Professional Documents
Culture Documents
USING
PRESTRESSED BEAMS
B A NICHOLSON
FOREWORD
For several years, the Prestressed Concrete Association has run a basic course on
prestressed concrete bridge design. This one day course has been held in various
parts of the country, and uses lecturers both from within the PCA member companies
and from outside consultants.
This book contains a development of some of the material presented in the course.
The original course notes were prepared by H. J. Lloyd, I. M. Gibb, and A. E. Gamble.
This book has been prepared by B. A. Nicholson based on the material in those
course notes.
The bulk of the book takes the form of a worked example of the design of the beams
for a simply-supported single span beam-and-slab deck. The final section is a partial
design example of an inverted T beam deck, included in order to illustrate the extra
calculations required for a solid slab deck.
1 STANDARD BEAMS
1.1 HISTORY
The use of precast prestressed beams in bridge decks in the post World War II era
owes its success in the main to the foresight of the Prestressed Concrete Development
Group, which in the 1950’s developed the first standard beam sections to be available
from the beam manufacturers.
This enabled factory production of the beams on a large scale, and, with the dawn of
major road construction in the late 1950’s and its philosophy of grade separation for
motorways and trunk roads, it gave bridge engineers scope to rationalise design
procedures using up-to-date load distribution theories.
The standard beam sections available at that time have of course themselves been
developed and modified, and in essence only one really remains today with any
significant usage. This beam, the inverted T beam, is used in bridge decks in spans up
to about 20 metres.
With the rapid development of the UK motorway network in the 1960’s, it was clear
that there was scope for a standard beam that would enable larger spans to be achieved.
Consequently, at the end of the decade a new beam was introduced for spans from
about 15 to 30 metres. This was designated the M beam, due to its width and intended
spacing. These beams were intended for use in pseudo-slab bridge decks with a
contiguous concrete bottom flange using transverse reinforcement located through
lower web holes at 600mm centres along the beams.
Eventually engineers realised that the M beam could be used more efficiently in beam
and slab decks by eliminating the bottom in-situ concrete and by spacing the beams
apart at up to 1.5 metre centres. The limitation on this type of use proved to be the
shear capacity of the beams, which have a web thickness of only 160mm.
Other beams developed around this time were the U beam for beam and slab decks
up to about 30 metre spans, and a U shaped variation of the M beam for use as edge
beams in M beam decks. Eventually, with the very popular M beam being used in a
manner somewhat different from its intended use, and bearing in mind the various
problems and limitations this presented, a new beam was developed by the Prestressed
Concrete Association in the late 1980’s. This was designated the Y beam.
The Y beam now has three variants: the TY beam, the Y beam, and the SY beam.
Together these cover all span ranges up to 45m. It is expected that in due course
inverted T beams and M beams will cease to be used in favour of the enhanced
properties of the Y beam ranges.
STANDARD BEAMS 3
Inverted T beam
M beam U beam
Concrete bridge superstructures using precast prestressed concrete beams fall into
three distinct types: slab decks, pseudo-slab decks, and beam and slab decks.
Slab Decks
Slab decks can be solid or voided, and provide simply supported spans of up to 20
metres. These decks use standard TY or inverted T beams placed side by side. The
space between them is then filled with in-situ concrete, and an overall covering of
75mm completes the deck.
Continuity of these decks can quite easily be achieved by using reinforcement in the
in-situ concrete over the supports. Suspended spans using TY beams or inverted T
beams can be lightened by introducing void formers into the space between the beams.
Pseudo-slab Decks
This type of bridge structure is currently not quite so popular. Precast beams are
incorporated into a voided slab type of deck by either adding an in-situ bottom flange
and top flange, as with the original M beam decks, or by using voided beams (e.g. box
beams).
A voided slab deck is thus created without the inconvenience of temporary works
and soffit shutters, and provides a torsionally stiffer deck than ordinary beam and slab
decks.
Spans for this type of bridge deck are usually limited by the length of precast beams
that can be transported to site, and therefore are rarely more than 30 metres.
The most common type of superstructure for small to medium span bridges, this type
of deck comprises individual precast beams at discrete centres with an in-situ concrete
top flange. M beams, TY beams, Y beams, SY beams, and U beams can all be used
in this form of construction.
With most of the standard range of precast beams the in-situ concrete top slab is cast
into permanent formwork which is located in recesses formed in the edges of the top
flanges of the beams. Typical spans for this type of deck are similar to the pseudo-
slab decks above, being limited in the main by transportable beam components.
Standard edge beams are available to complement the Y beam, TY beam, and M
beam ranges. These provide a vertical visible face, and have the capacity to carry the
extra loads from the parapet cantilever.
STANDARD BEAMS 5
For the types of superstructure indicated above, the beam manufacturers provide
standard details of the individual sections and their ranges together with an indication
of typical span ranges for decks incorporating these beams and carrying standard
highway loads.
There will obviously be situations where the choice of deck type is not clearly indicated
by the available span, and it is also inevitable that there will be areas of overlap where
the choice between inverted T beams in a slab deck or individual M or Y beams in a
beam and slab deck may not be clear cut. In this situation it may be necessary to
evaluate more than one solution, and the standard sections enable a swift selection of
the available ranges for comparative design exercises to be undertaken and cost
comparisons made.
It is also possible within the standard range of each beam type to be in a span range
that is covered by more than one specific beam unit. In this situation it is usually cost
effective to select the larger unit where there are no restrictive limitations on headroom.
Design
Although the various types of standard beam sections are well documented in terms
of dimensions and structural properties, it is important to point out that these factory
produced beams are standard only to the extent that they are manufactured using
standard shaped sections. The amount and magnitude of prestress applied to each
beam is dependent on its individual situation, and must be determined by the designer
prior to manufacture. The standard sections show where prestressing strands may
be located, but it is the responsibility of the designer to determine which of these are
to be used.
In their literature, the manufacturers give suggestions for good design details. These
should be adhered to, as they lead to economy and good workmanship.
Manufacture
Precast prestressed beams are manufactured in long lines of several units using straight
strands. These are debonded for varying distances at the ends of each beam within
the mould. This is necessary to maintain the stress in the beam at an acceptable level
as the self-weight bending moment reduces approaching the supports.
Once the concrete in the moulds reaches the minimum transfer strength, detensioning
can take place, the strands between the beam can be cut, and the beams removed to
the storage area prior to delivery to site.
STANDARD BEAMS 7
Span in
metres: 12 14 16 18 20 22 24 26 28 30 32
Y1
Y2
Y3
Y4
Y5
Y6
Y7
Y8
Beam selection chart for the Y beam range, taken from PCA literature.
1300
1200
1100
1000
900
800
260
210
160
110
60
0
Handling
Beams are usually manufactured with lifting loops, thus enabling on site lifting to be
achieved with either single or twin cranes to suit the site requirements. However, TY
beams and inverted T beams are usually lifted using a sling through the end web
holes.
Access to Site
It is of obvious importance that there is suitable access to the bridges in order for the
beams to be delivered and lifted off the trailer by suitably located cranes. Of course,
this also applies to the route to the construction site which must allow the delivery
lorries to manoeuvre their lengthy loads.
Camber
Variation in camber of prestressed beams is inevitable when one considers the tolerance
in prestress force and location, together with possible variation in concrete properties
with maturity and climatic conditions.
Edge Details
On site, construction of parapet string courses in one or more stages generally follows
the construction of the central deck slab area. This necessitates the formation of a
construction joint along the edge beam prior to constructing the fascia.
This section can be cast on site as a second stage after the rest of
the deck, or alternatively can be cast onto the UM beam by the
manufacturer so that the edge beam and parapet can be brought
to site as a single unit.
Construction joint
UM8 M8 M8
T2 T2 T2 T2
Cast by manufacturer
Skew
Although it is possible to manufacture precast beams with skew ends, the increase in
cost for each unit and the problems that skew presents should be considered in detail
at the design stage.
Firstly, it should be remembered that even a very small change in skew angle requires
a new stop end for the mould. A change from say 30° to 31° increases the width by
12mm for an M beam. To rationalise a range of angles with a variation of 10°, say,
would be a useful and economic possibility.
Structural problems created by skew in the ends of precast beams relate specifically
to the acute corner, where the formation of cracks can cause the corner of the flange
to spall when the beam cambers during transfer. Although not structurally significant,
this is undesirable, and is best prevented by blocking out the corner to give a local
square end.
An additional problem that presents itself with skew beams is that of locating transverse
reinforcement through the web holes. It is recommended that the standard web holes
permit reinforcement to be placed for skews up to about 35°. Higher skews than this
would require special non-standard web holes, which would increase the cost of the
beams significantly, and may even affect the shear capacity of the section. For high
skew bridges, it is normally better to place the transverse deck reinforcement at right-
angles to the beams rather than parallel to the abutments.
Transverse Reinforcement
For the transverse reinforcement through the web holes of precast beams, it is usually
better to use a number of smaller bars rather than a single large diameter bar, as lap
lengths are reduced and handling becomes easier. For some awkward skew situations
it may even be sensible to use untensioned prestressing strand threaded through the
web holes instead of reinforcing bars, as it is more flexible.
