Professional Documents
Culture Documents
Contents
Foreword xv
About the Authors xvii
Acknowledgements xix
6.7.1 Continous Deck with a Single Fixed Point Located at One of the
Abutments 203
6.7.1.1 General 203
6.7.1.2 Examples 204
6.7.2 Continous Deck with the Fixed Point Located on One of the Central
Piers 211
6.7.2.1 General 211
6.7.3 Simply Supported Spans Without Longitudinal Continuity, with a Fixed
Point on Each Span 211
6.7.3.1 General 211
6.7.3.2 Example 212
6.7.4 Fixed Points at the Two Abutments and a Structural Joint in the
Middle 212
6.7.4.1 General 212
6.7.4.2 Example 214
6.7.5 Deck Divided into Several Continuous Stretches, Each One Including
Several Spans and One Fixed Point 216
6.7.5.1 General 216
6.7.5.2 Example 217
6.7.6 Especial Situations 218
6.7.6.1 Seismic Design 218
6.7.6.2 Exceptional Geometries 226
6.7.6.3 Example of Exceptional Geometry 226
6.8 Example of Track–Structure Interaction 229
6.8.1 Verification of Stresses in the Rails 229
6.8.2 Verification of Horizontal Displacement at Abutment 2 Due to Braking
and Traction Forces 231
6.8.3 Verification of Horizontal Displacement at Abutment 2 Due to Vertical
Train Loads 232
6.8.4 Verification of Vertical Displacement at Abutment 2 Due to Vertical
Train Loads and Temperature Variations 234
References 235
Index 331
xv
Foreword
At the request of the authors, I have been given the honour of writing the foreword
to this book, which is devoted to railway bridges. It develops the aspects referring to
their structural conception, taking into account the characteristics of railway traffic:
actions, limit states, speeds, etc., and includes a detailed analysis of the superstruc-
ture of the track with its different components and singular elements (for example,
expansion devices) that allow the correct behaviour of the track.
In the following chapters, the knowledge and experience of the authors is passed
on. In this respect, I remember a technical conference that took place in the 1970s
at the Eduardo Torroja Institute, dedicated to bridges; at that time, the undersigned
engineer was assigned to the Renfe Bridge Division and attended it. Ramón
del Cuvillo, professor of Concrete at the School of Civil Engineering in Madrid,
presented a paper in which he focused on the defects and mistakes in design and
execution in projects and works in which he had been involved. His presentation
was the most applauded of the day’s and, personally, the one from which I learned
the most. I hope that reading this book will be useful to avoid the repetition of
problems that can be avoided, without having to wait for experience after the
execution of the works.
As the reader will appreciate, special emphasis is placed on the interactions
between the structure and the track, subjected to railway and environmental
actions, taking into account the requirements of their stability in different situ-
ations; solutions are also proposed and considered in relation to the transitions
between the bridge and the adjacent infrastructure (and track).
Special attention is paid to the dynamic nature of railway actions, studying the
dynamic response of the structure and its influence on the behaviour, also dynamic,
of the track and its components, with the repercussions that this may have on safety,
traffic flow quality, and maintenance needs.
To conclude, I would like to transmit here some ideas that the Emeritus Professor
of Structural Engineering of the University of Berkeley, Edward L. Wilson, sets out in
his book Static and Dynamic Analysis of Structures. In a section of Personal Remarks,
he relates that his first-year physics professor warned his students ‘not to use an
equation they could not prove’; he also advises, with respect to modern structural
xvi Foreword
engineering, ‘not to use a structural analysis program unless you fully understand
the theory and approximations contained in the program’. I fully agree with these
considerations; I therefore share them with the reader, in the hope that they will be
useful to them.
José Romo is Chief Executive Officer and partner of FHECOR, and also a bridge
engineer fully specialised in large-span bridges with more than 40 years of expe-
rience in bridge design, 35 of them working in FHECOR. He has vast technical
knowledge based on his design background complemented with his activity as
professor of concrete and steel structures at Madrid University, and his active
participation in national and international associations of bridge designers and
concrete and steel materials. He is a member of many scientific committees such as
Eurocodes, IABSE, and ACHE where he became president in 2014 and was awarded
with the honour’s medal in 2008. He is a fellow of the Institution of Civil Engineers
of UK. He has always worked as a bridge designer participating in innumerable
bridge projects in Spain and worldwide, and also in the construction engineering
for many of them. He has a great aesthetic vision that he applies to all the designs,
while having great concern for sustainability and the use of new materials and
construction techniques.
Acknowledgements
The authors would like to express their most sincere thanks to all the staff of
FHECOR and the Caminos de Hierro Foundation who gave us their support and
assistance in the creation and publication of this book.
We are particularly grateful to Francisco Javier Fernández Pozuelo, for his com-
mitment and dedication to this initiative, and to Eduardo Romo for his support from
the Caminos de Hierro Foundation.
The authors would like to express their deepest gratitude to Jorge Nasarre for his
help in the general approach to the publication and its subsequent technical review,
to Julio Sánchez for his advice and technical reviews, to Fabrice Leray for the prepa-
ration of the graphic material and the conception of the book’s design, to Eduardo
Conde for the layout work, and to Marta Heras for her coordination. We would also
like to thank all the FHECOR engineers who prepared the High-Speed Bridge Design
seminar, which was the seed of this book. Without their help and without the strong
support of FHECOR and the Caminos de Hierro Foundation, this book would not
have been possible. Many thanks to all of them.
1
must be analysed to calculate the dynamic response, as well as the way to consider
other aspects of the response, such as the irregularity of the track and the vehicle or
the interaction between the vehicle and the structure, are presented. The chapter
is completed with several practical examples and an appendix which includes
the theoretical aspects of general dynamics and their application to the analysis
of HSRB.
Chapter 6 is dedicated to the interaction between the track and the structure.
This section analyses this phenomenon and how to take into account the thermal
effects, traction and braking forces, vertical loads and rheological effects, in the case
of concrete decks. In addition to the analysis models, the checks to be carried out
to calculate stresses in rails and relative displacements are analysed. This chapter
also deals with the criteria for the placement of track joints, as well as the practical
application of the worked example.
Chapter 7 deals specifically with aspects linked to the conceptual design with
maintenance of bridges for high-speed rail lines in mind.
In addition to Chapters 1–7, the book includes two appendices. One is devoted
to a review of the general concepts of dynamics that the reader of Chapter 5 on the
dynamic behaviour of these bridges should be familiar with. The second appendix
includes a ‘register’ of high-speed railway bridges built in different parts of the world.
Figure 1.2 Span by span isostatic solution: China, China Railways (Courtesy of China
Railways).
braking force. This allows the elimination of joints in the rails but on the contrary,
there are structural expansion joints in all sections of the deck coinciding with the
piles. This solution, which might seem better from the point of view of track mainte-
nance, has the disadvantage of higher maintenance of the structure and in the case
of bridges in seismic areas, the lack of robustness in combination, which could cause
relative movements between adjacent decks during the seismic actions. This type of
design is being used very often in China because it is highly industrialisable and
because it allows for very flexible construction (Figure 1.2).
There are some intermediate alternatives that involve the construction of a series
of continuous sections having intermediate joints every certain number of spans.
This solution has recently been used in Germany [2] and [3], shown in Figure 1.3.
6 1 Introduction to High-Speed Railway Bridges
Figure 1.3 Gänsebachtal Viaduct (schlaich bergermann partner sbp), Germany, DB Netz AG
(Source: Störfix).
In summary, and from a typological point of view, high-speed bridges have some
specific aspects which make them, at least from a longitudinal perspective, different
from other bridges, as will be seen in Sections 1.4 and 1.5.
Horizontal Horizontal
fixed point fixed point
SEJ 162.5 SEJ 168.0 SEJ 299.0 SEJ 164.0 SEJ
SEJ = Structural expansion joint
TEJ = Track expansion joint
LFC = Longitudinal force coupler
Figure 1.4 Structural scheme bridge over the river Main at Gemünden (1984).
Figure 1.5 Structural scheme bridge over the river Main at Veitshöchheim (1987).
Figure 1.6 Structural scheme bridge over the Pfieffetal Viaduct (1989).
8 1 Introduction to High-Speed Railway Bridges
LFC LFC
Horizontal TEJ TEJ Horizontal
fixed point fixed point
Figure 1.7 La Grenette Viaduct with ‘inert’ section and double expansion joints of
structure and track.
TEJ TEJ
Horizontal
fixed point
one or two intermediate fixed points in the form of A-shaped piers that transmit the
braking loads to the ground (Figure 1.9).
On the other hand, Spain’s rugged orography has given rise to a series of long-span
bridges in which structural rigour has been combined with the aesthetic and land-
scape aspects of the works, as described below [7].
66.0
112.0
Structural expansion joints
Figure 1.13 Isostatic multi span bridge, China (Courtesy of China Railways).
Design
Name of the Main Year of Operation
Bridge Span (m) Typology Completion (km/h) Function* Country
Figure 1.14 Colne Valley Viaduct of HS2, England, UK (Courtesy of Knight Architects).
Figure 1.15 Fresno Viaduct in California, USA (Source: California High-Speed Rail
Authority).
1.4.3 Conclusions
In order to establish an analysis of trends in the design and construction of
high-speed bridges, it is first necessary to make a clear distinction between viaducts
1.5 The Landscape and the Design of High-Speed Railway Bridges 13
of more or less length on the one hand, and bridges requiring a long main span on
the other hand.
1.4.3.1 Viaducts
There are currently three types of solutions for viaducts with moderate spans and
shorter or shorter lengths.
In Europe, with the exception of Germany, continuous bridges are generally built
with a minimum number of structural expansion joints and corresponding track
expansion joints. The deck is supported by two devices per pier and per abutment.
The bridges on the US lines also generally follow these design criteria.
In China, on the other hand, isostatic solutions are being built with a complete
prefabrication in one piece of the deck of each span and therefore with an expansion
joint on each pier, but without track expansion joints.
As seen in Germany, semi-integral bridges are being designed and built, i.e. with-
out pier bearings, but with structural expansion joints, and in some cases without
track expansion joints.
What a HS2 user would see from the window of the train if a solid noise barrier is
arranged (static image on the left and train running at a speed of 340 km/h on the right).
Users wouldn’t even notice they are crossing the Colne Valley
Figure 1.16 Example of a standard anti-noise panel on the bridge (Courtesy of Knight
Architects).
However, it is clear that this is no longer enough. The 21st century bridge designer
must also consider the landscape that the traveller will be able to contemplate when
the train travels over the bridge being designed [11].
When this aspect is analysed, it is discovered that the structure itself rarely
obstructs the view from the train in any way. However, it is common that parts of
the bridge equipment, especially the anti-noise or wind barriers (when these are
opaque), disturb or limit the view of the landscape from the train (Figure 1.16).
If the bridge is short, the loss of vision caused by such panels would only be for
a few seconds. However, when the tracks run continuously through urban or
peri-urban areas, the tunnel effect can be annoying or uncomfortable for the user.
The same applies when the railway line passes through a point of outstanding
scenic beauty, such as a major river crossing, if the passenger’s view of the outside
is limited by some element of the bridge.
In such cases, it will be important for the viaduct designer to be aware of whether
the structure requires any type of panelling that will at least partially obstruct
the vision of the traveller. Whether it is necessary to install panels or whether
it is the structure itself that is disturbing, for example if the resistant section of
1.5 The Landscape and the Design of High-Speed Railway Bridges 15
What a HS2 user would see from the window of the train
with the transparent Specimen Design barrier (static
image on the left and train running at a speed of 340 km/h
on the right). At train speed, motion blur makes the vertical
elements of the edge condition almost completely invisible,
achieving an unobstructed view of the Colne Valley.
Figure 1.17 Study of the view from the train as it passes over the Colne Valley Viaduct,
England, UK (Courtesy of Knight Architects).
the deck is U-shaped, the project team must analyse whether it is possible to
reconcile functional requirements (noise emission control, wind safety, etc.) with
the possibility of the traveller being able to enjoy the landscape at least for a fleeting
glimpse (Figure 1.17).
number of viaducts and tunnels as soon as the terrain has some movement. Even
in flat terrain, it is common for modern high-speed lines to be built in structure in
order to maintain transverse territorial permeability under bridges. It is very impor-
tant in these cases to decide correctly on the level of the railway grade on the ground
because of its implications for the design of the viaducts.
The participation of bridge specialists in the early stages of the project is important
for the definition of the basic geometry of the line and to avoid starting the project
with initial conditioning factors that could damage the overall quality of the solution,
for example: the transverse permeability, the landscape implications of the design,
the technical quality, or the construction cost of the work.
Sections 1.5.2.1 and 1.5.2.2 analyse the landscape aspects of bridges and viaducts
on high-speed lines in different scenarios.
Continuous Bridges Continuous bridges have the advantage of reducing the number
of expansion joints in the structure. When these bridges are long, they require one
or more points to fix the deck longitudinally.
The first example of this way of solving the central connection by means of a sin-
gle element is the Pfieffetal Viaduct in Germany, 1989 (Figure 1.21). This bridge is
actually an isostatic span bridge, but because of its height the piers cannot carry the
braking load, which is transferred to a portal pier with two inclined piers. The shape
of the V-shaped valley makes the role of this central pier very clear.
18 1 Introduction to High-Speed Railway Bridges
Figure 1.22 Bridge over the river Main at Gemünden (1984), Germany (Courtesy of
Deutsche Bahn AG).
Another type of situation occurs when a long viaduct is required which can be
resolved with modest spans, but which presents a singular span due to having
to cross a major obstacle locally. This type of solution perhaps begins with the
Gemünden Bridge (Figure 1.22) , which serves as the cover of the most widely read
book on bridge aesthetics [12].
However, to return to very long viaducts that require a single span, the revolution
brought about in Germany by the Deutsche Bahn Guide [2] is worth mentioning. It
stipulates that long bridges for the Deutsche Bahn should generally be semi-integral
and as far as possible without a track expansion joint.
The first long bridges designed according to these guidelines are the Unstruttal
(Figure 1.23) and Gänsebachtal (Figure 1.24) viaducts [3]. Both bridges are superb in
terms of design, structural efficiency, maintenance of both bridge and track, as well
as structural innovation, with an obvious reading on the landscape to the trained eye.
Singular Bridges Another classic design situation occurs when the obstacle to be
overcome is significant and it is necessary to build at least one large span. This is
a situation that occurs when crossing deep valleys or when passing over very wide
and fast-flowing rivers or streams.
When crossing deep valleys, it is common for the bridge to be a short interval
between tunnels. This is the case, for example, with the colossal Beipanjiang Bridge
(Figure 1.25) on the high-speed line from Shanghai to Kunming in the Chinese
1.5 The Landscape and the Design of High-Speed Railway Bridges 19
Figure 1.23 Unstruttal Bridge (2012), Germany (Courtesy of Deutsche Bahn AG).
Figure 1.26 Alcántara Bridge (2019), Spain (Courtesy of Carlos Fernández Casado CFC &
ADIF).
Figure 1.27 Almonte Viaduct (2016), Spain (Courtesy of Arenas Asociados & ADIF).
province of Guizhou, which with its 445 m main span fits perfectly into the nar-
rowing of the gorge flanked by two tunnels [13].
In areas with a hilly but not necessarily mountainous topography, it may be neces-
sary to build a large main span accompanied by two long access viaducts. This is the
case, for example, of two Spanish bridges, the Alcántara viaduct (Figure 1.26) with a
span of 324 m [14] and the Almonte viaduct (Figure 1.27) with 384 m main span [15].
In both cases, the continuity of the deck and the careful design not only of the
main span but also of the access viaducts stand out, making these works excellent
examples of the concern for the visual aspect typical of Spanish engineering.
The above three are good examples of how well rigid arches fit into a rugged land-
scape when there is a major obstacle.
It should be pointed out here that, although at first sight the arch is a typology
which is not very suitable for use on the railway as it has to support significant vari-
able loads, in long-span bridges, the weight of the structure itself is the dominant
1.5 The Landscape and the Design of High-Speed Railway Bridges 21
gravity load compared to the railway overload and it is sufficient for the arch to be
sufficiently rigid for it to function correctly.
Another special design situation for high-speed line bridges is the crossing of
large rivers or estuaries. Especially when the crossing depths or navigation demands
require long spans.
The solutions in these cases are usually trusses, either with straight girders (for
moderate spans) or bow-string or cable-stayed bridges for large spans.
The Nantenbach Bridge (Figure 1.28) in Germany [16] and the Ulla Bridge
(Figure 1.29) in Spain [17] are good examples of the application of the variable-edge
straight-deck typology at river or estuary crossings. In both cases they are bridges
for two high-speed railways tracks.
When it is a question of covering very long spans and with many railway tracks,
the solution is usually a double-level cable-stayed bridge deck. This is the case, for
example, with the Hutong Bridge over the Yangtze River (Figure 1.30). The bridge
Figure 1.28 Nantenbach Bridge (1993), Germany (Courtesy of Deutsche Bahn AG).
Figure 1.30 Hutong Yangtze River Bridge, China (Courtesy of China Railways).
has a colossal main span (1092 m) in which the upper part of the deck serves a motor-
way while the lower part accommodates the high-speed railway.
In both cases the piers are bottle-shaped. An alternative to these piers is the one
proposed for the Third Tagus River Crossing in Lisbon (Figure 1.31), also designed
with a two-level deck. In this case, the pylon has A-shaped outer shafts, which give
the pylon a very clean form despite its robustness.
Fortunately, in the case of long-span or long bridges, the site and the functional
requirements they have to meet make each bridge a unique and unrepeatable
example. This should at least calm the desire for novelty [19].
This does not mean that designers should abandon the care for the formal aspects
of their bridges. On the contrary, they should strive to combine appearance and sub-
stance. An effort that should be directed not only towards achieving excellence in
singular works, but also towards seeking beauty, following Yanagi’s ideas [20], in
normal bridges, those that are seen every day.
The expert eye and perhaps also the non-specialist will appreciate the effort made
in many high-speed bridges to wisely reconcile these in no way contradictory aspects
of design. These exemplary works will surely become part of the collective heritage
and culture of the technical community and society in general.
Since its beginnings, the railway has marked the evolution of bridges. The condi-
tions required by the 19th century railway marked the development of structural
engineering, both in terms of materials and in the typology of the bridges themselves
(Figure 1.32).
It is the epic era of engineering marked by successes but also by great failures. The
dynamic loads due to the railway, the fatigue of the materials gave rise to resounding
accidents that served as a lesson to move towards safer structures.
The magnitude of the challenges and the integrity of the engineers had a strong
impact on the society of that time. The material and human development that
was achieved thanks to this new means of transport revolutionised the society of
that time.
Society as a whole and art did nothing but focus on the railway and its works.
Bridges appear in paintings first and then in photography creating a whole imaginary
regarding the progress created by the engineers.
Figure 1.32 Maria Pia Bridge (Ponte de Dona Maria Pia, 1878) Porto, Portugal
(Source: Seyring [21], photo: José Olgon).
24 1 Introduction to High-Speed Railway Bridges
Figure 1.33 Proposal for the Ulla Bridge, Spain (Source: FHECOR).
Perhaps the railway bridges where the first works in which society could admire
the plastic and formal value of the engineers’ work. The visual strength of these
structures undoubtedly had a favourable impact on a society that admired the value
of the technique with fascination [22].
1.9 Conclusions
HSRB are the best example of how a technology, the railway, is developed to reach
land speeds only previously imagined by man. The speed and the necessary safety of
the lines require highly demanding structures that have to meet strict design criteria.
The structural engineering of the moment has very powerful design and calculation
tools that allow the design of these bridges that were unthinkable a few decades ago.
References 25
References
1 Manterola J., Astiz M.A., and Martínez A. (1999). Puentes de Ferrocarril de Alta
Velocidad Revista de Obras Públicas n∘ 3386 Madrid.
2 Schlaich, J., Schmitt, V., Marx, S. et al. (2008). Leitfaden Gestalten von
Eisenbahnbrücken, 1e. Berlin: DB Netz AG.
3 Schlaich M. (2012). Integral Railway Bridges in Germany. 22nd Dresdener
Brückenbausymposium. Dresden.
4 Zellner W. and Saul R. (1991). Long span bridges of the new railroads lines in
Germany. IABSE Report 64.
5 Ramondenc P. and Bousquet C. (2004). The main bridges of the high speed line
HSL Méditerranée FIB Sympsium Avignon.
6 Chatelard P., Martin O., Roujon M. and Sayn P. (1998). Lot 2H – Les viaducs
d’Avignon. In: Travaux, n. 742.
7 Recuero, A. (2014). Viaductos singulares del Siglo XXI – Ferrocarril. Revista del
Ministerio de Fomento.
8 Yazbeck, N. (2010). LGV Rhin-Rhône, Branche Est – Les études du viaduc de la
Savoureuse. In: Travaux, n. 870.
9 Russell K. (1983). The Age of Sentiments Modern Age; Chicago, Ill. Vol. 27,
Iss. 3.
10 Dewey, J. (1934). Art as Experience. New York: Minton, Balch & Company.
11 Knight, M.; Beade, H. Colne Valley viaduct specimen design. Government UK
Publications.
12 Leonhardt, F. (1983). Bridges Aesthetic and Design. Deutsche Verlags-Anstalt
GmbH.
13 Chen, B., Su, J.Z., Lin, S. et al. (2017). Development and application of concrete
arch bridges in China. Journal of Asian Concrete Federation 3 (1): 12–19.
14 Manterola, J., Martínez, A., and Martín, B. (2015). Viaducto sobre el río Tajo
en el embalse de Alcántara para ferrocarril de alta velocidad. Revista de Obras
públicas n∘ 3562 Madrid.
15 Arenas, J.J., Capellán, G., Martínez, J. et al. (2016). Viaduct over River Almonte.
Design and analysis. Presented at:. In: IABSE Symposium: Challenges in Design
and Construction of an Innovative and Sustainable Built Environment, Stockholm,
Sweden.
16 Zellner W. and Saul R. (1991). Railway bridge with double composite action
across river Main IABSE Report 64.
17 Millanes, F., Ortega, M., and Matute, L. (2014). Viaduct over river Ulla: an
outstanding composite (steel and concrete) high-speed railway viaduct. Structural
Engineering International 24 (1): 131–136.
26 1 Introduction to High-Speed Railway Bridges
2.1 Introduction
Track presents a high level of interaction with the infrastructure components. This
deep relationship means that the optimum design is reached considering the civil
work components, as earthwork or bridges, and the track together along the design
process.
Both the short-term response of the infrastructure under traffic loads and
its long-term behaviour have an impact on railway ride safety and ride quality,
especially in the case of high-speed lines.
In this sense, it must be pointed out the importance of controlling the vertical track
stiffness, related to the short-term combined response of track and infrastructure,
and the evolution of track geometry quality, related to the long-term behaviour and
finally their influence on vehicle dynamic and, ultimately, on ride safety and ride
quality.
In the case of bridges, the design criteria related to the track have a high impact
on the final design of the structure. The present chapter is focused in describing the
track for a better understanding of this design impact.
Dead loads: x4
Figure 2.1 Differences in dead loads between road bridges and railway bridges.
Traffic loads: x3
Figure 2.2 Differences in traffic loads between road bridges and railway bridges.
That justifies the lower slenderness depth/span of railway bridge decks (about 1/12
to 1/15, depending on longitudinal scheme and the construction method, instead of
1/20 or lower of road bridges).
But not only are vertical loads higher, but also horizontal, due to transverse cen-
trifugal and nosing forces and longitudinal traction and braking forces are higher.
Beyond the magnitude of loads, it is also important the dynamic character of loads,
since it can produce resonance in the case of railway bridges, as it will be presented
in Chapter 5.
On the other hand, the fact that traffic loads are relatively important compared to
permanent loads makes the bridge subject to many cycles of high stress variability,
which can lead to fatigue problems.
But the question of what the main difference between railway bridges and road
bridges is would be still open, because the abovementioned difference in loads is
not the only significant difference. In fact, it might be said that it is not the most
important one.
The design of a railway bridge must never forget the importance of the interactions
between the bridge, rail, and train to result in an adequately functioning system. Due
to that interaction, vibrations shall be limited to ensure both track stability and ride
quality and safety; movements shall be limited to guarantee again a proper level of
track stability and track geometry quality, as well as a limitation of efforts on the
different track components; these aspects are critical.
Thus, the set of verifications related to the track, regarding deformations and vibra-
tions for railways bridges are extremely important, a set of limitations in which
criteria for traffic safety and ride quality are clearly identified. These criteria will
be presented in Section 0.
2.3 Description of the Track Superstructure 31
It must be pointed out that these limitations related to the track, regarding defor-
mation and vibration, are Serviceability Limit State (SLS) for the bridge, but Ultimate
Limit State (ULS) for the track and vehicle, as they are related to running safety
(Figure 2.3).
– softening the actions that vehicles cause on the track and transmit them to the
subgrade (distribute uniformly the load and vibrations, softening them, longitu-
dinally and transversally, from the sleepers to the subgrade);
– limiting the movements of the track, stabilising it on the vertical, longitudinal,
and transverse direction;
– providing elasticity and capacity to absorb energy to the track, making it a softener
layer for higher ride quality;
– enabling drainage of rainwater;
– protecting the subgrade materials from the action of the freeze;
– avoiding leakage of electric current (track insulation depends on ballast bed con-
tamination);
– enabling the establishment and recovery of track geometry through levelling,
aligning, and achieving the cant track, by tamping operations;
– reducing the noise generated by the running of the trains.
– elasticity enough to absorb the actions of vehicles and to distribute the loads;
effective resistance to prevent horizontal displacements of the track;
– appropriate porosity rate (without affecting elasticity) to allow storm water
evacuations;
– stable properties to deal with the action of water and ice;
– be easily compacted by mechanical means to support elastically the loads trans-
mitted by the sleepers and allow recovering the track initial geometry.
2.3.1.2 Sleepers
They are the cross-structural elements that, resting on the ballast, provide support
to the rails.
Sleepers have as fundamental missions:
– additionally, they contribute to preserve the road running surface during construc-
tion, in case of break of the rail, after derailments, and to absorb vibrations of the
track and lower the acoustic impact on the environment.
2.3.1.3 Fasteners
It defines a piece or a group of pieces of a track system that fix the rail to the sleepers
and keeps them in the required position, while preventing (or partially allowing)
vertical, cross, and longitudinal movements.
The performances required to these elements are of three types: (i) mechanical,
(ii) electrical, and (iii) functional.
Mechanical Allocate, absorb, and/or transmit the traffic loads (static and dynamic)
to the sleeper and keep rail attached to the sleeper with enough tightness in all pos-
sible cases.
Thus, these mechanical functions are:
– avoiding the overturning of rails (torsion);
– setting the rails against their support to restrict lateral movements and inclination
keeping the track gauge within a permissible tolerance;
– contributing to the vertical elasticity of the track together with the ballast, function
fundamentally allocated to the clips and elastic pads;
– decreasing the dynamic effects from the rails to the sleepers through the impact’s
absorption in the vertical direction of the loads and vibrations induced by the pas-
sage of trains, mission assigned to the elastic pads;
– preventing longitudinal sliding of the rail on the sleeper;
– controlling the forces on the rails caused by temperature changes (contraction and
expansion).
Functional
– preventing the components wear on the contact surface with the sleeper;
– getting a natural frequency of vibration greater than the rail one, to prevent
contact lost between rail and wheel;
– elastic deformation enough to maintain an elastic reaction under maximum defor-
mation.
2.3.1.4 Rails
The rail is the longitudinal metallic element (steel), which physically constitutes the
road running surface of the train wheels. Trains run along on these elements with
steel wheels, using the key feature of the railway that is the low friction between both
components.
34 2 Track for High-Speed Bridges
Rail must have several features that allow it to withstand a complex set of efforts,
which must fit its profile, its length, and its metallurgical composition.
It is therefore essential to ensure its optimum quality and conservation, which
involves a systematic and continuous analysis of its flaws, cracks, and wears.
The quality requirements are higher when the rail lines support passenger traf-
fic, particularly in the case of high-speed services. They are linked to the following
functions:
– unidirectionally guiding the rolling material with maximum continuity, both in
plan and elevation;
– serving as absorber, resistant and transmission element of the thermal stresses
and traffic loads – vertical, transverse, and longitudinal – to the sleeper and, thus,
to the infrastructure;
– acting as a conductive element for returning the electric current of the vehicle’s
traction, as well as the electric supply of signalling circuits;
– not rapidly aging by weather (oxidation).
To withstand stresses of different nature and high value, rails must achieve the
following requirements:
– high wear resistance;
– high compressive strength;
– high fatigue resistance;
– high yield, tensile strength, and hardness;
– high tear resistance;
– good welding properties;
– high degree of purity of the materials;
– good quality of the rolling contact surface;
– good flatness and fidelity profile;
– reduced internal strains after the manufacturing process.
(d) Rail expansion joints: is a track device allowing the relative longitudinal
movement between two adjacent rails while ensuring correct guidance and
support. These longitudinal movements can appear: at the end of continuous
welded rail (CWR), at the end of the structures, or at a combination of the two
precedent situations.
Such problems result in poor geometry of the track and rapid wear of the elements
which form the joints, reducing the speed of circulations and causing a lack of ride
quality, which would be lower, the smaller the lengths of the rails between joints are.
Considering these problems caused by rail joints, in the early twentieth century
CWR began to be used. It was first introduced in trams, which had rails embedded
in the pavement and thus hindered their movements on rails and tunnels which
suffered minor variations in temperature. Today their use has been fully generalised
with the improvement of design criteria and welding techniques.
(a) (b)
Figure 2.5 Ballast (a) and ballast detail (b) (source [2]).
2.3 Description of the Track Superstructure 37
2.3.3.2 Sleepers
Sleepers, being an artificial element designed specifically for railway function,
have evolved in parallel with the advance in technology and advances in the
understanding of the behaviour of the railway infrastructure.
The main types of sleepers that have developed throughout history are presented
below. It must be noted that all these types are still used in different worldwide
networks.
(a) Timber sleeper: Are wood elements, prism-shaped, with highly variable length
depending on the gauge and approximate cross-sectional setting section 25 cm
wide by 14 cm high. Those elements are protected against external actions by
impregnation with a product derived from the distillation of coal oil (Figure 2.6).
(b) Steel sleepers: Generally inverted U-shaped. These sleepers are light,
around 80 kg of weight, and have deterioration problems by oxidation in wet
environments. They allow an industrial design which drastically increases the
homogeneity between the different units and at the same time allows designing
rail support areas so that they are fixed (Figure 2.7).
(c) Bi-block concrete sleepers: With the emergence of quality-controlled and
suitable for prefabrication concrete bi-block prefabricated sleepers were
developed. They have two blocks of reinforced concrete and a steel stay that
maintains a fixed separation between the blocks. They have little horizontal
and vertical rigidity which reduces the rigidity of the track. Also, they bring the
problem of corrosion (Figure 2.8).
(d) Monobloc concrete sleepers: Finally, this is the most used typology in mod-
ern railways. They are made of reinforced concrete and prestressed concrete. Its
main advantages are the bigger weight, which increases the lateral stability, and
higher transverse and vertical rigidity and durability (Figure 2.9).
(e) Special sleepers: There are innovative designs that have been developed for
special situations or have not been successful because they produce significant
changes in the processes of construction, maintenance, and renewal of rail
infrastructure. One of the most important examples would be frame sleepers
developed in Russia. These sleepers achieve increased vertical and transversal
stiffness, thereby increasing the lateral stability. Besides the longitudinal
stiffness is homogenised because rail is supported longitudinally (Figure 2.10).
2.3 Description of the Track Superstructure 39
2.3.3.3 Fastenings
There are several types and classifications based on different criteria.
One of the most important classifications is direct, in which the rail and the
sleeper are directly connected through rail-pads and indirect, in which between
the rail and the sleeper there is the baseplate system, which includes a rail-pad and
a baseplate-pad.
Secondly, fasteners can be classified into rigid, semi-elastic, and elastic.
The most used types of fasteners based on both classifications are the following:
(a) direct rigid (with nail or lag-screw);
(b) direct elastic (Figure 2.11);
(c) indirect elastic (Figure 2.12).
(a) (b)
Figure 2.11 Elastic direct fastening. (a) Nabla, (b) Pandrol Fastclip (source [2]).
40 2 Track for High-Speed Bridges
(a) (b)
Figure 2.12 Elastic indirect fastening. (a) Pandrol. (b) Vossloh (source [2]).
2.3.3.4 Rails
The rail is made of steel since XIX. Initially it was given a profile with variable depth
gauge between supports made of stone dices (the ‘fish belly’ rail), leading to the first
type of rail that deserves to be mentioned (Figure 2.13).
