View PDF
View PDF
View PDF
Multiple Tee-beam
Bridges
G. Pircher
ii
Master Honours
2006
iii
iv
Synopsis
cast-on-site concrete slab are often a cost-effective way of constructing simply-supported and
multi-span bridge structures in many countries world-wide. For the design of these bridges
practical design work on this bridge type. A chapter on the modelling of various loading
bearing system, the Tee-beams, and the lateral load-bearing system, the roadway slab. A
summary of commonly used bridge systems in various countries is also included. All this
specifications.
The information included in this thesis has been used to define specifications for the
Material included in this thesis has also been published in the following conference
proceedings:
Pircher G., Pircher M. (2004) “Computer-aided design and analysis of multiple Tee-beam
v
vi
Acknowledgments
It has been a great experience for me to work on this thesis under the supervision of M.
Pircher, A.T. Wheeler and R.Q. Bridge. I have been encouraged and supported in the best
possible way over the past 3 years and I feel privileged to have had the opportunity of
completing my thesis as part of this team. I found it particularly motivating that the exchange
I need to thank my family for the support and understanding that long nightshifts became
Further thanks go to my father H. Pircher for giving me the chance to work on many bridge
projects over the years which allowed me to collect a certain degree of practical experience. I
never want to miss the efficient and meaningful technical discussions; they represent a major
vii
viii
Contents
Synopsis......................................................................................................................................v
Acknowledgments ....................................................................................................................vii
Contents .....................................................................................................................................ix
Notation ...................................................................................................................................xiii
1 Introduction ........................................................................................................................1
1.1 Goal of the thesis: .......................................................................................................1
1.2 The Pre-cast Multiple Tee-beam Bridge: ...................................................................2
1.3 History of pre-cast pre-stressed beams:......................................................................3
1.4 Advantages and disadvantages of pre-stressed girders.............................................10
1.4.1 Advantages of pre-cast pre-stressed girders. ....................................................10
1.4.2 Disadvantages of pre-cast pre-stressed girders: ...............................................11
1.5 Transportation and Construction Considerations .....................................................11
1.6 Assumptions for the geometry of Multiple Tee-beams ............................................13
1.7 Multi-Span Pre-cast Tee-Beams...............................................................................15
1.8 Erection Procedures..................................................................................................16
1.8.1 Typical erection procedures..............................................................................16
1.8.2 Erection of pre-cast beams: ..............................................................................16
1.8.3 Erection of cast in-situ Tee-beam bridges: .......................................................18
1.9 Codes requirements for design of multiple Tee-beam bridges.................................20
1.9.1 Internal forces:..................................................................................................20
1.9.2 The serviceability criteria: ................................................................................20
1.9.3 Ultimate moment capacity:...............................................................................21
1.9.4 Ultimate shear and torsion capacity: ................................................................21
1.9.5 Longitudinal shear: ...........................................................................................22
1.9.6 Other checks: ....................................................................................................22
1.10 Summary...................................................................................................................22
2 Loading.............................................................................................................................23
2.1 Introduction ..............................................................................................................23
2.2 Loading directions ....................................................................................................24
2.2.1 Vertical loading ................................................................................................24
2.2.2 Transversal loading ..........................................................................................25
2.2.3 Longitudinal loads ............................................................................................26
2.3 Permanent Loading...................................................................................................27
2.3.1 Self weight........................................................................................................27
2.3.2 Time-Dependent Effects...................................................................................27
ix
2.4 Additional loading.................................................................................................... 29
2.4.1 Wind loading.................................................................................................... 29
2.4.2 Traffic Loading ................................................................................................ 30
2.4.3 Temperature Loading....................................................................................... 31
2.4.4 Settlement......................................................................................................... 32
3 Pre- and post-tensioning................................................................................................... 35
3.1 Pre-stressing – principles, materials and applications ............................................. 35
3.2 Full and partial Pre-stressing.................................................................................... 38
3.3 Pre-stressing methods .............................................................................................. 39
3.4 Development length of pre-stressing strands ........................................................... 40
3.5 Pre-stressing losses .................................................................................................. 42
3.6 Primary and secondary effects ................................................................................. 44
3.7 Consideration of pre-stressing for SLS and ULS design code checks..................... 46
3.8 Precamber and application for pre-cast pre-tensioned members ............................. 48
4 Numeric modelling of the roadway in Tee-beam bridges................................................ 51
4.1 Introduction.............................................................................................................. 51
4.2 Modelling Systems................................................................................................... 51
4.2.1 Transversal beam elements – grillage model................................................... 51
4.2.2 Finite elements for the roadway slab ............................................................... 53
4.2.3 Finite elements versus grillage......................................................................... 54
4.3 Number of transverse elements per span ................................................................. 55
4.4 Stiffness of transverse elements in grillage models ................................................. 57
4.5 Principal stresses, shear and torsion in the roadway slab ........................................ 64
4.6 Connection of transverse to longitudinal members.................................................. 69
4.7 Summary .................................................................................................................. 73
5 Numeric modelling of the main Girders in Tee-beam bridges ........................................ 75
5.1 Basic Considerations................................................................................................ 76
5.1.1 The cross section of the main girder ................................................................ 76
5.1.2 The shear lag effect .......................................................................................... 78
5.1.3 The orientation of the principal axes in non-symmetrical cross-sections........ 82
5.1.4 Main girders in torsion..................................................................................... 84
5.1.5 The subdivision of the girder into structural elements..................................... 89
5.1.6 Connection between girder and supports ......................................................... 91
5.1.7 Continuity ........................................................................................................ 92
5.2 Composite Action .................................................................................................... 95
5.2.1 Change of cross-section properties in composite beams ................................. 95
5.2.2 Longitudinal shear in composite interfaces ..................................................... 98
5.3 Curvature in plan.................................................................................................... 101
x
5.3.1 Transportation and stability of pre-cast pre-stressed girders..........................102
6 Pre-cast Tee-beams worldwide.......................................................................................105
6.1 Introduction ............................................................................................................105
6.2 Great Britain ...........................................................................................................105
6.3 United States...........................................................................................................110
6.3.1 US – California (www.dot.ca.gov).................................................................112
6.3.2 US – Florida (www.dot.state.fl.us)................................................................117
6.3.3 US – Minnesota (www.dot.state.mn.us)........................................................119
6.3.4 US-Washington (www.wsdot.wa.gov)..........................................................121
6.4 Japan .......................................................................................................................124
6.5 Australia .................................................................................................................125
6.6 Malaysia and Indonesia ..........................................................................................129
6.7 Europe.....................................................................................................................132
6.8 Summary.................................................................................................................136
7 Conclusion......................................................................................................................139
8 References ......................................................................................................................143
xi
xii
Notation
The following short terms and expressions are used in the chapters of this thesis. All symbols
Abbreviation:
CP Connection Point defining the location where structural elements are joined
CS Cross-section
FE Finite elements
MG Main girder
PT Post- or pre-tensioning
xiii
Constants and Variables:
xiv
b Width of a cross section [m]
xv
xvi
1 Introduction
Multiple Tee-beams are often chosen as the preferable structural system for bridge girders.
The reasons for selecting this system are numerous and include savings in material and
reduction in self-weight. Generally, Tee-beams are easier to cast than other cross-sections
with the formwork quick and easy to assemble. An additional advantage is that the girder
height is generally small which is often an important argument for aesthetics in urban areas.
Changes in width of the roadway may also be implemented simply by changing the number of
Figure 1-1. KS7 Selzthal (Austria) – Transition from triple Tee to quadruple Tee-beam.
This type of construction is used world wide utilising numerous construction methods and
falling under the restriction of most major national design standards. To carry out the analysis
and design of these structures takes many forms, but the current trend is towards numerical
In this thesis a number of specific problems arising during the modelling and analysis of
bridges consisting of multiple Tee-beams will be investigated. The focus will primarily be on
1
In the numerical modelling of the bridges a number of aspects need to be addressed. These
include the modelling and design of the roadway slabs and the modelling and design of the
longitudinal girders. In looking at these aspects, the basic theory behind the choice between
the beam element or finite elements when analysing and designing multiple Tee-beam bridges
will be discussed. The other significant factor in the design of the Tee-beam bridges is the
effect of the prestressing, how this affects the behaviour, what needs to be considered during
modelling and the varying uses for the prestressing in differing environmental conditions.
A comprehensive study into the various types of multiple Tee-beams used internationally has
been carried out and a brief summary of the various aspects presented in Chapter 6.
