tlis report has been for static l o a d
However, as the live loading on a bridge is dynamic the following section outlines some of
research carried out with cyclic and rolling wheel type loads. One of the most influential series
of tests was carried out by Batchelor et a1 (197Sb). From the numerous results they obtained
the following relationship between the fatigue load and static failure load, where N is the
number of load cycles

This suggests that a load which is a fraction of the static failure load, such as the working bridge
load, needed to be applied for a large number of cycles to have any significant effect In their
tests the nun~berof load cycles varied between one and five million

Fang et a1 (1990) carried out tests on full-scale bridge models which incorporated fatigue tests.
A fatigue load of between 22kN to 115kN was varied sinusoidally and up t o 5 million cycles of
load applied. The same year Jackson (1990) referred to dynamic testing of bridge models in the
UK. He highlighted the greater effect of a rolling wheel load compared to a pulsating wheel
load. In Jackson's test, the rolling wheel load was applied for 5000 cycles Moussard (1993)
carried out dynamic tests on HSC specimens where two million load cycles of 100kN load were
applied at a frequency of 1Hz. He concluded that a dynamic coefficient of 1 3 was required for
comparison to static tests
The hlinistry of Transport in Ontario, have carried out numerous dynamic tests on both full and
part-scale models. In 1995 Bakht described tests conducted on third scale models, in which the
cyclic load was varied from OW to 40kN and applied for two million cycles. In 1995
Selvadurai and Bakht provided information on tests simulating a rolling wheel load on a hll-
A gitrde to comyressi~vn z e m h m r ~nciior2 irl bridge deck slobs 12

scale FRC deck They adopted Perdikaris's method of simulating a moving wheel load mhich
used a sequence of pulsating loads. A frequency of 3Hz was equivalent t o a wheel moving at
40krn/hour The deck slabs xvere siniilar to the fibre-reinforced slabs pioneered in 1993 by
Mufti et a1 and contained 0.88% polypropylene fibres. The decks behaved adequately under four
million passes at 98kN loading

3.11 Summary of methods

The review of literature presented an oven~iewof the various methods for predicting t l ~ e
enhanced strength of laterally restrained reinforced concrete slabs. The methods can be
categorised as hand-calculable elastic-plastic and plastic with non-linear finite element analyses
(NLFEA) being a computer based elastic-plastic method. Table 1 (p.27) summarises some of the
reported methods under these three categories.
The NLFEA computing techniques used to model arching action require both non-linear
material and geometric properties. Although the technology in this field has advanced greatly in
recent years, the programmes are overly complex- for the vast majority of bridge designers lo
use. Additionally, many of these NLFEA have been performed for slabs where the strengths
are known. This has often involved calibrating the model and in many cases the NLFEA
parameters differ for each experimental test model beins simulated. In this respect, it is not
feasible within the scope of this guide to summarise a NLFEA method for predicting the
strength of laterally restrained bridge deck slabs.
The plastic or yield line methods have generally developed from the work by Wood (1961) and
Park (1960 & 1980) Most methods require some mathematical manipulation which leads to
long expressions. The method proposed by Eyre (1994) predicts a safe load as opposed to an
ultimate load and the factor of safety is unknown. Eyre's method also assumes rigid restraint
which may not occur in structures in practice. For simplification and due to the similarities in
most of the plasticity approaches, the method by Park (19'80) has been presented in this design
guide
The elastic-plastic approach is typified by the Queen's University of Belfast method. However,
some of the approaches, such as Brotchie (1963), do not account for a varying amount of
external lateral restraint. Christiansen's semi-empirical method separates the arching and
bending phenomena This is similar to the method derived by Rankin and Long (1997) at
Queen's University which has been presented in this design guide
1.8 Srrnmmarg: of methods for pr-edictillg the strength of Eaterally restrained
bridge deck slabs
1.1 Introduction
In this section. an ovewiew is given of the various methods for predicting the strength of
iaterally restrained slabs Some distinction has been made between the type of analysis used
The University of Canterbury's (Park and Gamble, 1980) method has been used to demonstrate
a plastic approach and the Queen's University procedure represents an elastic-plastic method
Using the Queen's approach, the possible modes of failure for a bridge deck slab are established
and a distinction is made between the flexural and shear punching modes of failure.

4.2 Comparison to the existing code (BS5400)

In the British Standard, RS5400 (1990) and associated BD's, a bridge structure is designed to
resist the worst combination of HA, a uniformly distributed load representing normal traffic in
the UK, and HB loading, an abnormal vehicle unit loading in conjunction with other
superimposed loads such as wind and temperature The design code applies five separate load
con~binationseach comprised of other load combinations and envelopes. However, as discussed
in Section 2, the predominatins criterion for the design of the deck slab is the local effect under
a concentrated wheel load. In BS5400, the bendins capacity is described by the following
equation (with the safety factors removed):

The equation has been used to predict the flexural strength of the Taylor et al's (2001a) tests on
one-way spanning slab strips. As highlighted in Table 2 (p 29), the predicted values are highly
conservative for laterally restrained slab strips and are increasingly conservative at high concrete
compressive strengths This is demonstrated graphically in Fig4

FE - fixed end
S/S - simply supported
LiR - lateral restraint
K = stiffi~essof end restraint

0 20 40 60 80 100 120
Concrete Cube Strength (Nlmrn2)

Fig.4: Comparison of slab strip fzilmre loads to predicted values in design code
The test results clearly show an increase in ultimate strength nith increasing concrete
compressive strength but this is not recognised by the d e s i ~ ncode The test results also
demonstrate the increase in ultimate strength with an increase in the lateral restraint. This is also
neglected in the current code methods Conversely, the desisn code, with safety factors
removed, accurately predicts the strengths of the simply supported test models. A second series
of tests on bridge deck models has also been used for comparison (Taylor et al, 2001b) with the
current design code These tests were aimed at more realistically modelling the restraint
conditions in a real bridge deck slab by incorporating the supporting edge beams. The
supporting beams act as external restraint to the slab and by varying the width of the edge beams
the stiffness of the external restraint was varied. In order to relate the applied bending moment
to the flexural moment capacity, the relevant Pucher chart was utilised to assess the effects of
the concentrated load (similar to the local wheel load). This gave a relationship of applied load
to bending moment of

The punching shear strength is given in BS5400 by the following equation

The critical perimeter is assumed to be at a distance of I 5d fiom the edge of the loaded area. A
sumnary of the predicted ultimate flexural and punching shear capacities is given in Table 3
(p.29) and some of the results have been presented in Fig. 5 The test results clearly show an
increase in ultimate strength with increasing edge beam width which is equivalent to increasing
the external lateral restraint stiffness. This is not recognised by the BS5400. It can be concluded
that the current British codes do not accurately predict the strengths of the model deck slabs and
in reality there is a substantial enhancement in both the flexural and punching shear capacities
due to Compressive Membrane Action

trendline.
P"P
/-
/
/

0
-- --
I
--
BS5100 (flexural mode)

I I
-I
50 100 150 300 750

Supporting edge beam width (rnm)

Fig.5: Comparison of 113-scale model bridge deck slab failure

loads to predicted valares in design code
A mi& fo c o m v r e ~ s r l ~r e' m ~ ~ h r ~nctrm
m e rrz br-idye deck slobs 15

1.3 Qrreen9sUniversity af Beffast approach

4.1.1 Kirkpatrick, Rankin arrd Long

Kirkpatrick et a1 (1953) developed a simple means of assessing the punching capacity of
reinforced concrete bridge deck slabs by considerins arching action They assumed that a
bridge deck has an effecti17ely rigid restraint system and the mechanism of failure is a punching
shear mode Kirkpatrick believed that the diapliragms in his own tests had provided the
necessary restraint to prevent the transverse movement of the beams due to the arching thrust
The design charts derived by Kirkpatrick are illustrated in Fig.6

4 8 12 75 20 2i 28 32
L/h
Fig. 6: Kirkpatrick maximum arching moment carves

The design charts were used to establish the maximum arching moment in terms of the concrete
compressive strength and the span to depth ratio where.