The positioning of transverse deck reinforcement when using solid edge beams may
require the use of couplers at the edge beam interface.
Temporary Support
It is important to ensure that the beams are supported so that they cannot topple over
on site. Deeper beams, particularly when being jacked to their final level and during
bearing installation, must be assessed to eliminate this risk.
STANDARD BEAMS 11
Deck slab
Diaphragm
M beam
Diaphragm
800 wide
Sections 3 to 7 of this book consist of a design example of a beam and slab deck.
This design example shows the typical sequential calculations necessary for the full
design of a precast pretensioned concrete Y beam in a simply supported beam and
slab bridge deck.
The right hand pages show the numerical calculations involved at each stage, and the
left hand pages contain explanatory comments and further information.
The edge detail was chosen for aesthetic reasons, and the outer beams placed as near
to the edges of the bridge within this limit. This led to the beam spacing of 1.275m.
The span charts for Y beams give spans for beams at 1 and 2 metre spacings. It is
straightforward to interpolate from this information to make an initial selection of
beam size, in this case Y8.
Clearly alternatives would have been possible, for example eleven Y6 beams could
have been used, at about 1 metre spacing. However, it has been found that unless it
is necessary to make the deck as shallow as possible, it is usually preferable to use
fewer but larger beams.
DESIGN EXAMPLE 13
1275 1275
3 GRILLAGE MODEL
3.1 INTRODUCTION
Early bridge decks were analysed on a strip basis. Abnormal and wheel loads were
crudely distributed and conservative designs resulted. Experimental data became
available to determine the transverse load carrying characteristics of decks to determine
the correct level of transverse strength provision and to distribute load more logically
to the longitudinal members. For example, in the 1950’s Morice and Little developed
a Distribution Coefficient method which was a simple hand method based on
experiments which allowed for the overall distribution of loads on a plate structure
such as a bridge deck. It was satisfactory for skews up to 20°. This method was one
of several similar techniques extensively used in design offices for approximately 15
years, until the advent of computer techniques which enabled larger and more complex
structures to be analysed more accurately using grillage, finite strip and finite element
methods. Of these three methods grillages offer the widest range of structures which
can be analysed. Popular opinion suggests they are also the easiest to use and
understand.
No analysis method gives a rigorous solution, and some degree of error must be
accepted, usually ranging up to 10% or 20% depending on complexity. These errors
come from several sources, including the idealisation of the geometry and material
properties, and idealisation of the structural behaviour.
Grillage analysis has found favour as a bridge engineer’s design tool because it is
perceived to have the following advantages:
The method can be used for structures with beam and slabs decks, voided slabs or
solid slabs. It can be used for simple and continuous bridges, and allow for elastic
supports and settlement. It is suitable for right, skew and curved decks. This range
covers hundreds if not thousands of bridge decks designed in recent times, and certainly
covers all bridges with prestressed beams.
GRILLAGE MODEL 15
GRILLAGEANALYSIS
The analysis will be performed using the computer program “STAAD III / ISDS” from Research
Engineers (Europe) Limited.
Sign Convention
This analysis uses a right-handed orthogonal system of axes. The grillage lies in the X-Z plane,
and the Y-axis is vertically upwards. Loading will be applied in the negative Y direction.
Local axes are used for bending moment in members, etc. In this case the X-axis runs along the
length of the member, and the Y-axis is still vertically upwards. The Z-axis lies in the plane of
the grillage.
16 SIMPLE BRIDGE DESIGN USING PRESTRESSED BEAMS
Longitudinal grillage beams are placed on the line of the physical beams, and represent
the composite action of the beam and its associated section of slab. Longitudinal
beams are also positioned along the parapet edge beam. Transverse elements represent
the top slab. There are no end diaphragms in this bridge, but when these are present
they must also be represented by appropriate transverse elements.
This type of grillage model is suitable for beam and slab decks using M-beams and
Y-beams.
Because of the usually large number of beams in a T-beam deck, it may be preferable
to model two or three beams by one grillage member. Transverse elements represent
transverse solid infill elements. Because of the non-uniform shape of these elements
as they pass over and through the beams their depth is normally taken to the centre
line of the transverse holes. The wider spacing of model elements does not materially
affect the transverse element idealisation since the structure acts as a true slab.
However, care is needed when evaluating design moments shears and reactions due
to the combination of several physical elements into single model elements.
U-beam decks, although basically beam and slab decks, behave differently because
the transverse stiffness alternates across the deck between stiff through the beams
and flexible between the beams. The beams are positioned to try and equalise the top
slab spans between and across beams. One method of modelling a U-beam deck is to
place longitudinal elements on the centre line of each web. The longitudinal properties
for each grillage beam are then taken as half that of the composite box section. As
with the inverted T-beam decks, care is required in evaluating the output since there
are now two longitudinal elements representing one physical beam.
GRILLAGE MODEL 17
The in-situ downstand fascia and the parapet beam are both discontinuous and so do not
contribute to the structural action. The cross-section of the structural elements of the bridge is
therefore:
Grillage beams will be placed on the lines of the nine pretensioned beams, and nominal edge beams
will be placed along the parapet beams. Thus the grillage representation of the cross-section is:
Transverse members will be provided at 1.90m intervals to represent the slab. This divides the
length of the deck into 14 equal sections. The nodes of the grillage will generally be on a grid of
1900 x 1275, which is well below a 2:1 aspect ratio and therefore satisfactory.
18 SIMPLE BRIDGE DESIGN USING PRESTRESSED BEAMS
Grillage analysis idealises a deck into a grid of interconnected beams. The real
dispersed effects of bending, shear and torsion are assumed to be concentrated in the
nearest equivalent grillage beam.
Variations from the true behaviour arise because the real slabs element equilibrium
requires torques and twists to be identical and in orthogonal directions, but in grillages
the joints can rotate differently. However, if a slab is modelled by a sufficiently fine
grillage mesh these anomalies are smoothed out to become almost insignificant. Again,
moments in grillage beams are proportional to the beam curvature in that direction.
In real slabs, moments also depend on the orthogonal direction curvature, but this
error is also sufficiently small to be ignored.
• Place the grillage beams coincident with the physical beams or lines of
designed strength.
• Where possible, lay out the grillage to capture all the load, and for ease
of shape generation and section property calculations.
• Transverse elements should be spaced to try and reflect the aspect ratio
(length/width) of the whole deck.
• Skew decks can be analysed by orthogonal or skew meshes. If the skew
exceeds 20°, the model should be laid out within 5° of the real skew.
• Generally, transverse members should be orthogonal to the longitudinal
members, particularly when skew exceeds 20°.
• Bearing positions should be represented faithfully, and in skew bridges
the vertical stiffness must be modelled with care as they can have significant
effect on theoretical distribution of load.
Once the grillage model has been set up, it is recommended that an initial test load is
applied (such as a uniform UDL), to verify that it is behaving correctly. The test load
case should be checked against some simple hand calculations (e.g. wL2/8) to make
sure that the results are reasonable.
GRILLAGE MODEL 19
Grillage model
The first diagram shows the node numbers, and indicates the supports with a circle:
1 15
16 30
31 38 45
53
46 60
61 68
75
76 90
91 105
106 120
121 135
136 150
151 165
193 206
207 220
221 234
235 248
249 262
263 276
277 290
136 137 138 139 150
291 304
20 SIMPLE BRIDGE DESIGN USING PRESTRESSED BEAMS
Since the precast and in-situ concrete strengths do not differ by more than 10 N/mm2,
Clause 7.4.1 permits a modular ratio of 1.0 to be used. However in the example a
more accurate value has been calculated taking into account of the different concrete
strengths.
The Y beams have standard notches 50 mm deep along the top edges. These allow
formwork to be placed between the beams to support the deck concrete. In this case,
20 mm thick permanent formwork is used, so that the beam protrudes 30 mm into the
deck slab. The overall height of the section is 1.590 m. The composite section
properties are calculated by assuming the section is made up from the Y8 beam, a
rectangular slab which overlaps it by 30 mm, and the small overlap area which must
be subtracted as it has been counted twice.
The code permits stiffnesses to be represented on the gross concrete section ignoring
the reinforcement or strand. This is the most straightforward, since the amount of
reinforcement and strand has not yet been accurately determined at the analysis stage.
In some situations, such as continuous bridges at supports, the transformed section
may be important and should be used.
Under transient applied forces the short term elastic modulus should be used, and
under applied deformations or long term loads the long term modulus should be
used. To save analysis time for these two situations a value between long and short
may be chosen, ideally reflecting the proportion of permanent to transient effects.
Almost all analyses are executed on elastic models, even though the code allows
plastic methods with the approval of the bridge authority. An elastic analysis is
appropriate for the serviceability limit state, which is the most important for the design
of the pretensioned beams. The use of an elastic model for the ultimate limit state is
simple and conservative. It is a lower bound solution in which the structure is in
equilibrium and yield is not reached.