The aim of presenting this information in this thesis is so that it can be used to assist the
design engineer in the development of a software solution for both the modelling and design
The option of using pre-cast members is often advantageous when the overpass is to be
erected over areas that temporary structural supports are not applicable, such as roadways
under traffic conditions. A survey by Slatter (1980) for the 11th IABSE Congress in Vienna
found that 71% of multiple Tee-beams were in fact double-Tee-beams, the typical multiple
Tee-beams have spans of 20m to 35m and span-to-depth ratios of no less than 1/20. In many
cases the geometric situation at the bridge site did not allow the bridge support axes to lie
perpendicular to the longitudinal axis. This was further complicated by the fact that skew ends
produce Tee-beams of differing lengths within the same span, or complicated geometric
situations at the ends that must be appropriately modelled. These variations provide
significant challenges when comparing the design process for that of a hollow box bridge
2
Multiple Tee-beam bridges built as pre-cast girders are usually produced in plants using high
grade concrete under controlled conditions, this generally results in higher-quality products
with longer life expectancy. The added advantage is that using this method it is possible to
cast the entire structure in the plant and transport it to the site for an erection as a whole. This
method was used for the construction of the 24 mile Lake Pontchartrain Bridge near New
Orleans, Louisiana, US. Shown in Figure 1-2 this bridge is considered to be the longest bridge
in the world, with numerous spans each 19.0m long and 11m wide, pre-cast in a yard and
The first pre-stressed pre-cast girders date back to 1886 when P.H. Jackson patented the
system in San Francisco, California. In 1888 the German engineer C.E.W. Doehring
independently obtained a patent for pre-stressed concrete slabs with metal wires.
However, these early attempts were not successful because the pre-stressing losses due to
creep and shrinkage of the concrete were significant. The credit for successfully developing
the modern concept of pre-stressed concrete goes to the French engineer Eugene Freyssinet,
who demonstrated the usefulness of pre-stressing using high-strength steel to control pre-
stressing losses in 1941 [Steinman and Watson, 1957]. Freyssinet started in 1941 with a 60m
3
segmentally constructed, two-hinged, portal-framed bridge over the Marne in Luzancy in
France and followed by five other nearly identical bridges, Freyssinet proved the effectiveness
In the US the first major pre-stressed concrete bridge was the three span (25, 54, 25m) precast
Figure 1-3 and Figure 1-4 show different views of the bridge which is still in service after a
general renovation in 1993. Following its construction the Bureau of Public Works, now the
Bridges” which was revised in 1954. This can be considered as the first design code for this
type of construction.
Figure 1-3. Overall view of the Walnut Lane Bridge in Philadelphia, Pennsylvania.
Since then pre-stressed concrete bridges have been progressively replacing reinforced
concrete bridges and steel girder bridges as the choice in the small and medium span bridges.
According to the National Bridge Inventory (NBI) pre-stressed concrete bridges are the most
commonly built bridges today, almost 50% of all bridges built in the USA are of this type
[Dunker and Rabbat, 1990]. Consequently, the authorities in most states in the US have
developed specific rules and guidelines for the use of this bridge type. Today the name
4
“AASHTO” girder is a common expression for pre-cast and pre-tensioned bridge in Asian
countries.
Figure 1-4. Support detail of the Walnut Lane Bridge in Philadelphia, Pennsylvania.
The examples for the use of multiple Tee-beams are numerous and the method of using pre-
cast beams has been brought to new level producing spectacular structures. The following
figures show a selection of bridges illustrating that although the span length is limited when
using pre-cast beams the resulting bridges can be architectural pleasing and big in size.
Opening for traffic June 1997, the Confederation Bridge (See Figure 1-5) links New
Brunswick to Prince Edward Island across the Northumberland Strait. It is currently the
5
longest bridge in the world to cross ice-covered salt water. All the beams and piers were pre-
cast on site. At 12.9 km in length the Confederation Bridge was planned and designed in
Calgary with strong links to the University of Calgary. The design also included onsite
A more recent example of a large Tee-beam bridge is the Kien - Bridge in Vietnam (see
Figure 1-6). For this project pre-cast beams have been used for the approach spans. The
Figure 1-7. Bottom flanges of pre-cast beams of the Mullingar Ballymaon Road Realignment
Shandonagh Bridge
Figure 1-7 and Figure 1-8. This bridge is a typical example for the use of pre-cast girders in
6
bridges with no specific aesthetic requirements. The support axes are skew, the bridge is
almost as wide as the span. In this case the use of a different erection sequence would have
Figure 1-8. Skew support axes of the Mullingar Ballymaon Road Realignment Shandonagh
Bridge
The Mullingar Ballymaon Road Realignment Shandonagh Bridge is a 3 span structure with
Typically the pre-cast beams are of lengths equal to the spans, this is the case for the Pont
d’Ouche in France (see Figure 1-9) For the Pont d’Ouche bridge the ends of the beams lie on
the support structures as shown in Figure 1-10. The slab however is continuous which is
achieved by an additional pre-stressing passing through over the support and through the cast
7
Figure 1-9. Overall view of the Pont d’Ouche structure in France.
A slightly different example is the CPCI bridge in Spain shown in Figure 1-11. The bridge
represents a construction method that uses two different types of pre-cast beams in the one
bridge structure. The first and third beam overhangs the span by 2/10 into the adjacent spans,
the second span is therefore a different type being much shorter that drops between and rests
8
Figure 1-11. CPCI bridge in Spain
The use of pre-cast beams is manyfold and there are almost no limitations for the structural
systems. All structural parts can be built using pre-cast beams and structures consisting of
pre-cast columns or piers, pre-cast diaphragms at the support axes and pre-cast beams for the
spans have been released. Such an application is the Queretaro Bridge in Mexico City built
Precast
footing
Figure 1-12. Queretaro Bridge in Mexico City composed of pre-cast elements.
The 3 span bridge (spans of 15, 24 and 15m) contains a total of 36 pre-cast elements. The use
of pre-cast elements enabled the contractor to complete the structure in only 90 days.
9
1.4 Advantages and disadvantages of pre-stressed girders
The multiple Tee-beam structure is a very popular application for the use of pre-cast beams
with cast-in-situ concrete slabs. The pre-cast beams can be prepared in a convenient way in
specific castings yards. When considering the use of this type of structure the following
• The members may be cast in a controlled environment and then moved to site
• The system of pre-cast beams and cast in situ concrete deck are quick and easy to
erect.
• The pre-stressing has the advantage that tension cracking can be eliminated in a pre-
stressed structure. The uncracked structure leads to a more efficient design than that
the entire section, not just the uncracked section, is made to resist compression.
• Pre-stressing reduces the diagonal tension. Use of inclined tendons reduces the shear
carried by the webs, thus producing lighter sections and thus savings in transportation
• The smaller girder depths that are possible with pre-stressed concrete members.
• Pre-stressed concrete girders with large top flanges provide working space during
• Pre-stressed concrete have relatively longer service life as outlined by Jerzy Zemajtis
(1998).
10
1.4.2 Disadvantages of pre-cast pre-stressed girders:
• Aesthetically pleasing shapes are limited, with the geometry and spans limited by the
casting yard.
The dimensions of the pre-cast beams are generally defined by two parameters the length and
cross-section. Typically a maximum length of up to 40m can be transported with trucks (both
weight and length are an issue), while the cross-section including the width and height give
While not used extensively, the multiple Tee-beams do not have to utilise pre-cast beams, but
may also consist of cast in-situ bridges. While these bridges are outside the scope of this
thesis Figure 1-1 shows a typical example where the geometry given i.e. the widening of the
deck in this case, does not allow for the easy use of pre-cast beams thus cast in-situ is a viable
alternative. Another example of this type of bridge is the Muuga bridge in Estonia shown in
Figure 1-13. The bridge completed in 2005 was erected using the span-by-span erection
11
Figure 1-13. Formwork of the Muuga Bridge in Estonia
The majority of the multiple Tee-beam bridges utilise the cast and transport technique. A
good example of where the pre-cast members were advantageous is the Katharine River
Bridge in Australia as shown in Figure 1-14 For this bridge girders were cast in the casting
yard (see Figure 1-15) and then transported to the remote location of the bridge and put in
place.
Figure 1-14. Placing final span Katherine River Bridge July 2002
12
Figure 1-15. the pre-cast beams used for Katherine River Bridge.
All bridge structures have a three-dimensional geometry and any structural model set up for
the analysis must reflect this. The simplifications into two-dimensional models often neglects
important effects and are therefore not appropriate, or conservative assumptions are made to
approximate effects thus giving ineffective structures. A modern software tool for bridge
engineers should provide the means to define a detailed 3-D representation of a given bridge
structure that models all geometric and material effects. In order to model the exact geometry
The plan view and the elevation of such bridges are governed by the specific geometric
requirements of the road alignment; these requirements have repercussions on the cross-
section along the bridge. A cross-fall of the road slab leads to inclined top slabs and the
possibility of Tee-beams with different height levels. Additionally, skew ends are often found
in such bridges and attention must be paid to model these support conditions correctly as
detailed later. These geometrical requirements often have significant structural effects and
should all be considered in an appropriate structural model. Furthermore, the modelling of the
13
geometry of off-ramps and road widening also poses challenges to the development of an
accurate model.
The difference between the actual cross section geometry and the cross section geometry in
the model represents another source of potential modelling error. All cross sections of bridges
have a cross fall, while it is common practice to consider the cross section as being perfectly
Actual cross
section shape
Figure 1-16: Real cross section with cross fall and simplified cross section for structural model
As a consequence some of the non-symmetric behaviour of the bridge will not be modelled. A
similar problem occurs for vertical alignment, the longitudinal slope is in most cases not
considered when modelling the bridge deck. The typical longitudinal slope is in the range of -
5% < slope < +5% and the effect on the structural analysis can be neglected. When the slopes
exceed these limits, then the slope should be a consideration within the structural model as
detailed by Arthur Nilson (1978). The simplification however can be crucial since dimensions
14
Tangent point
Longitudinal slope
Longitudinal slope
Rounding“R“
Structure considered
flat for overall analysis.