The manufacturing difficulty of these rail resulted since then to prism-shaped
profiles to roughly double T. Numerous profiles that were discarded rapidly until
the appearance of double-head rail Bull-Head whose section had double symmetry
were developed. Thus, manufacturing is simplified and it could be reused turning
it. However, when trying to reuse, the lower head was warped which prevented it
reuse. Further advances in scientific knowledge about rolling forced a joint design
with the profile rail in coordination with the profile of the wheels resulted in
the appearance of the Vignoles-type rails, composed of three distinct parts, head,
web, and base. Some types of Vignoles rail type standardised by UIC are shown in
Figure 2.14.
(a) (b)
Figure 2.14 Vignoles rails; (a) UIC 46 E1; (b) UIC 60 E1.
Groove
Checkrail
Web
Foot
Finally, for cases in which compatibility of rail and road traffic must be assured,
appropriate rails, as grooved rails, to be embedded in the general pavement are used
(Figure 2.15).
Figure 2.16 Ballasted high-speed track. (a) Track in execution. (b) Madrid – Valladolid line
(source [2]).
4.650
4.000
3.350
1.818
1.100
E.E VIA
0.06
3
0.45 0.40 0.80 2
5% Ballast
0.30 Sub-ballast
5%
0.29
2
1 0.60
Blanket
5%
a reliability perfectly adapted to the life of railway lines which is estimated in about
100 years (Figures 2.16 and 2.17).
It also allows the maintenance and renewal processes, processes that can be per-
formed over the lifetime of the line without major affection to operation of the lines.
A typical ballasted track section in the case of a modern high-speed line would
consist of:
– heavy Rail, the UIC 60 type (60 kg/m) that allows a good distribution of verti-
cal and lateral forces. The geometric quality of the rails and their welds must be
especially careful to avoid the effects of wave formation;
– prestressed monoblock concrete sleepers with a length ≥ 2.50 m to ensure good
anchorage of the track in the ballast;
– the fastener would be formed by anchors, elastic pads, and resilient clips type
Skl-1 or similar, that guarantee, without maintenance, an excellent conservation
of tightening torque and an adequate vertical stiffness;
2.3 Description of the Track Superstructure 43
– a ballast with high hardness and important thickness (30 cm beneath sleeper),
which ensures a good resistance to wear and crushing, produced as a result of
dynamic overloads.
Rail on Rail on
resilient baseplate resilient baseplate
Derailment Derailment
upstand upstand
(optional) (optional)
Reinforced concret
e track slab
Bridge deck /
tunnel invert sla
b / etc.
The concrete slab may be built ‘in situ’, or using precasted parts. The asphalt base
is constructed under continuous compaction.
In some cases not only the ballast has been replaced, but also the sleepers. All
functions of the ballast and/or sleepers must be assumed by some of the components
of the ballastless track. The aim is to replace ballast by other material assuring a
greater stability track.
Thus, the fundamental components of the ballastlees track are:
– rail in long welded bars with elastic fastener (special for ballastless track);
– main slab (reinforced concrete or asphalt, on which rests the track formed by
sleepers, fasteners, and rails);
– base slab (can be gravel-cement or mass or reinforced concrete);
– subgrade (including protective layer and eventually anti-frost layer);
– anti-noise elements (optional).
In the case of ballastless track on a bridge, the deck will substitute the base slab
and the subgrade, as shown in Figure 2.18.
The classification of the different types of existing ballastless track can be made
according to Figure 2.19.
This classification considers on the one hand the elastic characteristics of the sys-
tem, and on the other the features of rail support.
Families
SLABTRACK CLASSIFICATIóN
1 2 3 4 5
Other Fixation Continuous Discrete
characteristics system
Rail levelling Every rail independently Using precasted elements
Precast
No No or block Sleper Slab
elements
Constructive Embedded Sleepers on Precast
Embedded rails Without sleepers
Elastic levels method sleepers support layer slab
EDILON PACT
1 elastic level on CDM porotrac TRANOSA JNR*
fastening TRANOSA
THYSSEN KRUPP
THYSSEN KRUPP
A Züblin BTE RHEDA
Sistema IVESA Heilit W. BES RHEDA berlin ATD BÖGL
2 elastic levels on APPITRACK RHEDA 2000 BTD ÖBB-
fastening CrailsheimFCC RHEDA Diwidag GETRAC PORR*
Rasengleis HEITKAMP IPA*
Hochtief/SM ZÜBLIN
NS blokkenspoor
Elastic level SONNEVILLE STEDEF
between block or B Bloques EDILON SATEBA
sleeper and slab Bloques TRANOSA
RAIL TECH
both track stability and ride quality and safety by the limitation of deformations
and vibrations of the track.
Regarding railway vehicle dynamics and then ride safety and passenger comfort, it
is interesting to review the classic Prud’homme formula for computation of dynamic
vertical overloads for short wavelength track defect:
√
𝜎ΔQ = V ⋅ A ⋅ mun ⋅ k
where
𝜎 ΔQ is the standard deviation of dynamic overload;
V is the running speed;
mun is the unsprung mass (wheelset);
k is the track vertical stiffness;
A is a parameter dependent on track geometry quality.
Thus, these vertical overloads are dependent on the speed, the track geome-
try quality, the unsprung mass, which is the mass that vibrates with the track
(the wheelset, considering that the rest of the mass of the train is connected to the
wheelsets through suspensions that represent a dynamic isolation), and finally the
vertical track stiffness. Track stiffness and track geometry quality will be then two
fundamental parameters of the track–infrastructure integrated behaviour.
One of the main objectives of track structural design reviewed in Section 2.3
is to provide adequate vertical stiffness of the track. A certain value of vertical
flexibility is required to distribute loads, and because one of the main aspects of
the mechanical behaviour of the track system is its dynamic response, and so, the
46 2 Track for High-Speed Bridges
Hypothetical
Q wheel force = 100 kN
Rail
Rail pad
Z
Sleeper CT describes
Sleeper pad the stiffness below
the wheel relative
CT Ballast or concrete slab
to a wheel force
Ballast mat
of 100 kN
Track stiffness
Substructure/subgrade or
concrete slab/base
For track geometry quality in Europe the specification in use is the standard EN
13848 [4, 5]. The main purpose of the standard is to define a minimum track geom-
etry quality to ensure safe operation of trains based on the experience of various
infrastructure managers.
Three indicators can describe the track geometric quality:
Alignment, cross-level, and track gauge variations are the major causes of lateral
vibration in railway vehicles, whereas vertical profile is the more important parame-
ter for vertical vibrations, through the parameter A in the Prud’homme formula seen
in Section 2.4.1.
Three main levels must be considered to make decision about maintenance
actions:
● Immediate action limit (IAL): refers to the value which, if exceeded, requires
taking measures to reduce the risk of derailment to an acceptable level.
● Intervention limit (IL): refers to the value which, if exceeded, requires correc-
tive maintenance.
● Alert limit (AL): refers to the value which, if exceeded, requires that the track
geometry condition is analysed and considered in the regularly planned mainte-
nance operations.
And, on the other hand, it is intuitive that not only the value of the defect is impor-
tant, but also the length in which we have the value. That it is not the same, a defect
of 1 mm in one meter than in one thousand meters. To describe a geometric error by
its wavelength (𝜆), three ranges are used:
Thus, limit values are given as a function of the wavelength range considered and
of speed, being the latter an important factor for the evaluation of track geometry
quality. As an example, we can see in this table the alert, intervention, and immediate
action limits for longitudinal level isolated defects (Table 2.2).
It can be seen how low those limit values are. For instance, for high speed a defect
of only 6 mm in 25 m is already an alert limit.
48 2 Track for High-Speed Bridges
Table 2.2 Limit values for longitudinal level – Isolated defects (Mean to peak value)
(source [5]).
Speed (km/h) D1 D2 D1 D2 D1 D2
(b) The placement of an elastic mat between deck and ballast amplifies the acceler-
ations in the ballast and therefore is harmful for this phenomenon. This impor-
tant conclusion has been obtained from the results of the tests commissioned by
the ERRI D214 committee [7].
(c) Accelerations of the order of g, even on decks without ballast, can decrease the
contact forces between rail and wheel (even causing take-offs) to inacceptable
limits.
(d) Regular and repetitive distribution of axes of the branches high speed can
result (at certain speeds) to resonant situations of the decks with important
amplifications both in vertical deformation and accelerations, as can be seen in
Figure 2.21.
Consequently, to ensure traffic safety, where a dynamic analysis is necessary
(usually for high speed), the verification of maximum peak deck acceleration due to
rail traffic actions shall be done. The goal is to avoid ballast instability on ballasted
bridges and unacceptable reduction in wheel rail contact forces on slab track bridges.
The maximum peak values of bridge deck acceleration calculated along each track
shall not exceed the following design values:
(a) 𝛾 bt for ballasted track;
(b) 𝛾 df for direct fastened tracks with track and structural elements designed for
high-speed traffic;
SPAN: L = 30 M
Ref. Freq.: f = 3 Hz
Vertical displacements Ref Mass.: 25.000 kg/m
22
20 Thalys
18
Displacements (mm)
16 Ice-2
14
12
Eurostar
10
8
ETR-Y
6
4
2 TALGO_AV
0
16 18 20 22 24 26 28 30 32 34 36 38 40
Speed/f (m)
Vertical accelerations
6
5 Thalys
Acceleration (m/s2)
4 Ice-2
3 Eurostar
2 ETR-Y
1
TALGO_AV
0
16 18 20 22 24 26 28 30 32 34 36 38 40
Speed/f (m)
for all members supporting the track considering frequencies (including considera-
tion of associated mode shapes) up to the greater of:
(a) 30 Hz;
(b) 1.5 times the frequency of the fundamental mode of vibration of the member
being considered;
(c) the frequency of the third mode of vibration of the member.
The dynamic analysis shall be undertaken using characteristic values of the load-
ing from the Real Trains specified. The selection of Real Trains shall consider each
permitted or envisaged train formation for every type of high-speed train permitted
or envisaged to use the structure at speeds over 200 km/h. In any case, the individual
project may specify the characteristic axle loads, spacings, and total number of axles
or train length, for each configuration of each required Real Train.
In some cases, a dynamic analysis is required for a Maximum Line Speed at the
Site less than 200 km/h (this and other aspects of dynamic analysis will be presented
in detail in Chapter 5).
Additionally, in the case of lines integrated into de trans-European rail system
network, these checks must be made for the 10 trains constituting the universal
High Speed Load Model (HSLM-A) defined by the ERRI D214 committee, collected
by the Eurocodes and the European Technical specifications for interoperability [8].
The HSLM-A covers the effects of existing and future high-speed trains that fulfil
the requirements established in Annex E of EN 1991-2.
On the other hand, dynamic acceleration calculations generally are made with
the assumption of track and wheels without irregularities. To quantify the effects
of these irregularities the maximum calculated acceleration must be multiplied
by (1 + 𝜑′′ ) for track with standard maintenance, or (1 + 0,5𝜑′′ ) for carefully
maintained track. The following expression is recommended for the computation
of 𝜑′′ :
[ ( )2 ( ) ( L )2 ]
v L
− Φ n0 LΦ − Φ
𝜌′′ = 0.56e 10 + 0.5 − 1 e 10
22 80
where
Considering that if v > 22, v = 22 will be taken, and 𝜑′′ = 0 if the result of the
computation is negative.
As abovementioned, more details will be presented in Chapter 5.
2.4 SLS Related to the Track 51
3m
Deck Twist Deck twist limits ensure that the four-wheel contact points of a bogie
(or of a two-axle vehicle) are not too far from a plane, assuring that rail-wheel
contact forces are controlled. In other case, the risk of derailment could be
increased.
The twist of the bridge deck shall be calculated considering the characteristic val-
ues of Load Model 71 as well as SW/0 or SW/2 as appropriate multiplied by the
dynamic coefficient Φ and the classification coefficient 𝛼 and Load Model HSLM
including centrifugal effects, all in accordance with EN1991-2, 6.
Twist shall be checked on the approach to the bridge, across the bridge, and for
the departure from the bridge.
The maximum twist t (mm/3m) of a track gauge s (m) of 1.435 m measured over
a length of 3 m (Figure 2.22) should not exceed the values given in Table 2.3.
The recommended values for the set of t are:
t1 = 4.5
t2 = 3.0
t3 = 1.5
Values for a track with a different gauge may be defined, but the following criterion
could be adopted:
t1 = 4.5 𝛽
t2 = 3.0 𝛽
t3 = 1.5 𝛽
where 𝛽 = 1.78r 2 /(r + c)2 and c = 0.5 m. Justification of this value is included in [9].
Vertical Deformation of the Deck Vertical deflection of the deck must be limited to
ensure acceptable vertical track radii and generally robust structures.
EN 1990 Annex A2 [10] establishes as general limitation, for all structure con-
figurations loaded with the classified characteristic vertical loading in accordance
V ≤ 120 t ≤ t1
120 < V ≤ 200 t ≤ t2
V > 200 t ≤ t3
52 2 Track for High-Speed Bridges
with EN 1991-2, 6.3.2 (and where required classified SW/0 and SW/2 in accordance
with EN 1991-2, 6.3.3) the maximum total vertical deflection measured along any
track due to rail traffic actions should not exceed L/600.
Additional requirements for limiting vertical deformation (that can be different
for ballasted and ballastless bridges) may be specified. In the case of ballastless
bridges, it must be considered that, in the absence of ballast, the possibilities of
long-term correction of track geometry are limited to the adjustment capacity of the
fastening system. Thus, long-term deformation of the deck can affect track geometry
quality and should be added to deformations due to traffic load in the verification
abovementioned. The final goal is to assure that track geometry quality can be
maintained in any circumstances including long-term deformations of the deck.
On the other hand, the vertical deflection of the end of the deck beyond bearings
must be limited:
The criteria for the limit values will be determined by the tensile forces in the
fasteners. In this sense, the German standard for slab tracks indicates that the max-
imum traction in the fasteners on both sides of the joint must not exceed the value
that produces the decompression of the rail pad, a value that depends on the charac-
teristics of the fastening system and can be determined by standardised tightening
force test [11–13].
UIC leaflet 776-3R [14] indicates a design criterion regarding the limitation of
stresses in the rail, indicating that when there are direct fasteners on both sides of
the joint, the variation of angle and lifting should be limited to not exceed a stress
of the order of 80 N/mm2 in the rail. This stress should be calculated considering the
real position and stiffness of the fasteners. For rail type UIC 60 E1, this maximum
stress is equivalent to a maximum bending moment of 26.8 kNm.
2.4 SLS Related to the Track 53
From the above, to determine the maximum rotations and vertical relative dis-
placements the following criteria should be considered:
– comply with the explicit requirements indicated by standards;
– limit the bending moment in rails to assure that a stress of 80 N/mm2 is not
exceeded, as indicated by UIC leaflet 776-3R [14];
– limit the tensile force on the fasteners so that the clamping force is not exceeded,
determined in a standardised test, as indicated by the German standard.
Finally, rotation of the ends of each deck should be verified in the presence of rail
expansion devices, switches, crossings, etc. to limit angular discontinuity.
Transverse Deformation and Vibration of the Deck Transverse deformation and vibra-
tion of the deck must be limited to ensure acceptable horizontal track radii and to
avoid dynamic interaction with the vehicle.
According to EN 1990 Annex A2, the transverse deflection at the top of the deck
should be limited to ensure:
– a horizontal angle of rotation of the end of a deck about a vertical axis not greater
than the values given in Table 2.4, or;
– the change of radius of the track across a deck is not greater than the values in
Table 2.4, or;
– at the end of a deck the differential transverse deflection between the deck and
adjacent track formation or between adjacent decks does not exceed the specified
value.
Again, as in the case of vertical deformation, in the case of ballastless track the
effect of long-term deformation should be considered, if the lateral adjustment
capacity of the fastening system is not enough to correct these lateral deformations.
On the other hand, lateral deformations can influence the dynamics of vehicles.
As in the case of vertical deformations, in some viaducts an analysis of the dynamic
effects of lateral deformations may be necessary.
Standards, as EN 1990 Annex A2, establish a minimum value for the first natural
frequency of lateral vibration of a span of 1.2 Hz. This limit was proposed in by the
ERRI D181 Committee [15], dedicated to the analysis of lateral forces laterals on
V ≤ 120 𝛼1 r1 r4
20 < V ≤ 200 𝛼2 r2 r5
V > 200 𝛼3 r3 r6
54 2 Track for High-Speed Bridges
railway bridges, and was justified in order to avoid lateral resonance phenomena in
railway vehicles circulating on the bridge, considering that, in general, the lateral
vibration frequencies of railway vehicles are below 1.0 Hz.
However, it should be emphasised that this limitation refers to ‘the first natural
frequency of lateral vibration of one span’, which can be interpreted as the lateral
deformation of the deck assuming that the piers represent fixed points in lateral
direction. However, this assumption cannot be realistic in some cases, as, for
instance, the case of long viaducts with high piers. In this case, the lateral deforma-
tions occurring during a train pass-by can be significant and the natural frequencies
of the first mode of vibration of the deck can be very low. It is not clear whether the
required verifications must be applied to spans considered independently, to several
successive spans, or to the whole viaduct.
Standards do not propose any analysis methodology to assess this situation and
validate the design of this kind of viaducts. In some special cases, the consideration
of lateral dynamics for the response of railway vehicles on bridges will require the use
of three-dimensional models, including degrees of freedom for lateral displacement,
rolling, and yawing in the vehicles [16, 17].
– 5 mm for continuous welded rails without rail expansion devices or with a rail
expansion device at one end of the deck;
– 30 mm for rail expansion devices at both ends of the deck where the ballast is
continuous at the ends of the deck;
2.4 SLS Related to the Track 55
– the relative longitudinal displacement between the end of a deck and the adjacent
abutment or;
the relative longitudinal displacement between two consecutive decks.
For vertical traffic actions, up to two tracks loaded with load model LM 71 (and
where required SW/0) 𝛿 H (mm) shall not exceed the following values:
– the longitudinal displacement of the upper surface of the deck at the end of a deck
due to deformation of the deck.
Vertical acceleration
Level of comfort inside the coach (m/s2 )
The levels of comfort and associated limiting values may be defined for any indi-
vidual project, but recommended levels of comfort are given by EN 1990 Annex A2
(see Table 2.5).
As abovementioned, to limit vertical vehicle acceleration to the values given in
Table 2.5, limit values for the maximum permissible vertical deflection along the
centre line of the track of railway bridges were established as a function of the span
length, the train speed, the number of spans, and the configuration of the bridge
(simply supported beam, continuous beam). These values can be determined
through Figure 2.23.
The values in Figure 2.23 are given for a succession of simply supported beams
with three or more spans. For a bridge comprising either a single span or a succes-
sion of two simply supported beams or two continuous spans the values should be
multiplied by 0.7. For continuous beams with three or more spans the values should
be multiplied by 0.9.
It must be considered that the limiting values are given for an acceleration
inside the coach of 1.0 m/s2 which may be taken as providing a ‘very good’ level of
3.000
V=
350
2.500
V=
V= 300
280
2.000 V=
250
V=
V= 220
L/δ
1.500 200
V=
160
1.000
V=
120
500
0
0 10 20 30 40 50 60 70 80 90 100 110 120
L (m)
Figure 2.23 Maximum permissible vertical deflection for railway bridges with 3 or more
successive simply supported spans corresponding to a permissible vertical acceleration of
1 m/s2 in a coach for speed V (km/h) (source [10]).
References 57
comfort. For other levels of comfort and associated maximum permissible vertical
accelerations the values may be divided by the maximum permissible vertical
acceleration in m/s2 .
To be coherent with the computation done by ERRI D190 committee, the vertical
deflections should be determined with Load Model 71 multiplied by the factor Φ and
with the value of 𝛼 = 1, in accordance with EN1991-2, and, for bridges with two or
more tracks, only one track should be loaded.
It must be pointed out that the values of L/𝛿 given in Figure 2.23 are valid for span
lengths up to 120 m and for a speed up to 350 km/h. For longer spans or higher speeds
a special analysis should be necessary. Similarly, for exceptional structures, e.g. con-
tinuous beams with widely varying span lengths or spans with wide variations in
stiffness, a specific dynamic calculation should be also necessary.
The limits established for vertical deflection can be considered as too conservative
in some structures. In this case, as for the structures not covered by these require-
ments, complex analyses following the principles adopted by the ERRI D190
committee can be carried out subject to the agreement of the corresponding
infrastructure manager. To be able to justify a given design, these analyses must
consider all the types of actual trains proposed for operating the lines, as well as
others envisaged to be operated in the future.
It must be pointed out that the limits proposed by the ERRI D190 committee are
related only to comfort. The compliance with these limits favours compliance with
the limits of acceleration of the bridge deck and other dynamic aspects related to
security. However, compliance with these limits should not be considered as a guar-
antee that the safety criteria linked to resonance are met. The review of work of
ERRI D190 committee indicates that, for certain bridges (for example continuous
beams of length less than 40 m with a single loaded track) the deck acceleration
limit of 3.5 m/s2 could be exceeded if the travellers’ comfort criteria were consid-
ered to govern the design. Likewise, it is likely that the bridges meeting the safety
criteria defined in this report have acceptable characteristics for traveller’s comfort.
However, compliance with the criteria related to vertical acceleration of the deck
cannot be considered a guarantee that the travellers’ comfort criteria will be satisfied.
References
3.1 Introduction
The first question that can be asked when designing a bridge for a high-speed railway
line is: What’s so special about a high-speed rail bridge?
Railways bridges have to meet strong structural requirements. The live loads in
those bridges are substantially higher than those of road bridges.
In the case of bridges for high-speed lines, the mass and speed of the trains interact
with the structure resulting in significant dynamic excitation of the structure, with
the consequent danger of resonance. Therefore, high stiffness is required to control
vertical acceleration to ensure the comfort of the passengers as well as the safety of
the trains. For this reason, the vertical and transverse deformations of the deck must
be controlled.
Other special conditions such as high braking forces or the interaction between
track and bridge deck are specific in HSR bridges as the use of continuous welded
rail CWR is indispensable to maintain the regularity of rolling [1].
Good design should meet structural and functional requirements with a high
visual quality. To achieve that goal, the design of HSR bridges has to assure a sub-
stantial vertical structural stiffness without using massive elements. Furthermore,
high-speed railways lines are the best example of the high technology of the 21st
century; HSR bridge’s design should meet public expectation of those especial
elements.
Throughout this chapter, first the specific conditions of this type of bridges are
presented as well as the implications they have on the design of the structure. After
that, the standard design situations are then presented and the different structural
types associated with these situations are explained. Finally, a description is made
and a design methodology of the specific structural elements of this type of bridges
is presented.
3.2.1 Introduction
– The specific aspects that influence the design of bridges for HSR are as follows.
Vertical stiffness to control the vertical acceleration produced by the coupling of
the vibration of the mass of the structure and the trains.
– Vertical stiffness to control the relative rotation at expansion joints.
– Longitudinal fixation of the deck to transmit the huge braking forces to the
foundation.
Special provisions to limit the stresses in the rails due to the track to bridge deck
interaction.
Although all of these aspects are dealt with extensively in other chapters, they are
briefly listed here for a better understanding of the conceptual design of this type of
structures.
In Sections 3.2.2–3.2.6 all those aspects are analysed as well as its implications in
the conceptual design of this type of bridges.
14.00
12.00
Force in fixed point (mn)
10.00
8.00
6.00
4.00
2.00
0.00
0 100 200 300 400 500
Viaduct length (m)
Another condition to assure the safety of the trains is the maximum relative defor-
mation between deck and infrastructure. The value of this relative movement is as
low as 5 mm when braking forces arc, which in practice entails that the deck has to
be longitudinally fixed to one rigid point of the infrastructure, normally one of the
abutments. That fixed point has to pass the large braking and traction forces from
the deck to the infrastructure. It is therefore necessary to design at least one element
of the substructure, either an abutment or an intermediate support, capable of trans-
mitting these large horizontal loads in the longitudinal direction (see Section 2.4).
Figure 3.3 Long viaduct with a longitudinal fixed point in a central pier.
Figure 3.5 Ultra-long and continuous viaducts with two intermediate fixed points.
joints are placed consecutively, in order to reduce their movement and thus make it
feasible to have two consecutive structures of more than 1 km each. This solution,
which is perfectly valid from a structural point of view, has the disadvantage for
rail traffic of having two consecutive rail joints, which are always delicate elements
from perspective of maintenance and train traffic safety.
The practical consequence of those main possibilities described depends on
whether the track is ballasted or not. The following sections show the possible
scenarios for each of these track types.
For bridges in seismic zones, the design for braking loads, and therefore the need
for a fixed point, has to be reconciled with the need to dissipate energy. This means
that longitudinal dampers have to be provided on these bridges to reduce the forces
caused by an earthquake [1, 3]. Those aspects are described in Section 3.7
100 m < L
Track
100 m < L < 1000 m expansion joint
Structural Structural
expansion joint expansion joint
Figure 3.6 Schemes configuration of continuous concrete decks with different length.
66 3 Conceptual Design of High-Speed Railway Bridges
All those requirements have to be considered during the conceptual phase. They
are very strong from a structural point of view and have a significant impact in the
visual aspect of the bridge. It is extremely easy to end up with a heavy structure if the
design does not take care of the visual aspect of the bridge. The balanced equilibrium
of all those aspects requires a careful design. In the following sections a discussion
of the different design situation is presented.
In all the cases, a coherent approach with the structural performance as a clear
driver of the solutions has to be used.
14.40
7.60
Long viaducts are required for crossing of long valleys or just the design of the
railways just a few meters over the ground surface to avoid placing the line over a
filling.
~L/17 to ~L/14
~L/17 to ~L14
~L/12
~L/10
In all those cases, the repetition of the spans could result in a viaduct with a kind
of millipede’s appearance (Figure 3.12). Furthermore, it is very usual to increase
the vertical clearance under the deck by using short spans which increase the
monotonous view of the structure, which also creates a visual barrier when the
bridge is observed in a skew angle.
A possible way of improving the aesthetic of those ‘millipede’ bridges is increasing
the span by using a continuous frame structure that also decreases the number of
foundations (see Figure 3.13).
Other alternative when the designer wants a more openness under the bridge
could be the use of an ‘inverted frame’ system. In that case, the deck should be in
steel and the result is a more open solution that could be used for instance when the
line has to cross over a highway (see Figure 3.14).
From the structural point of view both solutions can fulfil the structural conditions
related to HSL bridges without problems.
In the first case, the frames have to be supported by pot bearings (Figure 3.15).
Those elements transmit the vertical forces to the foundation and allow the hori-
zontal movement to the fix point usually placed in one abutment.
In the case of the inverted frames, the deck and the upper part of the columns
transmit the vertical forces to the part of the supports under the deck by pot
Bearing
Bearing
Figure 3.16 Bearing and pier layout for continuous inverted frames.
bearings. In the same way, the deck could be fixed monolithically to one abutment.
In Figure 3.16 a possible bearing layout in a standard pier is shown.
If those structures when the length of the deck is longer than 1000 m, a central
fixation is necessary as it was discussed previously. That central point attracts the
braking and traction forces, and therefore it is very difficult to resist that force by
the bending of the columns. There are several ways of designing those elements as
it can be seen in Figure 3.17. In all those cases special attention to the layout and the
connection of the different elements should be carried out to maintain the visual
quality of the structure but endowed with an appropriate control of the forces.
In some cases these long, low viaducts need a special, wider span to overcome a
special obstacle such as a motorway or river.
In such cases, it is both structurally and visually desirable for the deck to be con-
tinuous, although it will logically require an additional superstructure to bridge the
special span(s). The figure shows the studies carried out for the resolution of such a
2 pots
Pl – 14000
2 pots
Pl – 14000
2 neoprenes 2 neoprenes
400 × 400 400 × 400
2 neoprenes
400 × 400
30° 30°
3.0 0
0 3.0
4 cables 4 cables
2 × 4 cables
19T15 19T15
19T15
R1.
2.20
1.95
95
1.8 0
0 1.8
type of crossing. In that case, a solution with a variable upward depth, a truss, and
an inverted frame was analysed (Figure 3.18).
anchoring becomes more critical because of its visibility (Figure 3.20). Again, there
are some possibilities and when it is possible the use of one singular point of the
topography as the cross of a river or a higher intermediate point in the landscape
could be used to locate the central point [7].
The German railways have developed a solution for bridges with moderate pier
heights: 15–20 m, with a series of fixed points (approximately every 110 m) with
intermediate structural expansion joints between the sections, but without rail
expansion joints in the track [8]. An example of this solution is the Gänsebachtal
viaduct [9], shown in Figure 3.21. The bridge has a succession of flexible piers and
more rigid diapason piers that make a local fixed point.
Figure 3.22 Simply supported launching beam system for 32 m prefabricated concrete
decks, China (Courtesy of China Railways).
the point of view of railway operation. The biggest disadvantage is its low structural
efficiency and the higher maintenance requirements for the bridge as there is a struc-
tural joint in every span.
It should also be noted that in a seismic zone this solution has the risk of loss
of support for the deck. Likewise, the lack of transverse continuity between spans
considerably increases the risk of rail traffic during the action of the earthquake.
This solution is the most widely used in some of the Southeast Asian countries.
Specifically, railway lines in China are being built with this type of solutions. In this
country, a complete prefabrication solution has been developed for a section that
also allows assembly from the previously built deck (Figure 3.22).
Also in Spain, isostatic solutions consisting of two prefabricated U girders for
spans of 30–35 installed with a launching beam are common.
14.00
0.20
Figure 3.23 Cross section slab solution.
A-1 P-1 P-2 P-3 P-4 P-5 P-6 P-7 P-8 P-9 P-10 P-11 P-12 A-2
402.00
25.00 32.00 32.00 32.00 32.00 32.00 32.00 32.00 32.00 32.00 32.00 32.00 25.00
Fixed point
Expansion
rail joint
Table 3.1 Summary of dimensions, quantities, and loads of slab concrete decks 14 m wide.
Span range 25 to 35 m
Total depth related to span length L L/14 to L/17
3 2
Concrete deck average equivalent thickness m /m 0.75 to 0.95
Average prestress weight in deck per surface deck unit kg/m2 18 to 30
2
Average steel rebars weight in deck per surface deck unit kg/m 80 to 105
Standard self-weight (kN/m) 260
Standard mean dead load (ballasted track) (kN/m) 170
Standard Permanent Load with ballasted track (kN/m) 430
If the height above the ground does not exceed 15 m, it is usual to build with
ground-supported formwork span by span (Figure 3.25).
For higher heights or if the terrain has a low bearing capacity, the construction
can be carried out by means of movable a scaffolding system supported by the bridge
piers.
Preparation of the
next falsework Falsework
phase
Figure 3.25 Example of slab-type deck with span-by-span construction with conventional
formwork.
In the case of redundant structures the common practice is to give continuity con-
necting adjacent spans by posttension bars. Normally, the cross section is formed by
two U-shaped beams located under the tracks and an in situ concrete slab built over
them (Figures 3.26 and 3.27).
Table 3.2 summarises the main dimensions, quantities, and weights of this type of
decks.
14.00
1.75 10.50 1.75
2.35 2.35
0.30
0.44
2.50
2.73 2.73
Figure 3.26 Standard solution with two U-shaped beams with structural continuity over
the piers.
Table 3.2 Summary of dimensions, quantities, and loads of precast concrete decks 14 m
wide.
14.00
1.75 10.50 1.75
2.35 2.35
0.30
2% 2%
3.00
By increasing the number of beams, it is possible to reduce the depth of the deck.
Figure 3.28 shows a section formed by three U-shaped beams.
The most common system for the assembly of this type of solutions is by means
of cranes (Figure 3.29). One of the advantages of this prefabricated solution, when
assembled with the help of cranes, is that it allows the independent construction of
each span in the order that is most convenient for the work.
An alternative to cranes is the use of a beam launcher. This solution is the most
reasonable possibility when the deck is at a very high level from the ground, or when
the prefabricated beams cannot be transported to the vertical of the area from which
they are to be lifted (Figure 3.30).
The continuity of the deck is carried out in the second phase, often by means of
post-tensioning, to make the connection section work as a pre-stressed joint with
limited crack opening. Figure 3.31 shows a typical pre-stressed connection in pre-
fabricated U-beams.
3.5 Structural Types 77
4Ø40
0.06
Prestressed
Joint filled rebars
0.14 with grout
6Ø50
6Ø50 Prestressed
Prestressed rebars
rebars
6Ø50 6Ø50
Prestressed Prestressed
rebars rebars
14.00
7.00 7.00
2% 2%
0.20
0.40
4.40
4.26
0.55 0.55
0.30
77
°
3.35 0.90 5.50 0.90 3.35
0.40
0.40
0.40
0.40
1.50
1.25
1.25
3.40
3.40
4.40
4.40
3.70
4.40
3.70
3.70
2.15
2.15
0.80
0.60
0.60
0.80
0.30
0.30
1.50 1.00
0.30
Table 3.3 Summary of dimensions, quantities, and loads of concrete box decks 14 m wide.