GRILLAGE MODEL 21
Section Properties
0.220
Slab concrete fcu = 40 N/mm2, E = 31 kN/mm2
Beam concrete fcu = 50 N/mm2, E = 34 kN/mm 2
∴ Modular ratio = 31/34 = 0.91 1.370
The values for A(y--y)2 in this table can only be filled in after y has been calculated.
= 0.2429 m 4
Most decks have edge stiffening for the parapets. Many deck arrangements cause
the edge beam to be the most heavily loaded. In many instances these effects are
complementary, the extra stiffness reducing stresses from the extra loading. However,
the combined stiffness of the edge beam with the parapet upstand may be significantly
higher than the stiffness of the internal beams, and so may attract high unwanted
loads into the parapet upstand. Care must be taken in the computer modelling to
ensure that the stiffness allocated to the edge beam is not unrealistically high, as this
would aggravate the problem.
• model the edge upstand as a separate element (usually with less inertia
than the main elements). This reverses the trend of attraction.
• calculate the whole deck inertia including the upstands, and also excluding
the upstands. Then allocate half the difference to each edge member.
This is likely to be significantly less than the inertia calculated for the
discrete edge shape.
• make the edge upstand discontinuous. This is becoming more popular as
it also reduces thermal and differential shrinkage cracking. It does,
however, require careful detailing.
In the design example, the third method has been adopted. The downstand fascia and
parapet are cast after the deck has been completed, and are both cast in sections
about 2.5m long separated by a narrow gap.
The stiffness of the parapet cantilever is included with the outer Y beam. Edge
longitudinal elements are included in the model along the line of the parapet support
upstand. These elements are given very small section properties so that they do not
contribute structurally to the grillage. They are only included to give the grillage
model a tidy appearance the same size and shape as the bridge deck. They could be
omitted without affecting the results of the analysis. If nominal members are used in
this way, however, loads should not be applied directly to these members.
GRILLAGE MODEL 23
Edge Beam:
Slab members:
These represent a 1.900m section of slab:
= mbd 3/12 = 0.91 x 1.900 x 0.220 3 /12 = 0.00153 m 4
There is no end diaphragm, so the end slab members simply represent 0.950m of slab:
= mbd 3/12 = 0.91 x 0.950 x 0.220 3 /12 = 0.00077 m 4
24 SIMPLE BRIDGE DESIGN USING PRESTRESSED BEAMS
3.7 TORSION
C = k1b3d
This formula should not be used for elements which represent sections of a wide slab.
In this case, the value used for k1 must be reflect the whole slab action, and should not
be calculated for the individual elements. Slabs twist in both longitudinal and transverse
directions, so the value of C is halved for each direction to reflect this double action.
Additionally, the slab elements should be transformed in accordance with the modular
ratio. The torsional inertia of slab elements is thus given by:
C = 1/6 mb3d
Torsionless Design
For many composite beams, as here, the torsional inertia is an order of magnitude
less than the bending inertia. The analysis of such bridges can be simplified by ignoring
the torsion constraints. In other words, torsionless design can be used.
The resulting load distribution is less effective and this gives rise to slightly increased
bending moments. The correspondingly increased design strength is considered
adequate to carry the torques which would be associated with a full torsion model.
Torsionless designs should not be used for significant skews or box beam decks which
may be chosen for their high torsion stiffness properties or where torsional strength is
a significant requirement.
Torsion should also not be ignored in UM beams and thick edge beams such as YE
beams, even if internal beams are considered torsionless. Edge beams can be subjected
to considerable torsion due to loads from the parapet cantilever, and cracking of
these beams could occur if torsion is ignored in the design.
GRILLAGE MODEL 25
Torsion
Estimate torsional stiffness of composite section by idealising the section as three rectangles:
1 1.275 x 0.220
2 0.340 x 1.080
3 0.750 x 0.290
The bending inertia is clearly very much larger than the torsional inertia.
Torsion will therefore be neglected.
26 SIMPLE BRIDGE DESIGN USING PRESTRESSED BEAMS
4 CALCULATION OF LOADS
4.1 INTRODUCTION
The Departmental Standard BD 37/88, Loads for Highway Bridges, is currently used
to determine the loading on UK bridges. BD 37/88 effectively supersedes
BS 5400 Part 2 in the UK, pending revision of this Standard, and it is used throughout
this design example.
The loads generally specified in the Standard are nominal loads appropriate to a
return period of 120 years. Design loads will be obtained later by multiplying the
nominal loads by load factors γfL given in the Standard. An additional factor, γf3, is
also introduced to obtain the design load effects (moments, shears, etc.) from the
design loads. Values of γf3 are given in BS 5400 Part 4 for concrete bridges.
4.2 DEFINITIONS
It is worthwhile clarifying a few definitions, as they may differ from those used with
other structural design work:
CALCULATIONOFLOAD
DeadLoad
Dead load will be carried by the beams acting alone, with no composite action.
Carriageway: Asphalt surfacing - for simplicity assume maximum thickness of 165mm over
whole carriageway. This includes allowance for waterproofing protection boards.
Verge: The weight of the footpath, and non-structural (discontinuous) string course and fascia
will all be taken as SDL.
Notional lanes
For the purposes of calculating the loads to be applied to the bridge deck, the
carriageway is split into notional lanes. In this context, the carriageway is taken as
the distance between raised kerbs, thus including the hard shoulders (see Clause
3.2.9.1). Clause 3.2.9.3 then defines how the carriageway should be split into notional
lanes. Note that in this example there are three notional lanes for loading purposes,
even though the deck will be marked out for only two lanes of traffic.
HA Loading
For loaded lengths up to and including 50 m, the UDL expressed in kN per linear
metre of notional lane is given by the equation:
W = 336(1/L) 0.67
where L is the loaded length (in metres) and W is the load per metre of notional lane.
The KEL per notional lane is always taken as 120 kN.
The UDL and KEL are uniformly distributed over the full width of the notional lane
to which they apply. However, not all lanes carry the full HA load at the same time,
and this is dealt with by means of lane factors. These are functions of the loaded
length and the lane width, and are specified in Table 14 of the Standard.
The single 100 kN wheel load alternative to the UDL and KEL can be placed anywhere
on the carriageway, and occupies either a circular area of 340mm diameter or a square
area of 300mm side. The single wheel load is only significant in the local analysis of
the deck slab, which is not covered in this design example.
CALCULATION OF LOADS 29
Highway Loads
Carriageway width = 1.0 (hard strip) + 7.3 (two traffic lanes) + 1.0 (hard strip)
= 9.3 m
HB Loading
For all public highway bridges in Great Britain the minimum number of units of type
HB loading that must normally be considered is 30, but this number may be increased
up to 45 units.
For this design example, the client has specified 37.5 units of HB load.
The HB vehicle as defined in the Standard represents four axles with four wheels per
axle. One unit of load represents 10 kN per axle. Thus the full 45 units maximum is
equal to 450 kN per axle or 112.5 kN per wheel.
The distance between the central two axles is variable. For simply supported spans,
the smallest figure is obviously the most critical.
As with the HA wheel load the contact surface may be taken as circular or square
with a contact pressure of 1.1 N/mm2.
Note that in this example the HB wheel load is less than the HA wheel load. For slab
design the HA wheel will therefore be critical.
For this simply supported bridge, the shortest wheelbase will be critical.
Thus distance between central axles of the HB vehicle will be taken as 6m.
Horizontal Loads:
The nominal transverse load due to skidding is a single point load of 300 kN,
acting in any direction (parallel to the road surface).
32 SIMPLE BRIDGE DESIGN USING PRESTRESSED BEAMS
Methods of calculating wind loads are given in Clause 5.3 of the Standard.
Combination 2 loading (see page 38) is not significant in its effect on a large proportion
of bridges, such as concrete slab or beam and slab structures 20m or less in span, 10m
or more in width and at normal heights above ground. Wind load therefore does not
need to be calculated for most bridges designed using prestressed beams.
For loaded lengths of 36m and under, the nominal pedestrian live load is a uniformly
distributed live load of 5.0 kN/m2.
For superstructures carrying both highway and pedestrian loading, a reduction factor
of 0.8 is applied to the nominal pedestrian live loading specified for footbridges alone.
Thus, in this case the pedestrian live load is 4.0 kN/m2.
Temperature effects produce two aspects of loading, namely the restraint to the overall
bridge movement due to the temperature range, and the effects of temperature
differences (or gradients) through the depth of the bridge deck.
Temperature Range
The temperature range for a particular bridge is obtained by first determining the
maximum and minimum shade air temperatures for the location of the bridge from
isotherms plotted on maps of the UK, and shown in Figures 7 and 8 in the Standard.
As these isotherm maps are derived from Meteorological Office data relating to a
return period of 120 years (the bridge design life), it may be necessary to adjust the
temperatures for a return period of 50 years for certain applications such as footbridges
and carriageway joints. This is achieved by a straightforward increase or reduction in
temperature as indicated in Clause 5.4.2 of the Standard.
Maximum and minimum effective bridge temperatures are then derived from Tables
10 and 11 in the Standard. Prestressed beam bridges will always be type 4.