The most common way of assembling multi-span Tee-beam bridges is to use simply
supported pre-cast Tee-beams supported between piers then connected by a cast in-situ
concrete slab. The casting sequence of the concrete slabs is often geared towards reducing the
locked-in stresses in the final composite system. Shear connectors between pre-cast beams
and concrete slab ensure that forces are transmitted fully between the slab and the pre-cast
components.
In Australia, the “Super-Tee” system is an example for this method, these bridges are widely
accepted by the industry as the most efficient means of construction for most bridges and
consequently the design follows proven standard procedures. These design procedures are
based on a number of assumptions and it is possible to improve the efficiency of the structures
by utilising existing modelling software, but in developing the model the designer must
• The change in structural system at the piers where the pre-cast beams are simply
supported and made partially continuous by the concrete slab at a later time. Full
15
• Differential creep and shrinkage in the individual parts of the concrete-concrete cross-
section.
• Change in cross-section properties including the shift in the centre of gravity due to
While not discussed in detail in this thesis it was thought prudent to include a brief outline of
the erection procedure to obtain an overall understanding of all considerations for this type of
structure. When dealing with pre-cast elements and in particular with multiple Tee-beams an
There are two principle erection methods for multiple Tee-beam structures: either as cast in-
situ bridges, (see Figure 1-13), or the erection as composite structures consisting of a number
of pre-cast beams with a cast in-situ concrete deck (see Figure 1-7). The most economical
system is the pre-cast beams used with the cast in-situ concrete for the roadway deck/slab. In
this case the cross-section consists of parts with different concrete quality and different
Generally speaking the erection procedure is dependent on the design engineers and the
constructor. However, some regulatory bodies do give clear advice of how pre-cast Tee-beam
16
Figure 1-18. Typical construction sequence for 2 simply supported pre-cast beam (proposal
from Washington State Department of Transportation, US)
A typical example of this is the Washington Department of Transportation, in which they give
guidelines for the construction of single span simply supported bridges with pre-cast beam.
They also provide guidelines for two span simply supported and continuous bridges with pre-
requirements for the erection of multiple simple supported pre-cast girders, also provided are
As comparison also included in this document are the guidelines for the bridge that is
continuous over the centre support. In Figure 1-18 the continuity over the supported is not
achieved, while there is a cross beam closing the gap between the ends of the pre-cast
girder,
17
Figure 1-19. Typical construction sequence for 2 span continuous pre-cast beam (proposal from
Washington State Department of Transportation, US)
but there is no reinforcement or pre-stressing passing over the support for creation of a
continuous girder. However, in Figure 1-19 a similar procedure is shown, but in this case the
guidelines allow for a layer of reinforcement passing over the support cross girder creating a
The most frequently used erection method when casting multiple Tee-beam bridges on site is
the span-by-span method as shown in Figure 1-20. Each span is cast individually often with a
cantilever in the adjacent span to reduce the sagging moment in the span thus economising the
required cross-section.
18
Age 1
My span1 = My summation
Age 2 Age 1
My span2 My summation
My span3 My summation
My span4 My summation
When the bridge cast in-situ, creep and shrinkage play a significant role in the structure and
have a tendency to change the forces within the structure. Consequently the designer must
consider the implication of each stage of these effects to minimise any detrimental effects.
Figure 1-20 also shows how the bending moment changes during the various stages of
A combination of both pre-cast and cast in-situ methods is also be found in practise and the
methods and possible variants are numerous and well outside the scope of this thesis.
19
1.9 Codes requirements for design of multiple Tee-beam bridges
All international design standards require that any structure be checked under both
serviceability and ultimate strength conditions. Several criteria must be met, these include
stress limits and maximum deflection, and all structural components with in the structure need
to be designed accordingly.
The challenging goal to all engineers is to find the best possible solution. Consequently, it is
common that a number of iteration occur as the optimum solution is found considering
Once the basic parameters such as span length and cross-section shapes are defined the design
Determine all relevant internal forces due to all appropriate loading cases. These forces
are to be made available for each load case with and without multiplication factors for
For each national standard an allowable stress for the concrete under service load is
specified. The stress distribution is considered to be linear over the cross-section height
and the stresses for all relevant construction stages and load situations are checked on
The serviceability check is also done for the tendons for which the actual stresses under
service load need to be evaluated and compared to the allowable steel stresses.
20
1.9.3 Ultimate moment capacity:
partial factor for strength (some codes call it a material factor γm).
When calculating the ultimate moment capacity the cross-section internal equilibrium
considering all material components at their ultimate strength and the actual cross-
established. In Figure 1-21 the principle of the ultimate moment capacity calculation is
shown.
ε0
Reinforcement
and pre- ≥
stressing
tendons.
Figure 1-21. Ultimate moment capacity and comparison to the external moment My.
Closely related to the shear stresses resulting from flexure in beams are those that are
the result of torsion action. Torsional shear stresses also produce diagonal tension cracks
in the concrete. Torsional reinforcement is similar to the shear reinforcement, and they
are then in many cases combined with each other since the shear stresses produced by
load actions are usually much below the direct shear strength of the concrete. The real
concern is with the diagonal tension stress in the concrete produced by shear stress
21
1.9.5 Longitudinal shear:
For the slab-web interface of composite decks such as pre-cast beams + cast in-situ
concrete slab the longitudinal shear needs to be checked and the structure needs to be
designed for.
Depending on the selected code additional checks are required such as fatigue and crack
width.
1.10 Summary
The analysis and design of multiple Tee-beam bridges requires the consideration of several
construction and design issues such as construction stages, change of structural system, time
dimensional structural model. For the engineer the amount of work and the aspects to be
considered are not much different compared to the design work for a long span balanced
cantilever bridge. This stands in contradiction to the relatively easy and quick erection of
The following chapters will highlight issues for design and construction for multiple Tee-
22
2 Loading
2.1 Introduction
The longitudinal Tee-beams in a multiple Tee-beam bridge can be assumed to display beam-
like behaviour in the longitudinal direction. This assumption implies that cross-sections of
these Tee-beams remain plane and undistorted. Therefore, vertical loading on these
longitudinal Tee-beams causes bending in the longitudinal direction and depending on the
location of the load the possibility of torsion. The roadway slab connects the individual
this roadway slab is that of a two-dimensional plate. Bending of the roadway slab in the
lateral direction causes torsion in the longitudinal girders and vice versa. The load-carrying
Loads acting on bridge structures cause internal forces in the structure. These internal forces
must be transferred to the foundations by the individual structural elements. Loads may act
directly on primary structural elements, such as the main girders in a multiple Tee-beam
bridge, or on secondary structural elements such as the roadway slab (Figure 2-1). In a
numerical structural model the loading occurring on the physical structure must be
approximated in a realistic way. It is common that the national design codes give guidance on
the particular nature of these load models. One common way of classifying loading types is
the differentiation between permanent and transient (or non-permanent) loads. The difference
between these loads is covered later in this chapter. Main girders of a multiple Tee-beam
structure therefore have two major load applications: firstly loads acting directly on the
girders; and secondly, loads being transferred to the main girders by secondary structural
23
Loading on Loading on
roadway slab ma in girder
Three loading directions are normally considered for a bridge deck - longitudinal, transversal
and vertical as shown in Figure 2-2. In a structural model, concentrated loads, line loads and
surface loads are typically applied in these three directions to simulate the loading of the
physical structure.
longitudinal loads
(e.g. braking force).
CGs
Vertical loads acting directly on the Tee-beams cause longitudinal bending, transversal
bending and torsion loading to be applied to longitudinal members. The road way slab linking
24
the longitudinal members transfers the load to the adjacent longitudinal sections in the bridge.
Figure 2-3a shows a typical situation where a vehicle load acts on a member. The magnitude
of this load transfer to adjacent members depends the stiffness of the roadway slab and special
focus needs to be given to the correct modelling of the roadway slab during the analysis and
design process.
Loading on the roadway slab between the longitudinal girders causes the slab to deflect
(Figure 2-3b). This deflection causes the Tee-beam girders on either side to rotate which in
turn induces torsion into these longitudinal members. This transfer of the vertical loads also
It should be noted that vertical loading can also induce transversal bending into the Tee-beam
girders when the principal axes of the girders are not perfectly vertical – ie. in non-
Load – vehicle
(a) l d
rotated
deflected roadway slab cross-section
vertical and
transversal bending
(b)
Transversal loading such as wind load or centrifugal forces due to traffic impose transversal
bending on the bridge deck as a whole. For this loading condition the roadway slab can be viewed
25
as a stiff member (eg shear wall) connecting the individual longitudinal members and distributing
the horizontal load to all longitudinal members. The horizontal loads acting away from the shear
centre of the Tee-beams again introduce torsion into the longitudinal beams. The roadway slab
deflects and transfers this torsional action from one Tee-beam to the other. This behaviour of the
bridge deck is sometimes approximated by modelling the shear wall effect with a Vierendeel truss
MG1 MG1
MG3
MG2
MG2
Figure 2-4. Grillage under horizontal load acting as a truss with rigidly linked members (Vierendeel – truss).
