P. Jackson Giord and Partners
Introduction
Solid slab bridges
Single spans
abutments or bank seats (Department of Transport, 1995). Apart from the durability advantages, this can lead to saving in the deck due to the advantage of continuity. On short span bridges with relatively high abutment walls, being able to use the deck to prop the abutments can also lead to significant savings in the abutments. However, this normally depends on being able to build the deck before backfilling behind the abutments. When assumptions about construction approach such as this are made in design, it is important that they should be properly conveyed to the contractor, normally by stating them on the drawings.
‘accurate’ than in conventional structures. Neither the normal at-rest pressure behind abutments nor the resistance to movement is ever very accurately known. It might be argued that, because of this, designs should be done for both upper and lower bounds to soil properties. In practice, this is not generally done and the design criteria used have sufficient reserve so that this does not lead to problems.
Multiple spans
multiple simply supported spans are now used.
Voided slab bridges
Above a span of about 10–12m, the dead weight of a solid slab bridge starts to become excessive. For narrower bridges, significant weight saving can be achieved by using relatively long transverse cantilevers giving a bridge of ‘spine beam’ form as shown in Figure 1. This can
extend the economic span range of this type of structure to around 16m or more. Above this span, and earlier for wider bridges, a lighter form of construction is desirable.
therefore be provided with drainage holes at their lower ends. It is also important to ensure that the voids and reinforcement are held firmly in position in the formwork during construction. This avoids problems that have occurred with the voids floating or with the links moving to touch the void formers, giving no cover.
slab. That is, without considering either the reduced transverse shear stiffness or the local bending in the flanges. Unlike the previous British code, EN 1992-2 (BSI, 2005) does not give specific guidance on voided slabs. However, some is provided in the accompanying ‘PD’ pub-
lished by the British Standards Institution (BSI, 2008a).
composite flexural properties are used and the extra defor-mation under non-uniform bending is represented by calcu-lating an equivalent shear stiffness.
by flexure in the flanges, that is by the section acting like a vierendeel frame as shown in Figure 2.
of concrete is that any shape can be formed. Figure 3 shows a voided slab bridge of more imaginative appearance which carries main line rail loading. To make most efficient use of the curved soffit varying depth section, different sizes of void were used across the width.
Beam and slab bridges
slab structures have therefore been less common. However, there is no fundamental reason why they should not be used and there are thousands of such structures in service.
with each change of the loading specification, leaving the abnormal load the same until the most recent change which could actually make it less severe for shorter spans. However, reinforced concrete beam and slab bridges do not appear to have increased in popularity as a result. They are more popular in some other countries.
‘intermediate diaphragms’ is
extremely inconvenient and therefore expensive so they have become unusual. However, in an in situ structure which has to be built on falsework, it makes relatively little difference and is therefore more viable.
where the main beams are steel or prestressed concrete and this aspect will now be considered.
‘local moments’). These moments were obtained from Westergaard (1930) albeit usually by way of tables given in AASHTO. British practice also uses elastic methods to obtain local moments, usuallyeither Westergaard or influ-ence charts such as Pucher (19). However, the so-called
‘global transverse moments’, the moments induced in the deck slab by its action in distributing load between the beams, are considered. These moments, obtained from the global analysis of the bridge, are added to the local moments obtained from Westergaard (1930) or similar methods. Only‘co-existent’ moments (the moments induced in the same part of the deck under the same load case) are considered, and the worst global and local moments often do not coin-cide. However, this still has a significant effect. In bridges with very close spaced beams (admittedly rarely used in North America) the UK approach can give twice the design moments of the US approach.
Longer span structures
In modern practice, purely reinforced structures longer than about 20m span are quite unusual; concrete bridges of this size are usually prestressed. However, there is no funda-mental reason why such structures should not be built.
The longest span reinforced concrete girder bridges tend to be of box girder form. Although single cell box girders are a well-defined form of construction, there is no clear-cut distinction between a ‘multi-cellular box girder’ and a voided slab. However, the voids in voided slab bridges are normally formed with polystyrene or other permanent void formers, whereas box girders are usually formed with removable formwork. The formwork can only be removed if the section is deep enough for access, which effectively means around 1.2m minimum depth. Permanent access to the voids is often provided. In older structures, this was often through manholes in the top slab. This means traffic management is required to gain access and also means there is the problem of water, and de-icing salt where this is used, leaking into the voids. It is therefore preferable to provide access from below.
with the greater advantage of saving weight near mid-span, encourages longer span bridges to be haunched. Haunching frequently also helps with the clearance required for road, rail or river traffic under the bridge by allowing a shallower section elsewhere.