The effective bridge temperature range is then used for designing the bearings and
expansion joints, or if this movement is restrained then in determining the stress
resultants in the structure.
CALCULATION OF LOADS 33
Wind load
Wind load will not be specifically calculated for this bridge. It is assumed that Load Combination
2 will not be critical.
Footpath Loads
Temperature Range
Temperature Difference
Positive temperature differences occur within the superstructure when conditions are
such that solar radiation and other effects cause a gain in heat through the top surface
of the deck. Conversely, reverse temperature differences occur when conditions are
such that heat is lost from the top surface of the bridge deck as a result of re-radiation
and other effects.
Temperature gradient diagrams for each of these states are shown on Figure 9 in the
Standard. For surfacing of thickness other than 100mm these can be modified by
reference to Appendix C.
The coefficient of thermal expansion for concrete and steel is taken here as 12x10-6.
For concrete with limestone aggregates, a reduced coefficient of thermal expansion
of 9x10-6 can be used.
If the deck were fully restrained at each end, stresses proportional to the temperature
at each point in the deck would arise. These temperatures and stresses are shown in
the top line of diagrams opposite. The stress at the top of the slab, for example, is
calculated as:
In a simply supported deck there is no axial restraint at the ends, and no moment
restraint. The axial and moment components of these stresses will be relieved by
overall lengthening and hogging of the deck. A self-equilibrating set of internal stresses
will remain; they will exist without any external forces or reactions on the deck.
These internal stresses are calculated by subtracting the axial and moment components
from the stresses calculated for the fully restrained condition.
It is worth noting that the serviceability limit state stresses determined from these
temperature difference diagrams are subject to a load factor of 0.8.
CALCULATION OF LOADS 35
Temperature Difference
h3 = 0.20m
2.5°C 1.02 N/mm 2
A σ y Aσ Aσy
1 1.275x0.15 1.12 0.626 0.214 0.134
2 1.275x0.15 1.95 0.651 0.373 0.243
3 1.275x0.07 0.96 0.516 0.086 0.044
4 0.47 x0.18 0.44 0.391 0.037 0.015
5 0.75 x0.20 0.51 -0.822 0.077 - 0.063
Axial force = ΣAσ = 0.787 MN
Moment about centroidal axis = ΣAσy = 0.373 MNm
In this simply supported bridge, neither axial force or moment are in fact restrained, so locked in
stresses are calculated by subtracting these effects from the stress diagram above:
— — =
4.7 SHRINKAGE
When the in-situ top is cast on the precast beams some of the shrinkage of the beams
has already occurred. Hence differential shrinkage occurs between the precast and
in-situ concretes, and this results in the development of a pattern of internal stresses.
Clause 7.4.3.4 states that the Table 29 shrinkage values may be adopted. It is reasonable
(and usual) to assume that half of the beam shrinkage has occurred at the time of
casting the top slab. Hence the differential shrinkage assumed in the calculation is
half of the Table 29 shrinkage value.
The effects of differential shrinkage will be reduced by creep. Allowance is made for
this in the calculations by using a reduction coefficient, . A value of 0.43 is normally
used for this coefficient, as given in Clause 7.4.3.4.
Shrinkage
Differential shrinkage between slab and deck creates internal stresses. It is assumed that half
the total shrinkage of the beam has taken place before the slab is cast.
150 µε
1.480m
— — =
-0.54 N/mm 2
-0.68 N/mm 2 +1.22 N/mm 2
5 APPLICATION OF LOADS
BD 37/88 considers five combinations of loads. These are listed in detail in Table 1
of the Standard, which also gives load factors to be used in each case. The five
combinations can be summarised as follows:
Comb. 1: Permanent loads plus primary live loads. (For railway bridges, secondary
live load is also included.)
Comb. 2: Wind load, plus loads in Comb. 1 (but with some reduced load factors).
Comb. 3: Temperature effects, again combined with loads from Combination 1.
Comb. 4: Secondary live loads (each considered separately), in combination with
permanent loads and the associated primary live load.
Comb. 5: Bearings friction, together with permanent loads.
For bridges in the UK, the requirements of BS 5400: Part 4 must be modified according
to Departmental Standard BD 24/92, The Design of Concrete Highway Bridges and
Structures, Use of BS 5400: Part 4: 1990. The most important change this introduces
relates to the Combination 1 loading. The beams must comply with Class 1 SLS
stress limits for a modified version of Combination 1. BS 5400: Part 4 calls for a
maximum of 25 units of HB load for this condition, but BD 24/92 reduces the live
loading to HA alone for this condition. This design example follows the requirements
of BD 24/92.
Positioning of the loads to obtain maximum bending moment elsewhere in the span,
or on skew bridges, is not so easy. The arrangement of loads which give maximum
effects in the various beams can be found by trial and error. Alternatively, some
software packages will automatically analyse a multitude of different possibilities and
report the maximum effects.
The temperature loads in Combination 3 do not cause any bending moments in the
beams, and so will not have a significant effect at ULS. Only Combination 1 therefore
needs to be analysed at ULS.
APPLICATION OF LOADS 39
Load Cases
Load cases must be selected for input to the grillage analysis. For the design of the prestressed
beams, only the maximum moments (which will occur at midspan), and the maximum shear at
the ends of the beams and at quarter span are needed. Moments are required both at SLS and
at ULS. Only ULS is required for the shear calculations, but the SLS condition will also be
analysed to give maximum loads on the bearings.
Below is a summary of the load cases to be analysed. This has been based on Figure 13 of
BD 37/88. Note that the HB vehicle is wider than a notional lane. When the HB vehicle
straddles the adjacent lane, the KEL is omitted from that lane, and the lane factor for the HA
UDL is based on a notional lane width of 2.5m, giving a lane factor of 0.789 (see Clause
6.4.2(b)).
HB vehicle
HA with 37.5 units HB:
Combination 1 at SLS and ULS HA, 2
= 0.789
Combination 3 at SLS HA, 1
= 0.9, KEL
Loading to grillages can be applied in the form of member loads and joint loads.
Member loads can include UDL, varying UDL, point loads and torque. These are
normally applied with reference to the local member axes. Joint loads can include
both forces and torques, and displacements (normally used for settlement of a support).
These are usually applied with reference to the global axes. In this example, only
UDL member loads and vertical joint loads will be used.
For many grillage software packages a pre-processor routine handles the application
of loads to the grillage model. In the example given here, however, a manual method
of distribution is given to illustrate the technique.
Uniformly distributed loads are generally applied as member loads along the main
elements. A patch load is statically distributed across the members beneath the patch.
Line loads can be applied as member loads, but the HA knife-edge load in this
example is applied as joint loads. Similarly, the HB wheel loads are statically distributed
between the nearest joints, and applied as joint loads.
HA loading alone
In the example opposite, Load 1 is the HA UDL loading for lanes 1 and 2, multiplied
by the lane factor of 0.9 for these lanes. Elements 179 to 234 represent four main
longitudinal beams, and receive member loads representing a width of deck of 1.275m.
The basic HA UDL is 37.3 kN/m on a notional lane width of 3.1m, so the load
applied to members 179 to 234 is:
The remaining load is allocated to the beams represented by members 165 to 178,
and 235 to 248. So for members 165 to 178:
The last section of data shows how the individual loads are combined with factors
equivalent to the factors γfL x γf3 as given in BD 37/88 Table 1 for Combination 1
at SLS.
APPLICATION OF LOADS 41
The input data for some of the Combination 1 load cases is listed below.
LOAD 4 FOOTWAY
MEMBER LOAD
165 TO 178 277 TO 290 UNI GY -5.10
151 TO 164 291 TO 304 UNI GY -0.90
LOAD 5 SURFACING
MEMBER LOAD
179 TO 276 UNI GY -5.1
HA with HB loading
The page opposite shows the input data for load cases in which the HB vehicle is
positioned for maximum edge zone longitudinal bending. Note the absence of KEL
in the straddled lane next to the HB lane. A similar load layout is also required with
the HB vehicle in the central lane.
The data for Load 1 is for 37.5 units of HB (375 kN per axle, or 93.75 kN per wheel).
The wheel loads are distributed between the nearest grillage joints, and applied as
point loads. Loads 2 to 7 are similar to Loads 1 to 6 on the previous page. Loads 8
and 9 represent Combination 1 at SLS and ULS respectively.
As before, the combinations are made from applying the γfL and γf3 factors from Table
1 of BD 37/88. Note that for ULS loadings, γf3 = 1.1. Thus the factor applied to
Load 1 in the final set of data is:
For reference, the diagram of the grillage model is repeated here, showing the relevant
member and joint numbers, the notional lanes, and the position of the HB vehicle:
151 164
165 21 22 25 26 178
179 36 37 40 41 192
193 51 52 55 56 206
207 66 67 70 71 220
221 234
235 98 248
291 304
Grillage output
Each software package has its own convention for output display. Generally the
member end moments at the same joint in any element string (eg. main longitudinal
member) are averaged to find the maximum design moment at that joint. The sign
convention may give opposite senses for each moment in a sagging effect, or they
may be given the same sign, so care is therefore necessary in the interpretation. The
design shear forces and twisting moments are evaluated in the same way.