Span range 40 to 70 m
Total depth related to span length L L/12.5 to L/16.5
Deck average equivalent thickness m3 /m2 0.75 to 0.90
Average prestress weight in deck per surface deck unit kg/m2 22 to 40
Average steel rebars weight in deck per surface deck unit kg/m2 140 to 150
Standard self-weight (kN/m) 280
Standard mean dead load (ballasted track) (kN/m) 170
Standard Permanent Load with ballasted track (kN/m) 450
3.5 Structural Types 79
14.00
1.75 10.50 1.75
2.35 2.35
0.22
0.41
3.00
6.00
Table 3.4 Summary of dimensions, quantities, and loads of steel double I girder 14 m wide
decks.
Span range 40 to 70 m
Total steel beam depth related to span length L L/14.5 to L/17
Total depth related to span length L L/13.5 to L/15
2
Deck average steel weight kg/m 270 to 300
Deck average equivalent thickness m3 /m2 0.40
3
Average steel rebars weight in deck per concrete volume kg/m 140 to 150
Standard self-weight (kN/m) 180
Standard mean dead load (ballasted track) (kN/m) 170
Standard Permanent Load with ballasted track (kN/m) 350
In the specific case of piles over 80 m high, the pushing of a deck with double ‘I’
steel beam is much safer than that of heavy concrete decks, so composite solutions
can be very competitive for those cases (Table 3.4).
Composite bridges with a double ‘I’ steel beams can be built as well span by span
using cranes if access is possible and the deck height is moderate (up to 35–40 m)
and by launching in other cases.
14.00
C
L
Figure 3.39 Example of semi-through deck with lateral hollow steel beams.
82 3 Conceptual Design of High-Speed Railway Bridges
Table 3.5 Summary of dimensions, quantities, and loads of steel semi-through 14 m wide
decks.
Span range 40 to 70 m
Total depth related to span length L L/10 to L/12
Deck average steel weight kg/m2 400 to 500
Concrete deck average equivalent thickness m3 /m2 0.40
Average steel rebars weight in deck per concrete volume kg/m3 250
Standard self-weight (kN/m) 215
Standard mean dead load (ballasted track) (kN/m) 170
Standard Permanent Load with ballasted track (kN/m) 385
system formed by the two main steel beams located at the edges of the platform.
The usual characteristics of this type of decks are summarised in Table 3.5.
One of the usual advantages of this solution is that it usually allows the push from
one of the abutments, avoiding the need for intermediate supports for the construc-
tion of the deck and minimising the effect on the obstacle to be spanned.
Table 3.6 Summary of dimensions, quantities, and loads of steel truss decks.
a) Deck depth depends on the relative span between vertical supports on the arch.
B2
B2
B2
B1 =
B2
B1 >
B2
B1 ≫
B2
Figure 3.42 Geometry of the arches in the spring sections as the span increases.
SECTION A–A
A
that makes them appropriate to HSR Lines. Furthermore, the stays have a small
variation of the stress due to the live load which entails a good behaviour against
fatigue (Figure 3.45).
In Table 3.9 a summary of the main dimensions of this type of bridges is presented.
All the references included come from the Japanese High-Speed Lines.
Extradosed Bridges have a major application in the case of multi-span structures
where they exhibit a substantially higher stiffness compared with the traditional
Cable-Stayed Bridges. More extradosed bridges for HSRL are likely to be built in
the coming years.
120
120
the deck as a truss steel system. With that kind of solution, the vertical deformation
could be controlled.
For this reason, the decks have to have a significant stiffness. One way of provid-
ing sufficient rigidity without excessively increasing the weight of the structure is to
make a steel truss in which the tracks are located inside the structure.
This solution also allows for the incorporation of traffic on the upper level, thus
generating a two-level structure (Figure 3.46).
As mentioned before, the key problem for HSL Bridges is the relative rotation
at the expansion joints. A way of tackling that problem is by introducing an extra,
shorter lateral span. With that the rotation at the expansion joint will be drastically
reduced (see Figure 3.47).
In Figure 3.48, the proposal for the crossing of the HSL over the Tagus River in
Lisbon is a good example of the importance of the proper shaping of the elements
of such structures. Here, the use of continuous approach viaducts minimises the
rotation problems at the expansion joints. The double-level deck gives the required
stiffness of the system. The minimisation of the masses decreases the seismic loads
and contributions to the general slenderness of the proposal in spite of their colossal
dimensions.
r1
r2
r2 ≪ r1
Figure 3.47 Control of the relative rotation at the expansion joint by introducing an
additional lateral span.
88 3 Conceptual Design of High-Speed Railway Bridges
Figure 3.48 Tagus River Crossing proposal, 550 m main span, DJV
ADF – FHECOR – IDEAM.
17.50
Figure 3.49 HSL Wufengshan Yangtze River Bridge, deck’s cross section.
in some intermediate area of the deck is a specific feature of long high-speed viaducts
and is therefore analysed with special attention in this subchapter. Although not so
specific to HSR bridges, the usual types and configurations of the piers and bear-
ing devices common to this type of bridges are described. Finally, a description and
analysis of the special anti-seismic devices is included.
3.6.1 Abutments
There are basically three types of abutments on HSR bridges.
– With an expansion joint in the deck but without a track expansion joint.
– With an expansion joint in the deck and with a track expansion joint.
– Fixed: without a structure expansion joint.
In all cases, the abutments are designed to be closed to ensure the stability
of the approach wedge. Likewise, except in the case of clear rock, the abutments
for HSR bridges are usually built on piles to avoid significant vertical movements
in the deck – abutment transition. The most significant aspects of each of the
above-mentioned abutment types for HSR bridges are reviewed below.
3.00 17.00
Ballast
Continuous rail Sub-ballast
0.35
0.40
0.35
Shape layer
0.30
0.60
Structural
expansion 1 1
joint 1 1
Waterproofing
sheet
Ballast
Sub-ballast
0.30
Expansion rail joint 3.00 17.00 Shape layer
0.60
Waterproofing
sheet
1
1
Cement Gravel Embankment
treated
gravel
Drain
tube
bearing bearing
1.11
9.36
9.32
7.39
0.50
0.50
0.10
0.10
2.50
2.50
2.50
2.50
8.10 12.75 2.80 2.80 12.75 8.10
0.10
0.10
0.10
0.10
0.10
L pile
VAR. 1.10
VAR
Pot type
bearing
VAR
7.37
0.50
2.50
Ø1.80 Ø1.80
0.10
L pile
Longitudinal section
Figure 3.55 Lateral elevation and centreline cross section of an abutment with track
expansion joint.
– Fixing the deck to the abutment by using POT-type supports, one fixed and the
other transversally guided. This solution is feasible up to horizontal design loads
of 3000–3500 kN. In these cases, the bearings must be taken from the catalogue as
a function of the vertical design load, and the horizontal forces to be resisted must
be specified in the drawings (Figure 3.58).
92 3 Conceptual Design of High-Speed Railway Bridges
14.00
0.30
2% 2%
3.40
10.00
4.80
0.50
8.72
4.25
12.02
5.25
1.15 1.15
13.80
1.50
26.50
3.00
4.50
(MIN.)
1.80
16.50
4.50
1.80
1.50 4.50 4.50 4.50 1.50
16.50
Elevation
0.60
4.50
1.50
1.50 4.50 4.50 4.50 1.50
16.50
Plan
Figure 3.56 Front elevation and plan view of an abutment with track expansion joint.
Ballast Ballast
Sub-ballast
Continuous
Shape layer 21.00 3.00
0 .3 0
rail
0. 60
Waterproofing
1 Sheet
2 1
2
Drain tube
Embankment Gravel Cement
treated
gravel
– Fixing the deck to the abutment by using brackets that are lowered under the deck
and that bump against the abutment (Figure 3.59).
– Fixing the deck to the abutment by means of bars or prestressed tendons. This
type is usually used in viaducts where the final fixed point is located at interme-
diate points of the viaduct and it is necessary to fix the deck to the abutments
temporarily during construction (Figure 3.60). (In Spain, the Railway Authority
ADIF prefers the use of prestressed tendons for permanent fixing).
Other aspects to take into account in the design of those elements are as follows:
If the fixed point is realised with POT bearings, there must be sufficient horizontal
3.6 Structural Elements – Substructure 93
Pots
Pot
transversally Fixed pot
guided
Pot
Fixed pot
transversally
guided
Pots
Neoprene
Neoprene
Neoprene
Prestressed rebars
Prestressed rebars
Neoprene
Neoprene
Pots
Pot
longitudinally Pot free
guided
Neoprene
reinforcement under these bearings capable of introducing the horizontal force into
the body of the abutments.
If the fixed point is materialised with brackets, the space designed to house the
longitudinal stopper must allow the replacement of the neoprene that is placed in
the bracket to ensure proper contact between the bracket and the body of the abut-
ment. One possible method is to place the neoprene on a plate that is screwed to the
abutment, so that the screws are accessible from the sides of the bracket.
When designing the deck span next to the fixed abutment, the moment at the
end due to the eccentricity of the horizontal load must be taken into account. This
moment is attenuated until it becomes practically nil in the second pier.
Since the horizontal loads to be transmitted to the ground are very important,
it may be interesting to use a friction slab (see Figure 3.61). In the case of abut-
ments with direct foundations, it may be necessary to use a friction slab to avoid
over-dimensioning the footing. If the foundation is deep (most common), a friction
slab is usually used so as not to penalise the piles which, if not available, should
be able to transmit all the horizontal force to the ground. For the verification of
stability against sliding, it is usually considered that the friction slab rubs only on
its lower face. The combined operation of slab and piles is complex. A common
way to define the length of the friction slab required is to design it to resist all the
horizontal force in SLS, considering only the combined effect of the slab and the
piles for fulfilling the ULS.
For maintenance details of abutments see Chapter 7.
3.6 Structural Elements – Substructure 95
VAR.
5.40 5.40
1.30 1.80 1.80 1.30 1.80 1.80
VAR. (3.01–3.15)
3.40
VAR.
2.60
12.16
5.75
3.15
6.00 5.40 5.10 6.00 5.40 5.10
2.00
2.00
1.00
3.00
3.00
0.10 100
(MIN.)
(MIN.)
(MIN.)
0.10
0.10
0.10
(MIN.)
1.50 4.50 4.50 4.50 1.50 1.50 4.50 4.50 4.50 1.50
10.00 16.50 10.00 16.50
13.80
0.50
3.00
5.70
6.00
12.02
5.48
3.01
11.40 5.10
2.00
1.00
3.00
(MIN.)
(MIN.)
0.10
0.10
Elevation
3.6.2 Piers
Due to the magnitude of the vertical loads, railway bridge piers are often very robust.
In general, the piers have to transmit a significant longitudinal force, which corre-
sponds to the longitudinal deck force produced by the friction of the support devices
located at their head. Transversally, the governing load is usually the transverse wind
or the earthquake in the case of bridges located in areas with seismic activity.
Except when no fractured rock appears on the surface, in which case the founda-
tion of the piers is normally by footings, the foundation of piers of HSR bridges is
pile-caps with large-diameter piles.
In the case of a concrete slab or prefabricated deck, which corresponds to the
low span range, the piers are usually solid, as they rarely reach 20–25 m in height
(Figure 3.62).
For larger spans, associated with higher piers, those elements are usually hollow,
with constant cross sections up to 25–30 m in height and with variable dimen-
sions for higher values. See details for inspection and maintenance in Chapter 7
(Figure 3.63).
3.6.3 Bearings
Except for very special cases, the support devices for HSR’s bridge decks are of the
‘POT’ type made of neoprene confined in a metal cup and with a Teflon finish so that
the friction transmitted to the infrastructure is as small as possible (Figure 3.64).
96 3 Conceptual Design of High-Speed Railway Bridges
A A
5.20 1.90
Pot type Pot type
bearing bearing
Hf
Hf
3.40 5.20 3.40 3.25 1.90 3.25
2.70
2.70
0.10
0.10
Ø1.80 Ø1.80 Ø1.80 Ø1.80 Ø1.80
5.20
1.56 1.04 1.04 1.56
Space for
0.10 jacking
0.10
12.00
1.40
1.50 4.50 4.50 1.50
0.10
1.50
0.10
Pot type
.80 .8 0 .8 0 bearing
Ø1 Ø1 Ø1 1.90 1.90
2.70
1.90 1.90
8.40
1.90
1.40
5.20
2.70
2.60 2.60
Pot type bearing
5.20
.80 .80 .80
Ø1 Ø1 Ø1
1.50
1.90
1.40
0.25
0.20
Figure 3.62 Example of solid pier for a concrete slab deck type.
3.51 8.53
4.75
4.75
1.20
1.20
1 1 1 1
50 50
17,5 17,5
H=32.484
H=32.484
4.81 12.24
4.00
4.00
0.10
0.10
16.00
8.00 8.00
Bearing
0.60
12.24
0.40
7.00
1.20
0.40
14.00
VAR.
0.60
4.81
3.51
0.60
0.40
7.00
0.40
8.53
0.60
Inspection
VAR. hole
Cross section
Plant
In both piers, the deck must be transversely constrained, so the support arrangement
is as shown in Figure 3.66.
In the case of redundant decks, the support in piers is only transversal, so one of
the bearings must be guided longitudinally and the other free. In the case of short
bridges without earthquakes, where the fixed point on the abutment can be mate-
rialised by ‘pot’ supports, one of them will be fixed and the other will be guided
transversally (see Figure 3.67). If the longitudinal loads to be transmitted to the fixed
point on the abutment are large, the support configuration will be: one free and one
longitudinally guided support.
98 3 Conceptual Design of High-Speed Railway Bridges
Leveling
B Leveling
B Slide plate
mortar
mortar Slide plate
Slide plate Slide plate
Guide
Leveling
Leveling Mortar
A mortar A A
SECTION B-B SECTION B-B
GUIDE
Leveling Leveling
Slide plate Slide plate
Mortar mortar
B
B Leveling Leveling
SECTION A-A Mortar
Plant Plant SECTION A-A mortar
Generally mobile multidirectional pot bearing Unilateraly mobile unidirectional pot bearing
Leveling
Mortar
A A
H
Leveling
Mortar
Leveling Leveling
Slide plate
mortar mortar
Leveling Leveling
mortar mortar
SECTION A-A SECTION A-A
Slide plate
A A A A
Plant Plant
Expansion joint
resist the seismic load corresponding to a single span. Therefore, it is the longitu-
dinal behaviour that is fundamentally different in each of the situations as will be
explained below.
Adjustable
expansion joint
Deck Abutment
2 Pot bearings
Cement
2 Viscous treated
fluid dampers gravel
Abutment
only in the case of ultimate or collapse earthquakes does the damper system come
into operation by reducing the seismic load on the fixed point.
Earthquake
Openin g
Pounding g Openin Pounding
100% of the vertical load can be reached. This means that in the case of bridges
with significant transversal loads (i.e. seismic loads), transversal stoppers are
necessary to transmit these loads, as the bearing devices do not have sufficient
capacity.
Figure 3.70 shows an example of a central transversal stopper in the case of a
viaduct with a deck formed by two u-beams
In any case, in structures in a seismic area it is essential to have these transversal
stoppers in order to avoid the exit of the deck support in its support in piers.
Neoprene
Transversal
Stoppers
Steel connection
Transversally
Free bearing
Transversal damper
Figure 3.72 Special transverse reinforced bearing device with guaranteed maximum
horizontal force seismic devices.
devices. In these cases it is very important to design these elements or side stoppers
that work as a fuse so that once the threshold of the design force defined by the
designer is exceeded, the system is released from these elements and the transverse
dampers start working (Figure 3.72).
There are a variety of devices for damping bridges. In general, there are two main
groups of devices: those in which the damping is done in the same bearings and
those that are independent elements of the vertical support of the bridge.
The first include devices such as High Dumper Rubber Bearings, Lead Rubber
Bearings, Hysteretic damping with steel damping devices, and Pendulum devices.
The second type is the viscous hydraulic dampers devices.
Section 3.7.4.4 explain the practical application of these types of systems to HSRB
Damping through the bearings.
Neither High Dumper Rubber Bearings nor Lead Rubber Bearings are usually used
in HSRBs as these bearings have a shorter service life than POT-type bearings. Steel
damping and pendulum devices could be used but in combination with transversal
stoppers (Figures 3.73 and 3.74).
As explained above, transversal stoppers can be designed in two ways. Either
they are completely rigid, in which case there is no transversal damping, or they
function as a fuse that allows, in this case, the damper-supporting device to move
laterally therefore dissipating seismic energy and modifying the lateral flexibility of
the system.
104 3 Conceptual Design of High-Speed Railway Bridges
FD
FY
SY SD
2500
2000
1500
e=D
1000
Keff
500
V V 0
–800
–600
–400
–200
200
400
600
800
0
–500
–1000
–1500
–2000
–2500
From a longitudinal point of view, the damper devices are usually placed in one
of the abutments. In this case, it is necessary to have an elastic system to achieve the
re-centring of the deck in the post-seismic situation. If it is not used, the deck could
be displaced from its position before the earthquake.
One of the usual ways of solving the problem in long viaducts is to link the deck to
one or two central piers by means of fixed supports so that the piers act as an elastic
re-centring element. If the central piers are significantly higher than the rest, it is
usual to connect two piers to the deck. On the other hand, if the supports are rigid,
only one of them can be connected. In this case, it is necessary to check that the
flexibility of the pier is such that it does not absorb an excessive longitudinal force
that makes the solution unviable (Figure 3.76).
If this solution were not feasible, it would be necessary to have some neoprene
in addition to the general system of vertical bearings that would allow the elastic
re-centring of the system (Figure 3.77).
Bearings
Hydraulic
dampers
Figure 3.76 Example of re-centring with flexible – fixed piles in the centre of the bridge.
Figure 3.77 Example with re-centring by additional neoprene (plan and lateral view).
106 3 Conceptual Design of High-Speed Railway Bridges
3.8.1.1 Topography
The attached Figure 3.78 defines the plan and longitudinal profile of the route in the
area where a viaduct is planned to be built.
Elevation
P.C. +640.00 528 + 700 528 + 800 528 + 900 529 + 000
E-1
P.K. 528 + 631
14.00
6.25 6.25
5.70 5.70
5.05 5.05
2.35 2.35
0.78
3
0.14
h 3
2
0.82
2
2% 2%
0.40
0.56
3.8.2 Methodology
A logical sequence to determine the position of the bridge on the ground and define
its spans and dimensions of the main elements could have the following steps:
– Analysis of existing information
– Preliminary location of the abutments: definition of the total length of the viaduct
and position of the longitudinal fixed point
– Distribution of spans
– Deck pre-dimensioning
– Bearings
– Pre-dimensioning of pier elevations and foundations
This information constitutes the minimum data necessary to carry out the prelim-
inary design of the structure.
Later, for the development of the detailed design, this information will have to
be completed in some aspects such as the specific conditions of each particular
foundation of the structure.
529 + 000
368.00
E-1 P-1 P-2 P-3 P-4 P-5 E-2
P.K. 528 + 631 55.00 P.K. 528 + 685 65.00 P.K. 528 + 750 65.00 P.K. 528 + 815 65.00 P.K. 528 + 880 65.00 P.K. 528 + 945 55.00 P.K. 528 + 999
RIO
P.C. + 640.00 528 + 700 528 + 800 528 + 900 529 + 000
368.00
As the main span is 65 m, the two most feasible alternatives are a prestressed con-
crete deck with a box section or a double steel beam with a concrete slab deck.
Of the two solutions, the pre-stressed box solution was chosen because, as the
height of the piles is moderate, the friction during a possible push does not produce
excessive longitudinal moments in the pier. Furthermore, as the structure is not in a
seismic zone, the advantage of the decrease in mass associated with the steel solution
is not a great advantage.
For this reason, a prestressed concrete section with a deck of 4.45 m has been
planned, which means a span-to-deck ratio of L/15, which is very suitable for con-
struction using the push technique (Figure 3.83).
The standard cross section of the deck chosen is the typical one for a
post-tensioned concrete box with a constant deck. The bottom width of the
box is 6.00 m, with sloping side webs and 3.00 m long end cantilevers.
The box section is made up of a lower slab 0.36 m thick in the centre of the span,
which increases to 0.75 m in the area on piers with gussets at the junction with the
cores. The upper slab is of variable thickness due to the transverse slope with a min-
imum of 0.35 m in the centre of the slab. This slab is limited in its union with the
webs. The thickness of the webs is 0.50 m, which allows for parallel pre-stressing
cable pairs (Figures 3.84 and 3.85).
The estimated average dead weight of the deck is 297.5 kN/m, and the dead loads
to be considered are those included in Table 3.10.
14.00
6.25 6.25
5.70 5.70
5.05 5.05
2.35 2.35
0.78
3 3
0.14
2
0.20
0.82
2
2% 2%
0.40
0.56
0.40
0.21
0.13
0.58 0.58
3.60 1.22 3.20 1.22 3.60
3.70
4.36
0.55
0.35
5.70
1.12 3.00 1.12
0.40
0.62
6.50
14.00
0.14
0.20
0.40
0.13
0.21
0.21
1.80 3.20 1.80
0.55 0.55
3.41
4.50
4.36
0.71 3.20 0.71
0.35
0.62
0.40
3.60 0.55 5.70 0.55 3.60
14.00
0.14
0.20
1.00
0.13
2.98
3.41
4.50
4.36
2.75
1.80
1.80
2.10
0.75
0.62
Ballast
Max 10 × 0.60 × 20 × 1.3 = 156.0
Med 10 × 0.60 × 20 = 120.0
Min 10 × 0.60 × 20 × 0.70 = 84.0
Total
Max 198.1
Med 162.1
Min 126.1
112 3 Conceptual Design of High-Speed Railway Bridges
3.8.7.2 Bearings
The transversal loads are going to be resisted in all the piers and abutments, and the
torsional forces are also resisted in each support (piers or abutments). Therefore, in
each pier or abutment two bearings are used.
One is unidirectional, longitudinally guided POT, to transfer the transversal
forces and the other is a multidirectional POT-type device, free in both directions,
longitudinal and transversal (Figures 3.86 and 3.87).
3.8.7.3 Abutments
The abutments are of the closed type. Due to the length of this viaduct, a track
expansion device has been planned for Abutment number 1 (Figure 3.88).
In Abutment number 2 the fixed point of the deck is materialised by connecting
the deck to the abutment wall by pre-stressed bars. A friction slab has been installed
to prevent the sliding of the abutment (Figure 3.89).
Pot free
Longitudinally guided pot
D1 D2
ZS
D1 D2 ZNMB
ZS
A C
E.sup.
ZNMB
E.sup.
C
GUÍA
C
E.inf.
E.inf. ZAP.
ØA
D1
ZAP. 50 50
Section C–C
D1
ØA
50 50
B B Section A–A
ØA D ØA
D
D
E
ZS GUÍA E
ZNMB ZS
E.sup. ZNMB
E.sup.
D2
D2
C
E.inf.
E.inf.
ZAP.
ØA ZAP.
ØA
50 Section B–B 50
50 50
Section D–D
A C
Plan Plan
Longitudinally guided pot Free pot bearing
Figure 3.87 Longitudinally guided POT and multidirectional free POT bearing.
3.8 Worked Example 113
bearing
1.11
9.32
7.39
0.50
2.50
2.50
8.10 0.10 12.75 2.80
0.10
0.10
L pile
Ø1.80 Ø1.80
23.65
Ballast Ballast
Sub-ballast
Shape layer Continuous rail
0.30
21.00 3.00
0.60
Waterproofing
1 sheet
2 1
2
Drain tube
3.8.7.4 Piers
The piers on which the deck rests are made up of a single shaft in a box section with
external dimensions of 6.00 m wide and a variable deck, the minimum value of this
being 2.40 m at the crown and increasing with height at a rate of 4 cm/m. The wall
thickness is 0.40 m. At the top, the section has a solid zone of 1.50 m. The height of
these piers is between 17 and 24 m.
The foundations of piers 1 and 5 are direct by footings, while piers 2, 3, and 4
have deep foundations, consisting of a pile cap of 6 piles each that rest on rock
(Figures 3.90 and 3.91).
114 3 Conceptual Design of High-Speed Railway Bridges
A
6.50 A
Pot free 2.00 2.50 2.00 2.40
1.30
1.30
1.50
1.50
Longitudinally
guided pot 6.50
0.20
0.20
0.20
0.40
0.05
0.05
0.05 0.05
0.40
Variable
Variable
0.40
0.05
0.05
0.05 0.05
0.20
0.40 0.40
B B B B
Section B-B
Scale 1: 50
H
H
6.50
0.20 6.10 0.20
0.20
2.55 1.00 2.55
2.40
2.00
0.20
2.20 2.20
1.00
2.50
2.50
1.50
1.50
(Min.)
0.10
(Min.)
13.00 10.50
6.50
2.00 2.50 2.00 2.40
1.30
1.30
1.50
1.50
0.20
0.20
6.50
0.20
0.40
0.05
0.05
0.05 0.05
VARIABLE
VARIABLE
0.40
0.40
0.05
0.05
0.40
0.20
H
SECCIÓN B-B
50
6.50
0.20 6.10 0.20
0.20
0.40
0.20
2.20 2.20
VISTA POR A
3.00
3.00
50
(MIN.)
(MIN.)
0.10
0.10
References
1 Manterola J., Astiz M.A., and Martínez A. (1999). Puentes de Ferrocarril de Alta
Velocidad Revista de Obras Públicas n∘ 3386 Madrid.
2 ADIF. Norma ADIF Plataforma NAP 2-0-0.1. Puentes y viaductos ferroviarios. 2a
Edición 2019.
3 Romo, J., Corres, H., and Pérez, A. (2005). High speed railway bridges in seismic
areas. In: IABSE Symposium: Structures and Extreme Events, Lisbon, Portugal.
4 Rui-Wamba, J. (2020). Teoría Unificada de las Estructuras y Cimientos. Barcelona:
Editorial Reverté.
5 Romo J. (2010). Riudellots High Speed Line Bridge IABSE Symposium: Large
Structures and Infrastructures for Environmentally Constrained and Urbanised
Areas, Venice, Italy.
6 Leonhardt, F. (1983). Bridges Aesthetic and Design. Deutsche Verlags-Anstalt
GmbH.
7 Zeller, W. and Saul, R. (1991). Long span bridges of the new railroad lines in
Germany IABSE Reports.
8 Schlaich, J., Schmitt, V., Marx, S. et al. (2008). Leitfaden Gestalten von
Eisenbahnbrücken, 1e. Berlin: DB Netze AG.
9 Schlaich, M. (2012). Integral Railway Bridges in Germany. 22nd Dresdener
Brückenbausymposium. Dresden.
10 Martin J.L. (1999). Viaduc sur l’A7 à Bonpas. Ponts métalliques Bulletin 19
OTUA.
11 Plu B., Durot F., and Teisseire J. (1999). Viaduc de la Toulubre. Ponts métalliques
Bulletin 19 OTUA.
12 Priestly, M.J.N., Seible, F., and Calvi, G.M. (1996). Seismic Design and Retrofit of
Bridges. John Wiley & Sons Inc.
117
Design Basis
José Romo
4.1 Introduction
This chapter sets out the specific actions for rail bridges in general and for bridges
for high-speed lines in particular. It also indicates the criteria for combining these
actions for the various checks to be carried out.
In general, the different design regulations for bridges are based on a series of
general criteria and on the verification of the structure according to the theory of
Limit States: Service and Ultimate. The former (S.L.S.) are those which, if exceeded,
will cause the structure to cease to fulfil the purpose for which it was designed
(whether for functional, personal comfort, or aesthetic reasons), without this lead-
ing to its collapse, while the latter (U.L.S.) are those which, if exceeded, will lead to
the exhaustion or collapse of the structure or part of it.
In addition, and specifically for railway bridges, special checks must be carried
out to ensure the safety of the railway traffic running on the bridge. These checks
are described later in this chapter.
Likewise, and specifically for accidental load situations, the design of the struc-
ture must be robust enough so that damage from these accidental actions, such as
those corresponding to a derailment, does not cause disproportionate damage to the
bridge.
This chapter includes the specific loads and actions of railway bridges in general and
bridges for high-speed railway lines.
Firstly, permanent actions are defined which include, in addition to the own
weights of the elements, some special loading situations such as the partial removal
of the ballast in an area of the bridge and the actions due to overhead line equipment
and other railway infrastructure and equipment.
Secondly, actions due to railway operations are explained, including:
– vertical loads,
– dynamic effects,
– centrifugal forces,
– nosing forces,
– traction and braking forces,
– aerodynamic actions from passing trains.
Finally, accidental actions as the effect of traffic derailment are presented.
(the one indicated in the drawings of the project) and a lower and higher value
have to be considered. Thus:
– Nominal value, determined by the theoretical thicknesses defined in the
project. In general, in the absence of specific data, a thickness of 0.50 m can be
considered for UIC track widths.
– Lower value (Gk,inf), obtained by decreasing the nominal value by thirty percent
(30%).
– Higher value (Gk,sup ), obtained by increasing the nominal value by thirty per-
cent (30%).
– Rails: frequently of the UIC-60 type: According to Table A.6 of EN 1991-1-1 [1], a
load of 1.2 kN/m must be considered and per track.
– Prestressed concrete sleepers: According to Table A.6 of EN 1991-1-1, a load of
4.8 kN/m per track is to be considered.
– Ballast walls: Load obtained from their geometry.
– Sidewalks: Load obtained from its geometry and density of materials.
– Piping gutters: According to experience a load of 3.0 kN/m and track should be
considered.
– Catenary: Normally the load is transmitted pointwise coincident with the cate-
nary poles. It is not a critical load for the bridge and its value is of the order of
0.2–0.3 kN/m for a two-track bridge.
– Railings: Load obtained from the geometry and density of the materials. In the
case of a steel-type railing, a load of 0.35 kN/m per side can be considered
(Figure 4.1).
Figure 4.4 Example of loads representing normal traffic on main lines (source EN
1991-2) [2].
qvk q
vk
a c a
Load qvk a c
model (kN/m) (m) (m)
SW / 0 133 15,0 5,3
SW / 2 150 25,0 7,0
– Loads representing normal traffic on the main lines. For example the load model
71 in EN 1991-2 (Figure 4.4).
– Heavy load models: As is the case with the SW/2 in EN 1991-2 for instance
(Figure 4.5).
Passenger train loads at speeds exceeding 200 km/h, such as HSML freight models.
These are the load models used in dynamic analyses.
– Load model ‘train without load’ representing the effects of an unloaded train.
When combined with the action of wind, for example.
Depending on the line loads, the corresponding load values are usually amplified
or reduced (e.g. in EN 1991-2 loads 71 and SW/2) with a classification coefficient.
4.3 Rail Traffic Actions and Other Actions Specific of Railway Bridges 121
Qvi
qv1 + qv2 (1)
Qv1 + Qv2
qv1QV1 qv2QV2
Qvi/2
4:1
e Qvi/4 Qvi/4
(2)
r a a b
Figure 4.6 Transverse eccentricity, load distribution in sleepers, and load diffusion through
the ballast (source EN 1991-2).
The codes indicate the eccentricities to take into account possible transverse load
imbalances between rails. For example in EN 1991-2 an eccentricity e = r/18 is
indicated, where r is the distance between rails which in the case of 1435 mm results
in 80 mm. The standards also indicate how the point loads are distributed between
the individual sleepers, as well as the distribution of the loads on the ballast. For
example, in EN 1991-2, 50% of the load is considered for the sleeper underneath
and 25% for the adjacent sleepers. In this standard, the load distribution through
the ballast is considered to be at a 4 : 1 gradient (see Figure 4.6)
the supports; the rest of the force is transmitted through the track and resisted behind
the abutments. The fraction of the force transmitted through the deck to the supports
should be determined taking into account the interaction of track–structure.
Braking and starting forces not only condition the bridge substructure but also
increase the stresses on the rails. To determine the value of these stresses it will be
necessary to carry out an analysis of the track–structure interaction as described in
detail in Chapter 6.
Qtk Characteristic values of the centrifugal force corresponding to the vertical loads
(kN, kN/m).
Qvk Values of vertical loads, not affected by any impact coefficient, in (kN, kN/m).
v Maximum speed (m/s).
R Radius of curvature (m).
f Reduction factor
The reduction value f takes into account the influence length Lf of curved track
on the deck which is most unfavourable for the design of the element under consid-
eration, as well as the speed of the train (see EN-1991-2).
For example, when designing a bridge pier with continuous spans of 40 m,
Lf = 40 m, the value of the reduction factor f for a speed of 350 km/h is 0.40. If the
radius is 8000 m (standard value for a speed line of 350 km/h) the centrifugal force is:
(350∕3.6)2
Qtk = ⋅ (0.40 ⋅ Qvk ) = 0.048 Qvk (4.2)
g ⋅ 8000
Therefore the centrifugal force is close to the 5% of the vertical live load acting on
the pier.