’s Falls Bridge shown in Figure 4. The true arch form suits reinforced concrete well as the compressive force in the arch rib increases its flexural strength. As a result, the form is quite efficient in terms of materials.
efficient forms, such as prestressed cantilever bridges, that can be built with less temporary works. However, they may still be economic in some circumstances, particularly in countries where the labour required to erect the falsework is relatively cheap. A further factor may be local availability of the materials in countries where the prestressing equip-ment or structural steelwork required for other bridges of this span range would have to be imported.
bers so that the bridge, including the columns supporting the deck and at least main longitudinal members at deck level, can be built bay by bay behaving as a truss until it is joined up.
hence flexural capacity of, the rib. The design code’s lesser load factor (normally 1.0) for ‘relieving effects’ should be applied to dead weight when this arises. However, the letter of many codes only requires this to be applied in certain cases which are defined in such a way that it does not appear to apply here. This cannot be justified philosophi-cally and the reduced factor should be used.
possibly even after the falsework has been struck so that this does not have to be designed to take the full load.The deck structure is then much like a normal viaduct supported on piers from the arch rib and the rib has to take significant moments.
Design calculation
Geometry
–15 but continuous or integral bridges can be shallower. Because the concentrated live load (i.e. the wheel load) the deck has to carry does not reduce with span, the span-to-depth ratio of short span slabs tends to be towards the lower end of the range.However, deck slabs of bigger bridges often have greater
span-to-depth ratios than slab bridges. This is economic because the dead weight of the slab, although an insignifi-cant part of the load on the slab, is significant to the global design of the bridge.
cantilevers at the edges tend to look shallower than vertical sided bridges even if they are actually deeper.
Ultimate strength in flexure and torsion
as crack width or service stresses, a design has to be assumed and then checked. This makes the design process iterative. A first estimate is required to start the iterative procedure and the ultimate strength design provides such an estimate.
is using Mohr’s circle. Theoretically, it is most efficient to use orthogonal reinforcement placed in the directions of maximum and minimum principal moments. Since there is no torsion in these directions, torsion does not then have to be considered. However, it is not often practical to do this as the principal moment directions change with both position in the slab and load case.
that enable them to give these corrected moments, com-monly known as ‘Wood–Armer’ moments, directly. To enable them to do this, it is necessary to specify the direc-tion of the reinforcement.
–Armer approach leads to excessive requirements for transverse steel. When assessing existing structures, this problem can be avoided by using alternative analytical approaches. However, in design it is usually preferable to avoid the problem by avoiding the use of very highly skewed reinforcement. The disadvantage of this is that it makes the reinforcement detailing of skew slab bridges more complicated. This arises because the main steel in the edges of the slab has to run parallel with the edges. Orthogonal steel can therefore only be achieved in the centre of the bridge either by fanning out the steel or bypro-viding three layers in the edge regions. That is, one parallel to the edge in addition to the two orthogonal layers.
Ultimate strength in shear
slabs, such as in box culverts or the deck slabs of beam and slab bridges, and many codes do not allow them to be considered effective. The shear strength rules can therefore be critical in design.
The shear strength of concrete members without links does not increase with size as normal dimensional analysis would suggest; the stress at failure reduces as the depth of section increases (Bazant, 1984). This is also allowed for by many modern design codes. However, older codes did not consider either this effect or the effect of main steel area on shear strength. Even now, some codes (including the American AASHTO (2002) code) do not consider the effect of scale. AASHTO also only considers a very limited effect of reinforcement percentage.
Codes which did not consider steel area or scale effect often gave significantly higher allowable shear stresses than modern codes in deep sections with low main steel percentages. Test results suggest that these stresses cannot be justified. Even the current AASHTO can apparently be unsafe in some situations (Collins and Kuchma, 1999). However, fortunately, the maximum shear stresses nor-mally arise close to the supports. In these situations, the actual shear strength increases due to so called ‘short shear span enhancement’. This effect is not considered in the older codes and is under-estimated by AASHTO. The theoretical worst case of these codes is therefore unlikely to arise in practice.
‘short shear span enhancement’ to be
considered in design. It arises because shear failures in reinforced concrete normally occur on planes inclined at a shallow angle (typically around tan ?1 0.4) to the horizontal. Anything that forces the plane to be steeper results in a higher failure load. One disadvantage to the designer of
being allowed to use this effect is that it makes it less obvious where the critical section for shear will be. A sim-pler approach which is sometimes used is not to consider short shear span enhancement directly but to check shear an effective depth from the support, rather than at the
support.