LOAD 1 HB LANE 1
JOINT LOAD
66 67 70 71 FY -93.75 The data on this page refers to
51 52 55 56 FY -140.625 this arrangement of loads:
36 37 40 41 FY -93.75
21 22 25 26 FY -46.875 HB vehicle
HA, 2
= 0.789
LOAD 2 HA LANE 2 ( = .789)
MEMBER LOAD HA, 1
= 0.9, KEL
221 TO 234 UNI GY -12.1
207 TO 220 UNI GY -8.66
235 TO 248 UNI GY -8.66
LOAD 5 FOOTWAY
MEMBER LOAD
165 TO 178 277 TO 290 UNI GY -5.10
151 TO 164 291 TO 304 UNI GY -0.90
LOAD 6 SURFACING
MEMBER LOAD
179 TO 276 UNI GY -5.1
6.1 GENERAL
The left hand pages in this section are in the form of a commentary on the example
design calculations on the right hand pages.
All references to clauses and Tables in this section refer to BS 5400: Part 4: 1990.
The units used in the calculations are Meganewtons and metres. These have been
chosen in order to avoid the need to use large number of powers of ten, and to avoid
the need to convert between different units within complicated expressions. It should
be noted that the unit of stress of MN/m2 is numerically the same as N/mm2, and so
stresses have been quoted in the latter unit.
The grillage analysis does not include the dead load of the beams and the in-situ slab,
as these loads are carried by the prestressed beams alone. These loads must be added
by hand. Appropriate load factors for the different load combinations have been
included in the grillage analysis, but need to be applied to the dead loads at this stage.
The γfL values are obtained from BD 37/88 Table 1. Note that a value of 1.2 has been
taken for the dead load at ULS. The Standard permits a lower value of 1.15 to be
adopted, but the onus is on the designer to ensure that the assumed unit weight of
the structure is not exceeded. Consequently most designers prefer the conservative
value of 1.2 at this stage of the design.
The value for γf3 is taken as 1.1, as given in Clause 4.2.3 for methods of analysis
other than plastic methods, for which a value of 1.15 should be adopted.
PRESTRESSED BEAM DESIGN 45
The calculations which follow are for one of the seven internal beams. These will all be identical,
so the design is for the worst case.
The dead load bending moment is carried by the precast beam acting alone.
Loads are:
Total DL = 20.48 kN/m
DL precast = 14.03 kN/m
DL in-situ = 6.45 kN/m
Midspan moments for the SDL and LL are obtained from the grillage analysis. Appropriate load
factors have been included in the analysis, so the grillage output gives the factored moments. In
the table below, the bending moments for load cases 3 and 4 are taken from the grillage output
for midspan of an internal beam. Each case includes SDL, footpath loads, and HA load. Bending
moments in the table are all in MNm.
SLS ULS
Combination 1 Combination 3 Combination 1
Load case Nominal γfL Factored γfL Factored γfL x γf3 Factored
moment moment moment moment
The basic prestress design is carried out at SLS. Midspan stresses are calculated for
the various load combinations at the top and bottom of the beam, and at the top of the
slab. Prestress must be provided to ensure that the stresses limits are satisfied at the
time of transfer of the prestress (when the load on the beams is at a minimum), and
under maximum loadings.
The table in the calculations opposite list the relevant stresses due to all the different
loadings that need to be considered. Load cases 3 and 4 refer to the grillage load
cases for which the moments are listed on the previous page. Note that these all
include SDL, footpath loads and HA loads in the appropriate combination. The
stresses in the table for these load cases are simply obtained from the bending moments
by dividing by the section moduli.
The following loadings, and the limiting stresses which apply to them, are specified in
Clause 4.2.2 (as modified by BD 24/92) and tabulated opposite:
Tension:
Load Combination 3 : Clause 4.2.2(b) states that the section should be checked
as Class 2 or 3, but BD 24/92 specifies that Class 2 should be used, for which
Clause 6.3.2.4(a)(2) gives a limiting tensile stress for pretensioned concrete of
-0.45 √f cu = -3.2 N/mm2.
Compression:
Clause 4.2.2 (para 3) specifies that the section should be checked in compression
under the full Load Combination 1 and 3 loadings. Table 22 gives the allowable
compressive stress as 0.4 fcu. This gives 16 N/mm2 for the slab and 20 N/mm2
for the precast beam. However this value can be increased by 25% for the
upper surface of the precast unit in contact with in-situ concrete (Clause 7.4.3.2.)
provided that failure would be by tendon yield, because the in-situ concrete
confines the precast concrete. This increase is not used in this example; it will
be seen in the ULS bending check later in the calculations that this condition is
not satisfied.
PRESTRESSED BEAM DESIGN 47
Stresses at Midspan
The bending moments need to be converted into stresses in order to carry out SLS design.
DL moments are resisted by the beams alone, so stresses at the top and bottom of the beam
can be obtained using section moduli for the Y8 beam:
Grillage moments are resisted by the composite section. For example, stresses for SLS
Combination 1, HA alone, are calculated from the moment of 2.540 MNm given on page 45,
and the composite section moduli given on page 21:
Temperature difference stresses are included in Combination 3. A load factor γfL = 0.8 is applied
to the stresses calculated on page 35. Internal stresses due to shrinkage are included in all
combinations, if they have an adverse effect; these stresses are as calculated on page 37. All
stresses in the table below are in N/mm2. Figures in brackets do not have an adverse effect, and
are therefore omitted from the totals.
Dyform strand is used in the design example, initially stressed to 70% of its
characteristic strength. This gives an initial tension of 210 kN for each 15.2 mm
diameter strand.
It has been reported that problems have occurred when stressing Dyform strands
to 75% of their characteristic strength. Therefore, although Clause 6.7.1 allows
up to 75% to be used for pretensioning, some manufacturers elect only to
stress Dyform strand to 70%. This reduced value is used in the design example.
The chosen strand pattern should always have strands located at the
manufacturers’ standard positions.
PRESTRESSED BEAM DESIGN 49
Prestress Design
Prestress must be added to the net stresses in the table to comply with the stress limits.
For tension at the bottom of the beam, Combination 1 with HA alone is critical:
σbbeam = -19.59 N/mm 2
For Class 1 prestress (ie zero tension), the required prestress after losses must at least
overcome this stress. Taking a trial eccentricity of 370mm, and assuming 30% losses:
19.59 P Pe P P x 0.370
= + = +
0.7 A Z 0.5847 0.186
∴ P = 7.565 MN
For 15.2 mm Dyform strand, stressed to 70% characteristic strength, initial tension per
strand is 210 kN.
4 x 1200 mm = 4800
2 x 950 mm = 1900
8 x 160 mm = 1280
12 x 110 mm = 1320
11 x 60 mm = 660
37 9960
Centroid = 9960 / 37
= 269 mm from soffit
= 0.269 m
y = 0.639 m
∴ eccentricity = 0.639 - 0.269
= 0.370 m (as assumed)
50 SIMPLE BRIDGE DESIGN USING PRESTRESSED BEAMS
The prestressing force does not remain at its initial value. The prestress
transferred to the beams a few days after they are cast is less than the force
initially jacked into the strands. Prestress losses which occur at or before transfer
are due to:
(i) Relaxation of the strands.
(ii) Elastic shortening of the beam under the prestessing force.
Clause 6.7.2.2 states that the relaxation loss should be the 1000 hour value
obtained from BS 5896, which is 2.5%. An examination of strand manufacturers
data indicates that it is reasonable to assume that between 25% to 50% of this
occurs prior to transfer for long line pretensioning operations. For ease of
calculation, it is usual to assume 50% relaxation before transfer, and 50% after.
The elastic loss at transfer is calculated at the centroid of the tendons, and is
due to the compression arising from the prestressing force, after the initial
relaxation loss prior to transfer, and the self-weight of the beam. The latter is
included because the beam cambers during transfer, and hence has to carry its
self-weight as a simply supported beam.
The net force after transfer = initial force - relaxation loss - elastic loss.
Stresses at transfer due to the prestress alone are calculated here. These must
be added to the dead load stresses from page 47 to give the actual transfer
stresses in the precast beam.
The stress limits for compression are given in Clause 6.3.2.2(b) and Table 23,
and for tension in Clause 6.3.2.4(b)(1). The strength of the concrete at transfer
is referred to as fci which is less than the final strength fcu. In this case it is
assumed that fci = 40 N/mm2.
PRESTRESSED BEAM DESIGN 51
Loss of prestress
Stresses at transfer are now calculated and checked against allowable stresses.
Transfer stresses due to prestress alone:
σt = Pt - Pt e = 6.960 - 6.960 x 0.370
A Zt 0.5847 0.156
= 11.90 - 16.51 = -4.61 N/mm 2
Similarly
σb = 6.960 + 6.960 x 0.370
0.5847 0.186
= 11.90 + 13.85 = 25.75 N/mm 2
Check SLS stress limits at transfer for Class 1. The only stresses acting at this time are
beam DL and prestress. DL stresses are taken from page 47.