RI,, =kf,. hZ and k = N . R i , ( Mr = 4 for rigid plastic material rEqn.21

4 <4 for elastic-plastic material)
When the maximum arching moment has been calculated it can be related to the equivalent area
of flexural reinforcement, p,, using the following.
p, = kf,. hz [Eqn.8]
240d2

This was than substituted into Long's (1975) equation for the shear punching strength i.e: for a
circular load:
Ppt=1.52.(++d).d.&.(100~,)"" I%n-91

or more generally described by:

u here rt; shape factor = 1.0 (circular load) or 1.15 (rectangular load) and critical perimeter is
taken at 0.5d from the face of the loaded area.
 mide lo contpres.sr~en7embrzrlte crctio~zin bridge deck slabs
.-I 16

However the archmg capacity cannot exceed the maximum capacity as represented in Fig 7 and
by the following equation

ann = Id2

Fig. 7 : R/Lar;irnumarching capacity of a slab

Kirkpatrick et al's method showed good asreement between predicted and actual failure loads
for several reported tests on bridge deck s l d x However, the predictions were slightly unsafe
where a flexural failure occurs, such as in lightly reinforced slabs with a moderate degree of
external lateral restraint In this case the following refinements were developed

4 . 4 2 Rankin and Long

Rankin and Long (1997) first developed their approach for the strength of laterally restrained
slab strips as part of their rational approach to predicting the enhanced punching strength of slab
and column specimens The method separates the bending and arching components In reality
the two mechanisms are combined and the compression in the concrete is due to both the action
of arching and bending. Rankin extended the theory of McDowell, McKee and Sevin (1956)
which focused on the geometry of deformation of laterally restrained masonry walls Two non-
dimensional parameters, R and u, are used t o describe the geometry of deformation.

'R' is a measure of the elastic deformation and 'u' is a non-dimensional measure of the
deflection of the slab strip. Rankin mathematicaIIy manipulated these terms to derive an
expression for the arching moment ratio where.

(i) R > 0.26: RZ, = 0.3615 and u = 0.31 (constant) [Eq n. 131
R

(ii) 0 < R < 0.26: M = 4.3-16.1d3.3~10-~ + .IBGR

1'
~1 = - 0.15 + 0.36 J0.18 + 5 . 6 R [Eqn.l-%]
In predicting the arching strength of restrained reinforced concrete slabs it is, therefore,
necessary to establish the depth of the arching section, ?dl, and the plastic strain value, E, The
depth of section available for arching is firstly estimated from the overall depth minus the depths
of the con~pressionzones due to bendins The plastic strain value is the value of strain uhen
hi1 plasticity is first achieved This was established using an equivalent trapezoidal stress block
%here the base length of the rectangular portion is equivalent to F times the ma~iniumstrain
value (Fig 8) The value of the plastic strain is then given as
sC= 2 ES. ( I - P ) [Eqn.lS]

Fig. 8: Eqrrivalent stress conditions zt ultimate b a d under proposed methad

The strength of laterally restrained slabs is highly dependent upon the degree of external lateral
restraint Therefore, the restrained slab system with finite restraint stiffness was equated to a
rigidly restrained slab, i e infinite stifiess, using the three-hinged arch analosy as discussed in
Section 3.4. The solution to the equilibrium equation is outlined in Appendix B of Rankin's
thesis (1982). In summary, the longer equivalent rigidly restrained slab has been used to
describe the load-deformation response of a shorter finitely restrained slab. The equivalent
length is then given b y
A grride to conipressi~vmenibrnne nctiori ill brigye deck slabs 18

T h s provided a simple analytical expression 11-hich could be incorporated into the non-
dimensional parameters used to describe archiny moment. The stiffness of the slab strip has
been based upon the axial stiffness where
K = E,AIL, where A = a.4 per unit width, and Er = 4.23 fCtto' [Eqn.l7]

'A' is the depth of the contact zone as described in Fig3 and the \value for the elastic modulus
has been based upon Hognestad's (1956) relationship. Hence, a means of obtaining the arching
moment of resistance for a slab strip with finite lateral restraint had been achieved Howe\.er,
both the length of the equivalent rigidly restrained strip and the contact area are dependent upon
the degree of lateral restraint and this method requires an iterative process to determine a
constant value of dl To save mathematical iterations Rankin and Long (1997) suggested that it
is usually sufficient to approxin~ate'A' as half the depth of section available for arching, d l ,
times the width

4.4.3 Taylor, Rankin and Cleland's approach

The British Standard, BS5-100, assumes one value for the ultimate strain and describes the
parabolic stress block by an equivalent stress block equal to 0 . 9 (Fig
~ 10). It has been proven
that the value of ultimate strain changes with concrete strength and a report by the Concrete
Society (1999) proposed a series of stress-strain relationships for concrete with varying
compressive strength up to 1 1 5 ~ 1 r n m The
~ . stress-strain relationship was used t o develop an
equivalent rectangular stress block Taylor et a1 (2001a) have outlined the adaptation to the
n-~aterial properties to incorporate both normal strength concrete WSC) and high strength
concrete (HSC) The variation in the ultimate strain can be represented by the following
equation
E,, = 0.0043 - [ &, - 60)'2.5 x I o - ~ ) ] [Eqn.24]
The depth of the stress block is also given by:
p = 1 - 0.O03fcu
where p is the depth of the equivalent rectangular stress block as a fraction of the depth of the
neutral axis from the compression face, x (Fig. lo).

'+
proposed
BSj4OO
stress stress
block block
Fig.10 : BS5400 conditions at ultimate flextrral load for SSC (fr,< 6 0 ~ 1 r n r n ~ )
2nd proposed stress block
A YIIICJC fo C O ~ I J ; I I ' C . S . . T ~nI ~wC n ~ h m mn c f r o r l i i l br.i&e deck slabs 19

The method outlined in the previous section predicts the enhanced flexural capacit), of laterally
restrained slab strips but is not directly applicable to the punchins strength of bridge deck slabs
A means of assessing the degree of lateral restraint inherent in such a system is critical to the
prediction and in the past this has not been achieved satisfkctorily. It has been established that
the ~bidthof the edge beams had a significant influence on the ultimate strensh of the slabs
tested by the authors (Taylor et al, 2001b) This has been clearly demonstrated in (Fig 5) By
equating the bending of the wall support to a hypothetical spring of an equivalent stiff~iessin the
analysis of the strip tests by Lahlou and Waldron (1995), accurate predictions of the failure
loads were achieved It seemed reasonable to employ a similar hypothesis for the bridge deck
slabs The refinement proposed by Taylor (2000) used a model restraint system where the
supporting edge beams, end diaphragms and surrounding area of unloaded slab were equated to
a spring of an equivalent stiffness (Fig. 9).