Transverse bending moment obviously occurring under the described transversal loading condition
is often accompanied by torsion due to eccentricities of the load resultant in relation to the shear
centre. This torsional action on the Tee-beams causes bending in the roadway slab which in turn
Forces acting in the longitudinal direction of the bridge are especially important with regards
to the design of the bearings. Additional, bending moments are introduced into the bridge
deck due to the vertical eccentricity of the load application points in relation to the CG. When
the piers are rigidly connected to the bridge deck the braking forces become a governing
26
factor for the design of the foundation since the lever arms of the bridge piers generate
Self weight is often a dominant component of the accumulated bridge loading and the
accurate consideration of this loading type in a structural analysis is of great importance. The
self weight of the structural components can be computed easily by multiplying the volume of
these components with the specific weight of the associated material. For modelling reasons,
some overlaps between various structural elements may exist in a computer model and care
must be taken that the loading model is not doubled-up in the overlapping regions.
Additionally, the self weight of non-structural bridge components must be considered. These
loading items are often applied on the actual structure as surface loads (eg. pavement), line
loads (eg. barriers) or concentrated loads (eg. lamp posts). Self weight of the actual structural
components and additional dead loads are often grouped together individually to account for
importance in the design of pre-stressed concrete structures. Partial loss of pre-stressing force
and significant changes in deflection and stress distribution are often caused by these time-
dependent effects
Creep strain for concrete has been found to depend on time, on the mix proportions, humidity,
curing conditions and the loading history of concrete among others (CEB/FIP (1990)). Creep
strain is nearly linearly related to stress intensity. It is therefore possible to relate the creep
strain to the initial elastic strain by a creep coefficient. Typical values of the creep coefficient
27
range from about 1.6 to 3.2, the lower coefficient corresponds to the higher concrete
compressive strength. In pre-stressed concrete members the compression causing creep varies
over time because of relaxation of steel pre-stressing stress, shrinkage of the concrete and
member length changes associated with creep itself. This interdependence can be adopted to
lead to a step-by-step approach in calculating time dependent losses in which stresses acting
at the beginning of a specific time interval, causing the next increment of deformation reflect
all losses that have occurred up to time (Pircher 1994). For the present purpose losses are
treated individually in order to appreciate the role of each effect. Practical calculations are
often carried out on this basis as well. For pre-cast members, that later are connected to a cast
in-situ slab the moment of inertia of the composite section should be used in calculating the
stress caused by loads applied after the cast in-situ concrete has hardened.
Creep calculation according to the CEB/FIP (1990) require a careful handling of system data.
The following short example should illustrate the amount of data. Figure 2-5 shows the
development of the creep coefficient φ for a structure being erected in 3 construction stages.
The graph shows only one element of the structure. At the time of first load application, the
starting time for creep, the concrete is 14 days old. For this element for the first stage a creep
coefficient φ14 results. The coefficient changes during the stage 1 over a time of 28 days. At
the end of stage 1, new elements are added to the structure and new loads applied (PT = pre-
stressing, SW = self weight). Due to these new loads a new creep coefficient for the same
element needs to be determined. The two creep coefficients for one element now require
monitoring and undergo further changes as the structure changes. After new elements and
new loads are applied to the structure in stage 3 the previously existing creep coefficients are
changed again and a third one for this element is added. This procedure needs to be done for
all elements in a structure and for all ages and all loading cases if the analysis is to comply
28
REFERENCE – TIME AXIS [Days]
[ϕ]
ϕ t0
tcurrentl ..... Creep coefficient
ϕ 14
Shrinkage of concrete causes a reduction of strain in the pre-stressing steel equal to the
shrinkage strain of the concrete. The resulting steel stress loss is an important component of
For pre-tensioned construction pre-stressing often takes place as early as 24 hours after
casting and a high percentage shrinkage may therefore affect pre-stressing losses. Post-
tensioned members are less affected since pre-stressing is usually applied at a later stage
(Nilson 1987).
Wind loading is often approximated as lateral surface loading acting on the exposed side-
areas of the deck with amplification factors on the static loading to account for the dynamic
nature of wind gusts. Many design codes consider various wind loading cases for different
situations – eg. wind loading for the deck with traffic on the bridge, and wind loading without
traffic, or wind loading with different wind velocities for different load combinations. If the
cross section depth varies along the length of the bridge the resulting lateral load has a
29
longitudinal variation that needs to be considered. Oftentimes wind loading is assigned to the
longitudinal Tee-girders directly with loading acting on all exposed webs. It is important to
note, that depending on the consideration of traffic, the distance of the resulting load vector to
the shear centre varies and that torsional moments may be introduced by wind loading. The
effects and probabilistic properties of this type of bridge loading. However, the spans of
multiple Tee-beam bridges are typically sufficiently small to make the described
As with all other loading types the exact nature of a traffic loading model for a particular
country is determined by the respective national design code. However, some general
characteristics common to most of these codes can be summarised in the context of this
chapter.
The actual loading is usually given by the definition of loading vehicles, mostly consisting of
a series of concentrated wheel loads. These loading vehicles are often combined with
distributed loading components. Some codes (eg. British standards BS5400 (1990)) require a
variation of loading intensity depending on the loaded length of a girder based on the shapes
of the influences lines and the influence surfaces. Other codes introduce variabilities in the
distances between the individual concentrated wheel loads (eg. Australian Standard AS5100,
AS 2004). Traffic loading usually applies in all three mentioned directions; vertical loading
due to the weight of the passing vehicles, and longitudinal loading to account for forces
30
The location of traffic loading is usually driven by the setup of traffic lanes on the bridge
deck. Some codes allow the placement of the wheel loads along the centre line of a given
traffic lane (DIN1045 (2004)), other codes require the exact two-dimensional positioning of
the individual wheel loads (BS 5400 (1990)).Traffic loads act on both the main girders
directly and on the roadway slab depending on the setup of the traffic lanes in relation to the
The most detrimental position for the placement of one or more loading vehicles within a
given lane can then be determined for each member from the result matrix. It should be noted
here that the governing traffic position for one design force often differs from the governing
traffic position of another design force. This fact complicates traffic load evaluations in
Results for the loading of individual lanes are combined to give the result envelopes for the
complete bridge deck. Design codes give combination factors for these lane combinations
Railway bridges commonly have fewer traffic lanes than road bridges and are generally with
predefined alignments, but the loading magnitudes for railway bridges are significantly
higher.
constant temperature change of the whole deck, linear temperature gradients along a vertical
cut through the deck, and non-linear gradients in the same direction. Some design codes
specify different gradients for enclosed areas such as box-sections and plates (NZS 3100
(1995)) while other codes do not require the consideration of non-linear gradients (BS5400
(1990)).
31
In composite cross-sections the designer must consider that even a constant temperature
loading causes bending in the bridge deck if the expansion coefficients of the materials of the
Non-linear temperature distributions as shown in Figure 2-6 cause a theoretical dilemma since
many design codes are based on considerations based on beam theory. Beam theory however
postulates to remain plane with a linear strain distribution which stands in direct opposition to
distribution.
2.4.4 Settlement
Geological and code-specific data usually determines the assumptions of the extent of
settlement to be taken into account in a structural analysis for each foundation in a bridge
structure. Design codes also give different specifications with regards to the combination rules
For continuous girders hogging moments can be introduced in the deck above piers. These
hogging moments can become an important design issue when combined with pre-stressing
but no stresses. However, uneven settlement in simply-supported bridges with skew piers or
curved alignments cause considerable stresses in the deck as illustrated in Figure 2-7.
MG2
Bending moment due
to pier settlement. MG1
Figure 2-7. Twist of roadway slab due to pier settlement for skew support axis.
The typical shape of the moment diagram due to pier settlement is shown in Figure 2-8. For
each pier an individual loading case is calculated. All loading case results together give the
envelope representing the most unfavourable situation of all possible settlements. The
individual loading cases are combined according to the relevant design code definitions. Some
codes allow only one settlement at a time, other require considering the possibility that several
Loading case
combination –
min/max My
envelope.
Figure 2-8. Bending moment for pier settlement and secondary PT.
33
34
3 Pre- and post-tensioning
tension it is left floundering. The use of pre-stressing can ensure that concrete members
remain within their tensile and compressive capacity thus increasing their versatility. The
bridge types and in the construction techniques utilised for bridges. Pre-stressed concrete is
frequently chosen for bridges with spans ranging from 20m up to 350m (Hewson 2000).
The following chapter describes some principles of pre-stressing, the application and the basic
considerations for handling the pre-stressing effects in the design of multiple Tee-beam
bridges.
tendons to the surrounding concrete. Tendons are placed within the concrete members as
either grouted internal tendons or as external tendons. It should be stressed that the geometric
position of the tendon within the cross-section is of great importance for the design.
35
Consider a simply supported pre-stressed beam with a rectangular cross-section. A pre-
stressing tendon is placed in the centroidal axis and stressed the force F, as shown in Figure
3-1. This beam is also loaded by a uniformly distributed load. The tensile pre-stressing force
in the tendon produces a balancing compressive force in the concrete. In the case illustrated,
the pre-stressing force is acting in the CG of the cross section and the stresses in the extreme
σ= ±
F My
(3.1)
A I
In this case the stresses in the bottom of the member are less than those calculated at the top
of the cross-section.