4.3 Rail Traffic Actions and Other Actions Specific of Railway Bridges 123
Surface of
structure
q
1k
Surface of
structure
5.00
±q1k
≤ 5.00
5.00
a
g
q1k
the mass and suspension characteristics, the presence of regularly spaced supports
of the deck slab or track, the vehicle imperfections, and the dynamic characteristics
of the track.
(1) (1)
(2) (2)
α × 0,7 × LM 71 α × 0,7 × LM 71 α × 0,7 × LM 71
(1)
1.80 TYP
outside of the deck, applied at a height of usually 1.80 m above the running surface
of the rail closest to the element.
These loads can be critical in the case of non-massive elements, such as cables,
diagonals, and other truss elements.
The values of the longitudinal loads to be considered depend on the different
codes, reaching 10 MN in the longitudinal direction and 3.5 MN in the transversal
direction if the distance between the element and the axis of the nearest track is less
than 3 m. These values are zero when the element is located more than 5 m from the
axis of the nearest track.
The importance of these loads means that in general, if the bridge has sensitive
elements, the design must either move these elements away from the tracks (more
than 5 m), or use containment measures or increase the resistance of the element
itself of the structure.
Although these loads are accidental they are usually combined with permanent
loads and live loads. In particular, it is usually established that after the impact and
if the damaged element is an element of a certain strength, the bridge must be able to
withstand the permanent actions and static live loads, considering that the impacted
element has lost half of its resistance capacity. In the case of less robust elements,
such as cables, it is usually required that the bridge is able to withstand the per-
manent actions and static use live loads without counting on the resistance of the
element on which it has been impacted (Figure 4.10).
Impact on Bridge Substructure Elements On railway bridges over railway tracks, the
piers faces and other deck supports located less than 5.0 m from the nearest track
axis should support forces of similar values as indicated in Section 4.3.4.1.
4.4.1.2 Load Situations for Limit State and Associated Acceptance Criteria
In the case of checks relating to rail traffic safety and passenger comfort, the codes
usually establish the number of loaded tracks to be considered in each case.
In general, when it comes to checking the safety of rail traffic, all loaded tracks
are considered, while for the comfort of users, only one track is usually considered
loaded.
Thus, for example in the Eurocode EN 1990 [4], the traffic safety checks include:
deck twist, vertical and horizontal deformation of the deck checks is carried out
considering the worst situation even those where all tracks are loaded even if there
are three or more.
In the same code, for the verification of the combined response of the structure and
track to variable actions including limits to vertical and longitudinal displacement
of the end of a deck a maximum of two loaded tracks are considered. On the other
hand, when checking the vertical acceleration of the deck and the comfort of the
passengers, this standard considers only one loaded track.
In addition to these checks relating to the safety of rail traffic, other aspects must
be considered, such as those relating to passenger comfort, which are specified by
establishing a limit for vertical accelerations and vertical deformations of the deck.
These special checks related to traffic safety and passenger comfort are included
in Chapters 2 and 5.
368.00
E-1 55.00 P-1 65.00 P-2 65.00 P-3 65.00 P-4 65.00 P-5 55.00 E-2
RIO
14.00
6.25 6.25
0.13
0.21
0.55
0.35
5.70
1.12 3.00 1.12
0.40
0.62
6.50
4.7.2 Actions
4.7.2.1 Vertical Loads
Permanent Loads The permanent loads are, on the one hand, the own weight of the
deck and, on the other hand, the weight of the track and other elements of the railway
superstructure.
For the self-weight a typical cross section area of 12.22 m2 is used; therefore
SW = 12.22 ⋅ 25 kN/m3 = 305.6 kN/m.
For the deck loads (ballast, track, etc.) the maximum and minimum values
considered are: DLmax = 198.1 kN/m and DLmin = 126.1 kN/m.
Live Loads According to the EN 1991-2 for lines with no heavy trains load types 71
and SW/0 have to be considered in the design of the bridge (Figures 4.13 and 4.14).
qvk qvk
a c a
Load qvk a c
model [kN/m] [m] [m]
SW / 0 133 15,0 5,3
Due to the length of the span a single track with load type 71 has a total bending
moment of M = 50 150 kN/m and a maximum shear of V = 3060 kN whereas for the
load type SW/0 the total bending moment is M = 44 650 and the shear V = 2900 kN.
Therefore the load type 71 is the governing live load.
For the design of the bridge a classification factor of 1.21 has been consid-
ered as it is a line with slightly heavier traffic than a normal line. Therefore the
LL1 = 1.21 ⋅ 80 ⋅ 2 = 193.6 kN/m.
For the sidewalks a live load of 5 kN/m2 has been considered. In combination with
the railway load this live load in the sidewalks has a combination factor of 0.50;
therefore, the load to be considered is: LL2 = 2 ⋅ 1.80 ⋅ 5 kN/m2 ⋅ 0.50 = 9.0 kN/m.
For the purpose of preliminary dimensioning, the locomotive weight included in
the 71 type load has been simplified by an equivalent point load of 1180 kN which is
used in conjunction with the linear load of 80 kN/m.
Therefore the equivalent load is a uniform load, in the whole span, of 202.5 kN/m
and a point load of 1180 kN.
Dynamic Factor 𝚽 To take into account the dynamic amplification of the loads the
impact coefficient has been considered according to EN 1991-2.
The equivalent LΦ is given by the expression: LΦ = k ⋅ Lm = 1.50 ⋅ (54 + 4 ⋅
65 + 54)/6 = 92 m.
The value of the amplification factor Φ2 assuming a carefully maintained the track
is according to EN 1991-2 (6.4.5).
Φ2 = √ 1.44 + 0.82 = 0.973, and therefore the value to be considered is Φ2 = 1.00:
92−0.2
no dynamic amplification for the resistance of the structural members has to be con-
sidered.
Live Load Eccentricity Eccentricity of the vertical loads LM71 and SW/0 is calculated
according to EN 1991-2 (6.3.5) (Figure 4.15).
1.45
e = 2.35 + = 2.43 m
18
2.35
1.435
The traction force considering La,b = 368 m (total length of the viaduct)
Qlak = 33 kN∕m ⋅ 368 m ≤ 1000 kN => Qlak = 1000 kN
And the braking force:
Qlbk = 20 kN∕m ⋅ 368 m ≤ 6000 kN => Qlbk = 6000 kN
Therefore considering that the train on one track accelerates while on the
opposite track it brakes, the value of the total longitudinal load is equal to:
Qlk = 1.21 ⋅ 7000 = 8470 kN.
Mean Wind Mean wind is determined using the expression: vm (z) = cr (z). c0 (z) vb
according to EN 1991-4-4 (4.3)
where
c0 (z) is the orography factor: in this case c0 (z) = 1.10 (valley)
cr (z) is the roughness factor
The terrain category of the site is type II.
Therefore kr = 0.19, zo = 0.05 m, zmin = 2 m.
Considering a height z = 25 m
z
cr = kr ⋅ ln = 1.18
z0
Wind turbulence factor
√ √
7kz 7 ⋅ 0.19
cg = 1 + = 1+ = 1.42
cz ct 1.18 ⋅ 1.1
14
cd = 2.5 − 0.3 ⋅ = 1.70
5.4
– With live load
d = 5.24 + 4.00 = 9.24 m
(deck height, ballast, rails, and UIC train height = 4.00 m)
14
cd = 2.5 − 0.3 ⋅ = 2.05
9.24
Unloaded state (Figure 4.16)
2a Railing:
Table 8.1 EN 1991-1-4 (open parapets on both sides)
A = 4.44 + 0.60 = 5.04 m < 5.24 (deck + ballast height + rail)
( ) ( )
1 2 1
F = cd ⋅ A 𝜌vc = 1.70 ⋅ 5.24 ⋅ 1.25 ⋅ 53.62 ⋅ 10−3
2 2
= 16.0 kN∕m
Loaded state (Figure 4.17)
( )
1
F = 2.05 ⋅ 9.24 ⋅ ⋅ 1.25 ⋅ 53.62 10−3 = 34.01 kN∕m
2
4.00
9.24
0.80 (0.60 balasto
+0.20 m)
4.44
Figure 4.17 Deck loaded state.
To obtain the load on each bearing due to the centred loads, the corresponding
axial force in the pier was divided by 2.
Nbearing = Npier ∕2
In the case of eccentricity due to vertical loads, the load on each bearing was
obtained as follows:
Nbearing = Npier ∕2 ± Mpier ∕4.00
Being 4.00 the distance between bearings.
QtK
1.80
0.60 + 0.20
4.44
Nbearing S Nbearing
4.00
If Qv is the vertical load associated with the centrifugal force, and being
Qtk = 0.034 Qvk
The axial force per bearing is N bearing = ±0.034 ⋅ 1.76 = 0.06 Qvk (Tables 4.2 and 4.3).
F = 34.01 kN∕m
9.24
h= = 4.62 m Mτ = L ⋅ 34 ⋅ 10−3 ⋅ 4.62 = 0.157L
2
N = ±2.55 MN (central piers)
N = ±2.34 MN (lateral piers)
N = ±1.06 MN (abutments)
Table 4.2 Reaction in pier heads due to permanent loads and transversally symmetrical
vertical live loads.
Table 4.3 Reaction in pier heads due to transversally asymmetrical vertical live loads.
Table 4.4 Maximum axial Reaction in pier heads due to transversally asymmetrical
vertical live loads.
Longitudinal
connection
VA
2.00
HT
c.o.g
4.75
Deck
3.00
1.85
7.90
9.90
HA
0.30
3.15
Ancho
3.00
15.50
1.00
A
20.00 ¿L?
A1 P1 P2 P3 P4 P5 A2
Friction forces
54.00 65.00 65.00 65.00 65.00 54.00
These forces are listed in Table 4.5. For the determination of the friction forces of
the POTs a friction coefficient of 3% has been considered and, therefore, the force
transmitted to the abutment is 3% of the total permanent load of the deck.
For the force due to the temperature variation between rail and track an approx-
imate value of 2 MN has been adopted. An accurate value of this force due to the
via-structure interaction can be seen in the example in Chapter 6.
F2
10.90
F1
σ0 σ1
@seismicisolation
4.7 Worked Example 139
Vertical (MN)
1 1
𝜎 1 = k0 γ h F1 = 𝜎 ⋅ b ⋅ h = γ ⋅ b ⋅ h2 ⋅ k0
2 1 2
𝜎 2 = k0 q F2 = k0 q ⋅ b h
where b = 14 m the width of the abutment, h = 10.90 the total height, g = 20 kN/m3
the weight of the soil embankment, and φ = 30 angle of internal friction of the soil.
The values obtained are:
F1 = 8.32 MN F2 = 2.73 MN F = F1 + F2 = 11.05 MN
Table 4.7 summarises the forces acting on the abutment.
The friction produced by the vertical loads acts as a stabilising force. A value of
0.45 has been adopted as the soil–slab friction coefficient, which is equivalent to an
angle of 24.2∘ .
If a sliding safety coefficient of 1.35 is assumed, the value of the minimum length
of the friction slab can be obtained from the following expression:
0.45 (63.49 + 3.46L)
1.35 =
25.92
And therefore the value of the length of the friction slab L will be 4.15 m. Adopting
a value of L = 4.50 m.
140 4 Design Basis
References
1 EN 1991-1-1 Eurocode 1: Actions on structures. Part 1.1: General actions.
2 EN 1991-2 Eurocode 1: Actions on structures. Part 2: Traffic loads on bridges.
3 EN 1337-1:2001 Structural bearings – Part 1: General design rules.
4 EN 1990 Eurocode 0: Basis of structural design.
5 EN 1991 Eurocode 1: Actions on structures – Part 2: Traffic loads on bridges, 2
Annex D Basis for fatigue assessment of railway bridges.
143
5.1 Introduction
Railroad loads are highly dynamic, and their effects must be determined accounting
for this fact. Isolated dynamic loads will produce deflections which can be higher
or lower than the maximum value of the load applied statically. Whether they are
higher or lower will depend on the variation of the load with time and on the ratio
between the load duration, td , and the structure’s first natural period of vibration, T.
The ratio between the dynamic and static deflection is known in literature as the
Dynamic Load Factor or DLF (also denoted as Φ). As a maximum, an isolated
load can produce a deflection equivalent to twice the maximum value of the load
(DLF = 2.0). For example, for a suddenly applied rectangular load, this value is
reached when the ratio td /T exceeds 0.5.
But, for the same type of loading, the deflection is only 60% of the static deflection
when td /T is equal to 0.1. The effect of other load variations with time will result in
smaller deflections. For an introduction to the dynamic response to moving loads in
SDOF systems please refer to Appendix A.
In railway bridges, the problem is complicated by the fact that the passage of a
train over a bridge is equivalent to applying a series of consecutive, periodic, isolated
loads. In such cases, resonance phenomena can occur. In a system without damping,
the response of an infinitely elastic structure can be unbounded (DLF = ∞) if the
period of the load coincides with a natural vibration period of the structure. For an
introduction to resonance effects please refer to Appendix A.
5.1.1 Resonance
For a railway bridge, the load is periodic due to the number of axles that go through
a bridge at a constant velocity. Referring to Figure 5.1, the period or frequency of the
d d d d d d
D D
OO OO OO OO OO OO OO OO OO OO
P P1 P2 P2 P2 P2 P1 P
D D D D D D D
D (m)
T (s) 18 27
action can be determined from the velocity, by accounting for the spacing between
axle pairs (or bogies):
2π D D
Ti = =i i = 1,2, 3 … → vres,min = (5.1)
𝜔i v Tstr
The first natural period of a typical concrete high-speed viaduct can be in the
range of 0.1 to 0.3 seconds. The spacing, D, between pairs of axles of typical trains
varies between 18 and 27 m for trains with bogies. Using extreme values of the
above parameters, the velocity for which resonance occurs can be estimated. This
is done in Table 5.1. The smaller resonance velocity occurs for more flexible decks
and smaller distances between axle pairs. These numbers help to explain why
resonance is not normally considered for velocities below 200 km/h, as detailed
in the following sections. There is also no experimental evidence of resonance
problems in railway bridges for speeds lower than 200 km/h.
The dynamic factor for real trains is divided into the sum of the two factors, one if
the track has perfect geometry, 𝜑′ , and a second factor to account for imperfections
in the track, 𝜑′′ .
𝜑′ is given as a function of K, which depends on the first natural frequency of the
structure, n0 , the determinant length, L𝛷 , and the maximum nominal velocity, v, of
the real train:
v
K=
2L𝛷 n0
⎧ K
⎪ 4
if K < 0.76
𝜑′ = ⎨ 1 − K + K (5.5)
⎪1.325 if K ≥ 0.76
⎩
Figure 5.2 represents the value of 𝜑′ as a function of K. Its maximum value is equal
to 1.325.
The additional dynamic effect due to the irregularities of the track is given by
( )
min
v[m∕s] [ ( )
L [m] 2
( ) ( L [m] )2 ]
22 − 𝛷10 L𝛷 [m]n0 [Hz] − 𝛷20
𝜑′′ = 56e + 50 −1 e (5.6)
100 80
2.00
1.80
1.60
1.40
1.20
1.00
φ′
0.80
0.60
0.40
0.20
0.00
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
K
1.00 40
0.90
35
0.80
30
0.70
25
0.60
n0 (Hz)
φ″
0.50 20
0.40
15
0.30
10
0.20
5
0.10
0.00 0
0 10 20 30 40 50 60 70 80 90 100
LΦ (m)
φ″ for Upper limit for frequency φ″ for lower limit for frequency
n0 Upper limit n0 Lower limit
Figure 5.3 𝜑′′ as a function of the determinant length L𝜱 (for v > 80 km/h).
Figure 5.3 shows how 𝜑′′ varies with the determinant length for the upper and
lower limit natural frequencies for a maximum velocity greater than 80 km/h
(22 m/s). For smaller velocities, the values should be multiplied by a factor of
v[m/s]/22. This component of the dynamic factor is most important for bridges
with small spans and for secondary bridge components.
The total dynamic factor is a function of the type of rail track maintenance. The
static load due to railway traffic should be multiplied by 1 + 𝜑′ + 0.5𝜑′′ for a carefully
maintained track and by 1 + 𝜑′ + 𝜑′′ for standard maintenance.
The total dynamic factor for real trains is given in Figure 5.4 as a function of the
determinant length for the upper and lower natural frequency limits and for stan-
dard and careful track maintenance. The higher values of the dynamic factor apply
to the smaller determinant length. For standard maintenance and the lower limit of
the natural frequency the dynamic factor provides values which are slightly higher
than 2.00.
2.00
1.80
1.60
1 + φ′+ (1.00 or 0.5)φ″
1.40
1.20
1.00
0.80
0.60
0.40
0.20
0.00
0 10 20 30 40 50 60 70 80 90 100
LΦ (m)
n0,max Standard track maintenance n0,min Standard track maintenance
n0,max Careful track maintenance n0,min Careful track maintenance
Figure 5.4 Total dynamic factor as a function of the determinant length for the upper and
lower limits of the natural frequency and for careful and standard track maintenance.
load factor to be applied to the real trains to assess deflections and forces on the struc-
ture. The structure is to be designed for the worst case of the critical train with its
impact factor which can now go over the value of 200 due to resonance, or the design
train (LM71, SW/0 or SW/2) increased by the envelope dynamic factor, as outlined
in Section 5.1.2. On the other hand, dynamic analysis is carried out to determine
the maximum accelerations which are used to assess problems of track instability,
deconsolidation of ballast, and passenger comfort.
Dynamic analysis is to be carried out for a selection of real trains which should
account for ‘each permitted or envisaged train formation for every type of high-speed
train permitted or envisaged to use the structure at velocities over 200 km/h’ [1].
In order to cover this range for international lines where European high-speed
interoperability criteria are applicable, EN 1991-2 proposes two families of trains:
HSLM-A and HSLM-B. The second family is only relevant for simply supported
structures with span lengths less than or equal to 700 m. The HSLM-A family
comprises 10 trains which have different numbers of intermediate coaches (N),
different coach lengths (D), different axle bogie spacings (d), and different axle loads
(see Figure 5.5 and Table 5.2).
The mass of the system should consider all permanent loads (self-weight plus
removable loads: ballast, track, communications, handrails, etc.), obtained using
nominal values of density. An underestimation of the mass will lead to locate the
resonance point at larger velocities, since the resonance velocity is inversely propor-
tional to the period of the structure (see Eq. (5.1)) and the period becomes smaller
when the mass is smaller. On the other hand, an overestimation of the mass, which
leads to greater inertia, will lead to reduced maximum accelerations. Because of this,
5.1 Introduction 149
D N×D D
3 11 3 d d d d d d d 3 11 3
3.525 3.525
A1 18 18 2.0 170
A2 17 19 3.5 200
A3 16 20 2.0 180
A4 15 21 3.0 190
A5 14 22 2.0 170
A6 13 23 2.0 180
A7 13 24 2.0 190
A8 12 25 2.5 190
A9 11 26 2.0 210
A10 11 27 2.0 210
two different assumptions regarding the permanent mass should be made when car-
rying out the dynamic analysis:
– a lower bound, designed to predict maximum deck accelerations, for which the
minimum ballast thickness and minimum ballast dry density (∼1700 kg/m3 )
should be used
– an upper bound, designed to determine the velocity at which resonant phenomena
occur using the maximum saturated density of ballast and accounting for future
track lifts.
With respect to stiffness, an overestimation of the stiffness will again lead to
an underestimation of the period of the structure and therefore (Eq. (5.1)) to an
overestimation of the resonance velocity. For this reason, the estimation of the
stiffness should tend to be a lower bound. Any potential cracking of concrete should,
therefore, be accounted for.
Regarding damping, the response of the structure in the range close to resonance
is sensitive to this value. EN-1991-2 Table 6.6 specifies the values of the damping
index to be used as a function of the material used for the deck of the structure (steel,
prestressed concrete, or reinforced concrete) and as a function of the span, as shown
in Table 5.3.
For bridges with two tracks, the dynamic effects should be estimated considering
only one track loaded.
150 5 Dynamic Behaviour of High-Speed Railway Bridges
Table 5.3 Damping index to be used for dynamic analysis (acc. to EN 1991-2 Table 6.6).
Dynamic analysis should be carried out for all HSLM trains plus real trains, if
required, at velocities going from 200 km/h to 1.2 times the maximum design speed
of the railway line. The velocity increment should not exceed 10 km/h.
The results of the dynamic analysis can be plotted in terms of maximum displace-
ment (see Figure 5.6) and maximum acceleration (see Figure 5.7) as a function of
the train velocity for each of the universal and real trains considered.
These figures are drafted for a simply supported prestressed concrete bridge with
a span, L, of 4500 m with a cross section having the following properties:
– Area, A = 7.96 m2
– Inertia, I y = 7.71 m4
– Modulus of elasticity of concrete (C45/55) Ec = 36.3 GPa.
Regarding the loads, the upper-bound weight due to ballast, track, communica-
tions, and handrails amounts to G2,sup = 137.6 kN/m (to be used for analysis of res-
onance effects), while the lower bound is G2,inf = 74.4 kN/m (to be used for the
determination of maximum accelerations).
It is interesting to estimate the first natural period of vibration of the structure
(T str ) and from its value infer the velocity for which resonance occurs by application
of Eq. (5.1). The period can be obtained by assimilation of the simply supported beam
to a SDOF system as outlined in [3]. By equating maximum kinetic energy (Ek ) and
maximum potential energy of the SDOF system, the maximum velocity of the system
is equal to the circular frequency times the maximum deflection:
√
1 1 KE
K y 2
= ME ẏ max → ẏ max =
2
y = 𝜔ymax (5.7)
2 E max 2 ME max
The equivalent mass is determined by the condition that the kinetic energy (K E )
of the SDOF system and that of the actual beam be the same (see Eq. (5.8)). 𝜙 is the
form function which describes the ratio between the deflection at a given coordinate
x and the maximum deflection.
2
⎛ ⎞
⎜ ⎟
1 1
L L
⎜ y(x) ⎟
2
M (𝜔ymax ) = m(x)(𝜔y(x)) dx → ME =
2
m(x)⎜ ⎟ dx
2 E 2 ∫0 ∫0 ⎜ ymax ⎟
⎜ ⏟⏟⏟ ⎟
⎝ 𝜙 ⎠
L
→ ME = m(x)(𝜙(x))2 dx (5.8)
∫0
5.1 Introduction 151
18
16
14
12
ymax (mm)
10
0
0 50 100 150 200 250 300 350 400 450
Train velocity (km/h)
Train A1 Train A2 Train A3 Train A4 Train A5
Train A6 Train A7 Train A8 Train A9 Train A10
0.30
0.25
Deck acceleration (g)
0.20
0.15
0.10
0.05
0.00
0 50 100 150 200 250 300 350 400 450
Train velocity (km/h)
Train A1 Train A2 Train A3 Train A4 Train A5
Train A6 Train A7 Train A8 Train A9 Train A10
Figure 5.7 Maximum acceleration as a function of train velocity for the 10 train types of
HSLM-A.
152 5 Dynamic Behaviour of High-Speed Railway Bridges
In this example, the equivalent mass is determined in Eq. (5.9) by using a sinu-
soidal form function for 𝜙(x):
M = mL = A ⋅ 2.5 ⋅ L + G2 ⋅ L = (7.96 ⋅ 2.5 + 13.76) ⋅ 45 = 1514.7 ton
L L( )
π 2
ME = m(𝜙(x))2 dx = m sin x dx
∫0 ∫0 L
( 2π )
L 1 + cos L x L 1514.7
=m dx = m = = 757.35 ton (5.9)
∫0 2 2 2
The equivalent stiffness of the SDOF is obtained by dividing the equivalent
load, PE , by the maximum deflection ymax . The equivalent load is determined by
establishing that the external work done by the forces in the SDOF system is equal
to the external work done by the forces on the real system:
L L
y(x)
PE ymax = p(x)y(x)dx → PE = p(x) dx
∫0 ∫0 ymax
⏟⏟⏟
𝜙
L
PE
→ PE = p(x)𝜙(x)dx → KE = (5.10)
∫0 ymax
For the given example:
q(x) = q
L L ( ) [ ( )]L
π L π 2
PE = q(x)𝜙(x)dx = q sin x dx = q cos x = qL
∫0 ∫0 L π L 0 π
= 0.64 qL
PE 0.64qL EI 36.2 ⋅ 106 ⋅ 7.71
KE = = 4
= 49.152 3 = 49.152 ⋅
ymax 5 qL L 453
384 EI
= 150 545.1 kN∕m (5.11)
The first natural period would then be:
√ √
ME 735.35
T = 2π = 2π = 0.44 s (5.12)
KE 150 545.1
Given that the axle pair spacing of the Universal trains (which is equal to the coach
length, D) is between 18 and 27 m, it can be expected, by virtue of Eq. (5.1), that
resonant peaks would occur at velocities ranging from 18/0.44 ⋅ 3.6 = 147 km/h to
27/0.44 ⋅ 3.6 = 221 km/h. This is what can be observed in Figures 5.6 and 5.7. This
implies that it is not necessary to cover the full range of velocities when trying to
determine the maximum Dynamic Load Factor. It would be sufficient to verify the
resonant frequencies and their multiples, as well as the maximum velocity.
The effect of track irregularities can be accounted for by adding to the dynamic
coefficient determined by dynamic analysis 0.5 𝜑′′ for careful track maintenance or
𝜑′′ for standard track maintenance. High-speed railways always meet careful main-
tenance criteria.
5.2 Methods for Dynamic Calculations and Structural Response 153
It can be demonstrated that the eigenvectors of M−1 K are orthogonal with respect
to both M and K (for a demonstration see the Appendix of reference [3]). This makes
matrices 𝚽T M𝚽 and 𝚽T K𝚽 diagonal.
154 5 Dynamic Behaviour of High-Speed Railway Bridges
1∑
n−1
𝜉i = a (𝜔 )2j−1 (5.19)
2 j=0 j i
The values ai can be determined for the system of Eq. (5.19), by imposing values
of the damping index for each vibration mode.
With this assumption Eq. (5.17) becomes a series of m independent SDOF differ-
ential equations, which are easy to solve. m is the number of eigenvalues considered
in the analysis. If the elements of diagonal matrix 𝚲K are named Mi∗ , the elements
of diagonal matrix 𝚲K are named Ki∗ and the elements of diagonal matrix 𝚲C , are
named Ci∗ , then the expression of Eq. (5.20) holds:
Ci∗ Ki∗ ∑
n
𝜙−1 ij
q̈ i + ̇
∗ qi + ∗ qi = Fj (t)jy
Mi Mi j=1
Mi∗
Ci∗ = 𝛟−𝟏
i C𝛟i
Mi∗ = 𝛟−𝟏
i M𝛟i
Ki∗ = 𝛟−𝟏
i K𝛟i (5.20)
In Eq. (5.20), jy is a directional coefficient which is equal to 1.00 if force F j (t) goes
in the direction of degree of freedom i and is equal to 0.00 if it does not.
The forces applied on the nodes F j (t) can be simulated by triangular pulses
(Figure 5.8). The time between pulses can be approximated as the spacing between
bogies, D, divided by the velocity of the train, v. The duration of the triangular pulse
would be equal to the sum of the distances to the adjacent nodes divided by the
velocity.
Once the independent SDOF systems have been solved, their effects need to be
superimposed. This can be done, of course by direct summation of the displace-
ments, or acceleration time histories. However, this procedure is time consuming.
For this reason, it is common practice to determine the maximum response of each
vibration mode at a given location and superimpose the maximum effect by using
a combination rule that considers that the maximum response from the different
vibration modes is not likely to occur simultaneously. A classical combination rule
is the Square Root of the Sum of the Squares (SRSS), given in Eq. (5.21):
√
√m
√∑
E = √ Ej2 (5.21)
j=1
5.2 Methods for Dynamic Calculations and Structural Response 155
450
Fj (t) Load applied on a given node (kN)
400 D/v
350
300
250
200
150
100
50 (li–1 + li)/v
0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
t (s)
Load vs time on a node
where:
This criterion can, however, be unsafe when the periods of the vibration modes
differ by less than 10%. In such cases the Complete Quadratic Combination (CQC)
should be applied. This combination criterion is defined in Eq. (5.22):
√
√m m √
√∑∑
E=√ Ek rkj Ej = ET rE
k=1 j=1
√ ( ) 3∕2
8 𝜉k 𝜉j 𝜉k + 𝜌kj 𝜉j 𝜌kj
rkj = ( )2 ( ) ( )
1 − 𝜌2kj + 4𝜉k 𝜉j 𝜌kj 1 + 𝜌2kj + 4 𝜉k2 + 𝜉j2 𝜌2kj
Tk
𝜌kj = ≤1 (5.22)
Tj
This expression is very general and accounts for the possibility of having different
damping indexes for the different vibration modes.
1.50
1.00
0.50
0.00
Φ(x)
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
–0.50
–1.00
–1.50
x/L
n=1 n=2 n=3 n=4
for example, a sine function can be adequately used to model the eigenforms (see
Eq. (5.23) and Figure 5.9):
π
𝜙(x) = 𝜙0 sin n x (5.23)
L
To maintain the criterion that the norm of the eigenforms is equal to 1.00, 𝜙0
should be fixed as follows:
L
∫0 𝜙2 (x)dx 𝜙20 ( L ) 𝜙2 L ( ( ))
π π
= 1.00 = sin2 n x dx = 0 1 − cos 2n x dx
L L ∫0 L 2L ∫0 L
𝜙20 [ L
(
π
)]L 𝜙2 √
= x− sin 2n x = 0 = 1.00 → 𝜙0 = 2 (5.24)
2L 2nπ L 0 2
In general, to satisfy varying boundary conditions (fixed displacements or rota-
tions), the eigenforms can take on the shape shown in Eq. (5.25) [3]:
nπ nπ nπ nπ
𝜙(x) = a cos x + b sin x + c cosh x + d sinh x
L L L L
𝜕 4 𝜙(x) n4 π4
= 𝜙(x) (5.25)
𝜕x4 L4
⏟⏟⏟
(an )4
Note the property shown in the second line of the equation (i.e. that the fourth
derivative of the eigenform with respect to x is equal to the eigenform times the
fourth power of the form coefficient an ).
An expression for the natural frequencies of the system can be determined from
dynamic equilibrium conditions. It is well known from the static sectional equilib-
rium equation of moments that the moment is proportional to the curvature and the
5.2 Methods for Dynamic Calculations and Structural Response 157
mдxÿ
M + дM V V M V V + дV
дx дx
L
Mn∗ = m (𝜙n (x))2 dx = mL
∫0
𝜕 4 𝜙 EI L 2 𝜕 4 𝜙 EI
Kn∗ = (𝜙 (x)) dx = L
𝜕x4 𝜙n ∫0 n
𝜕x4 𝜙n
L
c 𝜕 4 𝜙 EI 1
→ q̈ n + q̇ n + 4 qn = F (x, t)𝜙n (x)dx (5.29)
m 𝜕x 𝜙n m mL ∫0 j
⏟⏞⏞⏟⏞⏞⏟
𝜔2
For the case of a simply supported beam, subjected to a given generic load
F max f (𝜏), the application of Eq. (5.29) would result in:
nπ
𝜙n = 𝜙0 sin x
L
𝜕4 𝜙 n4 π4 nπ n4 π4
= 𝜙0 4 sin x = 4 𝜙n (x)
𝜕x 4 L L L
y(t) = 𝜙n (x)qn
L
c n4 π4 EI F
q̈ n + q̇ n + 4 qn = max f (𝜏, x)𝜙n (x)dx (5.30)
m L m mL ∫0
⏟⏟⏟
𝜔2
To obtain the deflection at a given abscissa, y(x), the value of q(x) must be multi-
plied by the value of the eigenform at that coordinate so that:
nπx
y(x) = q(x)𝜙0 sin (5.31)
L
Applications of this equation are given in Sections 5.2.2 and 5.2.3.