LL
deck slab
b = effective nidth e\teriial restramt
5 loaded slab -

-
b
\
supporting edge
bean nidt h

Fig.9: Restraint model proposed by Taylor (2000)

It \&asestimated that the influence of the arching force was sufficiently low at a distance equal
to the effective span plus the depth of the slab (LC+ h) from the face of the support This
observation agreed with others' findings, such as those by Fang et a1 (1994) Initially the value
of external lateral restraint stiffness was used in an elastic-plastic method for a two-way
spanning slab with the elastic and plastic moment factors ascertained from a finite element
analysis and yield line analysis respectively. The method, as outlined in Chapter 7 of Taylor's
thesis (2000) was cunlberson~eand takes considerable effort A simplification to the method to
facilitate its use by designers m-as considered to be of primary importance
A hndamental simplification in the preceding analytical approach was the assessment of the
degree of lateral restraint by a 'restraint model'. This gave reasonable predictions for the
strength of the bridge deck nlodels tested and the restraint model provided an adequate basis for
a simplified approach A t)lpical bridge deck restraint model is illustrated in Fig 9 An effectiire
~vidthof slab subjected to arching forces can be described by
b e e c, + 2.L, + 2h (where LS=L/2- c,/2) [Eqn.l8]
This corresponds to a slab axial stiffness of

The edge beam equates to a spring of equivalent stiffness and has an equivalent axial area
described by (as derived in Appendix 7.1 of Taylor's thesis).
3
Ab=.r.Le.13b/ beff
= constant
for support conditions of edge beam
= 550 (most bridges) [Eqn.20]
xvhich equates to an axial stiffness o f

If the area outside the effective width acts in parallel to the end diaphragms in resisting the
outward arching thrust, the areas are cumulative and can be summed to give a total effective
area, Ad, where:
Ad= area of diaphragms + area of slab outside the effective width

However the stiffi~essof the edge beams act in parallel, as opposed to additionally, to the
diaphragms and it is the flexibilities of the system which accumulate to give an overall lesser
restraint than each component as given by:

or K,= 1/(11Kb + 11 Kd ) [Eqn.23]

Hence the ratio of restraint stiffness to slab stiffness is calculated by (K,/
Ks) derived from the
previous equations By use of the restraint model, Rankin and Long's (1997) integrated
procedure was then directly applicable. However, the material properties and tests used to
validate this theory were based upon concrete with cylinder compressive strengths up to
70/mm2 A modification to the stress-strain relationship to incorporate concretes with high
compressive strengths (>70/mrn2) was considered.
A refinement to determining the shear punching capacity of a slab with less than infinite
restraint was also established by utilising the arching capacity determined from the 'restraint
model' method Similar to Kirkpatrick et al's (1934) approach, an 'equi\ra1ent7 area of
reinforcement was estimated by using

This quantified the shear punching strength in terms of the equivalent area of reinforcement due
to the combined effects of bending and arching i.e.:
0.43
Pp, = --vE.(criticd perimeter).d(lOOpt)a'I [Eqn.27]
5
where rf, shape factor = 1.0 for a circular wheel load and critical perimeter is taken at 0.5d from
the face of the loaded area. For con~parisonwith test results usins a rectangular patch load rf =
1 15. For pneumatic tyred vehicles, it is unlikely that stress concentrations would occur,
therefore rr = 1 0 would be satisfactory
The tno-phase approach predicts an ultimate strength awarding to the lesser of the flexural and
shear punching mode predictions i e

The following flow chart illustrates the steps to be taken in the calculation procedure for the
flexural and shear punching strengths of laterally restrained slabs.

1. Effective mid th of loaded sEab 1

1 3. Bending Capacity I

4. Arching Section
I

1 5. Equivalent rigidly restmined slab

Repeat u n t ~ l
constant
, \ d u e of a

1 9. Arching Capacity

1 11. Shear puncfling capacity 1

I
/ 12. Ultimate capacity
The subsequent equations present the detail of the procedure proposed by Taylor (2000) for
predicting the ultimate strengths of a laterally restrained bridge deck slabs,
Effective width of loaded slab:

Stiffness parameters :
E, = 4 23fc,,"
K, = E,A,l/L,
IXb=~ ~ ~ or1transformed
1 2 I-beam
< <
&= L?I,d be$ where = constant for support condition
< =114 5 (SIS) or 985 (FE)
Kb = Ab EJ Le
.Ad= area of diaphragms + area of slab outside the effective width
Kd =C &.El L,
KT= l/(l/Kb+ I/&))
K, = E, h.b,&e

3. Bending Capacity :
of stress block, P = 1 - O.OO3& but < 0.9
-Depth
uepth of neutral axis, x = J . As
0 67.fcuPb
Lever arm, z = d - 0 5P x
Mb = A. A,.z
Pb = kb.Mb
4. Arching section
2.dl = h - 2x.P
New d l from previous iteration

5. Afine strip

A = a.b.dl and L, =L,

6. Arching parameters
E,L;
E~ = 0 60) 2 5 x lo-'] but < 0 0043 and R=-
0043 - [(f,,,-
4 d?;
Ecz2Eu(1 - p)
7, Deformation
R>O26 u = 0 3 1 (constant)

8. Contact depth
a = 1 - u/2
2 a d 1 use for refined arching section above until value remains constant
9. Arelming capacity

-
Xi, 0 168.b,fc,,dl2 h/l, (LJL)
[For maximum arching L, = L, 3 Ma,= 0 16S.fc,, dl' hl, ]

10. Flexural prrnchirtg capacity

Pa + Pb

1 1 . Shear punching capacity

critical perimeter at 0.5d fi-om face of loaded area

12. IJltimate capacity

A worked example for the proposed method is given in the Appendix. A spreadsheet I\-as
adapted for this procedure and has been used to compare the predicted values with several tests
carried out on bridge deck models as shown ill Tables 3, 4, & 5 The method has also been
compared with Kirkpatrick's earlier method in Tables 3, 4, & 5 In Table 4 it can be seen that
the flexural failures exhibited due to the low external restraint stiffness were not accounted for
in Kirkpatrick et al's n~ethod

4.3 Bark's method

Park has been one of the major contributors to research into compressive me~nbraneand uses a
plastic theory for the load-deflection behaviour of a restrained strip at ultimate load The
following procedure is outlined in Chapter 12 of Park and Gamble's book (1980). A fixed-end
strip with plastic hinges is depicted in Fig.1 1 with partial lateral restraint allowing an outward
displacement of 't' at each end The section of the slab strip between the hinges is assumed to
remain rigid.

Fig.11: Park's plastic hinges for a restrained slab strip

.4 gl/ide to compressive mt?rnbrarvi?actiolr in bridge deck slabs 24

Park also assumes that the sum of the strains have a constant value, E, along the length of the
strip Referring to Fig. 12, the shortening in the middle section is, ~(1-2B)L,giving a total lensth
of portion I -2 as shown.

Fig.12 :Park's geometry of slab strip portion between yield line

sections 1 - 2 above

From equilibrium of forces, the following expressions for the membrane force and the moment
of resistance were developed

The sum of the moments at the yield section about the mid-depth was expressed as:

By considering strain con~patibility,equilibrium at the section and by equating the work done at
the yield sections with the ~ i r t u a lwork done by the load, Park obtained an expression for the
load, dependent upon the central deflection, 6 The terms required to develop this expression
are lengthy and the procedure has been simplified by means of a spreadsheet. An example of
slab strip S5 tested by Taylor et a1 (2001a) is summarised below and the comparison to the full
test series is given in Table 2 The following variables are considered.