Compressive stress limit, fci/2 = 20 N/mm2
Compression at bottom, σb = 25.75 (prestress) - 6.68 (DL) = 19.07 N/mm2
Tensile stress limit = -1.0 N/mm2
Minimum stress at top, σt = -4.61 (prestress) + 7.96 (DL) = 3.35 N/mm 2
∴ Section is satisfactory at transfer
52 SIMPLE BRIDGE DESIGN USING PRESTRESSED BEAMS
Further loss of prestress occurs with the passage of time. Long term prestress
losses are due to:
(i) Shrinkage of the concrete as it cures.
(ii) Creep of the concrete due to the permanent compressive stresses.
(iii) Further relaxation of the strands.
These losses need to be evaluated to find the final (long term) prestressing
force. This final prestress is usually about 30% less than the initial prestress,
and it is the final value that must be used in the calculations for the design of the
beams.
Creep strain is given in Clause 6.7.2.5 as 48 x 10-6 per N/mm2 for cube strengths
in excess of 40 N/mm 2. However, if the stress exceeds one third of the cube
strength, the creep is no longer proportional to stress, and has to be increased
by a factor varying between 1.0 and 1.25 depending on the stress. A modification
factor, k, is used:
k = 1.25 when σb = fci /2 = 20 N/mm2
k = 1.0 when σb = fci /3 = 13.33 N/mm2
In this case, σb = 19.07 N/mm 2
so k = 1.22 by interpolation
The final prestressing force is obtained by subtracting the shrinkage loss, the
creep loss, and the relaxation loss from the force after transfer.
The design stresses in the precast beam are the stresses due to the final prestress
after all losses, as calculated on the opposite page, plus the stress combinations
tabulated on page 47. These stresses are then compared with the allowable
tensile and compressive stresses, also given on page 47.
PRESTRESSED BEAM DESIGN 53
Final losses
The prestressing force and eccentricity previously calculated are for midspan
where the dead and imposed load moments are maxima. At other sections the
moments are smaller, and in order to satisfy the stress limits, the eccentricity
needs to be reduced. This can be achieved by either debonding or deflecting
some of the strands. In this example debonding has been adopted. Some
manufacturers prefer not to use deflected strands, and will ask to manufacture
a debonded alternative design.
Stress limits at transfer will be critical at all sections other than midspan. The
computer program specifies the debonding required to limit the transfer stresses
to fci /2 in compression and -1 N/mm2 in tension.
Debonding of strands
Debonding will be used to reduce the prestress away from the midspan position. A computer
program has been used to limit the stresses in the precast beam at transfer to fci /2 in
compression, and -1.0 N/mm2 in tension. The program determines a strand pattern (compatible
with the already determined midspan pattern) that fulfils these criteria. The results are as
follows:
11 60 mm
12 110 mm
8 160 mm
2 950 mm
4 1200 mm
Debonding Details
After the prestress design has been carried out at SLS, a check should be made
of the situation at ULS.
The calculation of the ultimate moment capacity of the composite section which
is presented on the following page adopts a commonly used approximate method.
In this method, the pre-strain in the concrete, due to the prestress, is ignored.
Since the ultimate capacity of the section is normally (as in this case) found to
be well in excess of the ULS moment, a more exact calculation of the ultimate
capacity is not considered to be necessary.
The pre-strain in the strands is determined for the condition after all losses have
occurred.
The strains in the diagram are the additional strains due to loads. Hence the
pre-strain has to be added to the strand strain at each level to obtain the total
strain.
Note that the strain at the top of the slab is taken as 0.0035, the failure strain of
concrete. Ultimate failure is by crushing of the concrete, not by yielding of the
strands.
PRESTRESSED BEAM DESIGN 57
σ
1820 /1.15 = 1583
1583
317
1266
0.0035
Prestrain in strands,
εpc = Pf / Es As
= 5.349 MN / (200 x 10 3 x 37 x 0.000165 m 2)
= 0.00438
3
2
1
58 SIMPLE BRIDGE DESIGN USING PRESTRESSED BEAMS
Knowing the total strains, the strand stresses at each level can be obtained from
the stress - strain curve.
For this calculation, a constant compressive stress block of 0.4 f cu has been
assumed in the concrete. However, BS 5400 Part 4 would strictly require the
parabolic - rectangular stress block in Figure 1 to be used, because the neutral
axis is not in the flange (Clause 6.3.3.1(b)). A computer programme would
be required to perform the calculation rigorously. The approach adopted in the
example is conservative and adequate for most situations.
If the error was significant, the calculation would have to be repeated with a
modified neutral axis depth.
Because the outermost tensions have not yielded, Clause 6.3.3.1 requires the
moment of resistance to be at least 1.15 times the design moment.
PRESTRESSED BEAM DESIGN 59
Forces:
T1 = 11 x 0.000165 x 1414 = 2.566
T2 = 12 x 0.000165 x 1403 = 2.778
T3 = 8 x 0.000165 x 1387 = 1.831
T4 = 2 x 0.000165 x 892 = 0.294
T5 = 4 x 0.000165 x 614 = 0.405
∴ ΣT = 7.87 MN
Maximum ULS moment = 1.639 + 0.753 + 4.261 (Combination 1, 37.5 units HB,
= 6.65 MNm from page 45)
∴ Moment of Resistance > ULS Moment and section is satisfactory in flexure at ULS .
60 SIMPLE BRIDGE DESIGN USING PRESTRESSED BEAMS
6.6 SHEAR
A prestressed concrete beam can fail in two basic modes (see Clause 6.3.4.1):
At the support, the section is flexurally uncracked so only failure under mode (i)
need be considered.
Clause 7.4.2.2(a) allows the designer either to assume that all the shear is resisted
by the precast beam acting alone, or to calculate to shear resistance based on
the composite section.
The expression for Vco in Clause 6.3.4.2 is based on the elastic distribution of
stresses in a rectangular beam, and gives the shear force for which the maximum
principal tensile stress reaches the tensile strength of the concrete. For a flanged
beam, the expression is a conservative approximation.
A partial load factor γfL of either 0.87 or 1.15 has to be applied to the compressive
stress at the centroidal axis due to prestress, f cp. Obviously, 0.87 results in a
lower value of Vco, and so this value is used in the calculation.
The full prestress is not developed within the transmission length, which is
given in clause 6.7.4. Hence, if the support is within the transmission length, a
reduced value of prestress should be used in the calculations. Clause 6.7.4
states that a linear development of stress within the transmission length should
be assumed.
Maximum total shear force at support, for ULS Load Combination 1 with 37.5 unit HB,
V = 0.247 + 0.114 + 0.584 = 0.945 MN
The shear force will be assumed to be resisted by the precast beam alone,
as permitted by Clause 7.4.2.2.(a)(1).
where
h = 1400mm
b = 216mm (min)
ft = 0.24 50
= 1.70 N/mm 2
fcp = prestress at centroid of beam
The 0.4 term in the expression for required shear reinforcement in Clause 6.3.4.4
allows for degradation of shear capacity due to fatigue effects.
The effective depth dt is the depth to the outermost strands from the compression
face. Clause 7.4.2.2(c) makes clear that this should be derived for the composite
section, even when the calculation for Vc is based on the precast beam alone, as
here. Thus d t = 1590mm - 60mm = 1530mm, or 1.530m.
At the quarter span point the section is cracked in flexure at ULS. Hence both
modes of shear failure should be considered.
As for shear at the supports, the calculation for Vco is based on the precast beam
acting alone.
PRESTRESSED BEAM DESIGN 63
Ultimate shear resistance Vco is less than applied ULS shear force V, so shear reinforcement
must be provided according to Clause 6.3.4.4:
Shear also needs to be checked at other positions along the beam. These are the ULS loads at
the quarter span position. The central portion of the beam will be designed for these loads, with
the end quarters reinforced as calculated above. The section at quarter span must be checked both
as cracked and as uncracked in flexure.
(a) Section uncracked in flexure: Vco will be calculated assuming precast beam acts alone.
The debonding calculations give stresses at quarter span due to prestress and DL. From these,
stresses due to prestress alone are calculated by subtracting the effect of the DL:
ft = 2.98 - 0.932 / 0.156 = - 3.00 N/mm 2
fb = 13.68 + 0.932 / 0.186 = 18.69 N/mm 2
Interpolating for stress at centroid of section, 0.639m above soffit:
fcp = 18.69 + (0.639 / 1.400) x (-3.00 - 18.69) = 8.79 N/mm 2
In the case of the shear resistance for a section cracked in flexure, the calculation
based on the composite section is not too complicated, and results in a higher
value of Vcr than if the beam assumed to act alone. The method based on the
composite section is adopted here.
Note that M cr is less than the total moment at ULS. Hence the section would
be flexurally cracked at ULS.
(b) Section cracked in flexure: Vcr will be calculated for composite section.
Mcr is the moment which causes cracking, assumed to be when the stress in the extreme tensile
fibre reaches (0.37 fcu + fpt )= 0.37 x 50 + 0.87 x 18.69 = 18.88 N/mm 2.