Anchor plates
Uniform load e.g. self weight and tendon
Concentric tendon
Pre-stressing Force F
σ= σ= ±
F F My
A A I
Stresses due to Stresses due to Stresses due to pre-stressing
pre-stressing external load and external load
If the tendon is now placed with an eccentricity (e) to the CG of the cross-section as shown in
Figure 3-2 the pre-stressing force (F) is applied. To calculate the stress the same components
as in Equation 3.1 are utilised but the eccentricity also introduces a bending moment into the
girder and the resulting stresses in the extreme fibres are given by
36
σ= ± ±
F Fey My
(3.2)
A I I
Unlike the concentrically stressed member the stresses in the bottom of the member are now
Anchor plates
Uniform load e.g. self weight and tendon
CG of cross-section
Eccentric tendon
Pre-stressing Force F
σ= ± ±
F Fey My
A I I
Stresses due to Stresses due to Stresses due to Stresses due to pre-stressing
pre-stressing pre-stressing external load and external load
eccentricity
The required strength of the concrete is determined by the compressive stresses generated in
the concrete by the pre-stressing and applied forces. A minimum strength of fcu equal to
45N/mm2 is typical for pre-stressed concrete, however it is more common to use higher
strengths with a fcu up to 60N/mm2 (Ryall et. al. 2003). Even higher concrete is used for
specific projects. At the time of pre-stressing a minimum strength of 30N/mm2 for the
concrete is often required although this might vary depending on the tendon and anchor
arrangements and the magnitude of the applied load. To minimize the losses due to creep and
shrinkage, and thus losses in the pre-stressing, care is required in the mix design and the
The stressing is achieved using high tensile steel in use as wire, bars or strands. The nominal
tensile strengths of these components vary between 1570N/mm2 and 1860N/mm2 for wire and
37
strands and between 1000N/mm2 and 1080N/mm2 for bars. Once the stressing load is applied
steel relaxation occurs, this results in the reduction of stresses in the tendons. The magnitude
of relaxation varies depending on the steel characteristics and the initial stresses. Typical
relaxation ranges from 2.5% to 3.5% when a stress of 0.70 fpu is applied. For an initial stress
of 0.50 fpu the relaxation reduces to about 1% (Ryall et. al. 2003).
depending on the design criteria. When cement grout is used it is pumped into the ducts to fill
the void in the ducts around post-tensioned tendons. Generally a water/cement ratio of
between 0.35 and 0.40 is typically used; admixtures are sometimes added to improve flow and
to reduce shrinkage. One of the major advantages of grouting is the protection of the tendons
from corrosion.
When stressing a member two design methods are used either Full or Partial Pre-stressing.
The complete elimination of tensile stresses in members at normal service load is defined as
full pre-stressing. Early designers focused on this type of pre-stressing in order to avoid
cracks caused by tensile stresses on either side of the member. As experience grew solutions
between fully pre-stressed and reinforced concrete were often favoured. Such an intermediate
solution in which concrete receives tensile stresses which are compensated by reinforcement
is called partial pre-stressing. The cracking due to tensile stresses under full service load is
Typically the fully pre-stressed girders are heavy with large cross-sections, producing an
objectionable large camber at more typical loads than the value service load. Partial pre-
stressing tends to results in more economical solutions by reducing the amount of pre-
stressing steel. When designing for partially pre-stressed girders, the design code generally
38
nominates whether reinforcement in the appropriate layer may be considered for tensile
forces, or whether the partially stressed steel can be considered as passive steel, using the
It has been observed that partially pre-stressed girders are more or less the state of the art in
almost all popular design codes. Certain codes, for example the Eurocode 2 (CEN 2002),
demand full pre-stressing for certain load situations and partial pre-stressing for other load
method.
There are two methods commonly used for pre-stressing members as shown in Figure 3-3:
produced in a casting yard by stressing tendons between external anchorages before the
concrete is placed in the formwork between these anchorages. As the concrete hardens it
bonds to the steel, when the concrete has reached the required strength, the anchorages are
released and the pre-stressing force is transferred to the concrete. Only linear tendon profiles
can be achieved.
Post-tensioned members are stressed inside ducts after the concrete has hardened to the
required strength. The pre-stressing is applied through jacks against the ends of the concrete
member and the ducts are usually grouted. Hardening of the grout creates the bond between
tendon and girder. Tendons may be bundled parallel wires, stranded cables, or solid steel rods.
A typical post-tensioning arrangement can be seen in Figure 3-5 where the geometry follows
the anticipated moment distribution due to self weight and traffic loading. Stirrups are
installed in regular distances in longitudinal direction supporting the duct and guaranteeing
39
A A-A
B Post-
Beam thickened at end for Multi strand tendons
anchor tensioned
strand
Tendon
anchorage jacks
Precast beam
Figure 3-4. Straight and polygonal tendon layout for pre-tensioned beam.
Dead
anchor
R[m] = curvature Concrete cover
min.120mm
Stresses are transferred from the pre-stressing strands to the surrounding concrete through the
bond between the two materials. The distance from the end of the member over which the
40
effective pre-stressing force develops is called the “transfer length”. The “flexural bond
length” is the additional bond length necessary to develop the strand stress from effective pre-
stress to the ultimate at the ultimate flexural strength of the beam. The sum of these two
lengths is called the “development length”. Kaar et. al. (1963) gives the following equation
Ld = D + ( f su* − f se ) D
f se
(3.3)
3
The first term, (fse/3)D, represents the transfer length. The second term, (fsu-fse)D, represents
the flexural bond length. Some codes like the AASHTO (1996) require a minimum
Ld = ( f su* −
2
f se ) D (3.4)
3
In both formulas D is the nominal strand diameter, fsu* is the average stress in the pre-
stressing tendon at ultimate load and fse is the effective pre-stressing steel force after all
losses. Research by Deatherage & Burdette (1991) showed that the transfer and development
Pre-cast beams can be created in a controlled environment and usually have high concrete
quality and good precision in reinforcement layout and pre-stressing arrangement. Stress
members the anticipated moment diagram due to dead load and traffic loading is often
counteracted by groups of tendons which are sleeved in a staggered pattern towards the ends
of the members.
41
Capacity
built up from
zero to max. Full capacity (σallowable x Areatendon)
For pre-tensioned members the initial pre-stressing force acts at the permanent anchorages of
the casting yard. The tension is constant for straight tendons, losses only occur when deviators
are installed to achieve a specific tendon geometry. The pre-tensioning steel force is reduced
by the anchorage slip at the anchorages, which is also a relevant type of pre-stressing loss for
post-tensioned tendons.
For post-tensioned girders the full pre-stressing force is applied at the anchorages and the pre-
stressing force along the girder is reduced due to friction losses. Friction between tendons and
ducts results in a loss of pre-stressing force. The total friction losses are a combination of
wobble friction caused by deviations of the tendon within the duct and friction due to
curvature. Both types of friction losses depend mainly on the tendon material.
Additionally, concrete members shorten due to the compressive force applied by pre-stressing
thus reducing the pre-stressing force in the tendon. This effect is termed elastic shortening.
For pre-tensioned members the change in steel strain is the same as the concrete compressive
strain in the steel centroid and the losses may be calculated accordingly. For post-tensioned
42
members the elastic deformation of the concrete takes place after applying the jacking force
and there is automatic compensation for shortening losses. If parallel or overlapping tendons
are post-tensioned sequentially the interaction between the elastic shortening losses for each
tensioning are combined losses in the pre-tensioned steel occurring due to the post-tensioning
Time-dependent effects also cause reductions of stress in the pre-stressing steel. Creep and
shrinkage in the concrete as well as relaxation for the pre-stressing steel contribute to this
group of losses.
Figure 3-7 shows the force distribution in a pre-stressing tendon considering losses due to
friction, wobble, anchorage slip, creep and shrinkage and steel relaxation in a post-tensioned
member.
All losses require consideration in the analysis and design process. The time dimension
structures. The described losses do not happen simultaneously but at different points in time
Jacking force
43
Instantaneous Time-dependent
Due to conrete
Elastic Shrinkage Creep
shortening
Total pre-
Friction stressing
losses
Due to steel
Anchorage
Steel
slip
relaxation
Causal relation
Effect relation
Figure 3-8. Interrelationship of causes and effects among pre-stressing losses [Naaman and Hamza, 1993]
When pre-stressing tendons apply load to the structure the resultant forces and moments
generated can be considered as a combination of primary and secondary (or parasitic) effects.
Primary effects are the moments, shears and axial forces generated by the direct application of
the force in the tendon on the relevant section. Secondary effects occur when the structure is
statically non-determinate and restraints on the structure prevent the pre-stressed member
from deflecting when the pre-stressing force is applied. For continuous decks the intermediate
supports restrain the deck from vertical movement and secondary moments and shear occur.
The combined primary and secondary affects as shown in Figure 3-9 are derived directly from
the analysis output. The forces and moments from the tendon are applied at each anchor
position. Along the structure equivalent loads are applied to the model wherever the tendon
geometry has a change of angle. Care needs to be taken when modelling the structure since
the example shows that the support definitions have a major influence on the behaviour of the
structure under pre-stressing load. The correct consideration of pre-stressing as well as the
44
subdivision of the pre-stressing results into primary and secondary effects is a basic
Mp
Fh
Fv Force in tendon
Equivalent forces Load intensity = Radius of tendon
Hogging moment
Resultant moment: primary + secondary Sagging moment
Figure 3-10 shows the principle of how the ratio between primary and secondary effects can
be estimated by comparing the areas between the tendons and the CG of the cross-sections.