1 .Stiffness parameters:
Concrete elastic modulus, E, =4 23 x 10 1 l o = 42 531<r\Umm2

Slab stiffness, K, = 42.54 x 475 x l50/7 12.5 = 4254kNlmm

External restraint stiffness, K, = 197hT/1nm(measured)

K,= 4 11.74N/mrn
per mm width
2. Forces:
From ACT code C, = 0 85f c p 1 c where Dl= 0 85 for f ,' <30Nimm2 and
reduces by 0 008 for each unit increase but > 0 65
p,= 0 85 -[(SO 9 -30)0.008]= 0 62 ... p1= 0 65
For unit uidth. As=O 7lmm?/mrn ST = 0 71 u 500 = 356 S4N/mm
For consewative approach assume T = C, , :. T = T' = C', = 3 56 83Nlrn1m

3. Strain value a ~ l d!eve[ of neeitrd axis

Assume 6 = 0 311 = 5 0 m l for example [note 'D' used for '6'in spreadsheet]

TI-T - C', +CL

2 4 46 1.7fvcP,

11,,6 = (C, + C, -T)6 =OSSfc p l c 6 = 0 8 5 v SO9 v 0 6 5 ..: 3 4 7 7 .r

50=77691N mm/mm
n1,, = 0 S5fc Pic (0 5h-0 5Plc) + C,(O 5h-d') +T(d-0 5h)
= 0 85 x 80 9 x 0 65 x 34 77(O 5x150 - 0 5x0 65x34 77) + 356 S4(O 5x150 -
16) + 356 84(104- 0 5x150) = 1 19677N m d r n m (from symmetry
m,,=m',,)
4. Energy prinacipal
:. (PI2 x L/2.)8 = (mu + m', - nu618
(P/L)/4 = 119677 + 119677 - 77694 = 161660Nmm per unit width
= 76.79 KN.m for slab strip
3 P = 76.79 x 411.415 = 217.lk;V (at 6 = 50rnm)

The procedure is repeated at various values of deflection to find the peak load. The
-
graph of load versus deflection is then plotted (from the spreadsheet in the
Appendix the peak load was found to be 23 1 9kN) Thus;

Ratio test/predicted load = 192/231.9 =0.83

A gzude to compre.ssive rnembrme nctiors in bridge deck slabs 26

5.0 Cancluding remarks

From the review of literature and from the summary of the proposed methods for
predicting the strength of laterally restrained slabs, a primary conclusion to be drawn
is that bridge deck slabs have strengths far in excess of those predicted by
conventional design methods which are based upon conventional flexural theory and
empirical shear strength equations for unrestrained slabs. The use of Pucher charts
or Westergaard's equations, as recommended by BS5400, underestimated the
ultimate strengh of laterally restrained bridge deck slabs as they take no account of
the in-plane forces set-up as a result of Compressive Membrane Action (CMA).
CMA has a more significant effect on the strength of laterally restrained slabs with
low span to depth ratios. As the span to depth ratio of bridge deck slabs is normally
less than fifteen, the strength of these slabs is considerably enhanced CMA is also
strongly influenced by the degree of lateral restraint. Lateral restraint is inherent in
bridge construction in the form of edge beams, end diaphragms and the surrounding
area of unloaded slab. It has been demonstrated that the strength of the laterally
restrained slabs also increases with increasing concrete compressive strength This
gave a higher degree of conservatism in the predicted strensths of the high strength
laterally restrained slab strips, using standard flexural theory. Consequently, the
utilisation of the arching phenomenon in high strength concrete bridge decks has the
potential for producing substantial economies in beam-and-slab bridge decks.
The methods based on arching theory gave more accurate predictions for the strength
of a wide range of laterally restrained bridge deck type slabs compared to the current
codes The Queen's University of Belfast approach gave more consistent and
slightly conservative predictions compared to the highly conservative predictions
using the code The plasticity approach developed by Park at the University of
Canterbury, Xew Zealand, also gave more accurate predictions than the current code
However, the predictions were slightly unsafe compared to the ultimate strength of
slab strips tested at Queen's University (Table 2)
The drafi Departmental Standard, BD81101, has partly addressed the use of CMA for
the design and assessment of bridge deck slabs in the UK; although it does not
address some of the limitations to its use; such as the boundary restraint and
particularly its requirement for intermediate diaphragms. The draft BD8l is based
on the Queen's University of Belfast approach which has been validated by hll-
scale testing of real bridge deck slabs. Future in~provementsin this approach could
be that the limitations may be reduced if flexural punching failure is taken into
account The approach would then encompass punching in slabs with a relatively
low degree of external lateral restraint.
One of the barriers to the use of compressive membrane action has been the
complexity of the calculations required. This report has addressed this issue The
second barrier is the accurate estimation of the effective stiffness of the external
restraint in real slab systems, such as bridge deck slabs. Based upon tests on
reinforced concrete bridge deck models, a method for evaluating the level of ex7ernal
restraint stiffness was established. However, the method has been validated by only
a small number of tests and krther tests on bridge deck slabs with varying geometry
and on steel composite bridge deck models may further endorse this method
Recognition of CMA in bridge deck slabs through the adoption of the proposed
BD81 could lead to many innovations in the way in which future deck slabs can be
reinforced This should lead to less reinforcement and increased cover, giving more
econon~icaland durable concrete bridge deck construction
Table I : Categorisatton of methods for predicting the e~rlzanced strength of
restrained slabs

Category Authors Year

H;and-cakalable Ockleston 1956
elastic-plastic Brotchie 1963
Liebenburg 1966
Christiansen and Frederikson 1963 & 1983
McDowell, McKee and Sevin 1956
Long, Rankin, Kirkpatrick, Taylor & 1975-present
Cleland (Queen's)

Wood
hlorley et a1 (Cambridge)
Braestrup
Kemp and Eyre (City College) 1967-1993
Park 1960-3980

Csnrperter based Cope, Rao and Clark

elmtic-plastic Morley et a1 (Cambridge)
(NLFEA)
Jackson 1987- 1990
Bakht, Jaeger, Mufti, Agarwal and 1980-1998
Thorburn (Canada)
Lahlou and Waldron 1992
Fang et a1 1990 & 1994
Azad et a1 1993 & 1993
Petrou and Perdikaris 1996
 membrme nctior~irz bridge deck slabs
.4 gttrde fo conrpressi~~e 28

Table 2: Comparison of ultimate strengths of one-way slabs tested at Queen's

University (Taylor et A, 20012) with those predicted by British design
code, proposed Queen's method and Psrk's method

Model Reinforce- JE, PT Pns, PtRm P, P-rRp ?park PtPp,,~

No. ment (~lmm') (W) (I&) (kN) O
m
05%T&B 112

0 5%T&B $0 8

0 j%T&B 61 5

0.5%T&B 82 2

C).5%T&B 101 1
05%B 318

0 5% B 91.0

0 5%B 100 1

0 5%T&B 89 3

none 90 5

0 5% centre 96 8

Oj%T&B 1010

fibres only 104.9

0.5%T&B 39.5

0.5%T&B 60.9

Average = 2.03

* '514 span loading

T = top layer of reinforce~nenr
R = bolt0111 layer of reinforcement
F/E= fixed end
LR- longitudinally restrained
SIS- simply supported
=I! 17 mil1 Load. PkN

I
I
i
1125mm clear span 91I
Table 3: Comparison of d t i m a t e strengths of bridge deck s h b s tested c2t Queen's
Urriversity (Taylor et al, ZOOlb) with those predicted by British
Standard, proposed Queen's method a n d Kirkpatrick's method

D3 65 00 - 00 - 466 139 -
D74 65 00 - 00 - 43 6 148 - -

D5 150 38 0 3 91 83 4 1 SO 151 1 0 99 268 0 0 56

D6 182 38 1 3 78 84 4 2 16 173 3 1 05 276 1 0 66
D7 135 381 354 851 159 925 146 2809 0 48
D8 157 387 40_5 847 1 1489 1 057740
A\.erage 4-02 1.86 1.19 0.57
St. deviation 0.20 0.06
Coeff. Vsr. 0.168 0.101

25x500 loading plate

Tests by Taylor et al(2001b)

 Table 4: Comparison of test results by Kirkpatrick with
predictions using two methods
Test Panel Pt P, Pt/Pp Pkrk Bt/Pihh
No.
A1 102.5
A2 114 0
A3 102 5
A4 108.0
A5 96.0
B1 108 0
B2 96.0
B3 100.0
B4 90.0
B5 91.0
Cl 106 0
C2 100.0
C3 102.0
C4 1100
C5 105.0
Dl 98 0
D2 108 0
D4 118 0
D5 107.5
Average
St. deviatioa
Coeff. Var.