𝛿(x) in Eq. (5.32) is Dirac’s delta function which is equal to infinity when x = 0 and
equal to 0 otherwise. It has the property shown in Eq. (5.33):
⎧ P −2𝜉𝜔𝜔f
⎧ 2𝜉m𝜔
⎪A = 𝜙0 m ( )2
⎪ ( ⏞⏞⏞ ⎪ 𝜔2 − 𝜔2f + 4𝜉 2 𝜔2 𝜔2f
)
⎪ A 𝜔2 − 𝜔2 + B c 𝜔 = 0 ⎪ ( )
⎨ f m f →⎨ (5.35)
⎪ ( ) ⎪ 𝜔2 − 𝜔2f
P
⎪−A c 𝜔 + B 𝜔2 − 𝜔2 = 𝜙 P ⎪ B = 𝜙0 ( )2
⎩ m f f 0
m ⎪ m
⎩ 𝜔2 − 𝜔2f + 4𝜉 2 𝜔2 𝜔2f
The values of coefficients C and D can then be determined from the initial condi-
tions of a system that is initially at rest:
−2𝜉𝜔𝜔f 𝜙0 mP
q(0) = 0 = A + C = ( )2 +C
𝜔 − 𝜔f + 4𝜉 𝜔 𝜔f
2 2 2 2 2
2𝜉𝜔𝜔f 𝜙0 mP
→C= ( )2
𝜔2 − 𝜔2f + 4𝜉 2 𝜔2 𝜔2f
P 2𝜉𝜔𝜔f
→ C = 𝜙0 ( )2
m
𝜔2 − 𝜔2f + 4𝜉 2 𝜔2 𝜔2f
̇
q(0) = 0 = B𝜔f − 𝜉𝜔C + 𝜔d D
160 5 Dynamic Behaviour of High-Speed Railway Bridges
( )
𝜔2 − 𝜔2f 𝜙0 mP 𝜔f 2𝜉𝜔𝜔f 𝜙0 mP
𝜔
→ D = −( )2 𝜉
𝜔d ( 2 )2
+
𝜔
𝜔2 − 𝜔2f + 4𝜉 2 𝜔2 𝜔2f d 𝜔 − 𝜔2f + 4𝜉 2 𝜔2 𝜔2f
𝜔f
( ( ))
2𝜉 2 𝜔2 − 𝜔2 − 𝜔2
P 𝜔d f
→ D = 𝜙0 (5.36)
m ( 2 )2
𝜔 − 𝜔2f + 4𝜉 2 𝜔2 𝜔2f
The position and velocity of the structure can be determined when the force exits
the structure as follows:
𝜙0 P
q0 = ( )2
m 𝜔2 − 𝜔2f + 4𝜉𝜔2 𝜔2f
( )
⎛−2𝜉𝜔𝜔f cos 𝜔f L + 𝜔2 − 𝜔2f sin 𝜔f L ⎞
⎜ v v ⎟
⋅⎜ ( ( ) )⎟
𝜔
⎜+e−𝜉𝜔 v 2𝜉𝜔𝜔f cos 𝜔d L + f 𝜔2 − 𝜔2 sin 𝜔f L ⎟
L
⎝ v 𝜔d f v ⎠
𝜙0 P
q̇ 0 = ( )2
m 𝜔2 − 𝜔2f + 4𝜉𝜔2 𝜔2f
( )
⎛2𝜉𝜔𝜔2 sin 𝜔f L + 𝜔f 𝜔2 − 𝜔2f cos 𝜔f L ⎞
⎜ f v v ⎟
⎜ L ⎟
⎜ ⎛2𝜉𝜔𝜔f cos 𝜔d ⎞⎟
⎜−𝜉𝜔e−𝜉𝜔 Lv ⎜ v ⎟⎟
⎜ ⎜ 𝜔f ( ( ))
L ⎟⎟
⋅⎜ ⎜+ 2𝜉 2 𝜔2 − 𝜔2 − 𝜔2f sin 𝜔d ⎟ ⎟ (5.37)
⎜ ⎝ 𝜔d v ⎠⎟
⎜ ⎛−2𝜉𝜔𝜔f sin 𝜔d L ⎞⎟
⎜ −𝜉𝜔 Lv ⎜ v ⎟⎟
⎜+𝜔d e ⎜ 𝜔f ( ( ))
L⎟
⎟
⎜ ⎜+ 2𝜉 2 𝜔2 − 𝜔2 − 𝜔2f cos 𝜔d ⎟⎟
⎝ ⎝ 𝜔d v ⎠⎠
The above equations can be applied for the number of eigenmodes considered. To
obtain the actual displacements, the generalised coordinates q, should be multiplied
by the value of the eigenmode at the given location (see Eq. (5.31)). Therefore, the
contribution to the deflection at centre span for pair values of n will be nil.
To see an application of this method, the same example already analysed in
Section 5.1.4 is considered. The example consists in a simply supported bridge with
a span, L, of 4500 m. The area of the section is 7.96 m2 and the inertia 7.71 m4 . The
elastic modulus of concrete is taken as 36.2 GPa. Additionally, an upper-bound
superimposed dead load equivalent to G2,sup = 137.6 kN/m is present. So, the total
upper-bound mass of the structure is M = (7.96 ⋅ 2.5 + 13.76) ⋅ 45 = 1514.7 ton. The
first natural period of this structure is T1 = 0.448 s. It is assumed that the train is a
high-speed train, and that the maximum velocity is 350 km/h (i.e., v ∼ 100 m/s) and
that the damping index is 1% as recommended by EN 1991-2 (see Table 5.3).
Figure 5.11 shows the response of the system in terms of maximum deflection
at centre span for the first four modes of vibration. It is very clear that in such a
5.2 Methods for Dynamic Calculations and Structural Response 161
4.00 0.03
3.00
0.02
0.00 0
0.00 0.50 1.00 1.50 2.00 2.50 3.00
–1.00
–0.01
–2.00
–0.02
–3.00
–4.00 –0.03
Time (s)
Figure 5.11 Deflections due to the first four vibration modes – plotted separately.
system, only the first mode of vibration is of significance (note that the deflection of
modes 2, 3, and 4 is plotted on the secondary axis on a different scale, roughly 100
times larger). Modes with pair values of n do not contribute at all to the deflection
at centre span since the deflection of the corresponding eigenmodes is nil at that
point. Figure 5.12 shows the deflection due to the superposition of the first four
modes compared to that of the first model only. The maximum deflection of the
superposed modes is 0.64% lower than that of the first mode only and the difference
can barely be seen.
Table 5.4 shows the natural frequencies of the first four modes, the maximum
centre span deflection due to each mode, the deflection due to the sum of the modes
and the deflection obtained by the SRSS criterion, which would be applicable since
the difference between the periods of the first four modes is much larger than 10%.
The SRSS criterion is safe sided for this case since the contribution of the third mode
is negative at centre span.
Mode 1 2 3 4
4.00
3.00
2.00
y (mm)
1.00
0.00
0.00 0.50 1.00 1.50 2.00 2.50 3.00
–1.00
–2.00
–3.00
–4.00
Time (s)
Figure 5.12 Deflections due to the additive effect of the first four vibration modes.
3.00
2.50
2.00
1.50
y (mm)
1.00
0.50
0.00
0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00
–0.50
–1.00
–1.50
Time (s)
Bogie #1 Bogie #2 Bogie #3 Bogie #4
10.00
5.00
y (mm)
0.00
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00
–5.00
–10.00
–15.00
Time (s)
1 Bogie 7 Bogies 14 Bogies
4.00
3.50
3.00
2.50
Φ
2.00
1.50
1.00
0.50
0.00
0 50 100 150 200 250 300 350
Train velocity, v (km/h)
Figure 5.15 Dynamic Load Factor (𝛷) as a function of the train velocity.
Figure 5.15 shows the Dynamic Load Factor as a function of the train velocity.
A very clear resonant peak can be observed at 177 km/h (corresponding to the first
natural period of vibration).1 Also shown in the figure is the effect of two bogies
(which is the maximum number that will be on the bridge at a given time). With
only two bogies, there is no resonant phenomenon and, in this case the response
increases with the velocity of the train. This is logical as more energy is transferred
to the structure and the ratio between the load duration and the natural period of
the structure is not too low (td /T ∼ 0.5 for a velocity of 350 km/h).
1 It is interesting to note that for such a speed, a dynamic analysis would not be required by
EN 1991-2:2003, despite the fact that resonance occurs. In this case, the Dynamic Load Factor
estimated according to Eq. (5.3) would be only 1.03. However, this factor is applied to the LM71 or
SW/0, SW/2 trains which are much heavier than the real trains. For the LM71 train, the maximum
static deflection would be 18.6 mm, much larger than the 13.44 mm of the real train accounting for
the dynamic behaviour. This comparison demonstrates that such cases are covered by the envelope
loading.
5.2 Methods for Dynamic Calculations and Structural Response 165
5.00
4.00
3.00
Φ
2.00
1.00
0.00
0 50 100 150 200 250 300 350
Train velocity, v (km/h)
0% damping 1% damping 2% damping
Figure 5.17 Dynamic interaction model considering the interaction between vehicle,
track, and structure [5].
dtd dBd
Secondary
suspension M, J
MB, JB MB, JB
Primary
suspension
MW
LB deB
L
Figure 5.18 Full vehicle–structure interaction model (Source: taken from [4]).
Kp Cp
Mu
(equivalent to the mass of the bogie supported by the axis plus the weight of vehicle
box supported by the axle), while M u corresponds to the mass of the wheels and the
axle itself. These models neglect the coupling provided by the vehicle box and the
bogies.
5.3 Interoperability
5.3.1 Introduction
The European Union has issued two directives and one amendment on the inter-
operability of European railways systems:
– Directive 96/48/EC on the interoperability of the trans-European high-speed
railway system
– Directive 2001/16/EC on the interoperability of the trans-European conventional
railway system
– Directive 2004/50/EC which amends the above.
The objective of the directives is to facilitate the circulation of railway traffic in the
different railway networks existing in Europe. To this end, the directives target the
design, construction, serviceability, and exploitation phases setting a set of common
standards.
Directive 96/48/EC defines interoperability as ‘the ability of the trans-European
high-speed rail system to allow the safe and uninterrupted movement of high-speed
trains which accomplish the specified levels of performance’. This objective is to be
achieved by meeting regulatory, technical, and operational conditions.
An important technical condition is related to the traffic loading. For this pur-
pose, EN 1991-2 defines a series of universal dynamic trains for which all bridges
of the European Networks must be designed. These are presented in Sections 5.3.2
and 5.3.3.
= = = = = = = = = d = = = = = = =
4.65
3.35
3.8
4.2
4.8
3.5
6.00 20
5.50
Spacing between load, d (m)
4.35
3.65
5.65
2.8
3.2
4.5
5.8
5.00 15
4.50
6.15
2.65
5.15
5.5
4.00 10
N
4.35
1.2
4.5
5.8
3.65
3.50
4.2
1.6
3.5
3.00 5
4.65
5.65
2.50
2.00 0
0.0 0.6 1.2 1.8 2.4 3.0 3.6 4.2 4.8 5.4 6.0 6.6
Span, L (m)
d (m) N
depends on the span (see Figure 5.20). Figure 5.21 provides the values of these vari-
ables as a function of the span.
mL/8 = M/8 mL/4 = M/4 mL/4 = M/4 mL/4 = M/4 mL/8 = M/8
θ1(t) θ5(t)
y2(t), θ2(t) y4(t) , θ4(t)
y3(t), θ3(t)
l = L/4
R = 2.5 Pi(t) R = 2.5 Pi(t)
mL/8 = M/8 mL/4 = M/ mL/8 = M/8 mL/8 = M/8 mL/4 = M/4 mL/8 = M/8
Bar 12 Bar 23 Bar 12 Bar 23
4
θ1(t) θ1(t) θ3(t)
y2(t), θ2(t) y2(t), θ2(t)
º
y (t)
3
l = L/4 l = L/4
R = 2.5 Pi(t) R = 2.5 Pi(t)
(a) (b)
Figure 5.23 Symmetric and antimetric structural models. (a) Symmetric vibration modes.
(b) Antimetric vibration modes.
For the symmetric case, the boundary conditions are that the deflection at node
1 and the rotation at node 3 are nil. To impose these conditions, all the items of
the row and column corresponding to these degrees of freedom are set to zero
except for the term on the diagonal which is set to 1 (see Eq. (5.42)). These rows and
columns can then be deleted from the matrix when determining the eigenvalues and
eigenforms.
⎡y1 𝜃1 y2 𝜃2 y3 𝜃3 ⎤
⎢ ⎥
⎢1 0 0 0 0 0⎥ ⎡ 𝜃1 y2 𝜃2 y3 ⎤
⎢ ⎥ ⎢ ⎥
⎢0 6 ⎢ 4 −6 2
4 − 2 0 0⎥ 0 ⎥
⎢ l ⎥ ⎢ l ⎥
⎢ 6 24 12 ⎥ ⎢ 6 24 12 ⎥
EI ⎢ 0 − 0 − 0 ⎥ → EI ⎢− 0 − 2⎥
l l2 l2 2 (5.42)
l ⎢ ⎥ l ⎢ l l l ⎥
⎢ 6 ⎥ ⎢ 6⎥
⎢0 2 0 8 −
l
0⎥ ⎢ 2 0 8 − ⎥
l
⎢ ⎥ ⎢ 12 6 12 ⎥⎥
⎢0 12 6 12
0 − 2 − 0 ⎥ ⎢ 0 − 2 −
⎢ l l l2 ⎥ ⎣ l l l2 ⎦
⎢ ⎥
⎣0 0 0 0 0 1⎦
For the antisymmetric case, the boundary conditions are that the deflection at
nodes 1 and 3 are nil (see Eq. (5.43)).
⎡y1 𝜃1 y2 𝜃2 y3 𝜃3 ⎤
⎢ ⎥
⎢1 0 0 0 0 0⎥ ⎡ 𝜃1 y2 𝜃2 𝜃3 ⎤
⎢ 6 ⎥ ⎢ 6 ⎥
⎢0 4 − 2 0 0⎥ ⎢ 4 − 2 0⎥
⎢ l ⎥ ⎢ l ⎥
EI ⎢ 6 24 6⎥ EI ⎢ 6 24 6⎥
→
l ⎢⎢
0 0 ⎥
l ⎢⎢ l ⎥
0 − − 0 (5.43)
l l2 l⎥ l2 l⎥
⎢0 2 0 8 0 2⎥ ⎢ 2 0 8 2⎥
⎢ ⎥ ⎢ ⎥
⎢0 0 0 0 1 0⎥ ⎢ 6 ⎥
⎢ ⎥ ⎣ 0 2 4⎦
6 l
⎢0 0 2 0 4 ⎥⎦
⎣ l
The mass will be concentrated at the nodes of the model. The concentrated mass
matrix is obtained by applying the following equations:
The translational mass for node i is given by:
(l1 + l2 )
mi = m (5.44)
2
The rotational mass for node i, is given by:
(l1 + l2 ) (l + l )
j i = 𝜌c I y + mSDL y2g 1 2 (5.45)
2 2
172 5 Dynamic Behaviour of High-Speed Railway Bridges
where:
l1 , l2 are the lengths of the two bars that converge on the node (one of which may be
zero if there is no bar)
m is the mass per length of the bridge including the self-weight and the superim-
posed dead load (ballast, track, barriers, communications)
𝜌c is the density of concrete
mSDL is the mass per length of the superimposed dead load (ballast, track, barriers,
communications)
yg is the distance from the centre of gravity of the superimposed dead load to the
centroid of the cross section.
The concentrated mass matrix can then be written as shown in Eq. (5.46):
⎡ y1 𝜃1 y2 𝜃2 y3 𝜃3 ⎤
⎢ l ⎥
⎢ m 0 0 0 0 0 ⎥
⎢ 2 ⎥
⎢⏟⏟ ⏟ ⎥
⎢ m1 ⎥
⎢ ( ) l ⎥
⎢ 0 𝜌c Iy + mSDL y2g 0 0 0 0 ⎥
⎢ 2 ⎥
⎢ ⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟ ⎥
⎢ j1 ⎥
⎢ ⎥
⎢ 0 0 ml 0 0 0 ⎥
⎢ ⏟⏟ ⏟ ⎥
⎢ m2 ⎥
⎢ 0 ( ) ⎥
0 0 𝜌c Iy + mSDL yg l2
0 0
⎢ ⎥
⎢ ⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟ ⎥
⎢ j2 ⎥
⎢ l ⎥
⎢ 0 0 0 0 m 0 ⎥
⎢ 2 ⎥
⎢ ⏟⏟⏟ ⎥
⎢ m3 ⎥
⎢ ( ) l ⎥
⎢ 0 0 0 0 0 𝜌c Iy + mSDL y2g ⎥
⎢ 2⎥
⎢ ⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⎥
⎣ j3 ⎦
(5.46)
Since the mass matrix is diagonal, its inverse, M−1 , will be the diagonal matrix
formed by the inverse values of the diagonal elements.
Now that matrices K and M−1 are known it is necessary to determine the eigenval-
ues of M−1 K, which will be equal to the square of the natural periods of the different
vibrations forms of the structure. The analysis for the symmetric case will provide
the eigenvalues of the symmetric eigenforms, while the analysis for the antimetric
case will provide the eigenvalues corresponding to the antimetric eigenforms.
Assuming that M is a diagonal matrix, multiplication by M−1 will just divide each
row of matrix K by mii .
5.4 Application Examples 173
1
⎡ 0 0 ⎤
0 ⎡ 4 −6 2 0 ⎤
⎢ j1 ⎥ ⎢ l ⎥
⎢ 1 ⎥ ⎢ 6 24 12 ⎥
⎢0 m 0 0 ⎥ ⎢− 0 − 2⎥
EI 2
M K =⎢
−𝟏 2 ⎥ ⎢ l l l ⎥
⏟⏟⏟ ⎢ 0 0 1
0 ⎥⎥ l ⎢⎢ 2 0 8 − ⎥
6⎥
𝛼A ⎢ j2 l
⎢ ⎥ ⎢ ⎥
⎢0 0 1 ⎥ ⎢ 12 6 12 ⎥
⎣ 0 0 − −
m3 ⎦ ⎣ l2 l l2 ⎦
⎡ 4 −
6 2
0 ⎤
⎢ j1 lj1 j1 ⎥
⎢ 6 24 12 ⎥
⎢− 0 − ⎥
EI ⎢ lm2 l2 m2 l2 m2 ⎥
l ⎢ 2 6 ⎥⎥
= (5.47)
8
⏟⏟⏟⎢ j 0 −
⎢ 2 j2 lj2 ⎥
𝛼
⎢ 12 6 12 ⎥
⎢ 0 − − ⎥
⎣ l m3 lm3 l m3 ⎦
2 2
⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟
A
11.25
m1 = m3 = 33.66 ⋅ = 189.338 t
2
m2 = 2m1 = 378.675 t
11.25
j1 = j3 = (2.5 ⋅ 7.96 + 13.76 ⋅ 0.752 ) = 155.475 tm2
2
j2 = 2j1 = 310.95 tm2
EI 36.2 ⋅ 106 ⋅ 7.71
= = 2.481 ⋅ 107
l 11.25
M−𝟏 K = 2.481 ⋅ 107
⏟⏟⏟ ⏟⏞⏞⏞⏟⏞⏞⏞⏟
𝛼A 𝛼
To solve for the eigenvalues of matrix A, it is necessary to obtain first the charac-
teristic equation. This is done in Equations (5.50) and (5.51).
| −1 |
| M K − 𝜆I| = 0
|⏟⏟⏟ |
A
| −1 |
| M K − 𝜆I| = 0
|⏟⏟⏟ |
A
→ (2.573 ⋅ 10−2 − 𝜆)(5.008 ⋅ 10−4 − 𝜆)((2.573 ⋅ 10−2 − 𝜆)(5.008 ⋅ 10−4 − 𝜆) − 4.831 ⋅ 10−6 )
+ 1.408 ⋅ 10−3 ⋅ −3.430 ⋅ 10−3 ((2.573 ⋅ 10−2 − 𝜆)(5.008 ⋅ 10−4 − 𝜆) − 4.831 ⋅ 10−6 )
⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟
−4.829 ⋅ 10−6
⎛ ⎞
⎜ ⎟
⎜ 2.577⋅10 −5 6.620⋅10 −4 2.508⋅10 −7 ⎟
⎜ ⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞ ⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞ ⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⎟
⎜ ⎟
+ ⎜ 2 ⋅ 2.573 ⋅ 10−2 ⋅ 5.008 ⋅ 10−4 + (2.573 ⋅ 10−2 )2 + (5.008 ⋅ 10−4 )2 ⎟ 𝜆2
⎜ ⎟
⎜ −1.254 ⋅ 10−7 − 4.829 ⋅ 10−6 − 8.272 ⋅ 10−5 ⎟
⎜⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⎟
⎜ ⎟
⎝ 6.003 ⋅ 10−4 ⎠
⎛ ⎞
⎜ ⎟
⎜ −6.631⋅10−7 −1.291⋅10−8 ⎟
⎜ ⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞ ⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞ ⎟
⎜ ⎟
⎜ −2(2.573 ⋅ 10 ) ⋅ 5.008 ⋅ 10 −2 ⋅ 2.573 ⋅ 10 (5.008 ⋅ 10 )
−2 2 −4 −2 −4 2
⎟
⎜ ⎟
⎜ 1.243⋅10−7 2.419⋅10−9 6,53⋅10−9 ⎟
+ ⎜ ⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞ ⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞ ⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⎟ 𝜆
⎜ +2.573 ⋅ 10−2 ⋅ 4.831 ⋅ 10−6 + 5.008 ⋅ 10−4 ⋅ 4.831 ⋅ 10−6 + 2 ⋅ 1.254 ⋅ 10−7 ⋅ 2.573 ⋅ 10−2 ⎟
⎜ ⎟
⎜ +4.829 ⋅ 10−6 ⋅ (2.573 ⋅ 10−2 + 5.008 ⋅ 10−4 ) + 2 ⋅ 8.272 ⋅ 10−5 ⋅ 5.008 ⋅ 10−4 ⎟
⎜ ⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟ ⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟ ⎟
⎜ −8
⎟
⎜ 1.267 ⋅ 10−7 8.285⋅10 ⎟
⎜⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⎟
⎝ −3.413⋅10 −7 ⎠
1.660⋅10−10 −6.225⋅10−11
⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞ ⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞
+(2.573 ⋅ 10−2 )2 (5.008 ⋅ 10−4 )2 −4.831 ⋅ 10−6 ⋅ 2.573 ⋅ 10−2 ⋅ 5.008 ⋅ 10−4
−8.302⋅10−11 −6.224⋅10−11 2.333⋅10−11
⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞ ⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞ ⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞
(−2.573 ⋅ 10−2 )2 ⋅ 1.254 ⋅ 10−7 −4.829 ⋅ 10−6 ⋅ 2.573 ⋅ 10−2 ⋅ 5.008 ⋅ 10−4 + 4.829 ⋅ 10−6 ⋅ 4.831 ⋅ 10−6
−2.075⋅10−11 1.037⋅10−11
⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞ ⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞
+1.555 ⋅ 10−11 −8.272 ⋅ 10−5 (5.008 ⋅ 10−4 )2 + 8.272 ⋅ 10−5 ⋅ 1.254 ⋅ 10−7 + 1.556 ⋅ 10−11
⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟
2.550⋅10−12
=0
6.00E-08 ω 32 = 419273
4.00E-08
Value of characteristic polynomial
ω 12 = 13769
2.00E-08 ω 12 = 197.2
0.00E+00
1.00E+00 1.00E+01 1.00E+02 1.00E+03 1.00E+04 1.00E+05 1.00E+06
–2.00E-08
–4.00E-08
ω 42 = 868317
–6.00E-08
–8.00E-08
–1.00E-07
λ = ω2
Figure 5.24 shows a plot of the characteristic equation where the 4 eigenvalues,
multiplied by EI/l are shown. These values correspond to the square of the first four
natural frequencies corresponding to symmetric vibration modes. The horizontal
scale is logarithmic. Figure 5.25 shows a detail of the first part of the curve where it
is easier to see the first two roots.
5.00E-11
ω 12 = 197.2 ω 22 = 13769
0.00E+00
1.00E+00 1.00E+01 1.00E+02 1.00E+03 1.00E+04
–5.00E-11
–1.00E-10
λ = ω2
Figure 5.25 Characteristic polynomial and its solutions (symmetric system) – detail.
5.4 Application Examples 177
ω 32 = 647517
2.00E-08
Value of characteristic polynomial
1.00E-08 ω 12 = 320037
ω 12 = 3076
0.00E+00
1.00E+00 1.00E+01 1.00E+02 1.00E+03 1.00E+04 1.00E+05 1.00E+06
–1.00E-08
–2.00E-08
–3.00E-08 ω 42 = 957630
λ = ω2
For the symmetric model, the natural modes of vibration can then be derived from
the system of linear equations shown in Eq. (5.53), by substituting 𝜆 by 𝜔i 2 l/EI. In
this way the constant EI/l can be ignored.
( )
2 l
(M K − 𝜆i I)yi = 𝟎 ⇔ A − 𝜔i I yi = 𝟎
−𝟏
EI
⎤ ⎡𝜃1 ⎤ ⎡ ⎤
0
⎡2.573 ⋅ 10−2 − 𝜆 −3.430 ⋅ 10−3 1.286 ⋅ 10−2 0
⎢ ⎥ ⎢ ⎥ ⎢0⎥
⎢ −1.408 ⋅ 10
−3
5.008 ⋅ 10−4 − 𝜆 0 −2.504 ⋅ 10−4 ⎥ ⎢y2 ⎥ ⎢ ⎥
2.573 ⋅ 10 − 𝜆 −1.715 ⋅ 10−3 ⎥ ⎢𝜃2 ⎥ ⎢⎢ ⎥⎥
= 0
⎢ 6.432 ⋅ 10 −3
0 −2
⎢ 0
⎣ 0 −5.008 ⋅ 10−4 −2.817 ⋅ 10−3 5.008 ⋅ 10−4 − 𝜆⎥⎦ ⎢⎣y3 ⎥⎦ ⎢ ⎥
⎣0⎦
(5.53)
For the antimetrical model this can be done by the same procedure from the
expression of Eq. (5.54):
⎤ ⎡𝜃1 ⎤ ⎡ ⎤
0
⎡2.573 ⋅ 10−2 − 𝜆 −3.430 ⋅ 10−3 1.286 ⋅ 10−2 0
⎢ ⎥⎢ ⎥ ⎢0⎥
⎢ −1.408 ⋅ 10
−3
5.008 ⋅ 10−4 − 𝜆 0 1.408 ⋅ 10−3 ⎥ ⎢y2 ⎥ ⎢ ⎥
2.573 ⋅ 10−2 − 𝜆 6.432 ⋅ 10−3 ⎥ ⎢𝜃2 ⎥ ⎢⎢ ⎥⎥
= 0
⎢ 6.432 ⋅ 10−3 0
⎢ ⎥ ⎢ ⎥ 0
⎣ 0 3.430 ⋅ 10−3
1.286 ⋅ 10−2
2.573 ⋅ 10 − 𝜆⎦ ⎣y3 ⎦ ⎢ ⎥
−2
⎣0⎦
(5.54)
The corresponding eigenvectors for the symmetric structure are the following:
2.00
1.50
Eigen vector shape
1.00
0.50
0.00
0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00 45.00
–0.50
–1.00
–1.50
–2.00
x (m)
First vibration mode Second vibration mode Third vibration mode
400 Force history applied on nodes 2, 3, and 4 for the first vibration mode
350
300
250
ϕj1Fj (kN)
200
150
100
50
0
0.00 0.20 0.40 0.60 0.80 1.00
t (s)
Figure 5.28 Load history for nodes 2, 3, and 4 as a single axle weighing 340 kN crosses
the bridge, weighted by the eigenform – First vibration mode.
The next step is to solve, for each eigenfrequency and eigenmode, i, the equation
of motion for a single load circulating on the bridge, and then obtain the response
to the train by summing up the effect of the consecutive axles.
Ci∗ Ki∗ ∑
n
𝜙ji
q̈ i + ̇
∗ qi + ∗ qi = Fj (t)jy (5.60)
Mi Mi j=1
Mi∗
For each node, the effect of the passing single load is modelled with a triangular
load. The load history of the three nodes with vertical movement of the model is
represented in Figure 5.28. The analysis can be further simplified by considering the
load applied on each node separately and then adding the deflections obtained from
the three equations:
The values of Ci∗ , Mi∗ , and Ki∗ can be determined by forward and backward
multiplication of the Damping, Mass, and Stiffness matrices by the corresponding
eigenvalue. Alternatively, and more easily, Ci∗ can be determined as:
Table 5.5 Parameters to be used for first bogie and the first mode of
vibration in the formulations of Appendix A.
The resulting values of coefficients Ci∗ (assuming all modes have the same 2%
damping index), Mi∗ , and Ki∗ for the symmetric model are given in Eq. (5.63):
⎡ 212.94 ⎤ ⎡ 380.19 ⎤ ⎡ 7.454 ⋅ 104 ⎤
⎢ ⎥ ⎢ ⎥ ⎢ ⎥
1.809 ⋅ 103 ⎥ 385.76 ⎥ 5.305 ⋅ 106 ⎥
C∗ = ⎢ M∗ = ⎢ K∗ = ⎢ (5.63)
⎢5.350 ⋅ 105 ⎥ ⎢2.064 ⋅ 104 ⎥ ⎢ 8.669 ⋅ 109 ⎥
⎢3.553 ⋅ 106 ⎥ ⎢9.538 ⋅ 104 ⎥ ⎢8.274 ⋅ 1010 ⎥
⎣ ⎦ ⎣ ⎦ ⎣ ⎦
The solution to the dynamic equation of a single degree of freedom system sub-
jected to a triangular load is solved in [3] and presented in Appendix A. However,
these formulations do not account for damping. Damping is important when there
is a series of axles crossing the bridge since it will significantly reduce resonance
effects. Therefore, a generalisation of these expressions is included in Appendix A.
The differential equations shown in Eq. (5.61) can be solved using the solution
detailed in Appendix A, applied to the values for t0 , Δt/2 and Δt shown in Table 5.5
for the first vibration mode:
To determine the maximum deflection at mid-span (which corresponds to the
third degree of freedom) the following expression will be applied:
∑
n
y3 = 𝜙3,i qi (5.64)
i=1
Observation of Eq. (5.58) leads to the conclusion that only modes 1, 3, 5, and 7 will
contribute to the vertical motion of the central node since the coefficients 𝜙i,j corre-
sponding to the other modes for the row corresponding to y3,i are nil. Additionally,
because the deflection of a given mode is inversely proportional to natural frequency
of the damped system, 𝜔d , because the damping of the system occurs faster when
the natural frequency is high, and because the frequency increases quite rapidly
(factor
√ of 8.36 – theoretically this should be 9 – from the first mode to the third:
13769
∼ 8.36), the contribution of the higher modes will be neglectable (see below).
197.2
Figure 5.29 shows the deflection in the structure when a 340 kN load crosses the
bridge at 177 km/h considering only the effects of vibration mode 1. The result is
the sum of the effect of the triangular pulse acting on nodes 2, 3, and 4 multiplied
by the modal factor 𝜙j,1 , where j = 2, 3, 4. In this case 𝜙j,1 = [0.707;1.00;0.707]. By
comparison with Figure 5.13, it is apparent that the same results are obtained with
the matrix formulation as with assumed sinusoidal eigenforms. However, the matrix
formulation is more general as it can be applied to any structure without previous
knowledge of the eigenform shapes.
182 5 Dynamic Behaviour of High-Speed Railway Bridges
3.00
2.00
1.00
y (mm)
0.00
0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00
–1.00
–2.00
–3.00
Time (s)
Figure 5.29 Effect of single load travelling over bridge at 177 km/h – first mode only.
2.50 0.05
2.00 0.04
1.00 0.02
0.50 0.01
0.00 0
0.00 0.50 1.00 1.50 2.00
–0.50 –0.01
–1.00 –0.02
–1.50 –0.03
Time (s)
Mode #1 Mode #3
Figure 5.30 Comparison of the contribution of the first and third modes.
Figure 5.30 shows a comparison of the contribution of the response of the structure
to the passing of one bogie of the train of modes 1 and 3. Note that the contribution
of mode 3 is plotted using the secondary vertical axis and therefore both contribu-
tions are not represented on the same scale. The maximum deflection from mode
1 is close to 2.7 mm while the maximum deflection from mode 3 is only 0.033 mm,
that is around 1% of the previous figure. It can be concluded as in Section 5.2.2 that
References 183
1.00
–1.00
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00
–3.00
–5.00
–7.00
–9.00
–11.00
–13.00
Time (s)
1 Bogie 7 Bogies 14 Bogies
Figure 5.31 Variation of the maximum deflection with time due to the passage of several
bogies.
sufficient accuracy can be obtained by considering only the first vibration mode for
the case of this simply supported beam.
By superimposing the effect of several bogies circulating on the bridge with a time
offset of D/v = 22/177 ⋅ 3.6 = 0.447 s, the effect of the train load can be determined.
Figure 5.31 shows the results comparing the effect of a single bogie to that of 7, to
that of 14. The results are nearly identical to those obtained with assumed sinusoidal
eigenshape functions (see Figure 5.14). The maximum deflection considering 14
bogies is 13.8 mm, while the maximum static deflection which occurs when two
bogies are on the bridge (since D = 22 m and L = 45 m) is 3.27 mm. This provides a
DLF of 4.2.
References
6.1 Introduction
This chapter deals with the interaction between track and structure. Due to move-
ments related to loads (vertical loads and traction and braking forces), tempera-
ture variations and the rheological behaviour of concrete, differential displacements
occur between the track and the structure. These differential movements generate
additional stresses in the rails, with respect to the situation of the rail on the ground,
which could lead to failure of the rails, and can affect the safety and serviceability
of the train circulation. This chapter describes how to analyse these effects, which
loads need to be considered, how they are combined, and which verifications need
to be made.