However, this estimation does not consider the fact that the pre-stressing force is not constant
relatively high. This position of the CG becomes especially important when defining the pre-
stressing of the girder. The eccentricity of the tendon in the span relative to the CG is bigger
than the eccentricity at the support. When considering the example in figure 3-9 and
accounting for the eccentricity relatively high secondary moments result. In some cases the
secondary moments can be of the same magnitude as the primary moment which means that
45
The remaining structural design component to be considered is the axial force which causes
compression in the concrete. The secondary effects can be reduced by shifting the tendon
geometry closer to the CG. Increasing the distance to the bottom fibre in the spans decreases
Rectangular
cross-section Area between Area between Section neutral axis
tendon and CG in tendon and CG at
spans (area A) supports (area B)
Tendon eccentricity e
Tee-beam
cross-section Section neutral axis
Area between Area between
tendon and CG in tendon and CG at
spans (area A) supports (area B)
Primary effects: Mp=P*e Secondary effects
Hogging moment
Sagging moment
Σareas B << Σareas A →Secondary effects are sagging
Figure 3-10. Quick estimation of primary and secondary effects in a continuous girder for a rectangular cross
section and a Tee-beam.
3.7 Consideration of pre-stressing for SLS and ULS design code checks
At SLS the stresses in the concrete must be kept within allowable limits for both compression
and tension. It is usual to keep the concrete in compression across the full section under
permanent load and to allow tension stresses up to 2 - 3 N/mm2 for live loading. The primary
and secondary forces both combine giving the resultant force distribution as shown in Figure
46
3-11. The corresponding stress distribution in the cross-section is taken into account for the
SLS checks. In many design standards the Primary and Secondary effects require different
+ =
e
Applied Resulting
Primary Secondary stresses
force
stresses stresses
The moment of resistance, or capacity, at a section is derived by comparing the balance of the
tensile force in the tendons and reinforcement with the compressive force in the concrete. The
strain distribution is considered to be linear across the section with the point zero being the
47
For ULS checks it needs to be considered that a big portion of the pre-stressing force is in the
actual tendon when grouting the duct and creating the bonded situation between tendon and
cross-section. All loads applied subsequently cause deflection resulting in a rotation of the
cross section. This rotation introduces a strain in the cross section. At the concrete fibre at the
same location as the tendon the strain for both the concrete and the tendon are the same due to
the bonded situation. This strain causes new stresses in the tendon which are added to the
stresses form the initial strain of the tendon. The total of the tendon stresses is therefore the
combination of initial stresses due to initial strain plus additional stresses due to additional
loads. For the ULS the section is assumed plane under bending. The initial stress and strain in
the tendon, the reinforcement and the concrete are generated by the pre-stressing and the
permanent loading. The moment resistance of any section along the deck must exceed the
bending moment generated by applied loading to give a sufficient factor of safety against
Permanent deflections of the concrete deck occur due to pre-stressing, due to self weight and
due to the weight of the permanently applied loads followed by further deflections due to
implemented in order to achieve a certain desired final deck geometry under permanent
loading. The pre-camber is affected by the construction sequence and the concrete properties.
In the lifecycle of pre-cast beams they run many different construction stages from the casting
yard to the final position as member of a bridge deck. Because of the limited lengths pre-cast
beams are often prepared as simple supported beams. The first time pre-cast beams take load
causing deflection is at the time of cutting off the tendons at the anchorages which introduces
the pre-stressing into the girder. Depending on the tendon geometry the girder gets under
48
compression and lifts away from the formwork. At this stage the pre-cast beams sit on the
Pre-stressing
Reinforcement bulged
Reinforcement released
∆2 ……> ∆1
Pouring of slab-concrete
Reinforcement bulged
∆3 ……< ∆2
Slab part active… Slab is stress free until
next load application
…often of variable depth due to precamber of precast part.
Later construction stages and then again loads bend the girder further again. The fact that due
to the casting of the deck a composite section is built up requires special consideration in the
overall design and in the precamber definition. The properties change significantly when
creating the composite cross-section and the stress state of the pre-cast beam is locked into the
structure.
49
Due to the precamber of the pre-cast girders the deck slab gets a variable depth in case the
precamber is not applied to the deck as well. The load is then not uniform along the girder
which is different to the calculation. The deflection caused by the load as well as the
subsequent creep and shrinkage are underestimated and the precamber is not fully abolished.
50
4 Numeric modelling of the roadway in Tee-beam bridges
4.1 Introduction
In a structural model, the roadway slab represents the link between the adjacent longitudinal
members. The roadway slab is responsible for both the structural interaction between the
longitudinal Tee-girders and for the transfer of loads applied on the roadway slab to the main
girders. The roadway slab is commonly simulated by beam, plate or shell elements in a
numerical model. In selecting the type of elements to be used consideration of the following is
required;
the principal stresses, shear and torsion in the roadway slab; and
Prior to discussing the considerations for the detail model, it is essential to understand the
basic advantages and disadvantages of grillage and finite elements systems when modelling
with both bending and torsional stiffness. At the connections of the longitudinal and
transverse beams, deflection and slope compatibility equations can be set up. Although the
method is generally approximated as a necessity, it has the great advantage that it can be used
in most situations. Ryall et. al. (2003) gives comprehensive modelling recommendations for
51
Figure 4-1. Grillage model of 2-span bridge with PT in the main girders. Plan view as system
line, plan view with solid deck, perspective view as 3D frame.
In the grillage system, the bending and torsional stiffness’s in every region of the slab are
assumed to be concentrated in the nearest equivalent beam. The longitudinal stiffness of the
slab is concentrated in the longitudinal beam elements; the transverse stiffness is concentrated
Ideally each beam in such a grillage model behaves in a manner that closely agrees to the
physical two-dimensional slab-beam-system of the roadway slab. The moments, normal and
shear forces should resemble the stress resultants and deflections at a any given position. The
grillage deflections should also resemble closely the deflections in the actual slab. There are,
however, a few shortcomings since the grillage is only approximation of the physical
structure.
52
4.2.2 Finite elements for the roadway slab
The finite element method forms one of the most versatile classes for modelling structures,
and the method relies strongly on the matrix formulation of structural analysis. The
application of finite elements dates back to the mid-1950s with the pioneering work of
Zienkiewicz (1991). The finite element method is based on the representation of a body or a
functions are chosen to approximate the variation of displacements over each element
The entire procedure of the finite element method involves the following steps:
3. element stiffness matrix is derived using variational principle of mechanics such as the
5. the algebraic equations thus obtained are solved to determine unknown displacements
and
6. element strains and stresses are computed from the nodal displacements.
In Figure 4-2 a typical multiple Tee - beam deck modelled with finite elements is shown. The
finite element method has certain limitations in connection with the problem of result post-
processing as the method produces stresses and strains, and typically the action effects such as
53
Figure 4-2. Typical multiple Tee-beam bridge in plan. Combination of beam elements for the
main girders and finite elements for the roadway slab.
Both the grillage and finite element methods have advantages and disadvantages. The choice
of either method depends on several factors including geometry, support conditions and/or
design requirements. The structural behaviour of the bridge deck in Figure 4-2 for instance is
certainly better described by finite elements. The deck widening as well as the skew support
axes introduce both a complex behaviour that is more difficult to represent with a grillage
model.
The main advantage for the use of grillage models is founded in the fact that results are
achieved in the form of internal section forces in a given direction – that is often conveniently
also the reinforcement direction. Finite element results on the other hand are stress
components in arbitrary directions which may be difficult to interpret for design purposes, this
made more difficult as most design codes are based on considerations stemming from beam
54
theory. Intensive post-processing is usually required to transfer results from a finite element
Primary, the displacement results for beam elements and for the finite elements are generated
in an analysis directly for all node locations. Force results for beam elements are then derived
in these node locations while stress results for finite elements are computed for the so-called
Gauss points of each finite element. Post-processing routines must be employed to generate
results for locations between the nodes or Gauss points respectively. Such post-processing is
possible for beam elements in an analytically rigorous manner while for finite elements
During the design of bridges many operations are based on the results of analyses of
individual loading cases – e.g. preparing design load combinations, evaluating traffic loading,
performing design checks for SLS and ULS etc. For various reasons these operations need to
be performed in discrete locations along the structural members and traditionally node
locations are chosen for these tasks. In Figure 4-3 bending moment results for a uniformly
loaded simply supported beam are connected with straight lines between the nodes and are
compared to the analytically derived parabolic function for this particular loading condition.
In Case (a) ten beam elements have been utilised in the model, while for case (b) only half the
number of elements are used. Obviously, the accuracy is improved for higher numbers of
elements, with the peak moment in case (b) being underestimated significantly as a result of
55
Uniform load
1 2 3 4 5 6 7 8 9 10
(a)
Polygonal
analysis results.
1 2 3 4 5
(b)
Polygonal
analysis results.