Beams 1 50mmx450mm deep

Tests by IGrkpatrick, 1884

.4 ailide to con~ures.si~.e
rnenlhrmre ncfiorz irr bridge deck slabs 31

Table 5: Comparison af test results by Bafchelar et a! with

predictions using two methods
R'lodel NO. Pt Bp Pt/pp P ~ r h Ptm~rk

5 A3
5B3
5 c;
7C1
SCl
5A4
5B4
5C4

A\.erage 2.04 0.89

St. deviation 0.11 0.16
Cocff. Var. 0.012 0.177

L=2Xrum
Channel 3 76OmmcIc
b,:~3O?mm
loading shoe 95 s 631nm elhpse
(equlr alelit circular drameter of [(63+95)/2 + d] used for critlcal perimeter)

Tests by Batchelor et al, 19'78

-4 gurcle to compres.wse r~re~?ihrme
action 11.1 reinforced concrete bridge decks 32

Example for predicted strength of a typical bridge deck slab.

Assuming: - 350mm x 750mm rectangular supporting beams
- bridge span = 13m (has negligible influence on the slab design)

.. :. equates to
.:. ..: / 7
2.0111

of width, b S ~

Fig.Al: Typical bridge deck slab

EXAR~PLE1 : Slab variables:
d = lOOmm Le= 850mm
berY= 2300mm (as below) f, = 460~lmrn~
h = 150mm , = 40~imn1~
L = 2000mm A,= 1.0%=2300m2
Note: clear. spa^ taken as 2. On?)

2. Effective width

2. Stiffness parameters :
3 . Bendir~gCapacity :
Depth of stress block. = 1 - (0 003 x 40) = 0 88
Depth of neutral axis, x = 460 x 2300 = 19.51mm
067~40xO8Sx2300
Leverarm, z = 1 0 0 - ( 0 . 5 ~0 8 8 1~9 5 l ) = 91 4 3 n m
hlb = 360 s 2300 s 91 42 s lo-" 96 72kN m
For fixed end Pb = ( 8 L ) Mb = 386.9kN

4. Arching section
2.dl = 150 - (2 x 19.51 x 0.88) 3 dl = 57 84

5. .&Emstrip
Assume a.=I for first iteration
A = 57.84 x 2300 = 133032mm2

6. Arching parameters
E, = 0 0043 - [(40-60) 3 . 5 x lo-'] = 0.0048 < 0 0043? use 0 0043
r, = 2 x 0.0033 ( I - 0.88) = 0.00103

7. Deformation

8. Coratact depth
a = 1 - 0.27112 = 0.865 3 ad1 = 50.0

Use for refined archin,o section and steps 3 - 7 repeated give

A = 1 15000mm2
= 1588mm
E, = 0 0048 > 0 0043 use 0.0043

E, = 0.0013

R = 0 1935
O<K<O26=> u=O256
a = 1 - 0 256'2 = 0 872
ad1 = 5 0 4
Cse for refined arching section and steps 4 - 7 repeated @\re
 .4 guide to cornpre.~si~.e
17ienrbrcmenctiori in reirgorced concrete bridge & c b 34

A = 1 l592Ornrn2
L, = 1592mm
E, = 0 0048 > 0 0043 use 0 0043

E,= 0 0013
X = 0 1955
0 < R < 0 2 6 = u=0256
a = 1 - 0 25612 = 0 872
a ad1 = 50.4 same value as previous hence use these values

9. Arching capacity
O < R < 0 . 2 6 h4 =4.3-16,1,/3.3x10~'+0.1243x0.1955 = I 7 7 3 3

10. Flexural punching capacity

PPf=386.9 + 97.9 = 484 SkN

1 1 . Shear punching capacity

13 Ultimate capacity
If PPf < P,,. 3 P, = P,r
If P,f > P,, =3 P, = P,,

2 P, = 414.5W (shear type faiture)

EXAMPLE 2 : Slab variables: as Example I except As = 0.5% = 1150mm2

1 . Effective width (as before)

2. Stiffness parameters : (as before)

3, Bending Capacity :
Depth of stress block, P = I - (0.003 x 30) = 0 88
Depth of neutral axis, x = 460 x 1150 = 9 75mm
O67x40xO8Sx3300
Lever arm, z = 100 - (0.5 x 0.88 x 9.75) = 95.7mm

For fixed end Pb = (SIL).Mh = 202.5kN

4. Arching section
2 d l = 150 - (2 x 9.75 x O.S8) 3 d l = 66.3

5. Affine strip
Assume a.=1 ibr first iteration
A = 66.4 x 2300 = 1 52720mrn2

6. Arching parameters
E,, = 0 0033 - [(40-60).2 5 x = 0 0048 < 0 "0",3? use 0 0033
E? = 2 x 0.0043 (1 - 0.88) = 0.00103

7. Deformation

8. C o n t w t depth
a = 1 - 0 23712 = 0.882 z a d 1 = 58 6
Use for refined arching section and steps 4 - 7 repeated give'

A = 134780mm2
L = 1662mm
E, = 0.0048 > 0 0043 use 0 0043
E, = 0.00103

R = O 16155
O < R < O 2 6 = > u=0.225
cx = 1 - 0 22512 = 0.888
3 a d 1 = 59 0
Use for refined arching section and steps 4 - 7 repeated give
A = 135700mm2
= 1665mm
E,, = 0.0048 > 0.0043 use 0 0043

r, = 0.00103
R = O 162
O<R<0.26= u=O225
a = I - 0 22512 = 0 888
=i a d 1 = 59 0 same value as previous hence use these values

9. Arching capacity
0 < R < 0.26 M r = 4 3 - l 6 l & . 3 x l 0 ~ ' +0.1243x0.162 = 1.9958

Ma = 0.168 x 2300 x 40 x 66.42 s 1 9958 (85011665) x lo-' = 69 5kN m

P, =(4/L).M,= 139 OkN

10. Flexural punching capacity

PPf=202.5 + 139.0 = 341.5kN

11. Shear punching capacity

12. Ultimate caprtcity

If PPf < P,,. =3 P, = P,f
If P,f > P,,. z3 P, = P,,

The examples given can be compared to serviceability tests carried out on a bridge with similar
diinensions in Northern Ireland. The test panel with an average of 4 2 ~ l m mconcrete ~
compressive strength and 0.6% top and bottom reinforcement had a 3SOkN test load applied
u-ith no adverse effects and a maximum midspan deflection of less than (spad700).
The two examples highlight the effect of reducing the percentage of reinforcement on a
typical bridge deck type slab. The proposed method predicts a flexural type failure in the
lightly reinforced slab compared to a shear type failure in the slab with 1% reinforcement
The method also predicts a decrease in the ultimate failure Ioad of only 17% for a reduction in
reinforcement of 50%.
Bibliography
ACT Committee 3 1 S (ACI3 1 8-89)! Rrtrll'rr~g code rmpri~cnwiit.Jiw rc.rrlfol-ced coilcretc trl~~J
comntetltmg, Anierican Concrete Institute, Detroit, IJSA, 1989
ACI Special Publication, SP3O- I4 Cr.nckrrrg, Dc?fkctrotl LIFTLZ' II/trrmlc ko~7dof CotIerefe dub
s j ~ t e n ~ sAnierican
. Concrete Institute, hlichigan, USA, 197 1
ACI, FIP,SINTEF (1993) Higl~-rtrwgtl~ corrcrete, International qmposium, Lillehammer.
Norway, (Editors- Holand, I and Sellevold, E), 1993
MP,AFREM,ENPC (1996) Tl7e Utrlizatiorz of high s t r ~ ~ g rhigh
l l ye~firrtmrlce colicrefe,
Fourth International symposium, Paris, France, 1996
American Association for of State Highway and Transport Officials (MSHTO. 1983) S~C~LJL~IZJ
hrgliit~ajlh~.idyi..s Washington, DC, 13"' Edition, 1983
rpecficatro~~for
Azad, A.K , Baluk, h4 H , Al-mandi, M. Sharif, A M and Kareem, K (1993) Loss of yrnlchirtg
cnpacigl o f bridge u'ak slabs from crack dmmge , AACI Structural Journal, Vol 90, No 1,
1993, pp37-41
h a d , A K , Baluk, h/I H , Ababasi, M S A and Kareem, K , (1993) Pu~icl~iirgccpacigqof deck
slabs ul gilder dab bi.idge,s ACI Structural Journal, Vo1 91, No 6, 1994, pp656-662
Bakht, B and Jaeger, L G (1992), Ulfrmnfeload tests or] slabs orz girder budges, ASCE
Journal of Structural Engineering, Vol 1 18, No6, 1992, pp 1608-1624
Bakht, B and Aganval, A C (1995), Deck ~lcrbsqf ske~clgrrder hrru'ges, Canadian Journal of
Civil Engineering, KO 22, 1995, pp5 14- 523
Bakht, B and Agarwal, A C (1 996) Pzti~chiirgsheor-farlines in curmete decks srlap rl71.01igl1
112 rrnbrl~ty ASCE Journal of Structural Engineering 1996