The total cracking moment is a combination of the DL moments acting on the beam alone,
and the moment acting on the composite section:
Vertical shear forces in the beams always give rise to longitudinal shear forces.
Indeed, according to elastic theory, the longitudinal shear stress is equal in
magnitude to the vertical shear stress at any point. The construction joint
between the precast beam and the in-situ deck slab forms a weak plane, and the
longitudinal shear along this plane should be checked. It is often found that the
reinforcement provided for vertical shear is adequate for the longitudinal shear
requirements of Clause 7.4.2.3.
The shear capacity depends on the contact surface at the interface. It is usual
to assume a “rough as cast” surface, which only requires cleaning of the surface,
but no special preparation. This is defined as surface type 2.
The value of fcu to be used in the calculation is that for the weaker of the two
grades of concrete at the interface. Thus, based on the in-situ concrete, fcu = 40
N/mm2.
Both at the supports, and at midspan, the reinforcement provided for vertical
shear has also been found to be just adequate for longitudinal shear at the
interface. Note, however, that the calculated reinforcement requirement for
vertical shear at midspan was only 0.000218 m2/m, which would not have been
enough for longitudinal shear. More reinforcement than this was provided in
order to avoid exceeding the maximum permitted spacing.
PRESTRESSED BEAM DESIGN 67
The longitudinal shear along the top of the precast beam, at the interface with the in-situ deck,
must be checked to Clause 7.4.2.3.
Longitudinal shear is only generated by the shear force carried on the composite section.
At the support, maximum vertical shear on composite section is given on 61:
V = 0.584 MN
A = 1.275m x 0.220m x 0.91 = 0.255 m 2
y = 1.370 + (0.220/2) - 0.889 = 0.591 m
= 0.2429 m 4 (from page 21)
Vl = V A y / = (0.584 MN) x (0.255 m 2) x (0.591 m)/(0.2429 m 4)
= 0.362 MN/m
The surface is assumed to be Type 2 “rough as cast”. Two checks are required:
(a) k1 fcu Ls = 0.09 x (40 N/mm 2) x (0.400 m)
= 1.44 MN/m > Vl so OK
Similarly, at quarter span the vertical shear force carried by the composite section is
V = 0.458 MN
∴ Vl = 0.284 MN/m
For the vertical shear reinforcement of T12 links at 800 mm centres, Ae = 0.000283 m
vl Ls + 0.7Ae fy = 0.291 MN/m > Vl so OK
7 FINISHINGS
7.1 INTRODUCTION
Bearings
Joints
Waterproofing and surfacing
Parapets
Although they round off the design and detailing of a bridge deck, they should be
given very early consideration as they can have an important effect on the bridge
analysis and performance. For example, the bearing position on an abutment can
affect the span, access for inspection, maintenance and replacement. The consideration
of carriageway surfacing profile affects loading (eg. crowning or vertical curves).
The parapet type needs to be adequately anchored into the deck edge.
7.2 BEARINGS
Although the technology of bearing design can be left to the specialist suppliers’
mechanical engineers, the bridge designer should develop an appreciation of the
different bearing types to avoid problems and failures.
There are two main types of bearing commonly used with precast beams, namely
elastomeric and pot bearings. Elastomeric bearings have been chosen for the design
example, and they are by far the most common, probably due to their lower initial
cost. However, they may prove with time to have a more limited life.
Elastomeric bearings are generally laminated, and consist of alternate layers of steel
and rubber bonded together. Capacities depend on the area, height, number and size
of the layers. They accommodate vertical loads by compression, translation by shear
deformation, and rotation by variable compression.
In pot bearings, the confined elastomer is heavily loaded and is assumed to act like a
fluid to accommodate rotation. Translations are enabled by a sliding interface of
PTFE and stainless steel.
For elastomeric bearings, there are generally two possible methods of resisting the
horizontal forces. A pin can be provided at one end and a guide at the other end,
which carry all the horizontal forces. Alternatively, as in the example, the forces can
be resisted by the horizontal stiffness of the bearings; this situation is known as a
‘floating deck’.
FINISHINGS 69
Articulation
A floating deck articulation system will be used, so all 18 bearings will be the same.
Bearing positions
All bearings should be bedded horizontally, even when the bridge deck has a
longitudinal fall.
Care is needed on skew bridges to keep the axis of the principal movements
and rotations in harmony with the bearing axes. This may require a wider
bearing shelf .
End
Diaphragm
Deck slab
Y8 beam
Access
Gallery
Abutment
FINISHINGS 71
Deck slab
Y8 beam
Elastomeric bearing
Abutment
Bearing plinth
150 mm high
Detail of one end of the bridge span, showing cross section through abutment, bearing shelf and
elastomeric bearing. A drainage channel is provided at the back of the bearing shelf. Access to
the bearing shelf for inspection and maintenance will be from underneath the bridge.
72 SIMPLE BRIDGE DESIGN USING PRESTRESSED BEAMS
The bridge designer should carry out calculations of the range of loads and
movements that the bridge bearings will sustain. These are then specified on a
bearing schedule, and a bearing complying with these specifications will then
be selected (often by the contractor) using data supplied by bearing manufacturer.
Nominal horizontal loads have been calculated for the design example at the
same time as other nominal loads. The maximum horizontal loads for each
bearing are now calculated from these loads.
Vertical loads are obtained from the grillage analysis, except for the dead load.
The permanent vertical load on the bearings needs to be specified, as well as
both the maximum and minimum loads. On continuous bridges, as well as on
some skew simply-supported bridges, the minimum load can be less than the
permanent load.
FINISHINGS 73
Loads and deformations in all directions at SLS need to be calculated for entry into the bearing
schedule. This bridge has nine identical elastomeric bearings at each end.
Loads
BS 5400: Part 9: Clause 5.15.1 gives general guidance on the forces and
movements to be considered for bearings together with the relevant limit states.
For the design example, reversible and irreversible translations and rotations at
SLS are required.
Clause 6.7.2 of BS 5400: Part 4 gives guidance on shrinkage and creep, with
Clause 6.7.2.4 referring specifically to shrinkage, and Clause 6.7.2.5 specifically
to creep.
FINISHINGS 75
Movements
Longitudinal translation: irreversible shortening of the bridge occurs due to shrinkage and creep,
and reversible movements occur due to thermal expansion and contraction.
Shrinkage of concrete
- assume that half total shrinkage occurs before beams are placed on bearings.
Shrinkage/unit length = 300 x 10 -6 total
∴ Total shrinkage = 300 x 10 -6 x 26600mm= 8.0 mm
12 shrinkage occurring after beams are placed = 8.0 / 2 = 4.0 mm
/
Creep of concrete
- again assume that only half total occurs after beams are placed on bearings.
Creep/unit length = 48 x 10 -6 per N/mm2
From calculations for prestressed beam design,
average stress at centroid 10 N/mm2
∴ Total creep = 48 x 10 -6 x 26600 mm x 10 N/mm 2 = 12.8 mm
12 creep occuring after beams are placed = 12.8 /2 = 6.4 mm
/
Assume that total of shrinkage and creep shortening is shared equally at each end
Nominal irreversible movement = 1/2 (4.0 + 6.4) = 5.2 mm
SLS irreversible movement = γfL x γf3 x 5.2
∴ Irreversible translation at bearings = 1.0 x 1.0 x 5.2 = 5.2 mm
Temperature range
Nominal temperature range= 47°C
Coefficient of thermal expansion= 12x10 -6 /°C
∴ Nominal range of movement = 47 x 12 x 10 -6 x 26600 mm
= 15.0 mm
Assume bearings are fixed in centre of range
∴ Nominal range of movement = ±7.5 mm
Temperature effects are the only effects (except transient traffic loads) causing reversible
translations. Assume these reversible movements are shared equally between the two
abutments:
Nominal translation at bearing = ±7.5 x 1/2 = ±3.8 mm
SLS reversible translation = γ fL x γf3 x (±3.8)
∴ Reversible translation at bearings = 1.0 x 1.0 x ( ±3.8) = ±3.8 mm
76 SIMPLE BRIDGE DESIGN USING PRESTRESSED BEAMS
Irreversible rotations due to the self-weight of the precast beams take place
before the bearings are fixed in position, so do not affect the bearings. However,
rotations due to the dead load of the slab and the superimposed dead loads
must be absorbed by the bearings, and are calculated opposite.
Dead load rotation must be calculated by hand, based on bending of the precast
section acting alone. SDL rotation is most conveniently estimated by applying
the SDL load to the composite section, and calculating the rotation by hand,
ignoring any load distribution between the beams.
Reversible rotations due to live loads can be extracted from the grillage analysis.
All grillage load cases in the design example include SDL, which must therefore
be subtracted to get the reversible effect only.
FINISHINGS 77
Rotation
The total rotation at the bearing is made up of a number of components. The camber of the
beams due to the dead load and prestress occurs before the bearings are installed, so this does
not cause rotation of the bearings. Sag of the beams due to the in-situ slab and SDL causes
irreversible rotations. Creep causes hogging of the beams in the long term, leading to irreversible
rotation in the opposite direction. Live load obviously causes reversible rotations.