Under-estimation of
critical result
Analytical solution (ql2/8 parabola)
Figure 4-3. Bending moment due to self weight in simply supported beam.
The number of elements along the longitudinal girders also determines the size of the
elements representing the roadway slab as a result of compatibility. The roadway slab
elements transfer all loading acting immediately upon them to the longitudinal girders, and
This effect is illustrated in Figure 4-4 where a point load acting on the roadway slab is
presented. In the grillage model, as shown in Figure 4-4(a), the loading location is somewhere
on a transverse-beam which deflects and transfer the resulting internal forces to the nodes on
the longitudinal main girders that it is connected to. The resulting moment diagram in the
longitudinal girder will have a triangular shape with the peak in the node were the transverse-
beam is connected. In a model with finite plate elements representing the roadway slab, the
load is distributed to four nodes as opposed to 2 nodes and therefore the resulting moment
diagram in the longitudinal Tee-beams is trapezoidal without the pronounced peak of the
grillage model.
56
load application
point
(a)
cross-beam axis
Triangular moment
distribution in main
girder. Tee-beam axes
load application
point
(b)
Figure 4-4. Moment diagram of longitudinal member due to concentrated load acting on the
roadway slab.
The stiffness of beam elements in the grillage model depends on the material and cross-
geometrically cover the area of the roadway slab. They should be orientated so that they lie
side-by-side, touching but not overlapping and with no gaps. The widths of these cross-
sections are governed by the distances between nodes on the longitudinal girders. Since these
distances are rarely constant the cross-section widths usually vary from cross-beam to cross-
beam (Figure 4-5) care must be taken when generating the model. Changes of deck depth in
the lateral and longitudinal directions also adds additional complexity since these changes
must also be accounted for in the models of cross-beams. Moreover, as illustrated in Figure
57
4-6 in the event of curved alignments in plan view cross-beams have different cross-sections
at start and end of the elements. Such a situation is drawn in detail in Figure 4-7.
L1
End node for the
201
cross-member L2
202
101
203
102
Cross-member element 204
L1/2=b1 103
L2/2=b2
104 CP2
CP1 MG2
L1/2 L2/2 L3
L4
cross section 1 =
cross-section 2
cross section 1
b1 = L3 /2 + L4 / 2
cross section 2
L1 L2 b2 = L1 /2 + L2 / 2
L1
L2
beginning and end of such elements. However, this leads to a miss-representation of the actual
stiffness distribution in the deck. For large radii the error introduced by this averaging
procedure is small but for small radii this error can become significant. This problem can be
circumnavigated by splitting cross-beams into several elements, but this requires additional
series of longitudinal connecting beams to maintain integrity within the grillage. A better
solution for these cases would be to implement beam elements with formulations that take
account for the variable stiffness and represent this particular situation adequately.
58
System line for
longitudinal beams
Cross section –
element end
While considering the element width it is also important that the element length the cross
section depth be considered also, as all these dimensions give the necessary geometrical
definitions needed to define the transverse element. In the event of a constant thickness of the
roadway slab between the webs of the Tee-beams, the cross section depth for the transverse
element remains equal to that of the roadway slab thickness as shown in Figure 4-8.
Figure 4-8. Four Multiple Tee-beam with constant roadway slab thickness between webs.
Alternatively when the roadway slab has a variable thickness between the webs, as shown in
Figure 4-9, two common methods of representing this situation in a global model are used.
59
The first approximates this variable to one equivalent cross-beam element; while the second
method uses multiple elements to model this situation in a more detailed manner. In the global
analysis the subdivision of the transverse beam into several elements does not necessarily
make any difference regarding the quality of results for the longitudinal Tee-beams. However,
if analysis results for the cross-beams are to be used for the design of the roadway slab then a
Variable thickness of
cross section
Variable thickness of
roadway slab
Figure 4-9. Double Tee-beam with variable roadway slab thickness between webs.
the following procedure can be used to find the correct cross-section depth. It is important to
note that the torsion of the longitudinal Tee-beams introduce bending in the cross-beams and
should be considered.
MG2
MG1 Cross section 1 Cross section 2 Cross section 1
Element 1
MG2
MG1 Cross section 2
Figure 4-10. Three elements and two different cross section per transverse member.
60
The stiff-ended actions for the haunched and the representative beam must be equal in order
to model this effect correctly in the simplified case (Figure 4-11). In Figure 4-12 the bending
moment diagrams for a uniformly distributed load with either one or three spanning elements
are shown.
Torsion Torsion
Mt F
Mt
d1 d2
Vy
L1 L2 L3
L = L1 + L2 + L3
Torsion Torsion
Mt Mt
F
My1 My2
Parabolically
Moment diagram in cross
interpolated
beam for one element with
moment – change
constant thickness. of stiffness is not
considered.
Figure 4-12. Bending moment in the cross member consisting of one and three elements.
Using equilibrium equations an expression for the depth may be derived as:
61
Vy = J= = → d = 3 12 *
F *l3 F *l3 b*d3 F *l3
and
192 * E * J 192 * E * V y 12 192 * E * V y
If detailed representations of haunched roadway slabs are chosen then the model size
connected longitudinally with the additional series of longitudinal beams (Figure 4-13). In
Figure 4-14 the roadway slab is haunched in the transversal direction calling for four
additional series of longitudinal beams, thus demonstrating the increase in model size. In
Figure 4-14 an example for a double hollow box is presented, a similar example for a double-
Tee beam is shown in Figure 4-15. In all cases the cross-sections of the transversal beams
MG1 MG1
MG3
MG4
MG5
MG2 MG2
Figure 4-13. Plan view of grillage model with only one transverse element and a combination of
horizontal and longitudinal elements for the roadway slab.
additional longitudinal
beams longitudinal beam 2
longitudinal beam 1
Figure 4-14. Example for element layout for cross section with variable roadway slab thickness
62
Centre of gravity of System axes of
transversal beams transversal beams
System axes of
Reference node at top of
longitudinal beams.
cross-section
Cross-sections of
transversal beams.
Cross-section widths of
transversal beams.
Another consideration when modelling the stiffness of transverse elements is the behaviour of
the bridge deck under lateral horizontal loading. Under this loading condition the bridge deck
acts akin to a shear wall. If a grillage model is employed then this shear wall behaviour must
also be represented adequately. When only two main girders are connected by single element
transverse beams connected by longitudinal-beams have been modelled then the stress flow
through the deck can be represented in more detail. Figure 4-16 illustrates this aspect for a
MG1 MG1
Ix, Iy, Iz Ix, Iy
MG3 MG3
Ix, Iy, Iz Ix, Iy, Iz (of total cross section)
MG2 MG2
Ix, Iy, Iz Ix, Iy
Mz for all Virendel members Mz for central beam only.
Figure 4-16. Plan view of grillage model with only one transverse element and a combination of
horizontal and longitudinal beams.
63
4.5 Principal stresses, shear and torsion in the roadway slab
Differences in normal force in neighbouring transversal elements in the grillage model are a
measure of the transverse shear force present in the bridge deck, this is illustrated in Figure
4-17 for the case of horizontal loading. In finite element models this shear force will be
longitudinal girders also introduces torsional action into the roadway slab (Figure 4-18),
which leads to the introduction of additional transverse shear forces into the roadway slab. It
is essential that the transverse shear in the roadway slab, from all actions (including torsion)
Vertical
support.
Cross beam
at Support
axis.
Vertical and
transversal
support.
Horizontal transversal load
Bending of longitudinal
girder introduces torsion
into the roadway slab.
Figure 4-18. Differential bending of longitudinal girders resulting in torsion of the cross-
members.
64
-
beams.
0 .
0
.
2
-
0
- -
5
.
0
2 0 0
. .
0
3
-
1 1
-
6
-
5 1
0
0 .
2 5 3
9 8 8
2
.
2
.
0
0 . 4 4 8 - . 3 6 2 - 0
- 0
1
0
6
-
. 0
. 0
0
.
0 3
4 - 0 . 5 3 5 0 -
5 3
0
7
7
6
.
5 2 1
9
2 7
2
0
0
.
. 0
0
0 2
3
. 3
0
3 0 0
0 .
9
.
7 0 9 . 0 . 4 9 2
0 5 1 .
1
2
. 9
0
2
1 6
5
7
5 1
8
3
5 0
-
0
-
.
0
-
2
-
- -
- 0
.
0 .
0 0
0
-
-
0 . .
.1
0
2 1
6 5
.
. 1 8
2
4
3
1 0
0
2
1
0 5 1
6
.
6
7
0 2
1
0 7
3 7 5 0 . - 2 7 6 - 0 .
4
- -
0 0
- 0 . 6 9 6
. .
4 2
0 0
-
0 5 6
0 9
.
1
-
9 3
5
0
1
.
0 0 3 9 2 0 .
0 . 0 5 7 6
1
.
0
0 0 5 0 0 .
0
.
9
0
0
. 2 1
8
4
4 8
1 8
3
-
0
-
-
.
-
-
-
3
- - -
- 0
0 0 0
.
0
0
0
2
. . .
3 3 2
. 2
.
.
.
7
1 1 6
2
0
2
2
6 4 4
3
8 4
0
3
4
6 3
0
4
8
-
0
0 . 4 2 9 -
6 5
- .