Bakht, B Jaeger, L G and hiufti, A A (199S), A'otes 017 .steel.fi.ee deck slrbs, Institution of
Civil Engineers (LK) Seminar, July, 1998
Batchelor, BdeV. and Tissington, I R(1976), Sl7ear str-er~glhof hc1o-lr8qh d g e slabs, ASCE
Journal of Structural Engineering, 1976
Batchelor, BdeV, He~vitt,B E , Csagoly P. and Holovvka, hT. (1978), Imvstigariou o f r l ~ citltmnte
sti.ertgt11 cf deck ~ l d ofs corry~osrleSteel Cmlcrete hixlgcs, TRRGG3, 1978, pp162-I 70
Batchelor, BdeV , Hewitt, B E and Csagoly P. (1978), A7 irnvstigntlor~of flzefatrgrte sfrmgtt'7
cf deck slabs. of con~positesteel comrefe bridge decks, Transport Research Bridge Engineering
Conference, Canada, 1978, pp153-161
BDl5192, Departmental Standard, Geuer-al yl-r~mylesfor the dcs~gnof and col7strr~tror-Iof
hrru'ges. Use of BSWOO Part 1. 1988, Highways Agency, December, 1992
BD2492, Departmental Standard, The u'es&n ofcollcrete hrglntny brrdges m d Stmcfwes. Use of
BS5100 Part 4 1990, Highways Agency, December, 1992
BD37101, Departmental Standard, Lo& .for Hrginr>qiBridges, (Supersedes BS5400 Part 2)
Highways Agency, August. 200 1
BD34I95, Departmental Standard, 7he crssessmwt of co~ic~vte Department of
l ~ r g h r qhri~iges,
*
Transport, Highway and Traffic, January, 1995
BDS 1I0 1 (draft), Departmental Standard, IJsc uf' Con?yressr~~e
A4e171b1.aneActioil 111 bridge decks,
Highways Agency, Nov 200 1
m ~ dY. ~A S~C Ei Journal of
Braestrup, h/l W (1 9SO), Dornc {ffect 111 I?(' .slobs: rrg~d-plart~c
Structural Engineering, Vol 106, No ST6, June 1980, pp1237-1253.
Braestrup, M.W and hlorley, C T (1980), Dome e@ct rrz RC slnbs: elnstic-plnslic nlmlysis,
ASCE Journal of Structural Engineering, Vol 106, No ST6, June 1980, pp1255-1262
British Standards Institute, BS 5400: Poi,/ 2, Britrsh Sim~7'nrd.forflre d e ~ i p rof .sfeel, C O I I C ~ C T ~
cmJ contyosife bridges, London, 1978.
British Standards Institute, BS 5400: Pmi 4, Brifrsh S f ~ ~ r d afor
r d [he design of steel, c o r u e f e cil~J
corrtposrle bridges, London, 1978 and 1990.
Brotchie, J.F. (1963 ), A i.ejcirted tl~eorji.for. .slubs, Journal of the Institution of Engineers,
Australia, Vol.3 5, 1963, pp292-296.
Christiensen, K.P. (1963) 7%e eflect of men~brarlestresses or2 the ultinmfe sfrerrgfli of nrterior
panel of remforced corlcrete slab Structural Engineer Vo141 No 8 Aug 1963 pp261-265
Christiansen, K.P. (1 983) Ulfirmfestrerrgth of coricrefe dabs ~ t ' i fI.7ori;orifal
h resfranrf.s, Danish
Society of Structural Ensineers, Vo1.53, No.;, 1983, pp61-86
Christiansen, K.P and Frederiksen,V.T. (1983), Experimerlfal iinlestigafiot~of recfarrgular
cor~cretedabs ~c'rtl?horimifal restraitlls, Materials and structures, Vol 16, May 1983, pp 179-
192
Concrete Society (1999), D e s i g ~Guidcmce .for. high strergfh colmefe, Technical Report No
TR49, Report for the Concrete Society Working Party, DETR
Cope, R.J and Rao, P.V. (1 977), Nn~~-lit~enr$~zife
elemerzf ardysis of slab sir-ltcfrrres, ICE
proceedings, Part 2, No.43, March 1977, pp 159-179
Cope, R J , Rao, P V and Clark, L. (1 979), A trorr-litlenr.d e s i g ~of cor~crefebridge slabr r~sirlg
)trite elcmenf pr.ocrd/m,s CSCE-ASCE Internationd symposiun~,University of Waterloo,
Ontario, Canada, 1979, pp379-405
Department of the Environment (Ni) Desig? Spec!ficafrorz .for Brrdge Decks, DOE, Northern
Ireland, 1990.
Eyre,J R and Kemp, K 0 (1984j, .4 graphical solutio~ipredictrr7g rl-rcrease rrl sfrwigrli uf
.slab,s Magazine of Concrete Research, Vo1.35, No. 124, Sep 1983, pp15 1-156.
Eyre, J R. (1990), Flow rule it1 elastrcally restrairtedom way sparrrtit7g slabs, ASCE Journal of
Structural Engineering , Vol.116, No. 12, Dec 1990, pp32.51-3267.
Eyre,J R and Kemp, K . 0 (1991), Iri-plnrie sr(ffi7es.s qf remfor-ced coricrete slabs under
contpmsi\v n m ~ b ~ m crcf~or~,
re Magazine of Concrete Research, 1993, pp67-77.
Famiyesin, 0.R. and Hossain K.M A. (1 W8), Opirnised desigt7 charts for firlly restrair~edslabs
by FE predictions, ASCE Journal of Structural Engineering, Vo1.124, No.5, May 1998, pp560-
569.
Fang, I.K., Worley, J and Burns, N H (1 990), Reha\~iowof rsofi.oyic R C bridge decks on .sfeel