AASHTO (2002). In a typical simply supported test speci-men loaded by a single concentrated load, the moment shear force ratio and the shear span-to-depth ratio are the same thing. However, in continuous beams they are not related and the moment to shear force ratio is low at
points of contraflexure where shear planes are not forced to be steep. The American code avoids allowing excessive shear at these points by limiting the moment to shear force ratio that can be used in calculations.
‘addition principle’. In this the shear strength of the beam is assumed to be the sum of the strength without links and the strength due to the links. The latter is calculated using 458 truss analogy where the member is assumed to be made up of the links acting as ties and concrete struts angled at 458.
‘varying angle truss’ approach. In this, all the shear is assumed to be taken by the analogous truss but the angle of the truss can be varied to give the greatest strength. A flatter angle truss enables a given area of links to take a higher shear force but this implies a greater force in the main tension steel. This does not affect the maximum area of main steel required in a member but does alter the design of any curtailments.
‘addition prin-ciple’. The behaviour of beams with links is fundamentally different from those without. Unless they have short shear span-to-depth ratios, beams without links fail as soon as the shear crack appears. However, the links can only be suf-ficiently highly stressed to be effective after the concrete has cracked. The addition principle is therefore a purely empiri-cal approach. It has no theoretical basis and is justified only by test results. Indeed, the code rules for the strength of
beams without links are themselves purely empirical.
‘effectiveness factor’(Batchelor et al., 1986; Hewit and Batchelor, 1975) to obtain the compressive strength to be
used for the inclined concrete struts. It is also necessary to limit the angle of the struts. The limits in EC2 are cot ?1 1 and 2.5.
principle implies that any amount of links would give an increase in strength. In contrast, with small areas of links,varying angle truss analogy gives a strength that is less than the strength without links. What happens in practice is that the section fails much as it would with no links
and small areas do not really increase strength. This is why codes specify a minimum link area for slabs as men-tioned above. It might be argued it is not required if varying angle truss design is used but EN 1992 does have such a limit.
these aspects. However, the upper or ‘web crushing’ limit in shear is rarely critical in solid slab structures. It is more likely to be critical in box and other flanged sections where it pays to make the webs as thin as practical.
Ultimate strength in punching shear
bars extend to some 3d from the load.
‘flexural’ shear check in slabs. This gives high concentrated values near concen-
trated loads. In a case that is anywhere near the ‘punching shear’ limit, these are invariably outside the code rules. The implication is that the code does not require you to check the elastic stress so close to the loaded area. The physical explanation of why this works is that a failure so close to the load would attract some short shear span enhancement.
Service stresses
bridge structure in which the reinforcement has yielded.However, checking such criteria as deflection and crack width for a structure in which the steel had yielded would be extremely difficult. Once a structure has gone outside the elastic range there is no guarantee that it will recover from deflection or crack opening caused by transient loads. It is therefore theoretically necessary to undertake a non-linear analysis of the complete load history to ensure the structure is satisfactory. This is not practical. The usual approach is to impose limitations on steel stress in tension and concrete stress in compression to ensure the material stays within the linear range. This enables crack widths and deflections to be checked without considering non-linear behaviour.
In EN 1992-2, steel stress always has to be checked but checks on concrete stress can be avoided.
Crack widths
widths than would be needed for aesthetic reasons are often imposed. Research does not support this relationship (Beeby, 1978) and there does not appear to be any justifica-tion for limits more severe than around 0.3mm.
very wide-spaced bars.
‘quasi permanent’ loads, which actually means (since bridge structures with high permanent load ratios are usually prestresed) that they are unlikely to be critical.
This is because the cooling surface layers are restrained by the core which stays hot for longer.
‘early thermal cracking’ and the reinforcement required to pre-vent excessive cracking due to it. More information is given in the water retaining section, EN 1992-3 (British Standards Institution, 2006). Fuller guidance is given in a CIRIA document (Bamforth, 2007).
Fatigue
’s curve fatigue assessment of reinforcement to be undertaken similar to that used for steelwork. However, a simpler approach of checking compliance with a live load stress range under service loads is also provided. The stress range currently specified in EN 1992-1 is extremely cautious, representing the extreme worst case. The UK published document (British Standards Institution, 2008a) provides more realistic ranges. Only the stress range due to normal traffic load(‘Frequent’ in the Eurocode) is considered because the number of cycles of the rarer load is much lower.
to be restrictive in the deck slabs of beam and slab bridges. These are, however, exempted
from the check by the UK National Annex (British Standards Institution, 2008b). The justification for this is that tests show that the stress range in the reinforcement is much lower than the calculations suggest as will be apparent from Hewit and Batchelor (1975) and Holowka and Csagoly (1980).