-
0 1
- 0
8 0
.
0 .
- 0 . 6 6 6 0 .
0
7 0
1
-
6
0 8
4
. 1 6
-
3
1
8
8
0
3
0 . 0 1 3 1 -
- 0 . 0 2 6 9
.
0 . 2 9 3
0 0 . 2 7 1
2
.0 0 . 1 8 8
6
2
2
6
-
0
-
.
-
-
-
-
-
0
0
- 0
3
- -
0
0
0 0
.
0
.
.
3 . .
.
.
. 2
3 3
6 1
7
2
4 0
6
0
2
7
6 7
0
1
7
0
4
6
7 8
4 3
9 1 3
-
0
0 . 4 1 7 -
- .
1 -
0
0 -
. 8 0
0 .
6 0 .
- 0 . 5 9 6 6 0
6
-
0
0 0
9
-
9
2
. 4
2
0
9
0 0 . 0 1 2 8 -
- 0 . 0 3 0 7
.
6 0
. 0 0 . 1 1 7 0 . 1 7 1 0 . 1 8 9
3
2
2
5
2 5
-
0
-
.
-
-
-
0
0
- 0
-
3
-
0
0
0
.
0
.
4
.
.
. 2
8
.
- 0
.
.
3
1 3
2
3
2
2
5
1
6
. 2
4
1
8
0
2
2
0 0
9 3
2
5 4
-
0
0 . 3 8 6 -
.
2
-
0 0
-
1 .
0 0
9
.
0
- 2
1
- 0 - 0
0
-
.5 6 . 0
4 0
2 .
2 5
0
4 1
2
9
- 0 . 0 1 5 8
.
0
0 0 4 5 0 . -
5 0 . 0 6 0 0 .
0 . 1 0 3 0 . 1 3 2
2
0
2 3
8
- 0
-
-
0
. 4
.
0
-
- 0
4
- 0
.
- 0
0
4 0
- 0
. 3
.
. 2
. 3
2
.2
3
6 3 -
1
3 0
7
. 2
5
2
9 .
0 2
4
0
5
1 6
9
3
6
-
0
3 5 3 - 0 . .
2
1 -
6 0
.
1
1
-
0 .
- 0
-
-
9
0 4
0
.
4
0
. 2
1
4
.
4
0 0 5 7 0 .
6 0 0 0
0 0
1 . 0 . 0 0
2
7 0 . 1 .0 .
6 0 2 . 0 0
5 8
1 8
2
2
3 7 2
8 6 7
9
3
2
1
- 0
-
-
0
0
.
. 4
- - -
4 0 0 0
- 0
- 0
- 0
7
.3
2
. . .
3 3 3
2
. 3
. 2
3
3
.2
4 4 3
0
7
2
9 8 0 -
2
0
1
3
.
2
2 -
0
.
2 -
-
- 0 2 0
0
. 3 .
8 .
1
2 0
4 4
7 - 0
5
9 . 0
-
-
2
0 5 8
-
3
0
0
.
2
.
0 1 2 9 0 .
. 0 0
4 5 0 0
3
.0 0
. 0 .0 .
2
7 1 0
2 0 4
0 6 . 5 3
0 8
2
1 1 5
6
7
9
2
- 0
-
0
. 4
.
4 - -
0 - 0 - -
- 0
4
0
9 0 0
. .
3 . 3
6
6 3 . .
.2
3 3
5 4
5 2 0
7
2 6
6
6 2
8
-
0
.
2
8 -
0 -
0
4
. - -
2 . 0
1 0 -
0
3
7 . .
1 0 .
9 0
0 1 6
3
3 6
0
-
-
9
9
0
-
0
0 6 0 . 3 -
0 .
-
2 0 0
0 1 7 0 0 .
.
. 7
0 . 7 .
4 .0
0
4 0 0
2
1 3 0 9
2
.
3
1 .
9
2 E
3
0 8
3
- 3
8 4
6
6
4
0
- 0
-
0
. 4
.
5 - - -
- - -
7
0 0 0
2 0 0 0
. . .
2
0 .
3 3 3 . .
3 3 3
6 6 4
5 2 0
5 1 7
9 8 9
-
0 -
-
. 0
0
2 . -
2 -
7 . 0
1 - 0 -
6 5 0 0
9 .
1 .
0 0 .
.
4 4 1 0
7
0 4
6
3
5
5
5
-
0 9
- 0
- .
2 9 6 0 . -
-
2
0 0
0
. 2
9
.0
0
0
.
.
4 .0
4
0
1 0
1 8
.
0 5 3
9
.
8
4 1
7 3 7
3
0
3
2
5 6
- 0
-
0
. 4
.
0
5 -
- - -
0
9
4 0 0 0
-
.
. 3
8
4 3
0
. . .
3 3 3
6
.
7
7 6 5 3
6
9
3
5 6 2
7
-
0
-
0
. - -
2
. 0 - 0
0
- 0
2 1
.
.2 - 0
.
1 8 1 . 0
1
7 . 2
4 9
2 6 5 0
1
3
5
7
9
-
0
- 0
.
2 9 3 0 . - 0
3 0
-
.2
.0
0
0 . 1
. 0
- 0
0
0
6
9
9 6 0
3
9
3
.
.
2
5
. 4
4
1
5
3
1
1
3
0
7 0
- 0
-
-
0
. 5
0
.
- -
2
5 -
0 0 - -
0
6
. 3
0 0
7
. .
8 3 3 . . .
8
3 3 3
8 8
8
2
7 6
8 9
3 3 2
-
-
0
0
-
. 0 -
2 . -
0
2 0
.
5 1 .
1
1 .
- 0
0 1
- 0
.2 8
6 6
4
. 2
8 3 7
3
7 4
3
- 0
0 2 9 8 0 . - 0
0
.
.0
-
3
. 2
.
0 2
- 0
0
2
0
0
. 0
0
3
8
9 7 .
0
. 4
8
.
1
1
6 6
2
1
3
1
9
8 6
- 0
-
0
. 5
.
5
5
9 - -
- - -
0 0
9
5 0 0 0
. 3
. .
. . 3 3 .
9
4 4 3
9 9
0 0 8
6
8 1
0 2 3
- - -
0 0
0 -
0 -
. . .
2 2 0
2
.
2 .
8 5 3
1
0
- 0
2 5
. 3 0 9
9
0 1
6
. 2 8 3 - 0
- 0
0 1 1 0 . 3 - 0 0
. .
.0
-
-
3 0
. 2
0 2
. 0
5 2
2
7 0
0
0
2
0 6 9
.
3
.
.
1
4
2
4
4
5
3 2
0 3
-
0
-
-
0
. 5
.
6
9
2 -
0 - - -
-
5
5
4
0 0 0
0
.
4 . . .
4 4 4 .
4
1
1 3
1 1 0
0
4
5 2 7
2
-
- -
0
0 0 -
-
. 0
0
. 2 .
3 2 .
9 2 .
2 7 2
5
8 4
0 6
7
0
- 0 . 3 3 5
3 0 1 - 0 .
- 0
. 3 3 1 - 0 0
0
0
.
. 0
. 0 0
-
-
3
. 3
2
. 0 2
4
7 0
8
1
4
0
0
2
1 .
8
1 1
.
.
4
5
4
4
7
4 9
2 3
-
-
0
-
0
. 6
.
6
3
5
- -
4
0 0 - - -
7
9
0 0 0
. .
4 4
. . .
4 4 4
3 1
2 2
2 2 1
8 7
4 0 8
-
0 -
- - -
0
0 0 0
.
3 .
. 3 . .
3 4 3 2
2
6 0 8
4
4
2 5 7
2 7 0 . 3 - - 0 . 3 7 1
3 6 0 - 0 .
- 0
- 0
0
0
.
0
. 0
0
-
.
2
3
. 0
. 3
-
2
7
8 0
9
6
0
5
3
.
0
9
4
6 1
.
4
0
.
2
4
4 4
6 8
- 0
-
-
0
. 6
0
.
7
9
.
0
1
-
4
0 - - - -
1
0 0 0 0
.
4 . . . .
4 4 4 4
5 1
4
4 3 3 3
4
0 6 1 0
-
-
0
- 0 - -
0 . 0 0
3 .
. 3 . .
4 9 3 3
7
0 3 5 3
3
9 4 4
3 6 4 - 0 .
3 9 9 - 0 . 0 . 4 1 4 -
0
0
- 0
-
. 0
0 0 . 3
0 1
-
-
. 0
. 0
.3
4
9
3
9
.
0 1
1 1
0
0
0
2
8 2
.
.
0
4
4
9 1
6 8
-
0 -
0
.
7 .
7
6
5
1
5 -
- -
0
0 0
-
.
0 4 . .
4 4
. 4 - -
4 3
4 0 0
6
5 1 9 . .
4 4
4
6 4 -
0
1 6 - -
0 0
.
4
. .
4 3
2
0 7
5
3 9
4 6 1 - 0 .
4 4 7 - 0 .
4 1 3 0 . -
- 0
0
0
-
- 0
.
0 0
. 0
. 0 0
-
4
3
. . . 3
7
-
6 4 0
. 3
-
1
4
0
5
4
2 9 0
0
.
0
0
3 5 6
.
6
1
.
4