g~rders,ASCE Journal of Structural Engineering, Vol. 116, No.3, March 1990, pp659-677.
Fang, 1K , Lee, J H and Chen C R. (1994), Behmiozir- ofpartially resfrnirred comrete slabs
rnru%rc o r t c e ~ 7 f r ad d ACI Structural Journal, V 91, NO 2, Mar 1994, pp133-139.
Guice, L K and Rhomberg, E.J. (1988), A4erubr~~r1e acfior?ir? pmficrlb~resfrarrred slob, ACI
Structural Journal, Title No 85-534, July 1988, pp365-373.
Gvodzev, A A. (1 936 & 1960) 7he De ferrnitrafiortof flw due of the collapse load for statically
irldeterwirmte .sjsfeni.s ~cudergoirlg ylnsfic defov~natior? (Translation from 1936 paper)
International Journal of Mechanical sciences, Vol. 1, 1960, pp 322-335.
E & F N Spon. London, UK
Hambly, E C. (1989), BmJge deck behm)lozir, ZradE~i~tror~,
Hewitt, B .Eand Batchelor, BdeV. (1 975), Pradi.irrg sltenr. sfr-e~gll?
qf resfrnirmi .slrbs, ASCE
Journal of Structural Engineering, Vol. 101, ST9, 1975, pplS37-1853
r.eirlfor.ct.J colrcrcfe 117 BSjJOO, Structural Engineer,
Jackson, P A (1987), The .s.r.ess l~rr~rfr,for,
Vol65a, No 7, July 1987, pp259-263.
Jackson, P A (19891, CTon~yres.sr\z
n~en~brn~re
rctlort IH brld,o &ck sl~lh.~,
PhD thesis, Uni\rersity
South West and BCA, 1989.
o f br.r&e deck sIczb,s Structural Engineer.
Jackson, P A (1990), The glohnl ard loccrl hehrn*ro~~r.
Vol 68, No 6, March 1990, pp I 12- 1 16
Jackson, P A and Cope, R S (1990a), 717e bel~m~io~tr
of br-rdge decks rtrlder fill1 glohnl load
Span Bridge '90 Conference, 1990, pp253-264
Jackson, P A and Cope, R S (1 990b), Sfrerrgfl~
~rscess;rrrerif
ntc.rlrodr$~-corrcrele brrdges Bridge
hlanagement Seminar, Ch 67, 1990, pp329-438
Jackson, P A. (1993), Sfrwgih assessmer~ffo keep brrdges 111 .sen7rce,Bridge hlanagement 2
Seminar, Ch 67, 1993, pp717-726
Kirkpatrick, J , Long, A E and Thompson, A (1 9821, of ,111-
Load dr,sfr.rb~lfiorrcImr~ncfer.~sfrcs
hcnnt hrrdge ui.cXx.r, Structural Engineer, Vol 60b, No 2, Jun 1982, pp34-13
Kirkpatrick, J , Rankin, G I B and Long, A E (1984%) Sfretrgfh e~lal~ctrfrorz
of hf-benit? brrdqe
~leck.~,
Structural Engineer, Vol 62b, No 3, Sep 1984, pp60-68
Kirkpatrick, J., Long, A E and Thompson, A (1984b), Load drsfr-ihiifronclmrncter-7~frcs
qf
sycrcedhd-beam hrrdge decks, Structural Engineer, Vol 62b, No 3. Dec 1984, pp86-88.
Kirkpatrick, J , Long, .A E and Rankin G I B. (1986), l7?e ~nflltemeofcon~pressn~c rnembr.me
aclior~or1 sen~rcenh~lrty
of heant m d d r h hr~dgedecks, Structural Engineer, Vo164b, No 1, Mar
1986, pp6-9
Kuang, J S. and hiorley, C T. (1992), P ~ m c l ~ mshew
g h e l m ~ o qf
~ ~i.estr.nrfled
r corlc. s1~rb.s(-
diirci~snorr),ACI Structural Journal, Val 89, No I . Jan 1992, pp 13-18
Kuang, J S and Morley, C T. (19931, A ylastrcrp modelfor the yrrrlchmg shem of Iafertri/~'
s e ~ f r m m dcorlrele slah.r ~trifhcon~yressrventernhr-nrw actroil, International Journal of Science,
Vol 35, No 5, 1993, pp371-385 (Discussion , Vol35, No 10, 1993 p 889)
Laldou, E W and Q7aldron, P (1992), Aternhrzrrre c~cfrorrrrl o r r c - ~ i qspcunlr~g
~ strrpr, ICE
Proceedings (Structures and Buildings), No 94. Nov 1992, pp4 19-428
Liebenberg, A C (1966) Arch nctrorl in corlcrefe s l ~ ~ (Pmfs
hs I R- 2), South Ahcan Council for
Scientific and Industrial Research, Report No 234, 1966.
Long, A E., Kirkpatrick, J and Rankin, G I B (198S), Ecorlon~rcdevg~rqf rpaced Ad-l-bemn
bridge decks, Technical Report Queen's University of Belfast, 1988
Long, A.E. and Rankin, G I.B. (1989), of RC s f r v c f ~ n - a ,
Real sfrw1gfi7a i d robr~strm~s
Conservation of engineering structures, Thomas Telford, London, 1989, pp47-58.
Long, A E , Rankin, G I B , Gilbert, S.G , Niblock, R A,, Skates,A and Ruddle, h4 (1WO),
S'frwgth uf cell~~krr.
corrcrcfe .rtri~t~tres, Concrete Offshore in the Yineties, Offshore
Technology Report, Department of Energy, UK, 1990, pp37-49
Long, A E , Kirkpatrick, J. and Rankin, G.1 B (1 9 9 9 , E~rt~arlcing ilrflitence of cony.wessi\v
n7embrnrle actiorl irr bridge decks, Bridge Modification Conference Proceedings, Ch 2 1, 1995,
pp2 1 7-227.
-4guide to cot?lpressi~*e action
~ne~~ihm n e III hr~dgedeck slnbs 40