Irreversible rotation at SLS is made up from DL and SDL rotations, with appropriate load
factors:
SLS rotation = (0.00125 x 1.0 x 1.0) + (0.00048 x 1.2 x 1.0)
= 0.00183 radians
From grillage analysis output, which includes SDL but not DL,
maximum rotation at SLS = 0.00281 radians
The reversible (live load) rotation is thus this value with the SDL subtracted:
Reversible rotation at SLS = 0.00281 - (0.00048 x 1.2 x 1.0)
= 0.00223 radians
Maximum rate will occur when the beam is acting alone, with no composite action. This is when
the bearing rotates under the weight of the in-situ slab:
Rotation = 0.00125 radians
Reaction = wl /2 = (6.45 kN/m) x (26.6 m) /2
= 85.8 kN
Rate = rotation /100 kN
= 0.00125 / 0.858
= 0.00146 radians /100 kN
78 SIMPLE BRIDGE DESIGN USING PRESTRESSED BEAMS
Bearing Schedule
The results of the calculations from the previous pages are entered into a bearing
schedule. BS 5400: Part 9.1 contains a standard form of bearing schedule.
The completed schedule for the design example is shown below, and the page
opposite shows how a bearing can be chosen from a manufacturers’ catalogue,
and checked for compliance with the schedule.
Bearing Selection
Check stiffnesses:
40 mm wearing course
60 mm regulating course
10 mm waterproofing
Two types of waterproofing are in common use, namely sheeting and spray membranes.
Although generally thought of as more expensive, the spray versions are gaining in
popularity due to their more comprehensive coverage, and due to the fact that they
reduce the overall dead load.
Drainage of the waterproofing in areas near joints or severe crossfalls where ponding
may occur will improve the durability of the surfacing in these areas.
Reference should also be made to the Highways Agency Standard BD 47/94. Further
information can be found in TRRL Research Report 185: “A field trial of waterproofing
systems for concrete bridge decks” by A. R. Price.
FINISHINGS 81
Deck waterproofing
Waterproofing detail
82 SIMPLE BRIDGE DESIGN USING PRESTRESSED BEAMS
7.4 JOINTS
The function of any joint is to bridge any gap or discontinuity between the deck and
the abutment, whilst allowing horizontal movement and rotation of the deck to take
place.
They are subject to all the effects applied to bearings, and are divided into two main
categories: buried joints and mechanical joints.
Buried joints have continuous surfacing over the structural discontinuity, and are
usually reserved for situations where the movements are less than 15 mm. The design
example uses asphaltic plug joints at each end of the bridge span. A steel plate
bridges the gap between the deck and the abutment, and a strip of flexible asphalt
placed over it along the line of the joint.
Larger movements are often accommodated with mechanical joints. In this case the
surfacing is discontinuous. An example is illustrated below.
Deck Abutment
wall
FINISHINGS 83
Steel plate
Surfacing Asphaltic plug joint
Deck slab
20mm PVC
Y8 beam drain
Abutment
wall
7.5 PARAPETS
There are five parapet groups for highway bridges referred to in Highways Agency
Standard BD 52/93:
Standard metal versions in steel or aluminium are available from the various suppliers
listed in BS 6779 Part 1. There are differences between the types which affect the
edge plinth detailing and spacing of posts. Reinforcement should be detailed to
accommodate standard approved anchorage cages. Most designers give the contractor
the choice of steel or aluminium parapets, although individual client authorities may
have preferences relating to maintenance aspects.
Concrete can also be used for the P1 and P2 containment requirements, sometimes
faced or sandwiched with other materials such as brick.
P6 high containment parapets are often provided in the form of precast concrete
panels, which are either cast into the edge of the deck, or bolted down onto it. The
edge of the deck must be designed to resist the high transverse bending moment that
a vehicle hitting the parapet can generate; see BS 6779: Part 2. The bending moment
is too high to be resisted by the bending strength of a standard deck slab, or the
torsional strength of M or Y beams. A torsionally stiff edge beam should be provided,
or the base of the parapet panels should span over two adjacent beams.
Precast parapet
panels, 2.5m long,
bolted to deck.
A typical edge detail for a bridge deck with a high containment parapet.
FINISHINGS 85
Mesh infill
Parapet details
86 SIMPLE BRIDGE DESIGN USING PRESTRESSED BEAMS
8.1 INTRODUCTION
The design of prestressed beams in a solid slab deck follows the same pattern as for
a beam & slab deck. This section consists of a partial design example of a solid slab
bridge deck, to demonstrate how to deal with the transverse moments and shears
using reinforcement threaded through the web holes of the precast beams.
Inverted T beams were chosen as the most suitable form of construction for this short
span. The beam selection charts indicate that T2 beams are appropriate for a span of
about 9m. The beams are placed side by side in this form of construction, so for the
width of deck required, eighteen T2 beams are needed. The bridge could alternatively
have been designed using twelve TY2 beams.
In-situ concrete provides the infill between the beams, and the topping over the beams.
The standard thickness for the topping is 75mm, and this is used here. The topping
should never be specified less than 75mm thick.
SOLID SLAB DESIGN EXAMPLE 87
The grillage used to analyse this bridge deck has nine longitudinal members, each
representing two inverted T beams with their associated in-situ concrete:
BS 5400: Part 4: Clause 7.4.1 does not require a modular ratio to be used for concrete
strengths varying by 10 N/mm2, and so no account has been taken of the difference
between the precast and the in-situ concrete in the analysis of this bridge. The section
properties of the internal longitudinal members are therefore based on a simple
rectangular shape.
The edge members have different properties, due to the geometry of the edge detail,
and the presence of service ducts under the footpath.
Transverse members have been provided to divide the span into eight equal segments
of 1.136m. The beams oversail the bearings by 0.605m on this bridge, so the members
along the line of the bearings represent a similar width of slab to the members within
the span.
Because the skew of 21° is not too large, the transverse members have been positioned
parallel to the abutments. This leads to a simple grillage layout, and is reasonably
accurate for small skew angles. Decks with a large skew should be analysed with a
grillage in which the transverse members are at right angles to the longitudinal beams.
Grillage Analysis
Span = 9.089 m
1.016
1.136
The difference between the strength of the precast concrete beams and the in-situ concrete is
10 N/mm2, and so a modular ratio of unity will be assumed.
Thickness of slab assumed for transverse members is down to web holes of T beams,
which are 0.175m above soffit of beams.
∴ Slab thickness = 0.495 - 0.175 = 0.320 m
Area = 0.320 x 1.136 = 0.364 m 2
= (0.320 3 x 1.136) /12 = 0.0031 m 4
C = (0.320 3 x 1.136) / 6 = 0.0062 m 4
90 SIMPLE BRIDGE DESIGN USING PRESTRESSED BEAMS
Transverse sagging reinforcement can only be provided through the web holes of the
precast beams, which are at 0.610m centres. The sagging moment is therefore
recalculated for a 0.610m section of slab.
Note that the depth to the reinforcement is the same as the depth assumed for the slab
in the analysis.
Transverse reinforcement of 3T16 bars through each web hole is used. This is preferred
to a single T25 bar as the extra flexibility will make it easier to thread the bars through
the web holes on site, particularly if these are not perfectly aligned.
Reinforcement of 4T16 through the web holes is often used for the larger inverted T
and TY beams.
Note that Clause 5.3.3.2 also has a requirement for additional longitudinal shear
reinforcement in the tensile zone. In this case, this only amounts to 83mm2 per 0.610m
section. This requirement is easily fulfilled by the excess of bending reinforcement
provided. Furthermore, the maximum bending and the maximum shear will not occur
at the same place, so this requirement will rarely actually result in extra reinforcement
being needed.
Clearly the concrete and the tensile steel have plenty of capacity to resist the shear
between the beams. The interface between the precast beams and the in-situ concrete
also needs to be considered. The shape of the precast beams allows shear to be
transferred by interlocking between the beams and the in-situ concrete. Additionally,
shear can be assumed to be transferred by dowel action of the reinforcement through
the web holes:
Dowel action shear capacity = 0.7 A s (0.87 f y ) = 0.7 x 0.000603 x 400
= 0.169 MN
This is well in excess of the ULS shear of 0.066 MN for a 0.610m section of deck.
SOLID SLAB DESIGN EXAMPLE 91
Reinforcement is provided through the web holes, which are at the standard spacing of 0.610m.
Transverse Shear
Hogging moments can occur in this bridge deck near the edge beams, if the
edges of the deck are loaded, but there is no load near the centre. Note that the
maximum hogging moment is much lower than the maximum sagging moment
on the previous page.
A142 mesh will be used in the topping throughout the deck. This nominal
reinforcement is used in the topping concrete which is predominantly in
compression. However, at the edges of the deck where hogging occurs, this
reinforcement is not enough, and T6 bars are added.
SOLID SLAB DESIGN EXAMPLE 93
The A142 mesh provides a reinforcement area of 142 mm2/m, which is insufficient for the
hogging at the edges of the deck. Additional transverse reinforcement of 2m lengths of T6 bars
at 200mm centres will therefore be added in the topping at each side of the deck.