If the rail stress verifications are not satisfied, then a track joint has to be intro-
duced. This is a measure that ought to be avoided as much as possible due to the
maintenance costs associated with track joints. However, in some cases they are
unavoidable. In such cases, the track joints need to be designed (allowed displace-
ment) and regulated. Installation and maintenance of track joints are also discussed.
Finally a series of considerations are given depending on the longitudinal layout of
the bridge with support from a series of design examples which are fully developed.
For more information cf. [1].
σs = EsαΔT
Jointless length
Track
Figure 6.1 Stresses in Continuous Welded Rail (CWR) placed ground between rail joints.
steel (see Figure 6.1). The stress in the rail due to the temperature difference with
respect to the mean ambient value will be given by Eq. (6.1). For a 50 ∘ C tempera-
ture difference the stresses in the rail are about 105 MPa. The track is designed to
withstand these stresses, residual stresses due to manufacturing process and those
coming from traffic loads, plus a limited overstress.
𝜎s = Es 𝛼ΔT (6.1)
When the railtrack crosses a bridge, this analysis is complicated by the relative
movements between track and structure. These relative movements will increase the
stresses in the rails and can lead to deconsolidation of ballast, in the case of ballasted
track, or excessive forces on other track components, such as fasteners, in the case of
ballastless track. The track–structure interaction analysis is carried out to verify that
the stresses remain within admissible ranges and that other effects due to relative
displacements are controlled.
The track and deck of a bridge are connected by a series of elements whose
behaviour is certainly non-linear (see Figure 6.2) and in reality very complex. For a
ballasted track, the rails are connected to the sleeper by a fastening systems which
includes clips, pads, tie plates, and bolts. The sleepers are embedded on a bed of
ballast with 30–70 cm of ballast below the bottom face of the sleeper.
The connection between deck and rail is strained for a number of reasons. First of
all the track and the deck of the bridge will be subjected to different imposed strains.
Due to temperature, while the rail will not move if there are no track joints, the
bridge deck will expand and contract due to temperature variations (which can reach
around 35 ∘ C). But even if there are track joints, because of the much greater thermal
conductivity of steel with respect to concrete, the rails will be subjected to larger
temperature variations (around 50 ∘ C) and therefore larger strain variations than the
deck, which will also be at least partially protected from direct sunlight by the track.
Additionally, the concrete deck will be subjected to strains due to the rheological
behaviour of concrete (creep and shrinkage) which will be absent from the rails.
Vertical traffic loads cause rotations in the deck which will increase relative dis-
placement between the deck top surface and the track, as the deck will rotate with
respect to its centroid or with respect to its bearings if the longitudinal displacement
of the bearing is fixed. Traction and braking loads will be applied on the rails and will
6.2 Problem Statement 187
7.00
2.80 4.20
1.30 1.50 1.85 2.35
Rail
Sleeper
Ballast
0.20
1.76
Figure 6.2 Half section of a high-speed railway bridge showing a ballasted track and the
deck.
be transmitted to the deck through the connection between rail and sleeper and then
will be transmitted horizontally by friction within the ballast to the bridge deck. The
bridge deck will then move depending on the stiffness of the piers and the presence
or absence of structural joints.
As a result of all these actions, overstresses will develop in the rails with respect
to a situation where the track is laid on the ground which does not move. If the
overstresses are too high there is a danger that the rails will fail, either in tension
or in compression by buckling, leading a catastrophic accident as a train crosses the
bridge.
Another issue that can cause significant problems is the relative displacements
between decks at structural joints and between abutments and the deck ends. Here
longitudinal and vertical relative displacements need to be limited. If these relative
displacements are too large some undesirable phenomena can occur. In the case of
ballasted tracks there is a risk of ballast deconsolidation that can lead to both safety
issues as well as to a deficient quality of the train ride. In the case of ballastless track
relative displacements can produce excessive efforts on the rail or fasteners, leading
to fatigue problems for these elements or problems for rail pads (that can get loose
if clamping forces are reduced).
If it is not possible to comply with the given limits to rail overstresses track joints
have to be introduced (see Figure 6.3). This measure, however, is to be avoided when-
ever possible due to the maintenance problems which track joints entail.
The importance of the effects of track–structure interaction will depend on the
structural configuration (expansion length, presence of joints in the deck), the
188 6 Longitudinal Track–Structure Interaction
mechanical properties of the deck (vertical flexibility, rotation at the ends, distance
between the centroid and the top section, distance between centroid and bearings,
longitudinal stiffness of the bearing + pier system) as well as on the type of track
(ballasted or ballastless track), the presence of rail expansion devices, and on the
properties of the track (axial stiffness of the rails, resistance of the rails and sleepers
to longitudinal movements within the ballast, for ballasted tracks, or resistance of
the rail fastenings, for unballasted tracks).
Pier
Figure 6.4 Possible model for the analysis of track–structure interaction (as suggested in
Figure 3.3 of reference [2]).
6.3 Model for Analysis 189
6.3.1.1 Rails
The rails are connected to the top of the deck by a rigid element in the vertical direc-
tion and by non-linear springs in the horizontal direction. These springs model the
response of the sleeper and ballast to horizontal loads (or the force-slip behaviour
for unballasted tracks). A very important feature of the model is that a significant
stretch of track outside the bridge needs also to be modelled in order to obtain real-
istic boundary conditions (fixity between track and ground outside the structure).
This length is recommended to be at least equal to 200 m, and, in any case, larger
than the breathing length. Additionally, [3] recommends that the distance between
nodes be limited to a maximum of 2.00 m.
The rails should be modelled as a beam element. All the rails on the structure’s
deck can be grouped into a single element having the sum of the area and inertia of
all the actual rails (except in the case of some exceptional geometries, as it will be
shown in Sections 6.7.6.2 and 6.7.6.3).
6.3.1.2 Deck
The deck is also modelled using beam elements with the area and inertia of the deck.
The top of the deck is connected by means of stiff vertical members to the centroid
of the deck. This allows to simulate the effects of rotations of the deck on the top of
deck, in terms of strains.
When modelling a series of simply supported precast spans with a continuous top
slab over several spans, a double-hinged element should be placed between spans
to model the longitudinal continuity of the deck at the level of the top slab, while
allowing deck rotations.
6.3.1.4 Bearings
At each support (abutments or piers) the centroid of the deck is connected to the
bottom of the deck where the structural bearings are. This point is connected to the
190 6 Longitudinal Track–Structure Interaction
70
Loaded
60
50 Ballasted
k (kN/m)
40
Unloaded
30
20
10 Without
ballast
0
0 2 4 6 8 10 12
u (mm)
Unloaded Loaded
top of the abutment or pier by a vertical and a horizontal spring, which are meant
to model the stiffness of the deck–support connection. These can be linear springs if
the supports are neoprene bearings or non-linear springs if the supports are sliding
bearings. They can also be rigid if there is a monolithic connection
6.3.1.5 Columns
The columns are modelled as beam elements with the stiffness provided by the pier
cross sections. In normal practice, the uncracked pier stiffness is considered when
modelling the track–structure infrastructure.
6.3.1.6 Foundations
The model can be further complicated, by considering the rotational and transla-
tional flexibility of the foundations. This can be done by adding linear or non-linear
springs at the level of the foundations. However, given the highly non-linear
6.4 Actions 191
behaviour and the large uncertainties involved in the characterisation of the soil,
such refinements may be more of a theoretical exercise than a reflection of the
structure’s actual behaviour and will not necessarily lead to greater reliability.
6.4 Actions
6.4.1 Temperature Variations
It is necessary to consider the two components of thermal variation:
– Uniform variation. The variation of temperature in the rail can be taken as ±50 ∘ C.
The uniform variation of temperature of the deck can be determined according to
EN 1991-1-5 [6]. EN 1991-2 allows to consider a simplified value of ±35 ∘ C
– Temperature gradient in the deck, which can also be determined according to EN
1991-1-5. This gradient will induce rotations which will produce displacements at
the level of the top deck surface.
The value of k to be used to estimate the effects of temperature should be that of
the unloaded track (ku ), since the rate of application of this load will be much slower
than the transit of a train through the bridge.
length of 300 m. For SW/2 loading, the value is increased to 35 kN/m. However, in
this case the loaded length is limited to 57 m only.
Regarding traction, and this time for all load models, including SW/2, the force
is defined as 33 kN/m times the length of application, but not greater than 1000 kN,
which effectively limits the length of application to 30 m.
In this case, the value of k should depend on whether or not a given zone of the
structure is considered to be loaded.
100
90
Equivalent temperature variation (°C)
80
70 45 °C
60
50
40 92 °C
30
47 °C
20
2 Years
10
0
1 10 100 1 000 10 000 100 000
Time (days)
Figure 6.6 Example of creep and shrinkage strain occurring before the installation of the
rail at t = 2 years.
Figure 6.6 shows an example. A deck of concrete class C35/45 with a notional
depth equal to 500 mm, in an environment with a relative humidity of 50% with a
mean axial compression due to prestressing equal to 6 MPa, would be subjected to an
equivalent temperature load of 90 ∘ C. If the creep and shrinkage happening in the
first two years is discounted, the equivalent temperature variation drops to 45 ∘ C,
that is, half the total value.
Additionally, in ballasted tracks, track maintenance operations to keep the track
layout constant involve local lifting of the track. Such operations liberate a part of
the stresses due to creep and shrinkage, so that considering the total value of creep
and shrinkage deformations may be over-conservative, even after discounting the
deformation occurring before fixing of the rails. However for an unballasted track,
no further reductions are possible.
For shrinkage and creep, the value of k to be considered is that of the unloaded
track (ku ) since, as happens with temperature variations, these actions develop
slowly with time and are not affected by the momentary increase in resistance
afforded by a passing train.
Envelope Envelope
Hypothesis Hypothesis (Alternate spans) (Positive + Negative)
Temperature Temp decrease + Traction
increase + negative positive gradient + Vertical and
gradient shrinkage + creep loads braking
Load 1 2 3 4
Comb 1 X X
Comb 2 X X
Comb 3 X X X
Comb 4 X X X
6.5 Verifications
6.5.1 Verifications in Terms of Stresses
EN 1991-2:2003 provides limits to the rail overstresses which are valid when the
following conditions are fulfilled:
In a bridge without a track joint the maximum variation of stresses in the rails
should be limited to 72 MPa in compression and 92 MPa in tension. These are over-
stresses, since the continuous welded rail (CWB) would be subjected to 105 MPa (see
Section 6.2) without the presence of a bridge.
If a track joint is present, the verification should be made in terms of total stresses
in the rail. In this case, the limits would be 72 + 105 = 177 MPa in compression and
92 + 105 = 197 MPa in tension.
– Less than 5 mm if the rail is a continuous welded rail or if there is a track expansion
device at one end of the deck
– Less than 30 mm if there are track expansion devices at both ends of the deck
Due to vertical loads only, the horizontal displacement (𝛿 H ) (see Figure 6.8) of the
deck at the end of the deck should be limited to:
– 8 mm if the deck rotation is determined accounting for the stiffness of the track
– 10 mm if the deck rotation is determined accounting only for the stiffness of the
deck.
This verification will be influenced by the typology of the bearing, since rotations
will produce larger displacements at the top of the deck if the bearings are fixed than
if the bearings are free to move longitudinally, as shown in Figure 6.8.
δB
δf
δi
δB δi
δf
δi
δB = ∣δf – δi∣
δB = ∣δf – δi∣
(a)
(b)
Figure 6.7 (a) Displacement between decks at expansion joint and (b) between deck and
abutment at bridge end.
196 6 Longitudinal Track–Structure Interaction
δH δH
(a) (b)
Figure 6.8 Horizontal displacement of deck due to vertical loads (a) fixed bearing;
(b) sliding bearing.
δv
δv
(a) (b)
Figure 6.9 (a) Vertical relative displacement between two deck or (b) between deck and
abutment due to longitudinal displacement coupled with longitudinal slope.
Due to vertical loads and temperature variations, the relative vertical displacement
between the end of two decks at an intermediate deck joint, or between the end span
and the abutment (see Figure 6.9) should be limited to:
– 3 mm if the maximum line speed at the site is less than or equal to 160 km/h
– 2 mm if the maximum line speed at the site is more than 160 km/h.
This discontinuity, in this case, is caused by the longitudinal slope of the deck, and
will, therefore, be sensitive to this variable.
Stock rail +
L
Switch
rail
Bridge
Stock rail
joint
72,03
5,3
14,56 Y
14.82
R200
R70
R200
R1
54
R90
6 1
3 R
R21 R2
R11
R1
1:20,17 R35 R35
1:20,17
R15
53,07
R3
R3
1:2, 75
75 1:2,
R19 R1
9
134,07
X X
1
28
R1
9
19
R
58,65
1:4
56
1:17
R4 44
+Y R4
20
20
3 R3
R +X
Y
55 2,67
140
switch rails of the rail expansion joint are fixed by means of specially designed base
plates and rail-fastening components.
When the bridge has long expansion section lengths and thus large longitudinal
movements, it will be necessary at bridge joints to either reinforce the rail or provide
additional supports to the rail. It would not be necessary for REJs with a maximum
expansion length of 300 mm, but it will be for 600, 1200 mm, or longer.
The most common solution is the second option, i.e. maintaining the standard
rail section but providing additional supports to the rail in the bridge joint. Thus,
the decking inside the bridge gap is supported by steel moveable sleepers arranged
in accordance with the gap width (one for a maximum expansion length of 600 mm,
two for 1200, and three for 1800). These sleepers are mounted on auxiliary beams
aside from the track which act as auxiliary bridges. A crossbar control lever mech-
anism (see Figure 6.13) is used to adjust the position of these sleepers according to
the bridge gap movement so that the running rails are deflected comparable to the
other parts of the track.
With this mechanism the gap-opening variation is distributed over the moveable
sleeper spacings (between the last two sleepers at both sides of the joint and the steel
moveable sleepers) to ensure that:
The maximum spacing between sleepers is compatible with rail, fastening, and
the railway line characteristics: see Figure 6.14 with an example for a 1200 mm REJ,
in which maximum spacing remains below or equal to 650 mm, in the case of max-
imum opening of the lever mechanism of 1950 mm.
The minimum opening of the joint is limited: in the example of Figure 6.15 it is
higher than or equal to 190 mm in the case of minimum opening of the lever mech-
anism of 750 mm.
The maximum difference between one gap width between moveable sleepers and
the average gap is limited: in the example of this mechanism assures that the gap
with will vary between 650 mm (for the maximum opening of the mechanism of
1950 mm) and 250 mm (for the minimum opening of 750 mm).
On the other hand, the design of the bridge must assure that limit values for
maximum angular rotation and maximum vertical relative displacement between
adjacent structural elements (deck and abutment or adjacent decks) at joints are
Steel
moveable Longitudinal
sleepers girder
Ilanta
Structural Lever
joint mechanism
(scissors)
Figure 6.13 Lever mechanism for moveable steel sleepers (Source: [12]).
200 6 Longitudinal Track–Structure Interaction
1390 mm 190 mm
(a) (b)
Figure 6.15 Example of results of computations of uplift forces on fasteners (a) and
bending moments on rails (b) due to angular rotations and relative vertical displacements
of the decks.
fulfilled, as explained in Section 6.5. These verifications will assure that traffic safety
and comfort cannot be endangered by the creation of unacceptable changes in
vertical track geometry. However, in the presence of REJs, and specially with direct
fastened track, additional verifications must be done. Thus, it will be necessary
to verify: (i) that the REJ fasteners are not subject to unacceptable uplift forces;
(ii) that stress on rails and longitudinal girders of the REJ is acceptable.
The German Requirements Catalogue for the construction of slab tracks states that
the maximum tensile force on fastenings on both sides of the joint shall not be greater
than the force that decompresses the rail-pad, which depends on the fastening and
can be determined with normalised test of clamping force.
UIC Leaflet 776-3 states that in the case of direct fastened track on both sides of the
joint, angular rotation and vertical relative displacement must be limited to assure
that rail stresses are limited to 80 MPa, taking into account actual values of vertical
tensile stiffness of fasteners.
Some computations have been made that state that for usual values of vertical
tensile stiffness and force that decompresses the rail-pad, the limits established for
angular rotation and vertical relative displacement in the Eurocodes and UCI leaflets
will assure the fulfilment of these to verifications [13], but specific computations
must be made in every specific case.
6.6 Rail Expansion Joints 201
6.6.2 Regulation
Although the REJs are usually delivered at their neutral position, they shall be regu-
lated, at the moment of its welding, and installed at a position that takes into account
the expected remaining bridge movement (due to creep and shrinkage, forecasted
temperature variation, and breaking and traction traffic loads) and rail movements
(due to foreseen variations of rail temperature with regard to its neutral temperature,
as REJs shall be welded to the standard track at neutral temperature of the rail), and
breaking and traction traffic loads.
Although the devices can be adjusted if necessary, after being installed on the
track, in practice it is a very laborious process; therefore, it is highly recommended
to carry out the regulation before.
To do this, the REJ is first placed on a sliding surface (the ideal is to deposit it on
the rails of a well-levelled track or on the embedded rails of the turnout pre-assembly
slab installed in fabrication workshop or the mounting base).
In order to do this regulation of the REJs it will be necessary to measure the gap and
ambient temperature. It is advisable to make several measurements at different tem-
peratures and to do it few days before the regulation of the device (see Figure 6.16).
Based on these measures, it is possible to foresee significantly what will be the
dimension of the joint on the day of assembly at the scheduled time, especially if
there are very recent measures.
The state and position of the rail expansion device shall be recorded at installation
compared to a fixed reference point (FRP), by adjusting a pointing device that will
indicate the longitudinal movements of the bridge from the moment of installation.
6.6.3 Installation
They are installed on the side of the moving end of the structure. The fixed part of
the REJ must be on the fixed part of the viaduct (abutment – embankment), and the
movable part of the REJ on the superstructure of the viaduct itself (deck).
Once the position in which the element is to be placed has been defined, the guard
element is first laid out, and then the ballast is spread in a uniform layer.
This guard element is the necessary system for interrupting the ballast bed
(Figure 6.17a) to prevent the ballast from being deconsolidated by the movements
of the structure (which can occur if the ballast is continuous over the structural
joint) and to prevent it from falling through the joint opening. At the outer edges of
the wall there are wing walls to prevent the ballast from spilling over to the sides
in this area. The top of the wall must reproduce the final cant that the track will
have. An elastic sheet is usually placed between the two facing walls to catch loose
stones and other objects that may fall from the track for any reason. This sheet can
be folded and unfolded following the movements of the deck (Figure 6.17b).
Once the ballast bed has been prepared, the expansion device shall be positioned in
accordance with the assembly drawings and then moved into its theoretical position.
The flatness of the bearing assembly shall be maintained at all times.
202 6 Longitudinal Track–Structure Interaction
Se = 790 − a
a (Tambient)
a
aminimum = 190 mm aneutral = 790 mm amaximum = 1.390 mm
Abutment Deck
a ≥ 790 mm Se ≤ 0
End of switch rail Extended gap: the REJs must be
FRP open by moving the switch rail
Switch rail towards the fixed part
Stock rail
Abutment Deck
a ≤ 790 mm Se ≥ 0
End of switch rail
Reduced gap: the REJs must be
FRP
closed by moving the switch rail
Switch rail Stock rail towards the moveable part
(a) (b)
Figure 6.17 (a) Ballast guard walls; (b) elastic sheet preventing ballast stones falling
(Source: [12]).
6.7 Longitudinal Schemes 203
6.7.1 Continous Deck with a Single Fixed Point Located at One of the
Abutments
6.7.1.1 General
This first solution can be applied for small, medium-size or even large deck lengths.
However, depending on deck length, the installation of expansion devices at the
other abutment could be necessary.
Usually, depending on the specific characteristics of the bridge, for lengths lower
than 100 m it is possible to avoid expansion devices.
On the other hand, for long bridges which need the installation of REJs at the
mobile abutment (Figure 6.18), it must be taken into account that the maximum
displacement of the device will increase with the expansion length of the deck. A
204 6 Longitudinal Track–Structure Interaction
REJ
Fixed point
Figure 6.18 Continuous deck with a single fixed point located at one of the abutments.
6.7.1.2 Examples
Example 1. Bridge needing REJ
The example corresponds to a ballasted double-track continuous high-speed railway
bridge with 7 spans, measuring a total length of 165 m (Figure 6.19).
The deck consists of a voided concrete slab with a constant height of 1.90 m and a
width of 14 m.
The connection between the girder and each pier is performed using pot bearings
sliding in longitudinal direction. One of the abutments of the viaduct (the one on the
right in Figure 6.20) is firmly connected to the deck and is the only fixed point of the
viaduct; on the other abutment the bearings are also free in longitudinal direction.
Figure 6.20 shows the stresses in rail computed for a seasonal variation of temper-
ature of plus 35 ∘ C in the deck. As can be seen, only under the action of the variation
of temperature, the maximum compression in rail here is greater than the allowed
limit of 72 MPa for bridges with ballasted track.
Thus, it will be necessary to install an expansion device in the track to limit this
rail stress (Figure 6.21).
Once decided to install the REJ, it will be necessary to make the analysis including
all the actions to be considered.
The combinations used in this case consider, in a first step, seasonal variation of
temperatures combined with equivalent shrinkage and creep long-term deforma-
tion, produced after the installation of the CWR [13], which is equivalent to an
additional decrease of temperature of 35 ∘ C.
In a second step stresses for the vertical traffic forces are computed, considering
trains on both tracks, and for horizontal braking and traction forces, considering
several situations of trains breaking or starting.
165 m
20 m 25 m 25 m 20 m
4.70
1.90
14.00
Figure 6.19 Longitudinal scheme and deck cross section of bridge for example 1.
6.7 Longitudinal Schemes 205
100
80
60
Stresses in the rail
40
20
0
–20
–40
–60
–80
–100
Expansion
165 m
device
20 m 25 m 25 m 20 m
Figure 6.21 Longitudinal scheme of example 1 with REJ at the mobile abutment.
Envelop
Tension limit = 105.0 MPa
Compression limit = –122.3 MPa Rail stresses (MPa)
250
Tension limit
200
150
100
50
0
–50
–100
–150
–200 Compression limit
Then the envelope of rail stresses for all the actions are computed. As can be seen
in Figures 6.22 and 6.23, those stresses are lower than the limit values (adding to the
limit values the rail stress due to rail temperature variations).
Also, the verifications of limit values for displacements are fulfilled.
206 6 Longitudinal Track–Structure Interaction
And finally, to design the required expansion device its maximum displacements
are computed. In this case a maximum displacement of 239 mm is computed, so an
expansion device with an allowable displacement of at least 300 mm is required.
Example 2. Bridge not needing REJ
The usual one for this type of viaducts, with a fixed point in one of the abutments,
implemented by means of anchoring deck to abutment.
The example corresponds to a ballasted bridge of the high-speed line
Madrid–Barcelona. It is a continuous viaduct with four spans and an overall
length of 110 m, with independent decks for each track. The bearings system is the
same that for Example 1, using pot bearings sliding in longitudinal direction, and
with a connection between the deck and one of the abutments (the one on the right
in Figure 6.24) which is the only fixed point of the viaduct; on the other abutment
the bearings are also free in longitudinal direction.
First a computation has been executed considering continuous rail at the two
ends of the bridge, with a combination of variation of temperature of deck and rail,
110
20 35 35 20
E-1 E-2
P1 P2 P3
deformation due to shrinkage and creep and traffic loads for different positions of a
train. Figure 6.25 shows the obtained envelopes of stresses in the rail. As can be seen
on the free abutment the limit of compression is slightly exceeded (72 MPa plus the
stress due to an increase of rail temperature of 50 ∘ C).
The situation which produces this maximum of compressions is due to the com-
bination of the following actions: decrease of temperature of deck equal to −58 ∘ C,
which is the sum of the seasonal maximum decrease of −35 ∘ C and the equivalent
temperature for the deformation due to shrinkage and creep of −23 ∘ C, and the train
located with its head on abutment number two and braking.
Considering that the limit of stresses is only slightly overpassed, but that the sim-
plified combination of actions used is usually conservative in terms of rail stresses,
for the same situation a more accurate step-by-step computation has been carried
out. To perform this computation, the first thing to do is to define a law for the lon-
gitudinal behaviour of the fastening–ballast system as a function of the vertical load,
as shown in Figure 6.26.
The maximum horizontal force F N will increase linearly with the vertical load on
the point, starting from a minimum value F 0 for a null vertical load, up to the max-
imum value of 60 kN/m according to standards for a maximum vertical load (that
can be considered as 80 kN/m per meter of distributed load according to the model
of UIC 71). The expression of F N will depend on the level of ballast maintenance as
shown in the figure.
Once these laws have been defined for non-linear springs the step-by-step compu-
tation proper can begin, where the first step shall be computing the situation after
deformations imposed in rail and deck.
After the first step the non-linear springs will be in different situations, as shown
in Figure 6.27, depending on the relative displacement that has occurred between
rail and infrastructure.
208 6 Longitudinal Track–Structure Interaction
t
N = 80 kN/m
60 kN/m
Increasing N
FN
N=0
F0
2 mm u
Increase rail temperature (50 °C). Increase deck temperature (30 °C)
–60
–80
–100
–120
–140
Rail stresses (MPa)
–160
ΔT rail = 50 °C
ΔT in deck = 30 °C
N=0
F0
2 mm u
Figure 6.27 First step of step-by-step computation (strains imposed on rail and deck).
From the situation after step 1, the vertical loads are introduced step by step to
represent the train running along the viaduct. For each one of these loading steps the
situation of non-linear springs will change depending on the vertical load and the
relative displacement between rail and infrastructure, according to the previously
6.7 Longitudinal Schemes 209
t
N2
F2 = FN2
N1
FN1
F1
F0 N=0
Δu1 Δ u2
2 mm u
Braking
t
N2
FN2
Δu2
N1
FN1
Δu1
F0 N=0
2 mm u
defined laws, with the corresponding increases of force and variations of stresses
in rail. For the last computation step the braking forces are entered at the desired
position. The new relative displacements will then imply a new state of non-linear
springs, with increase of force or not, depending on the initial situation (Figures 6.28
and 6.29).
210 6 Longitudinal Track–Structure Interaction
–5
Vertical
load
Figure 6.30 Additional rail stresses due to vertical traffic load for example 2, computed
with simplified combination and with step-by-step methods.
20
Braking
Stresses in rail (MPa)
Step-by-step Method
10
Simplified Method
–10
Braking
–20
Figure 6.31 Additional rail stresses due to braking load for example 2, computed with
simplified combination and with step-by-step methods.
6.7 Longitudinal Schemes 211
–60
Combination
–80
–100
Stresses in rail (MPa) Step-by-Step Method
–120 Simplified Method
Compression Limit
–140
–160
–180
–200
Braking
ΔT in deck
= –58 °C
Vertical
load
Figure 6.32 Envelope of rail stresses for example 2, computed with simplified
combination and with step-by-step methods.
6.7.2 Continous Deck with the Fixed Point Located on One of the
Central Piers
6.7.2.1 General
In this case, deck shall be free from longitudinal restraint at abutments and at the
rest of the pier.
Again, depending on deck lengths, the installation of expansion devices at both
abutments can be needed. This solution also implies transmission of the whole lon-
gitudinal braking/traction action to one pier, without help from abutments.
Fixed bearing
Sliding bearing
Free bearing
Figure 6.33 Usual bearing system for simply supported span in low-seismicity areas.
958 m
46 m 46 m 46 m 46 m
A–1 A–2
P1 P2 P18 P19
Figure 6.34 Longitudinal scheme and deck cross section of bridge for example.
6.7.3.2 Example
The example corresponds to a ballasted double track simply supported high-speed
railway bridge with 20 spans, and an overall length of 958 m (Figure 6.34).
The deck consists of a HP-40 prestressed concrete box with a constant height
of 3.80 m. Each span has a fixed neoprene support at one end and a sliding
neoprene–Teflon support at the other. The fixed support is the closest to abutment 1.
Thus there is a sliding bearing support at abutment 2.
The following figures show the stresses in rail computed for (from top to bottom):
(i) a seasonal increase of temperature; (ii) a seasonal decrease of temperature and
associated equivalent decrease due to creep and shrinkage; (iii) vertical traffic load
for several positions of the train; (iv) braking and acceleration forces for the same
positions of the train. Finally, the envelope of rail stresses is shown in the lowest
figure. As can be seen, limit values are not reached and, consequently, installation
of REJs is not needed.
Also, the verifications of limit values for displacements are fulfilled (Figure 6.35).
6.7.4 Fixed Points at the Two Abutments and a Structural Joint in the
Middle
6.7.4.1 General
This solution can be applied for long bridges when the following conditions exist:
– It is not possible the solution described in Section 6.7.1: with the fixed point at
one of the abutments and the structural joint at the other, the expansion length of
6.7 Longitudinal Schemes 213
P1 P2 P18 P19
the bridge would be too high and, then, the expected movements on the structural
joint would be too large to be accommodated by a REJ.
– It is not possible the solution described in Section 6.7.2: the characteristics of the
foundations or the geometry (important height) of the central piers make impos-
sible to install the fixed point in the middle of the bridge.
214 6 Longitudinal Track–Structure Interaction
The alternative solution in these situations is the one presented in this section,
with fixed points at the two abutments and two continuous stretches of the deck,
with a structural joint, and then a REJ, in the middle of the bridge.
This is a very common solution in France, in which a neutral central portal frame
is used, combined with a double REJ (common in the French railways but not used
for other railway administrators.
6.7.4.2 Example
The example corresponds to a ballasted double-track high-speed bridge with 25
spans of lengths between 31.0 and 48.0 m, and an overall length of 1147.0 m. The
deck has two continuous stretches: section 1 between spans 1 and 14, with a length
of 685.0 m; section 2 between spans 16 and 25, with a length of 462.0 m. There are
structural joints at both ends of span 15, which becomes this way the neutral central
span (see Figure 6.36).
The deck consists on a HP-40 prestressed concrete box with a constant height
of 4.00 m. At both abutments there is a connection between deck and abutment
by means of prestressing steel. All bearings are sliding in longitudinal direction
neoprene–Teflon, except in the case of pier 14, corresponding to one of the ends of
neutral span 15 that are fixed in longitudinal direction.
A first computation of rail stresses is made without considering the presence of
rail expansion devices, for a seasonal variation of temperature of plus 36 ∘ C in the
E–1 E–2
P–14 P–15
Neutral span
P–14 P–15
deck. Figure 6.37 shows the stresses in rail computed. As can be seen, only under the
action of the variation of temperature, the maximum compression in rail is around
±240 MPa at the neutral span, much greater than the allowed limit of 72 MPa for
bridges with ballasted track. It is then clear the need of installing at least one rail
expansion device to limit this rail stress.
A second computation is made considering the presence of a REJ at pier 14 (mobile
end of section 1 and fixed point of neutral span). In this case it will be necessary to
make the analysis including all the actions to be considered.
The combinations used in this case consider: seasonal variation of temperatures;
equivalent shrinkage and creep long-term deformation produced after the installa-
tion of the CWR; vertical traffic forces considering trains on both tracks; horizon-
tal braking and traction forces, considering several situations of trains breaking or
starting.
Then the envelope of rail stresses for all the actions is computed. As can be seen in
Figure 6.38, those stresses are lower than the limit values (adding to the limit values
the rail stress due to rail temperature variations).
However, in this case the computed maximum displacement of REJ in this case
would be higher than 1200 mm. At the time of the design of this bridge REJs with an
allowed expansion length higher than 1200 mm were not available and this solution
was not possible (nowadays REJs for expansion length up to 1800 mm are used in
Chinese lines).
Computations were then repeated but considering a double REJ at the neutral
span, with rail joins at two ends of this span. This is actually the more usual design
for the track in the case of bridges with a neutral central span. As can be seen in
Figure 6.39, in this case again computed stresses are lower than the limit values.
In this case the maximum expansion computed for every of the two rail joints
is more reduced, but even though in the case of the joint on pier 15 is lower than
250
Rail stresses (MPa)
Tension limit
200
Tension
150 Compression
100
50
0
–50
–100
–150
–200 Compression limit
REJ
E–1 E–2
P–14 P–15
Double REJ
E–1 E–2
P–14 P–15
Figure 6.39 Envelope of rail stresses with a double REJ at neutral central span.
600 mm, in the case of pier 14 it is higher, and then it would be necessary to have a
device that allows a maximum expansion of 1200 mm (as explained in Section 6.6,
usual allowed maximum expansion length are 300, 600, and 1200 mm). However, a
600 mm device could be installed under special supervision and taking into account
that it would be necessary to carry out adjustment operations on it.