Durability
Detailing checks
capacity with the concrete cracked and the steel yielding may be less than the moment required to cause the first crack. Theoretically, the steel area required to prevent this increases both as the effective depth to total depth ratio reduces and as the concrete tensile strength increases. In practice there is no control over the maximum concrete strength, so it is not possible to write a rigorously correct rule. It can be shown that the minimum steel areas in many codes are not theoretically adequate but few resulting problems have been observed. One reason for this is
because, due to the interaction between the code checks, a seriously inadequate ‘minimum’ steel area can only arise where the calculated applied moment is significantly less than the cracking moment. The section is therefore unlikely to crack. Although reinforced concrete is always designed to crack, many reinforced concrete structures do not crack; they actually work due to the tensile strength of concrete.
Conclusions
or prestressed concrete main elements. In this chapter the types of RC deck have been considered and the principal design checks for reinforced concrete design reviewed.
References
American Association of State Highways and Transportation Officials (2002) Standard Specification for Highway Bridges, 17th edition. AASHTO, Washington, DC.
Allenby N. and Ropkins J. W. T. (2004) The use of jacked-box tunnelling under a live motorway. Proceedings of the Institution of Civil Engineers Geotechnical Engineering, 157 issue GE4, October, 229–238.
Armer G. S. T. (1968) Correspondence, Concrete. August, 319–320.
Bamforth P. B. (2007) CIRIA C660 Early Age Thermal Crack Control in Concrete. CIRIA, London.
Batchelor BdeV.,GeorgeH.K.andCampbellL.T.I.(1986)Effec-
tivenessfactorsforshearinconcretebeams.JournaloftheStruc-
tural Division. American Society of Civil Engineers, 112, No.
ST6, June, 14–1477.
Bazant Z. P. (1984) Size effect in shear failure of longitudinally
reinforced beams. American Concrete Institution Journal,
September–October, 456–468.
Beeby A. W. (1978) Corrosion of reinforcing steel in concrete and
its relation to cracking. The Structural Engineer, 56, No. 3,
March.
British Standards Institution (and other member Institutions of
CEN) (2005) EN 1992-2 Eurocode 2 Design of Concrete
Structures; Part 2 Bridges. BSI, London.
British Standards Institution (and other member Institutions of
CEN) (2006) EN 1992-3 Eurocode 2 Design of Concrete Struc-
tures; Part 3 Liquid Retaining and Containment Structures.
BSI, London.
British Standards Institution (2008a) PD 6687-2 2008 Recommen-
dations for the Design of Structures to EN 1992-2-2005. BSI,
London.
British Standards Institution (2008b) UK National Annex to
EN 1992-2 Eurocode 2 Design of Concrete Structures: Part 2
Bridges. BSI, London.
Collins M. P. and Kuchma D. (1999) How safe are our large,
lightly reinforced concrete beams, slabs and footings? American
Concrete Institute Structural Journal, 96, No. 4. July–August,
482–490.
Department of Transport (1995) BA 57/95 Design for Durability.
Department of Transport, London.
Hewit B. E. and Batchelor BdeV. (1975) Punching shear strength
of restrained slabs. Journal of the Structural Division. American
SocietyofCivilEngineers,101,No.ST9,September,1831–1852.
Holowka M. and Csagoly P. (1980) Testing of a Composite
Prestressed Concrete Aashto Girder Bridge. Ontario Ministry
of Transport and Communications. Downsview, Ontario,
Research Report 222.
Jackson P. A. and Cope R. J. (1990) The Behaviour of Bridge
Deck Slabs under Full Global Load. Developments in short
and medium span bridge engineering 90. Montreal, August,
253–2.
Kirkpatrick J., Rankin G. I. B. and Long A. E. (1984) Strength
evaluation of M-beam bridge deck slabs. The Structural
Engineer, 62B, No. 3, 60–68.
Nielsen M. P. (1984) Limit Analysis and Concrete Plasticity.
Prentice Hall, Upper Saddle River, NJ, 205–219.
Ontario Ministry of Transportation and Communications (1983)
Ontario Highway Bridge Design Code. Downsview, Ontario.
Pucher A. (19) Influence Surfaces for Elastic Plates. Springer
Verlag, Wien and New York.
Westergarrd H. M. (1930) Computation of stresses in bridge deck
slabs due to wheel loads. Public Roads, 2, No. 1, 1–23.
Wood R. H. (1968) The reinforcement of slabs in accordance with
a pre-determined field of moments. Concrete, February, 69–76.