Masterton, D M and Long, A E (1974), n7e yztrzchr~?gsfrrrzgth o f ~ l n b sa .flexrrrzr/ npproncl?

rrrnrg.finrte elenler~ts, ACT Special publication SP-12 Shear in reinforced concrete, V 2, Part 3 ,
ACI, 1974, pp747-768
McDowell, E L , MvKee, K E and Sevin,E (1956), Arclirr~gnclmz theory of resir~~~lzed n7asoi1ry
~ r d l s ,ASCE Journal of Structural Engineering, VolS2, NO ST2, 1956, pp9 1 5- 1- 9 15- I8
Morley, C T (1967), I'reld h7e theory .for rnerrforced concrete dabs at n~odei.nfelj~
Inrge
dcflecfrorls, Magazine of Concrete Research, Vol 19, No 6 1, Dec 1967, pp2 11-222
Morley, C T and Olonisakin, A A (1995), The Galerkin mefliod uyylred to corq~re~sr~.e
n~enzbrarleactrorl, Development in computational techniques for structural engineering, Ch 47,
1995, pp 15-20
Moussard, M M. (1993), Rorze Rrver Experrn~enfalBridge, Third International Symposium on
The Utilization of High Strength Concrete, Lillehammer, 1993, pp.562-569
Mufti, A.A., Jeager, J. and Bakht, B (1993), Experrnienfcd rnrwfrgat~onof FRC deck slabs
~t~ril~ozrt
ntferr7al r f , Canadian Journal of Civil Engineering, Vol20, Pt 3, 1993, pp39S-406
Niblock, R. (1986), Con~pressi~ie nlen~braileactlor? aild Il?e zrlfrn~atecnpacitjl o f irngfon~~ljl
loaded
r.enforced coricrefe slabs, PhD thesis, Queen's University of Belfast, Northern Ireland, 1986
OckIeston, A.J. (195.9, Load tests or1 a 3-storey liC bnr~d~rrg
I H Johannesburg, Structura!
Engineer, Vo1 33, Oct 1955, pp304-322.
H ~ g h c n yBridge Design Code (OHBD), 1979 [amended
Ontario Ministry of Transport, Or~fcrrio
1983 & 19921
Park, R. (1964a), Ultlrmrate slrergfl-,qf recfntmgulnr coxrete ~ l a h szcrrder- ultrmnte load wrfh edge
resti-arrrf, Proceedings of the ICE, Vo128, June 1964, pp 125-150
Park, R. (1964b), Ulfrntate s~erzgthnrld lormg term hehmiottr of dtintafe load 111 hrto Ituy
cormcr-eteslnbs, Magazine of Concrete Research, Vol 16, No 48, Sep 1964, pp 139-152
Park, R (1965), The Iarernl sttfitess a ~ srrwtgrh
d reqcir-ed to eiwwe nwn?brme acfror? nt
ullznmfe loud, Magazine of Concrete Research, Vol. 17, No.50, March 1965, pp29-38
Park, R and Gamble, P (198O), Rt?irlforccJ cormete slabs, John Wiley & Sons, London, UK
1980
Peel-Cross, J (1996), Co~npressrwntentbrm?e a c f ~ olri
~ ?steel cor~tpo.s~tc
slabs, hlSc thesis,
Queen's University of Belfast, 1996
Petrou, M F. and Perdikaris, P.C. (1996), P r n d i q shenr.farlure irz concrete decks as n srmap-
sfabrlrty loads, ASCE Journal of Structural Engineering, Vol 122, No.9, Sep 1996,
tl~raoirgli
~ ~ 9 91005
8-
Peel-Cross, R.J., Rankin, G.1.B , Gilbert, S.G and Long, A E , (2001) Cor.r~prcssi~~e rnernbrme
~rctioilrr7 con~posirefloorslabs Irl the Car-diizglor~LBTFr, Proceedings of the Institution of Civil
Engineers, Structures and Buildings Journal, 146, Issue 2, May 2001, pp 217 - 226.
Pucher, A. (1964), Iq7riertce szrrfrces of elastic plates, Wien, New York, 1964
Rankin, G I B (1982), Prn~chirmg.farlrnemtd corizyresm,e ~mnlbrirrrencfiorl I?! reir?for-cedcorlcrefe
slabs, P l D thesis, Queen's University of Belfast, 1982.
Rankin, G I.B and Long, A.E. (1987a), Predicting p ~ c h r ~stre~gil?
g of c o r n w ~ l r ~ m
slab-
l
colzcrr~rlspecinms (& drscussioi~I988), ICE Proceedings -Structures and Buildings, No 82,
April 1967, pp327-346
Rankin, G I B and Long, A E (19S7b), Prm'fcttrrg rl~ee.rlhtmced ylmchtrig 1freirg111
of irlferror-
corrr~ecfro17r, ICE Proceeding -Structures and RuiIdings, No 83, Dec 1987
s/~rh-colr~nir~
pp1165-1 I86
Rankin, G I B , Niblock, R A Skates A and Long, A E (1991), C'onrprt.ssr~~e nrenrhimrc sirerg117
~1rl7mcerireiif
T I T rnrrform/j~loaded restr.almd slobs, Structural Engineer, Vol 69, S o 16, Aug
1991, pp2S7-295
Rankin. G I B and Long, X E (1997), illd7n1g crcfrnil ~ f r - ~ r ~er~17~mce1~1crrf
glh 111 I ~ ~ I L ' ~ c I / / J
r.esfr.cr~rwd~ l n h ~ICE
, Proceedings -Structures and Buildinss. No 122, No\: 1997, pp46 1-167
Roberts, E H (1 967), Load c a r g ~ i gccryncriy of slab st7-rps re rfmrrled ogczrtzst lorlg~trtdrrml
euparl~ro~Concrete, Vol 3 , No 9, Sep 1969, pp369-378
Ruddle, h4 (1990), i 4 ~ d ~ nctroi1 cayacrg c$relr!fi)i-ceLJ coi~crefebeams, PhD
~ r ~ ga17d the z~lfrmafe
thesis, Queen's University of Belfast, Northern Ireland, 1990
Selvadurai, A P S and Bakht, B (1995), Srrnulafrorl of~.ollrrlgwheel loads or1 an FRC deck,
Proceedings 2ndUniversity/lndustry workshop on FRC, Toronto, Canada, 1995, pp 14 1- 154
Skates, A (1 987), De~vlopmerlfof n de.sjp 117efhod~for
~.e.Wnif~ed
cot~crete~ l a h.s).sfcn~s
s~lhje~f
loc~drrlg, PhD thesis, Queen's University of Belfast, Northern
fo comeritrated n17d ru~;forn~
Ireland, 1987
Taylor, R. and Hayes, B. (1%5), Some tests OH the e#xt of ecfge re~ftzrrrrtor1 p m c l m g slmu- 111
RC slabs, Magazine of Concrete Research, Vol 17, No.50, March 1965, pp 39-44
Taylor, S E , Rankin, G I B and Cleland, D J., (2001a) Al-chnlg nctrorl I H high .ifreng/h coilcrefe
slctbs, Proceedings ofthe Institution of Civil Engineers. Structures and Buildings, Vol 146, Issue
4, Nov 2001 pp 353-362
Taylor, S . E , Rankin, G I B and Cleland, D.J. , (2001b) The trl/rn~afesfz-er~gthof high
p f o m n i l c e corlcrete brrdge deck slnbs Third International Conference on Concrete under Severe
Conditions - Environment and Loading to be held in Vancouver, BC, Canada June 18-21, 2001
Taj lor, S E , (2000) Corr7prc.ssiw mem hrnrre crcfruir Irr /rig-hsrrergfl?corrcrele budge deck
~lcrhs,PhD thesis, Queen's University of Belfast, Jan 2000
Taylor, S E , Rankin, G I B and Cleland, D J (199S), Corii)yressriv n~ernbmreucirort rr? hrg11
sfreng/h cor7crrfe hricfgc deck slabs, Proceedings from the Second International Conference on
Concrete under Severe Conditions - Environment and Loading, Trams@, Nonxay, June, 1998,
pp1703-1713
Thorburn, J and Mufti, A A. (1 995), Full-scale fesfrl~gofexter~~allyrezifoi'ccd FRC budge decks
of1 sfeel grr-ders, Annual conference for the Canadian Society of Civil Engineers, June 1995, pp
533-552
Turner, C A P (1 909), Concrete-steel construction Part I - Buildings ACI Special Publication,
SP-52, American Concrete Institute, h/lichigan, USA, 1909
Westergaard, I4 M (1930), Conrprrfnfiorlof sresses ill bridge slabs due fo >theelloads, Public
Roads, Vol 11, 1930
\17estergaard, H hl and Slater, W A (1921), A4ortrerifs a d sfr-esscsnr slabs, ,4CI Vol 17, 1921,
ppll5-538
Wood, R H. (1 96 I), Plastic arm' elnsfrc Je.ng-n ofslnbs ar7dylafe.r, Thames arid Hudsen, London,
196 1
Guide to Compressive Membrane Action

This technical paper provides a concise summary on the use of

compressive membrane action in concrete bridge decks. It is
based on work funded by the Concrete Bridge Development
Group as part of its research programme.

Guide to Compressive Membrane Action

Technical Paper No 3

First published 2002

O Concrete Bridge Development Group 2001

Published by the Concrete Bridge Development Group

Riverside House 4 Meadows Business Park Station Approach
Blackwater Camberley Surrey GU17 9AB UK
Tel: +44 (0)1276 33777 a Fax: +44 (0)1276 38899

All rights reserved. Except as permitted under current legislation no part of this work may be photocopied,
stored in a retrieval system, published, performed in public, adapted, broadcast, transmitted, recorded or
reproduced in any form or by any means, without the prior permission of the copyright owner. Enquiries
should be addressed to the Concrete Bridge Development Group.

Although the Concrete Bridge Development Group (limited by guarantee) does its best to ensure that any
advice, recommendations or information it may give either in this publication or elsewhere is accurate, no
liability or responsibility of any kind (including liability for negligence) howsoever and from whatsoever
cause arising it, is accepted in this respect by the Group, its servants or agents.