6.7.5.2 Example
The example corresponds to a double-track viaduct of a flying junction of a
high-speed line over a conventional existing railway line. To deal with the geometric
conflict between the position of the piers of the new viaduct and the tracks of the
conventional existing line, two different decks for every single track were necessary,
with lengths 630 m (left deck) and 608 m (right deck). The left viaduct consists of
21 spans, all of them 30 m long, whereas the right viaduct has 6 spans of 30 m,
10 spans of 28.5 m, one span of 23 m, and 4 spans of 30 m.
In both cases the deck consists in a HP-60 precast concrete girder with a depth of
2.0 m, and an in situ concrete slab HA-35 with a 0.25 m depth.
All the spans are simply supported. However, due to the geometrical difficulties
mentioned before, some of the piers do have not enough stiffness to become a fixed
point. Thus, the solution explained in Section 6.7.3 is not possible and it was nec-
essary to provide with axial continuity to some spans, by prestressing the joints, to
constitute continuous stretches in terms of longitudinal behaviour. The continuous
sections are:
Left deck:
Right deck:
In the case of spans without longitudinal continuity, there is a fixed point at one
end of each span. In these cases, the piers consist of a square-shaped pier shaft
founded on a four-pile cap.
In the cases of continuous sections, most of the supports consist on a portal frame
with two cylindrical piers, each of them founded on a single pile, under an HP-40
prestressed concrete beam with an inverted T section to support the decks. These
supports have not the required stiffness to become fixed points. The support for fixed
points is then constituted by a portal frame with two inclined piers on each side of
the existing conventional railway line, under the same HP-40 prestressed concrete
beam (see Figure 6.40).
Rail stresses are computed for the two viaducts without considering the presence
of rail expansion devices, for the following actions: seasonal variation of deck tem-
perature combined with equivalent shrinkage and creep long-term deformation pro-
duced after the installation of the CWR; vertical traffic forces considering trains on
both tracks; horizontal braking and traction forces, considering several situations of
trains breaking or starting. Results of the envelope of rail stresses for all the actions
are shown in Figures 6.41 and 6.42 for both viaducts.
As can be seen, limit values are not reached and, consequently, installation of REJs
is not needed.
218 6 Longitudinal Track–Structure Interaction
Pot
Continuity bearings
prestressing
Figure 6.40 Longitudinal scheme of continuous stretch for spans 11 to 13 of right deck,
with fixed point at pier 11.
–40
–120
–40
0.20
Greater seismic
0.00 forces
0 1 2 3 4
Period (s)
Figure 6.43 Influence of the requirements on stiffness for fixed points on seismic design.
considerations have an impact not only on the volume of material but may also rad-
ically influence structural design.
Track–structure interaction design considerations are stricter for seismic condi-
tions. To ensure proper viaduct structural behaviour against braking/traction hor-
izontal actions, limiting deck displacements and stresses on rail, fixed points must
be established with sufficient stiffness. The requirements on stiffness for fixed points
mean higher structure natural frequencies in longitudinal direction and thus greater
seismic factors than for railroad bridges. In the case of road bridges, the use of seismic
isolator, as elastomeric bearings, usually leads to natural frequencies on the lower
zone of the spectrum, whereas in the case of railway bridges natural frequencies will
usually be on the zone of higher ground accelerations. That, combined with heavier
decks due to requirements of in-service deformability, makes that seismic forces will
be much higher (Figure 6.43).
Therefore, seismic considerations may be most important, especially for long
viaducts and even restrict the type of viaduct.
The seismic design of a railway viaduct should be based on a criterion of
seismic stress balance over piers and abutments compatible with requirements
of track–structure interaction. Longitudinal action is the most difficult analysis,
especially for long viaducts, due to the coupling of three actions of different nature:
(i) the seismic action itself; (ii) strains imposed by temperature variations and
rheologic effects; (iii) braking/traction actions. Thus track–structure interaction
is a key requisite for selecting a structural type, especially as regards longitudinal
continuity of deck, location of fixed points to counteract temperature effects and
fixed points for resistance to stress and limitation of displacements due to braking
and traction. Main variables will thus be bridge dilatable length(s), and deformable
deck length, depending on longitudinal continuity of deck and fixed-point location,
where the difference between dilatable and deformable length is that fixed points
may not be the same for slow deformations (temperature, rheologic effects) or for
fast ones (braking/traction, seismic action) because of installed STUs or dampers.
220 6 Longitudinal Track–Structure Interaction
The three main options of design exposed in Section 6.7 must be considered:
– Continuous deck with fixed points that define dilatable and deformable length of
deck, where expansion devices may be eventually installed.
– Simply supported spans without longitudinal continuity, with fixed points on
each span, where interaction issues may be addressed without needing expansion
devices.
– Mixed solution, with the deck divided in several continuous stretches each of ones
including several spans and one fixed point.
However, for high-seismicity zones, deck continuity issues should consider seis-
mic actions and their coupling with braking/traction and imposed strains, and also
relative displacements of deck.
Continuous Deck A continuous deck implies a large mass fully excited by seismic
action (deck plus total viaduct dead weight). Longitudinally, there are four types of
resistance schemes to the seismic force associated with the vibration of this mass:
(a) The first solution is the one already exposed in Section 6.7.1, usually used for
low-seismicity areas or small deck lengths, with a single fixed point located at
one of the abutments. This is also usually the chosen with high piers that can’t
become fixed points. As already explained this solution implies, depending
on deck length, the installation of expansion devices at the other abutment.
The main drawback is that in high-seismicity areas the fixed abutment shall
support the whole seismic action.
This is usually the case when we have high piers that can’t become fixed points.
(b) A second solution would be similar to the one explained in Section 6.7.2, in
which fixed point is located on one of the central piers. In this case, deck shall
be free from longitudinal restraint at abutments, which would imply transmis-
sion of the whole longitudinal action (seismic plus braking/traction) to the pier
in which the fixed point is situated.
Due to the need of high stiffness of the fixed point, the central pier used to
establish it is usually an especial one, a very rigid A-type pier or similar, and
with a large foundation, as in the examples of Figure 6.44.
Figure 6.44 Examples of especial central pier for fixed point in Spanish HS railway.
6.7 Longitudinal Schemes 221
Figure 6.45 Continuous deck with several fixed points located at the central piers.
Figure 6.47 Continuous deck with several fixed points located at the central piers and
STUs on all or a few piers.
Elastic bearings
Shock Transmitter Units
for re-centering
It is the case of the viaduct showed in Figure 6.49, in which dampers were
installed on both abutments.
(e) Finally, the deck can be segmented into segments longer that one single span,
which is an intermediate situation between continuous deck and simply sup-
ported. This solution can result in different variants, dampers only on abut-
ments, on abutments and piers, always associated with an adequate bearing
system (to deal with braking/starting forces).
It is the case of the viaduct showed in Figure 6.50, in which dampers were installed
on both abutments.
Discontinuous Deck The most common situation for discontinuous decks is the case
of simply supported spans without longitudinal continuity in which fixed points
with a sufficient stiffness will be necessary on each span.
Usually, in low-seismicity areas, a four-bearing system is installed, longitudinally
fixing the deck to pier with restricted bearings while the other end is free with sliding
bearings, as shown in Figure 6.33.
However, for high-seismicity areas this bearing system has some disadvantages.
In the first place, the presence of different bearings restricting the same deck
displacement, as here, with a fixed bearing and a sliding bearing in transversal
direction, because they have different tolerances, may produce a longitudinal load
imbalance that can be critical in the design of the bearings, especially against
seismic forces [14, 15] (Figure 6.51).
224 6 Longitudinal Track–Structure Interaction
k (stiffness)
d
t (tolerance)
Fixed bearing
Figure 6.53 Bearing system for simply supported span in high-seismicity areas.
This is the case of several HS lines, as can be seen in Table 6.3, as for instance
the Taiwan HSL, in which a Type I earthquake was considered for a return period
of 950 years and PGA up to 0.4 g, and a Type II service earthquake with a third
of the previous PGA, or the California HSL, in which a Maximum considered
earthquake is computed for a return period of 950 years, and an Operating Basis
Earthquake for a return period of only 50 years. And similar values for France and
Morocco.
Track–bridge interaction analysis under Service seismic action must be performed
to confirm that relative displacement is under the limits. However, as explained
in Section 6.3.1.3, the non-linear nature of the rail–structure interaction makes
time-history analysis unavoidable, as modal-spectral analysis is not possible. Addi-
tionally, longitudinal law is also dependent on the track being loaded or unloaded,
which further complicates the computation when combining seismic and traffic
actions (Figure 6.54).
In any case, synthetic displacement or acceleration time histories must be gener-
ated and applied at the base of the pier foundations for interaction analysis.
This problem can be solved by using STU, to provide the deck with structural conti-
nuity against seismic actions, to avoid relative displacements between adjacent decks
during earthquakes (Figure 6.55).
USA
Country Taiwan (California) France Morocco
Ultimate earthquake RP = 950 years RP = 950 years RP = 475 years 1000 years
(type I) PGA = 0,4 g
Reduced earthquake RP = Undefined RP = 50 years RP = 20 to 50 Undefined
(type II) PGA = 1/3 years (0.582 ultimate
ultimate acceleration)
acceleration
226 6 Longitudinal Track–Structure Interaction
t
N = 80 kN/m
60 kN/m
increasing N
FN = F0 + μN
FN
N=0
F0
2 mm u
Fixed bearing
Figure 6.55 Improved bearing system for simply supported span in high-seismicity areas
with STUs.
Linear springs
Section type 1
Section type 2
Section type 1
Section type 1
Section type 2
Section type 1
in both directions Axis of track
for neoprenes Section of rails
Section type 5
Section type 6
Section type 5
Section
type 3
Section
type 4
Section
type 3
slab of variable thickness between 0.25 and 0.39 m. All the beams have laminated
elastomeric bearings of variable dimensions (Figure 6.56).
This is a clear case in which the use of a conventional 2D models is not suitable,
and a model as the one shown in Figure 6.57 is needed, to consider both longitudinal
and transverse deck deformations.
A simplification is made assuming that the two rails of each track can be consid-
ered as concentrated on the track axis has been made, as shown in the figure. In
this scheme it is observed that only the pergola part covered by an upper slab with
continuity throughout the width of the pergola has been idealised (see Figure 6.57),
which is a sufficient approximation to the deformation behaviour of the deck.
On the other hand, the interaction analysis takes into account the vertical loads in
order to assess the influence of deck bending on rail stresses. However, in this case,
the main direction of bending is the same as the beams and is almost at right angle
with track axis and therefore this effect can be neglected. Thus, the computation
model can be greatly simplified as a plane model can be sufficient. On one hand
strains and displacements can be considered on the structure plane and furthermore
all elements can be considered on the same plane without considering the different
228 6 Longitudinal Track–Structure Interaction
Figure 6.58 Case example: scheme of position of braking and acceleration forces on the
tracks.
heights of bearings, axis of deck, deck upper deck and track, as is usually done to
take into account effects of bending.
Another special consideration is related to longitudinal traffic load. The different
considered situation must differentiate the position of the train in the two tracks, as
shown in (Figure 6.58) for one of the considered cases.
Finally, as in the previous examples, seasonal variation of deck temperature in
both directions, longitudinal and transversal, is combined with equivalent shrinkage
and creep long-term deformation produced after the installation of the CWR. As
creep deformation is different in each direction, that leads to different final values
considered of deck temperature decrease. Thus, temperatures variation considered
are:
Results of the envelope of rail stresses for the two tracks considering all the actions
are shown in Figure 6.59.
As can be seen, limit values are not reached and, consequently, installation of REJs
is not needed.
6.8 Example of Track–Structure Interaction 229
100
Variation of stress in rail (MPa)
50
Free end
(no joint)
Fixed end
Bridge
0
–250 –150 –50 50 150 250 350 450 550
–50
–100
–150
x (m)
Stresses in rail Limit in compression Limit in tension Bridge position
Figure 6.60 Increase of stresses in rail due to temperature increase in the deck only.
225
175
(joint)
75
Fixed end
25 Bridge
–250 –150 –50 –25 50 150 250 350
–75
–125
–175
–225
x (m)
Stresses in rail Limit in compression Limit in tension Bridge position
Figure 6.61 Total stresses in rail due to increase in temperature in deck and rail plus
positive temperature gradient plus creep and shrinkage.
Figures 6.66, 6.67, and 6.68 represent the effect of the braking and traction
forces. They differ in the position of the trains, which are placed from abutment 1
towards abutment 2 in the first figure, centred on the bridge in the second, and
from abutment 2 towards the bridge in the last one. Differences are small but the
maximum stresses occur at abutment 1 when horizontal forces are acting from
abutment 1 towards the bridge. The maximum stresses are about ±11 MPa.
Finally, Figure 6.69 shows the envelope of rail stresses. Also represented separately
are the envelopes due to imposed strains (uniform temperature, temperature gradi-
ent, creep, and shrinkage), the vertical train load, and the braking and traction forces.
6.8 Example of Track–Structure Interaction 231
225
175
(joint)
75
Fixed end
25 Bridge
–250 –150 –50 –25 50 150 250 350
–75
–125
–175
–225
x (m)
Stresses in rail Limit in compression Limit in tension Bridge position
Figure 6.62 Total stresses in rail due to decrease in temperature in deck and rail plus
negative temperature gradient.
50
40
30 Free end
(joint)
20
Stress in rail (MPa)
Fixed end 10
Bridge
0
–250 –150 –50 50 150 250 350
–10
–20
–30
–40
–50
x (m)
Stresses in rail Bridge position
Figure 6.63 Total stresses in rail due to train load placed on odd spans.
In this case it is clear that the effect of imposed strains is the dominating factor. Also,
the maximum stresses comply easily with the set limits for total rail stresses, indi-
cating that a significantly longer bridge can be designed using a single track joint,
placed at abutment 2.
50
40
30 Free end
(no joint)
20
Stress in rail (MPa)
Fixed end 10
Bridge
0
–250 –150 –50 50 150 250 350
–10
–20
–30
–40
–50
x (m)
Stresses in rail Bridge position
Figure 6.64 Total stresses in rail due to train load placed on even spans.
50
40
30 Free end
(joint)
20
Stress in rail (MPa)
Fixed end 10
Bridge
0
–250 –150 –50 50 150 250 350
–10
–20
–30
–40
–50
x (m)
Stresses in rail Bridge position
Figure 6.65 Total stresses in rail due to train load placed on all spans.
20
Free end
10 (joint)
Stress in rail (MPa)
–10
Stresses in rail - Forces towards Ab. 2
Stresses in rail - Forces towards Ab. 1
Bridge position
–20
x (m)
Figure 6.66 Total stresses in rail due to braking and traction forces acting from
Abutment 1 towards Abutment 2.
20
Free end
10 (joint)
Stress in rail (MPa)
–20
x (m)
Figure 6.67 Total stresses in rail due to braking and traction forces acting centred on the
bridge.
Table 6.4 Displacements of the top of the deck due to train live load (LL) at abutment 2.
20
Free end
10 (joint)
Stress in rail (MPa)
–20
x (m)
Figure 6.68 Total stresses in rail due to braking and traction forces acting Abutment 1
towards Abutment 2.
225
175
75
Fixed end
25 Bridge
–250 –150 –50 –25 50 150 250 350
–75
–125
–175
–225
x (m)
the deck section from the top is 1.85 m. The maximum value of the horizontal dis-
placement is 2.73 mm, which complies with maximum limit of 8 mm for a model
accounting for the stiffness of the track (see Section 6.5.2).
Table 6.5 Calculation of vertical displacement between deck edge and abutment 2.
Displacement Rotation 𝜹H 𝜹V
Load comb (mm) (mrad) (mm) (mm)
of this vertical relative displacement are determined in Table 6.5, accounting for the
longitudinal gradient at abutment 2 of 1.5%. The maximum value of the relative ver-
tical displacement is 1.70 mm, which is smaller than the limit of 2.00 mm for a design
speed larger than 160 km/h.
References
1 Calçaza, R., Delgado, R., Campos e Matos, A. et al. (2008). Track-bridge Interac-
tion on High-Speed Railways. Boca Raton: Taylor & Francis.
2 Granell, I., Arrieta, J.M., Celemin, J., et al. (2009). Ejemplos de Aplicación de la
IAPF-07. ACHE Monografía M-15.
3 Union Internationale des Chemins de Fer (2001). UIC Leaflet 774-3R.
Track-bridge interaction. Recommendations for calculations.
4 European Committee for Standardization (CEN) (2003). Eurocode 1 EN 1991-2.
Actions on structures – Part 2: Traffic loads on Bridges.
5 Union Internationale des Chemins de Fer (1979). UIC Leaflet 776-1R. Loads to
be considered in railway bridge design.
6 European Committee for Standardization (CEN) (2003). Eurocode 1 EN 1991-1-5.
Actions on structures – Part 1-5: General actions – Thermal actions.
7 Cuadrado Sanguino, M. and González Requejo, P. (2004). Consideración de las
deformaciones por retracción y fluencia en el estudio de fenómeno de interac-
ción vía-tablero en el proyecto de puentes ferroviarios. Revista de Obras Públicas
3446: 45–51.
8 Granel, I., Arrieta, J., Celemin, J. et al. (2009). Interacción Vía-Estructura en
puentes ferroviarios. Algoritmos de cálculo paso a paso. Revista de Obras Públicas
3499 (156): 39–48.
9 Cuadrado, M., González, P., and Losa, D. (2010). New considerations on
track-structure interaction in railway bridges. In: IABMAS2010: The Fifth
International Conference on Bridge Maintenance, Safety and Management,
Philadelphia.
236 6 Longitudinal Track–Structure Interaction
7.1 Introduction
This section outlines some considerations that must be taken into account in the
structural design of the bridge in order to improve its durability and reduce mainte-
nance work during its service life.
Designers of such bridges should be acquainted of the most updated versions of
the catalogue of damages used by inspectors, in order to elaborate designs oriented to
avoid the defaults and deteriorations identified by experience. Furthermore, design-
ers must keep in mind the need of providing access to important elements, such
as anchor heads, bearing devices, or expansion joints. In particular, design must
take into account sufficient overall robustness in order not to depend on elements
or details that cannot be sufficiently inspected or monitored, such as hangers and
prestressing tendons, that could lead to brittle failure.
From a maintenance point of view, the optimal structures are those that are inte-
gral, i.e. where there are no support devices or expansion joints. However, except
for very short structures, it is not possible to build completely monolithic bridges.
Nevertheless, the trend in some countries, e.g. Germany [1], is to build semi-integral
bridges that concentrate in a few piers or abutments the points where the expansion
joints and bearings are located. It is also very important not to forget that for the gen-
eral maintenance of the construction site it is best to have the minimum number of
rail joints and this is only possible by having some expansion joints in the structure
every 120 m or so.
Once the design has been conceived in general terms, the following aspects must
be taken into account:
– Accesses to the various elements of the deck.
– Possibility to inspect and change bearings, expansion joints, or other structural
components especially susceptible to corrosion attack.
– Drainage system.
– Systems to avoid eddy currents.
All these aspects are developed in Sections 7.1–7.5.
7.2 Accesses
One of the most important aspects when considering the maintenance strategy is
to provide access to the various structural elements of the bridge. In principle, areas
hidden from inspection should be avoided and, therefore, the interior areas of decks
or piers and abutments should be accessible. It is also highly recommendable to
provide sufficient lighting so that inspectors can carry out their work properly.
7.2.1 Decks
The decks must be accessible from the inside. To this end, the corresponding
accesses must be arranged in such a way that any point of the deck can be accessed
without difficulty. The entrance can usually be made either from the abutment
or from the pile heads. It should be borne in mind that, in addition to allowing
inspection personnel to pass through, it is necessary to take into account the
possible material and equipment to be moved inside the deck: stressing jacks,
support devices, etc. (Figure 7.1).
7.2.2 Piers
The piers must be accessible from the top, not only to inspect the bearings, but also to
replace them. This may condition the dimensions of the pier head, since in addition
to the space needed to place the temporary jacks, the space needed to move them and
to remove and place the new bearings must also be taken into account. In addition,
B A
B A
SECTION A-A SECTION B-B
C C
B
SECTION C-C
in the case of hollow piers, the inside of the shaft must be accessible by means of an
internal access staircase (Figure 7.2).
7.2.3 Abutments
As in the case of piers, the bearings placed on abutments must be replaceable, so
the necessary access must be provided, both for personnel and for the handling of
the supporting devices and the jacks and other devices necessary for the dismantling
and assembly of the bearings.
242 7 Conceptual Design for Maintenance
0.05 1.80
0.40 1.00
5.25 0.50 5.00 0.40 0.40
0.85
0.35
0.30
0.47
0.30
0.47 0.85 10 %
Inspection
gallery
1.70
0.30
0.30
1.83
2%
Metal door for
0.30
1.25
access a hollow
gallery 170 x 85 cm. 0.90 1.25
on left wing
5.02
7.30 1.50 6.30 1.00 1.50
1.80
1.80
8.80 8.80
Elevation Section
7.3 Bearings
In general, high-speed rail bridges tend to have high-quality bearings. Typical bear-
ings are of the POT or spherical type, while the normal neoprene strapped bearings
are generally not allowed.
As indicated above, it is usual that the replacement of the supports is carried out
by entering from the deck at the top of the pier. For this reason, it will be necessary
to have an opening in the bottom of the deck of sufficient size to give access to main-
tenance personnel and to allow the passage of the devices necessary to carry out the
replacement of the bearings.
The reinforcement layout of the top part of piers and abutments shall take into
account the position and loading transmitted by the auxiliary jacks used during the
substitution of bearings (Figure 7.4).
7.4 Expansion Joints 243
D D
C C
A A
Pots
bearings Pots
bearings
Jacks for Jacks for
bearing B bearing
substitution substitution
SECTION C-C SECTION D-D
A 200
Plate e = 15
in channel area
Back plate
500
Plate e joint Bolt Ø20
of dilation in Plate e = 15
moving abutment in passage area
Drill Ø80 mm
with grid to the
entry cada 2.00 m "Concrete in-situ"
Precast
Betun Elastomer membrane edge beam
Non-affecting passable bolts reinforced with geotextile (4 mm)
to the safety of the circulation
of work vehicles
A
Sidewalk detail
15 mm S = 0.20 m
e 20 mm 0.20 m = S = 0.40 m
25 mm 0.40 m = S = 0.60 m
30 mm 0.60 m = S = 1.00 m
Movement
Smax+600
Plate e Plate e = 15 100
Bolts Plate e = 10
200
Ø20/0.20
Bolts
Ø20/0.50
100 200 100 S 100
Deck Fixed
abutment
B 200
Plate e = 15
in channel area
Back plate
500
Drill Ø80 mm
with grid to the
entry cada 2.00 m "Concrete in-situ"
Precast
Betun elastomer membrane edge beam
Non-affecting passable bolts reinforced with geotextile (4 mm)
to the safety of the circulation
of work vehicles
B
Sidewalk detail
Movement+100 S
Ø100 Ø9 0
e = 4–6 mm. e = 4–6 mm.
Welding
Railing joint detail
15 mm S = 0.20 m
e 20 mm 0.20 m = S = 0.40 m
25 mm 0.40 m = S = 0.60 m
30 mm 0.60 m = S = 1.00 m
100 Movement L
Plate e Plate e = 15 100
Bolts Plate e = 10
200
Ø20/0.20 Bolts
Ø20/0.50
200 100 S 100
Deck Fixed
abutment
0.05
0.32 0.45 0.43
Protective Protective
concrete plate Bitumen grout
0.06
0.25 0.20
Sand filling
T 120, galvanized
0.10
Mortar MGIII
Filter stones
0.42
7.5 Drainage
Adequate drainage is the cheapest and most effective way of preventing degra-
dation of structures. In this regard, it is strongly recommended to follow the
indications of documents like, among others, the CEB Durable concrete structures
CEB Design Guide [3], regarding the appropriate selection of shapes, presence of
gutters, etc.
7.5 Drainage 247
In general, water will be channelled through closed pipes and drainage will be
through the piers (if they are hollow and can be inspected) and/or through the abut-
ments. Figures 7.9 and 7.10 show examples of the shape of the drainage network of
a bridge.
In general, drain troughs like the one drafted in Figure 7.9 are highly recom-
mended, since they avoid water dripping or running down the front walls of
abutments and deck (Figure 7.10).
DUCT Ø300
Funnel
DUCT Ø300
Funnel
Tooth plate
Anchors
Drain trough
Elevation
7.6 Conclusions
Maintenance must be included in the initial design phases of a bridge. In fact,
in many cases, maintenance needs can shape both the structural solution and
the dimensions of the different elements that make up the bridge. In any case,
the maintenance design of a high-speed railway bridge cannot be independent
from the general track maintenance strategy. In many cases the prioritisation of
track conservation, e.g. by eliminating the rail joint, leads to a higher number of
structural joints which are contrary to the maintenance of the structure. On the
other hand, the construction of long viaducts with a single expansion joint requires
track joints, which are always a critical point for track maintenance. Perhaps the
intermediate alternatives, as proposed in the Deutsche Bahn Design Guide for
Railway Bridges [1], are the best solutions when analysing the conservation of the
railway infrastructure as a whole.
In the authors’ opinion, no important structural element or detail must be hidden,
inaccessible to inspection, monitoring, and potential replacement. In this regard,
the designed lifespan of the overall structure must be compatible with the likely
shorter life of such elements, as well as with secondary but important components
like drainage pipes, gutters, and drain troughs.
References
Appendix A
Factors A and B depend on the initial position of the system, with respect to its
position of static equilibrium y0 , and its initial velocity ẏ 0 as shown in Eq. (A.3):
⎧y = y → A = x
⎪ 0 0
t=0→⎨ ẏ 0
⎪ẏ = ẏ 0 → ẏ = −𝜔A sin 𝜔t + 𝜔B cos 𝜔t → B = 𝜔
⎩
ẏ
y = y0 cos 𝜔t + 0 sin 𝜔t (A.3)
𝜔
When the problem is one of forced vibration, the move-
ment will depend on the load-time law. If it is assumed
M
that a load is introduced suddenly and maintained con-
stant in time, the dynamic equilibrium equation can be
written as in Eq. (A.4).
K
M ÿ + Ky = F (A.4)
F
y = A cos 𝜔t + B sin 𝜔t +
K
⎧ F
⎪y = y0 → A = y0 − K
t=0→⎨ ẏ 0
⎪ẏ = ẏ 0 → ẏ = −𝜔A sin 𝜔t + 𝜔B cos 𝜔t → B =
⎩ 𝜔
( ) ẏ 0
F F
y = y0 − cos 𝜔t + sin 𝜔t + (A.5)
K 𝜔 K
For a system that is initially at rest and placed at the static equilibrium position
(i.e. y0 = 0 and ẏ 0 = 0), Eq. (A.5) simplifies to Eq. (A.6).
F
y= (1 − cos 𝜔t) (A.6)
K
⏟⏟⏟
yst
If plotted (see Figure A.2), it becomes apparent that for this case, the maximum
dynamic deflection reaches 2 times the static deflection yst = F/K. The ratio between
the dynamic deflection and the static deflection is known as the Dynamic Load
Factor (DLF or 𝛷). The case considered above, where the force is applied instan-
taneously and maintained indefinitely, is the worst situation for a non-periodic
load.
A more general expression may be obtained for the DLF when the load history is
less extreme. When the SDOF system is subjected to a dynamic load, F(t), having a
generic load-time law, this load can be seen as a superposition of impulses, that is,
forces that are applied during a very short time (see Figure A.3).
A.1 Dynamics of Single Degree-of-Freedom Systems 251
2.50
2.00
1.50
y/yst
1.00
0.50
0.00
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Time, t (s)
Figure A.2 Dynamic over static deflection resulting from a SDOF system initially at rest
when a load is applied suddenly and maintained through time. Resulting DLF = 2.
t=τ
t
dτ
The solution of the dynamic equation for a single impulse is straightforward. If the
impulse is assumed to be applied when y = 0, and the duration of the load is very
small (as per definition of impulse), it can be assumed that during its application
the y coordinate does not change, so the equilibrium equation simplifies as shown
in Eq. (A.7):
F F
M ÿ + K y = F → ÿ 0 ∼ = constant → ẏ 0 = d𝜏 (A.7)
⏟⏟⏟ M M
∼0
The general solution given in Eq. (A.3) then simplifies to Eq. (A.8):
Fd𝜏
y= sin 𝜔t (A.8)
M𝜔
252 Appendix A Basic Concepts of Dynamics
If the impulse is applied at a time 𝜏, Eq. (A.8) will have a difference in time phase
as shown in Eq. (A.9):
Fd𝜏
y= sin 𝜔(t − 𝜏) (A.9)
M𝜔
Assuming linear elastic behaviour, the solution of a SDOF system subjected to
an arbitrary load, varying in time, can be obtained by superposition of a series of
impulses as shown in Eq. (A.10):
t t
F(𝜏) Fmax
y= sin 𝜔(t − 𝜏)d𝜏 = f (𝜏) sin 𝜔(t − 𝜏)d𝜏
∫0 M𝜔 M 𝜔 ∫0
⏟⏟⏟
K
𝜔2
t
Fmax
= 𝜔 f (𝜏) sin 𝜔(t − 𝜏)d𝜏
K ∫0
⏟⏟⏟
yst
t
y
→ = 𝜔 f (𝜏) sin 𝜔(t − 𝜏)d𝜏
yst ∫0
( )
y
𝛷 = max (A.10)
yst
This equation can be easily generalised to account for damping. The general
expression in this case is given in Eq. (A.11):
t
y
𝛷= = 𝜔d f (𝜏)e−𝜉𝜔(t−𝜏) sin 𝜔d (t − 𝜏)d𝜏
yst ∫0
√
𝜔d = 𝜔 1 − 𝜉 2 (A.11)
where:
𝜉 is the damping index (ratio between the structural damping and the critical damp-
ing1 ), which, for reinforced concrete structures, is around 2%.
With this expression the DLF(𝛷) can be determined for other situations. As an
example, the DLF(𝛷) for a constant load that is applied for a time equal to 1/4 times
the period of the structure can be solved as follows:
The SDOF system will be subjected to two regimes:
– The first one is valid while the constant load is applied. In this case, Eq. (A.6) is
fully applicable:
T
y = yst (1 − cos 𝜔t) t< (A.12)
4
Of course, in this case, the maximum is reached for t = T. The maximum value of
this expression for t < T/4 corresponds to t = T/4 with y = yst .
– The second regime applies when t is greater than T/4, when the system vibrates
freely with the initial conditions provided by the first regime: initial position
(y0, 2nd ) and initial velocity (ẏ 0, 2nd ):
( ) ( )
T 2π T
y = yst 1 − cos = yst
(4) T 4
T 2π T
ẏ = yst 𝜔 sin = yst 𝜔
4 T 4
ẏ 0, 2nd
y = y0, 2nd cos 𝜔t + sin 𝜔t
} 𝜔
y0, 2nd = yst
→ y = yst (cos 𝜔t + sin 𝜔t) (A.13)
ẏ 0, 2nd = yst 𝜔
The maximum
√ of this expression is attained for 𝜔t = π/4, for which the deflection
would be 2yst , so the DLF would be 1.41.
It is easy to repeat this calculation assuming different durations for the application
of the force, td /T, (or different impulses). The result of this analysis is summarised
in Figure A.4. When the duration of the load reaches 0.5T, the DLF becomes equal
to its maximum value (i.e. 2.0). It can also be noted that when the duration of the
load is short, then the DLF can be smaller than 1.0.
Another case with a practical application for the design of railway bridges is the
triangular pulse. Figure A.5 shows an example of a triangular pulse of 380 kN ampli-
tude. The problem can again be solved by using Eq. (A.10). In this case, four time
intervals need to be distinguished:
– Before the impulse is applied, (t < t0 ) at which time the response of the system will
be nil
– The interval t0 ≤ t ≤ t0 + Δt/2, during which time the value of the force is
increasing
2.50
2.00
1.50 √2
DLF = y/yst
1.00
0.50
0.00 0.25
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
td /T
400
350
300
250
F(t) (kN)
200
150
100
50
0
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90
t (s)
Single triangular pulse
– The interval t0 + Δt/2 ≤ t ≤ t0 + Δt, during which time the value of the force is
decreasing
– After t0 + Δt when the system will behave as a freely vibrating system with initial
conditions determined by the position and the velocity of the system for t = t0 + Δt.
Therefore:
t
F(𝜏)
y(t) = sin 𝜔(t − 𝜏)d𝜏
∫t0 M𝜔
( t t )
2𝜔
= yst 𝜏 sin 𝜔(t − 𝜏)d𝜏 − t0 sin 𝜔(t − 𝜏)d𝜏
Δt ∫t0 ∫t0
([( ) ]t )
2𝜔 𝜏 − t0 1
= yst cos 𝜔(t − 𝜏) + 2 sin 𝜔(t − 𝜏)
Δt 𝜔 𝜔 t0
(( ) ( )
2𝜔 t − t0 t − t
= yst cos 𝜔(t − t) − 0 0 cos 𝜔(t − t0 )
Δt 𝜔 𝜔
)
1 1
+ 2 sin 𝜔(t − t) − 2 sin 𝜔(t − t0 )
𝜔 𝜔
2
= yst (𝜔(t − t0 ) − sin 𝜔(t − t0 )) (A.16)
𝜔Δt
For the third time interval,
t
F(𝜏)
y(t) = sin 𝜔(t − 𝜏)d𝜏
∫t0 M𝜔
( ( ) )
t0 + Δt
2 2(𝜏 − t )
t
0 2(𝜏 − t0 )
= yst 𝜔 sin 𝜔(t − 𝜏)d𝜏 + 2− sin 𝜔(t − 𝜏)d𝜏
∫t0 Δt ∫t0 + Δt Δt
( 2
)
t0 + Δt t
2𝜔 2
= yst (𝜏 − t0 ) sin 𝜔(t − 𝜏)d𝜏 + (Δt − (𝜏 − t0 )) sin 𝜔(t − 𝜏)d𝜏
Δt ∫t0 ∫t0 + Δt
2
Δt Δt
⎛ t0 + 2 t0 + 2 ⎞
⎜ 𝜏 sin 𝜔(t − 𝜏)d𝜏 − t0 sin 𝜔(t − 𝜏)d𝜏 ⎟
2𝜔 ⎜ ∫ t0 ∫t0 ⎟
= yst
Δt ⎜ t t
⎟
⎜ +(Δt + t ) sin 𝜔(t − 𝜏)d𝜏 − 𝜏 sin 𝜔(t − 𝜏)d𝜏 ⎟
⎝
0
∫t0 + Δt ∫t0 + Δt ⎠
2 2
⎛ [( ) ]t0 + Δt
⎞
𝜏 − t0 2
⎜ cos 𝜔(t − 𝜏) + 𝜔12 sin 𝜔(t − 𝜏) ⎟
2𝜔 ⎜ 𝜔 ⎟
[( ) t0
]
Δt ⎜⎜ ⎟
= yst t
Δt + t0 − 𝜏 ⎟
+ cos 𝜔(t − 𝜏) − 𝜔12 sin 𝜔(t − 𝜏)
⎜ 𝜔 Δt ⎟
⎝ t0 + 2 ⎠
( ) ( )
⎛ t + Δt
− t ( ) t −t ⎞
0 0 Δt
⎜
2
cos 𝜔 t − t0 − − 0 0 cos 𝜔(t − t0 )⎟
⎜ 𝜔 2 𝜔 ⎟
⎜ ( ) ⎟
1 Δt
⎜ + 2 sin 𝜔 t − t0 − ⎟
⎜ 𝜔 2 ( ) ⎟
2𝜔 ⎜ 1 Δt + t0 − t ⎟
𝜔(t 𝜔(t
Δt ⎜⎜ ⎟
= yst − sin − t ) + cos − t)
𝜔2 0
𝜔 ⎟
( )
⎜ Δt + t0 − t0 − Δt ( ) ⎟
⎜ − Δt 1
2
cos 𝜔 t − t0 − − 2 sin 𝜔(t − t) ⎟
⎜ 𝜔 2 𝜔 ⎟
⎜ ( ) ⎟
⎜ +
1
sin 𝜔 t − t −
Δt ⎟
⎝ 𝜔2 0
2 ⎠
( ( ))
2 Δt
= yst 𝜔(Δt + t0 − t) − sin 𝜔(t − t0 ) + 2 sin 𝜔 t − t0 − (A.17)
𝜔Δt 2
256 Appendix A Basic Concepts of Dynamics
Finally, for the situation after the impulse, the system would behave as a harmonic
system with the initial conditions of position and velocity which are derived from
Eq. (A.18) at time t = t0 + Δt (see Eq. (A.3)):
( ( ))
2 Δt
y(t) = yst 𝜔(Δt + t0 − t) − sin 𝜔(t − t0 ) + 2 sin 𝜔 t − t0 −
𝜔Δt 2
( )
2 Δt
y(t0 + Δt) = y0 = yst − sin 𝜔Δt + 2 sin 𝜔
𝜔Δt 2
( ( ))
2 Δt
̇ = yst
y(t) −1 − cos 𝜔(t − t0 ) + 2 cos 𝜔 t − t0 −
Δt 2
( )
2 Δt
̇ 0 + Δt) = ẏ 0 = yst
y(t −1 − cos 𝜔Δt + 2 cos 𝜔
Δt 2
ẏ 0
y = y0 cos 𝜔(t − t0 − Δt) + sin 𝜔(t − t0 − Δt) (A.18)
𝜔
As is logical, the maximum value expression of Eq. (A.18) depends only upon the
ratio of impulse duration Δt to first natural period and on the static deflection.
As an application, Figure A.6 shows the results for two systems subjected to
a triangular pulse (t0 = 0) for durations of the pulse equal to one-fourth and
five-fourths of the period. Vertical lines indicate the end of the applied impulse.
The maximum DLF for the first case is limited to 0.7 while for the second case it
reaches 1.4. These results are logical in themselves and, also, when a comparison is
done with the rectangular pulse shown in Figure A.4. Obviously, the constant pulse
ω = 14 rad/s
2.0
1.5
duration of pulse for Δt = 5T/4
1.0
0.5
DLF = y/yst
0.0
0 0.5 1 1.5 2 2.5
–0.5
–1.0
–1.5
Figure A.6 Time history of a SDOF system with a natural frequency 𝜔 = 14 rad/s
(T = 0.45 s) subjected to a triangular load as a function of the duration of the impulse
(Δt = 0.25T and Δt = 4/5T).
A.1 Dynamics of Single Degree-of-Freedom Systems 257
results in a larger DLF for the same ratio of pulse duration to natural period of the
system.
| |
From this equation it is easy to see that if the frequency of the periodic force is close
to the natural vibration period the displacements become very large. For a system
without damping, it becomes infinity when both frequencies are equal.
Eq. (A.20) can be generalised to account for damping. The resulting expression
is given in Eq. (A.21), where 𝜉 is the damping index, that √ is the ratio between the
damping coefficient c and the critical damping value cc = 2 km = 2m𝜔:
Kymax 1
= √ (A.21)
F0 ( )2 ( ( 𝜔 )2 )2
𝜔f
2𝜔𝜉 + 1− f
𝜔
As can be seen, for damped systems, the deflection is finite but still experiences
a significant increase as the period of the action is close to the first natural period
258 Appendix A Basic Concepts of Dynamics
35.00
30.00
25.00
Kymax /F0
20.00
15.00
10.00
5.00
0.00
0 0.5 1 1.5 2 2.5 3
ωf /ω
No damping Damping ratio = 1.5%
of the structure. For actual structures, the damping index varies from 0.5% for steel
structures to 2.0% for reinforced concrete structures. From Figure A.7, it can be con-
cluded that for such low damping indexes the effect of damping on the reduction of
resonance effects is modest. More important in limiting the effects of resonance in
railway structures is the fact that the number of cycles is limited, while the above
analysis considers an infinite number of cycles.
– Before the impulse is applied, (t < t0 ) at which time the response of the system will
be nil
– The interval t0 ≤ t ≤ t0 + Δt/2, during which time the value of the force is
increasing
– The interval t0 + Δt/2 ≤ t ≤ t0 + Δt, during which time the value of the force is
decreasing
– After t0 + Δt when the system will behave as a freely vibrating system with initial
conditions determined by the position and the velocity of the system for t = t0 + Δt.
b
⎡ ⎤
⎢ ⎥
⎢ −𝜉𝜔(t−𝜏) cos 𝜔d (t − 𝜏) ⎥ t
cos 𝜔d (t − 𝜏)
= ⎢e ⎥ − 𝜉𝜔e−𝜉𝜔(t−𝜏) d𝜏
⎢⏟⏞⏟⏞⏟ 𝜔 d ⎥ ∫t0 ⏟⏞⏞⏞⏟⏞⏞⏞⏟ 𝜔d
⎢ u ⏟⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏟⎥ du
⏟⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏟
⎣ v ⎦a v
[ −𝜉𝜔(t−𝜏) ]b t
e 𝜔
= cos 𝜔d (t − 𝜏) − 𝜉 e−𝜉𝜔(t−𝜏) cos 𝜔d (t − 𝜏)d𝜏
𝜔d a 𝜔d ∫t0 ⏟⏞⏟⏞⏟⏟⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏟
u2 dv2
[ ]b
e−𝜉𝜔(t−𝜏)
= cos 𝜔d (t − 𝜏)
𝜔d a
b
⎛⎡ ⎤ ⎞
⎜⎢ ⎥ ( ) ⎟
𝜔 ⎜⎢ −𝜉𝜔(t−𝜏) − sin 𝜔d (t − 𝜏) ⎥ − sin 𝜔d (t − 𝜏) ⎟
b
−𝜉 ⎜⎢ e ⎥ − 𝜉𝜔e−𝜉𝜔(t−𝜏) d𝜏 ⎟
𝜔d ⎜⎢⏟⏞⏟⏞⏟ 𝜔d ⎥ ∫a ⏟⏞⏞⏞⏟⏞⏞⏞⏟ 𝜔d ⎟
⎜⎢ u2 ⏟⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏟⎥ du2 ⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟ ⎟
⎝⎣ v2 ⎦a v2 ⎠
[ −𝜉𝜔(t−𝜏) ]b
e 𝜔
= cos 𝜔d (t − 𝜏) + 𝜉 2 [e−𝜉𝜔(t−𝜏) sin 𝜔d (t − 𝜏)]ba
𝜔d a 𝜔d
( )2 b
𝜔
− 𝜉2 e−𝜉𝜔(t−𝜏) sin 𝜔d (t − 𝜏)d𝜏
𝜔d ∫a
⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟
I1
b
e−𝜉𝜔(t−𝜏) sin 𝜔d (t − 𝜏)d𝜏
∫a
b
⎡ ⎤
⎢ ( )⎥
e −𝜉𝜔(t−𝜏) 𝜔
=⎢ ( (
⎥
)2 ) cos 𝜔d (t − 𝜏) + 𝜉 𝜔 sin 𝜔d (t − 𝜏) ⎥
⎢ 𝜔 d
⎢ 𝜔d 1 + 𝜉 𝜔 ⎥
⎣ d ⎦a
( )
e−𝜉𝜔(t−b) cos 𝜔d (t − b) + 𝜉 𝜔𝜔 sin 𝜔d (t − b)
( d
)
−e−𝜉𝜔(t−a) cos 𝜔d (t − a) + 𝜉 𝜔𝜔 sin 𝜔d (t − a)
( )2 )
d
= I1,a→b (t) = ( (A.22)
𝜔
𝜔d 1 + 𝜉 𝜔
d
260 Appendix A Basic Concepts of Dynamics
Integral I 2 :
b
⎡ ⎤
⎢ ⎥
b
⎢ cos 𝜔 d (t − 𝜏) ⎥
I2,t0 (t) = 𝜏e−𝜉𝜔(t−𝜏) sin 𝜔d (t − 𝜏)d𝜏 = ⎢𝜏e−𝜉𝜔(t−𝜏) ⎥
∫a ⏟⏞⏟⏞⏟⏟⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏟ ⎢⏟⏞⏟⏞⏟ 𝜔d ⎥
u dv ⎢ u ⏟⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏟ ⎥
⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟ ⎣ v ⎦a
I2
b
cos 𝜔d (t − 𝜏)
− (e−𝜉𝜔(t−𝜏) + 𝜏𝜉𝜔e−𝜉𝜔(t−𝜏) ) d𝜏
∫a ⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟ 𝜔d
du ⏟⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏟
v
b
⎡ ⎤
⎢ ⎥
⎢ −𝜉𝜔(t−𝜏) cos 𝜔d (t − 𝜏) ⎥ 1
b
= ⎢𝜏e ⎥ −𝜔 e−𝜉𝜔(t−𝜏) cos 𝜔d (t − 𝜏)d𝜏
⎢ ⏟⏞⏟⏞⏟ 𝜔 d ⎥ d
∫a
⎢ u ⏟⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏟⎥ ⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟
⎣ v ⎦a [ ]t
t
− e−𝜉𝜔(t−𝜏)
sin 𝜔d (t−𝜏) +𝜉 𝜔𝜔 e−𝜉𝜔(t−𝜏) sin 𝜔d (t − 𝜏)d𝜏
𝜔d t0 d ∫ t0
⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟
I1
b
𝜔
−𝜉 𝜏e−𝜉𝜔(t−𝜏) cos 𝜔d (t − 𝜏)d𝜏
𝜔d ∫a ⏟⏞⏟⏞⏟⏟⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏟
u2 dv2
b
⎡ ⎤
⎢ ⎥ [ ]b
⎢ cos 𝜔d (t − 𝜏) ⎥ 1 e−𝜉𝜔(t−𝜏)
= ⎢𝜏e−𝜉𝜔(t−𝜏) ⎥ + sin 𝜔 (t − 𝜏)
⎢⏟⏞⏟⏞⏟ 𝜔d ⎥ 𝜔d 𝜔d d
a
⎢ u ⏟⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏟ ⎥
⎣ v ⎦a
b
𝜔
−𝜉 e−𝜉𝜔(t−𝜏) sin 𝜔d (t − 𝜏)d𝜏
𝜔2d ∫a
⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟
I1
b
⎛⎡ ⎤ ⎞
⎜⎢ ⎥ ⎟
𝜔 ⎜⎢ −𝜉𝜔(t−𝜏) − sin 𝜔d (t − 𝜏) ⎥ − sin 𝜔d (t − 𝜏) ⎟
b
−𝜉 𝜏e −𝜉𝜔(t−𝜏)
𝜏𝜉𝜔e −𝜉𝜔(t−𝜏)
𝜔d ⎜⎜⎢⎢⏟⏞⏟⏞⏟ ⎥ ⎟
− (e + ) d𝜏
𝜔d ⎥ ∫a ⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟ 𝜔d ⎟
⎜⎢ u 2
⏟⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏟ ⎥ du2 ⏟⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏟ ⎟
⎝⎣ v2 ⎦a v2 ⎠
A.1 Dynamics of Single Degree-of-Freedom Systems 261
b
⎡ ⎤
⎢ ⎥ [ ]b
⎢ −𝜉𝜔(t−𝜏) cos 𝜔d (t − 𝜏) ⎥ 1 e−𝜉𝜔(t−𝜏) 𝜔
= ⎢𝜏e ⎥ + sin 𝜔d (t − 𝜏) − 𝜉 2 I1,a→b
⎢⏟⏞⏟⏞⏟ 𝜔d ⎥ 𝜔d 𝜔d a 𝜔d
⎢ u ⏟⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏟⎥
⎣ v ⎦a
b
𝜔 𝜔
+𝜉 [𝜏e−𝜉𝜔(t−𝜏) sin 𝜔d (t − 𝜏)]ba − 𝜉 2 e−𝜉𝜔(t−𝜏) sin 𝜔d (t − 𝜏)d𝜏
𝜔2d 𝜔d ∫ a
⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟
I1,a→b
( )2 b
𝜔
− 𝜉2 𝜏e−𝜉𝜔(t−𝜏) sin 𝜔d (t − 𝜏)d𝜏 →
𝜔d ∫a
⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟
I2
b
𝜏e−𝜉𝜔(t−𝜏) sin 𝜔d (t − 𝜏)d𝜏
∫a
[ ( ( ) )]b
e−𝜉𝜔(t−𝜏)
𝜔d
𝜏 cos 𝜔d (t − 𝜏) + 1
𝜔d
+ 𝜏𝜉 𝜔𝜔 sin 𝜔d (t − 𝜏) − 2𝜉 𝜔𝜔2 I1,a→b
d a
( )2 )
d
= (
1 + 𝜉 2 𝜔𝜔
d
( ( ) )
⎛ e −𝜉𝜔(t−b)
b cos 𝜔d (t − b) + 𝜔 + b𝜉 𝜔𝜔 sin 𝜔d (t − b)
1 ⎞
⎜ ( ( d d ) ) ⎟
⎜ −e −𝜉𝜔(t−a)
a cos 𝜔d (t − a) + 𝜔 + a𝜉 𝜔 sin 𝜔d (t − a) − 2𝜉 𝜔 I1,a→b ⎟⎠
1 𝜔 𝜔
⎝
( ( )2 )
d d d
I2,a→b (t) = (A.24)
𝜔d 1 + 𝜉 2 𝜔𝜔
d
(A.25)
A.1.3.2 Solution for the damped SDOF System Subjected to a Triangular Load
For the second time interval, t0 ≤ t ≤ t0 + Δt/2:
t t
F(𝜏) F
y(t) = e−𝜉𝜔(t−𝜏) sin 𝜔d (t − 𝜏)d𝜏 = max e−𝜉𝜔(t−𝜏) f (𝜏) sin 𝜔d (t − 𝜏)d𝜏
∫0 M𝜔d M𝜔d ∫0
Δt
if t0 ≤ t ≤ t0 +
2
Fmax t −𝜉𝜔(t−𝜏) 2(𝜏 − t0 )
y(t) = e sin 𝜔d (t − 𝜏)d𝜏
M𝜔d ∫t0 Δt
262 Appendix A Basic Concepts of Dynamics
( t t )
Fmax 2
= e −𝜉𝜔(t−𝜏)
𝜏 sin 𝜔d (t − 𝜏)d𝜏 − t0 e −𝜉𝜔(t−𝜏)
sin 𝜔d (t − 𝜏)d𝜏
M𝜔d Δt ∫t0 ∫t0
F 2
y(t) = max (I (t) − t0 I1,t0 →t (t)) (A.26)
M𝜔d Δt 2,t0 →t
For the third time interval, t0 + Δt/2 ≤ t ≤ t0 + Δt
Δt
if t0 + ≤ t ≤ t0 + Δt
2
Δt
Fmax t0 + 2 −𝜉𝜔(t−𝜏) (𝜏 − t0 )
y(t) = 2 e sin 𝜔d (t − 𝜏)d𝜏
M𝜔d ∫t0 Δt
( )
Fmax t (𝜏 − t0 )
+2 e −𝜉𝜔(t−𝜏)
1− sin 𝜔d (t − 𝜏)d𝜏
M𝜔d ∫t0 + Δt Δt
2
F ( )
2
y(t) = max I2,t0 →t0 + Δt (t) − t0 I1,t0 →t0 + Δt (t)
M𝜔d Δt 2 2
t ( )
Fmax (𝜏 − t0 )
+2 e −𝜉𝜔(t−𝜏)
1− sin 𝜔d (t − 𝜏)d𝜏
M𝜔d ∫t0 + Δt Δt
2
F ( )
2
= max I2,t0 →t0 + Δt (t) − t0 I1,t0 →t0 + Δt (t)
M𝜔d Δt 2 2
(( ) ( ) )
Fmax t0 1
+2 1+ I Δt (t) − I Δt (t)
M𝜔d Δt 1,t0 + 2 →t Δt 2,t0 + 2 →t
F ( ( )
1
y(t) = 2 max I2,t0 →t0 + Δt (t) − I2,t0 + Δt →t (t) − t0 I1,t0 →t0 + Δt (t)
M𝜔d Δt 2 2 2
( ) )
t0
+ 1+ I Δt (t) (A.27)
Δt 1,t0 + 2 →t
Finally, for the fourth time interval (t ≥ t0 + Δt), the system will be in free vibration
with the initial velocity and position attained at time t0 + Δt:
y0 = y(t0 + Δt)
F ( ( )
1
= 2 max I2,t0→ t0 + Δt (t0 + Δt) − I2,t0 + Δt →t0 +Δt (t0 + Δt) − t0 I1,t0→ t0 + Δt (t0 + Δt)
M𝜔d Δt 2 2 2
( ) )
t
+ 1 + 0 I1,t0 + Δt →t0 +Δt (t0 + Δt)
Δt 2
Fmax
ẏ 0 = 2
M𝜔d
( ( )
1 ̇
× I 2,2,t0→ t0 + Δt (t0 + Δt) − İ 2,t0 + Δt →t0 +Δt (t0 + Δt) − t0 İ 1,2,t0→ t0 + Δt (t0 + Δt)
(Δt ) )
2 2 2
t0
+ 1+ İ Δt (t + Δt)
Δt 1,t0 + 2 →t0 +Δt 0
( )
ẏ 0 + 𝜉𝜔y0
y(t) = e −𝜉𝜔(t−t0 −Δt)
y0 cos 𝜔d (t − t0 − Δt) + sin 𝜔d (t − t0 − Δt) (A.28)
𝜔d
Reference
Appendix B
B.1 Germany
A number of special bridges built in Germany are listed below. The first four
(Gemünden, Veitshöchheim, Pfieffetal, and Nantenbach) correspond to the con-
struction of the first German high-speed railways built mainly in the second part
of the 1980s and the first part of the 1990s. The rest are special bridges built after
German reunification in the first two decades of the 21st century.
The German bridges built in the first stage were designed so that any deck could be
replaced in the shortest period of time. Therefore, German bridges of that period are
generally isostatic, and in the case of continuous bridges, their length was limited
(to 400 m) to allow for a hypothetical rapid replacement.
Bridges built in Germany in the second decade of the 21st century have been
developed according to the guidelines of the DB Netze Railway Bridge Design Guide
2008. With the implementation of this guide, the condition that the decks should be
quickly replaceable has been removed and semi-integral bridges have been designed,
generally without track expansion joints and structure joints every 100–120 m or
so. Examples of such solutions are the Unstruttal bridge and the Gänsebachtal
viaduct.
B.1.1.2 Description
This bridge has a total length of 793.5 m and is divided into three continuous sections
of 164.0 – 299.0 – 330.5 so that it can be replaced in parts, with typical spans of
55.0 m. The braking is independently resisted in each sector. The lateral spans are
anchored to the abutments, which are the elements that take the braking loads. In
the intermediate section, V-shaped piers, the braking loads are transmitted through
the hinge located at the base of the V-shaped pier. The deck is 4.50 m in depth in
general, increasing to 6.50 m in the area of the V-shaped piers. The main span has a
length of 135 m. The track has only two expansion joints at the ends of the V-deck
section.
Further Reading
Leonhardt, F. (1986). Mainbrücke Gemünden – Eisenbahnbrücke aus Spannbeton mit
135 m Spannweite. Beton- und Stahlbetonbau 81 (1): 1–8.
Zellner, W. and Saul, R. (1991). Long span bridges of the new railroad lines in Germany.
IABSE Report 64 (1991).
https://structurae.net/en/structures/gemunden-high-speed-rail-viaduct.
B.1 Germany 265
B.1.2.2 Description
The bridge over the river Main in Veitshöchheim has a total length of 1264 m. It is
divided into five separate parts of 237.0, 369.5, 214.0, 160.5, and 281.6 m. The track
has rail joints in the two abutments and in the pier between the 214.0 and 160.5 m
long sections. The deck is anchored longitudinally at the arch keystone and at the
three piers located at the transition between the 160.5 and 281.6 m sections.
The 237.0, 369.5, and 214.0 deck sections are connected by shock transmitters
located in the transition pile sections that allow slow movements but are locked to
transmit the braking loads to the arch keystone.
Further Reading
Zellner, W. and Saul, R. (1991). Veitshöchheim viaduct: a concrete arch bridge with
162 m main span. IABSE Report 64.
Zellner, W. and Saul, R. (1991). Long span bridges of the new railroad lines in Germany.
IABSE Report 64.
https://structurae.net/en/structures/veitshochheim-viaduct.
B.1 Germany 267
Figure B.4 Veitshöchheim Bridge (Source: Störfix - Eigenes Werk, CC BY-SA 3.0,
https://commons.wikimedia.org/w/index.php?curid=916591).
268 Appendix B Singular Bridges for High-Speed Railway Lines
B.1.3.2 Description
This structure has a total length of 17 × 58 = 986 m. The deck is made up of 17 spans
of 55.75 m span and 5.30 m depth. The spans are isostatic so that they can be replaced
quickly and independently. Each deck span is connected to its adjacent spans by
pre-stressed cables to transmit the braking loads. Due to the considerable height
above the valley (approx. 95 m), a central A-pier has been provided to transmit the
braking loads to the ground. The shafts of the inclined piers are curved to eliminate
bending under their own weight. The track only has two expansion joints at the
abutments. The track has expansion joints only in the abutments.
Further Reading
Harries, H., Kinkel, H., and Petri, H. (1987). Die Rombachtalbrücke. Beton- und
Stahlbetonbau 82 (7): 179–185.
Zellner, W. and Saul, R. (1991). Long span bridges of the new railroads lines in
Germany. IABSE Report 64.
B.1 Germany 269
B.1.4.2 Description
The first section of the bridge is 320.10 m long and has spans of 50 m. The deck of
this part is a hollow box concrete section. The second sector is a 374.4 m long con-
tinuous mixed truss with spans of 83.2, 208.0, and 83.2 m. The deck depth is 8.50 in
the central area and 16.50 in the support sections of the main piers, giving a slender-
ness of 1/24 and 1/13, respectively. The truss has a double composite action, with a
bottom slab in the lower part of the support sections in the main piers.
Further Reading
Zellner, W. and Saul, R. (1991). Railway bridge with double composite action across
river Main. IABSE Report 64.
Zellner, W. and Saul, R. (1991). Long span bridges of the new railroad lines in Germany.
IABSE Report 64.
https://structurae.net/en/structures/railroad-bridge-at-nantenbach.
B.1 Germany 271
B.1.5.2 Description
This viaduct is made up of four sections of 580 m long, each consisting of 10 con-
tinuous spans. The spans have a typical length of 58 m, and the fixed points of each
section are materialised in the keystones of the arches that are located in the centre
of the span. The piers are very slender and all are connected monolithically at their
head to the deck. This results in a structure in which the only bearing devices of the
deck are in the abutments. There are track expansion joints at the same position as
the structural expansion joints.
Further Reading
Rimane, T. (2011). Der Bau der Eisenbahnüberführung Unstruttalbrücke. In: Deutscher
Bautechnik-Tag 11-13. Mai 2011 (ed. DBV), 69–70. Berlin: Tagungsband.
Schlaich, M. (2012). Integral Railway Bridges in Germany. 22nd Dresdener
Brückenbausymposium. Dresden.
https://structurae.net/en/structures/unstrut-viaduct.
B.1 Germany 273
B.1.6.2 Description
This viaduct consists of 10 sections of 112 m each separated by structural expansion
joints. Each segment has five spans with a rigid double pier which is the fixed point of
each segment. All piers are monolithically embedded at their heads in the deck, and
therefore there are only bearings on the abutments. The track is continuous along
the entire length of the bridge and there are no track expansion joints, not even at the
abutments.
Further Reading
Schenkel, M., Goldack, A., and Schlaich, J. (2010). Die Gänsebachtalbrücke, eine
integrale Talbrücke der DB AG auf der Neubaustrecke Erfurt-Leipzig/Halle. Beton-
und Stahlbetonbau 105 (9): 590–598. https://doi.org/10.1002/best.201000034.
Schlaich, M. (2012). Integral Railway Bridges in Germany. 22nd Dresdener
Brückenbausymposium. Dresden.
https://structurae.net/en/structures/gansebach-viaduct.
B.1 Germany 275
B.1.7.2 Description
The superstructure of the 18-span bridge consists of a western 12-span bridge made
of prestressed concrete, the 3-span current bridge made of steel, and the eastern
3-span foreshore bridge made of prestressed concrete.
The total length is 809.87 m. It consists of one section with 12 spans of 34.07 m
– 34.25 m – 34.16 m – 34.21 m – 34.20 m – 34.21 m – 34.20 m – 41.11 m – 45.18 m
– 45.06 m – 43.62 m, of the truss structure section with spans of 66.99 m – 105.77 m
– 66.9 m, and the third sector with spans of 40.66 m – 40.55 m – 40.52 m.
Horizontal braking forces are transferred to the east abutment and the pier
between the steel bridge and the west flood bridge. The track expansion joints are
located at the abutment west and above the pier between the steel bridge and the
flood bridge east.
Further Reading
Braun, M. (2013). Die Eisenbahnbrücke bei Hämerten in Sachsen-Anhalt. Bautechnik
90 (2): 113–119. https://doi.org/10.1002/bate.201200039.
Heinisch, R., Siegmann, J., Stuchly, H., and Witt, P. (2001). Ingenieurbauwerke.
Hestra-Verlag, Darmstadt (Germany): ETR - Eisenbahntechnische Rundschau.
https://structurae.net/en/structures/hamerten-railroad-bridges.
B.1 Germany 277
B.1.8.2 Description
The bridge connects the Boßlertunnel and the Steinbühltunnel. The structure con-
sists of two single-track, parallel prestressed concrete bridges, 485 and 472 m long,
respectively. The spans of the western structure are 44, 95, 150, 93, 58, 45 m and of
the eastern structure 44, 95, 150, 88, 50, 45 m. The two bridge superstructures are
designed as single-cell box girder cross-sections prestressed of 9.20 width. They are
monolithically connected to the bridge piers. On each abutment, there is an expan-
sion joint and the deck is resting in bearings. The frame construction is referred to
as semi-integral construction.
Further Reading
Angelmaier, V. (2004). Neubaustrecke Wendlingen – Ulm. Die Filstalbrücke als Teil des
"Albaufstiegs". Umrisse 4 (2): 13–17.
Schlaich, M. (2012): Integral Railway Bridges in Germany. 22nd Dresdener
Brückenbausymposium, Dresden.
B.1 Germany 279
Figure B.16 Filstal Bridge (Courtesy of Deutsche Bahn AG, photo: Arnim Kilgus).
B.2 France 281
B.2 France
Unlike the first generation of German bridges, French bridges were not designed so
that their decks could be replaced quickly if necessary. This has led to bridges on
high-speed lines in France having continuous decks. The use of continuous decks
means that for long viaducts it is necessary to have track expansion joints.
In these cases, French bridges use a so-called neutral or inert span. This is an
independent span, simply supported, and usually located in the centre of the bridge,
which allows the effective expansion length of the bridge to be divided in two. Thus,
the two long side decks are usually fixed to the respective abutments and two rail
joints are placed using the two expansion joints of the central neutral span.
In the case of very long viaducts, instead of anchoring the decks to the abutments,
they have been anchored to intermediate piers so that the fixed points are approx-
imately in the centre of each section. In these cases, track expansion joints are
arranged to coincide with the expansion joints of the structure at the abutments.
In addition to the aspects relating to the long viaducts, French bridges have the
particularity of the joint work between engineers and architects on the unique
bridges, with a result that is not always fortunate.
282 Appendix B Singular Bridges for High-Speed Railway Lines
B.2.1.2 Description
The flagship structure of the Lyon-Mediterranean high-speed line, it is 324.60 m long
consisting of four spans, including two 52 m approach spans and two main bowstring
spans of 110.30 m. A central arch joins the two bowstring spans and slightly increases
the overall rigidity. The steel arches support a concrete deck.
The two arches have a height above the deck of 15 m, giving a rise/span ratio of
1/7.7, while the maximum height of the central arched part has its highest point at
a distance of 24 m above the deck. The central pier supporting the two arches is the
fixed point of the bridge and has been designed to withstand the braking loads.
Further Reading
Mimram, M. Le viaduc de La Garde-Adhémar. In: Formes et Structures, n. 127.
Ramondenc, P., Plu, B., Pourcelot, M., and Hoorpah, W. (1992). La Garde-Adhémar
Bolletin 19. Pont métalliques OUTA.
https://structurae.net/en/structures/garde-adhemar-viaduct.
B.2 France 283
B.2.2.2 Description
This bridge is divided into three 450 – 650 – 400 m sections. The first two sections
have 100 m spans and the last one has 50 m spans. The connection between the first
two sections is made by means of a half corbel joint. The first section is fixed to
the two piers closest to the half corbel joint, while the central section of the deck
is joined to two piers centred on the section of considerable height. The track has
expansion joints coinciding with the expansion joints between sections and with
the abutments.
Further Reading
Blassel, J.F., Desvignes, M., and Virlogeux, M. Les viaducs sur le Rhône à Avignon. In:
Formes et Structures, n. 127.
Chatelard, P., Martin, O., Roujon, M., and Sayn, P. (1998). Lot 2H - Les viaducs
d’Avignon. In: Travaux. n. 742.
B.2 France 285
B.2.3.2 Description
The access spans to the bow-string are composite concrete slabs with double “T”
steel beams with spans ranging from 45 to 60 metres. One of them is the inert span
that houses the track expansion devices. The inert span is right next to the bow-string
section. In order to improve the performance of the system, the overlapping arches
are connected, in addition to a system of general bars, by means of solid zones in the
keystones and at the start of the arches.
Further Reading
Ramondenc, P. and Bousquet, C. (2004). The main bridges of the high speed line HSL
Méditerranée. In: Presented at: fib Symposium, Avignon.
Vallée, P., Roujon, M., and Sayn, P. (1998). Lot 2C - Le viaduc de Mornas. In: Travaux,
n. 742.
B.2 France 287
B.2.4.2 Description
The structure is composed of a succession of mixed-frame portals of a constant and
deliberately short length (66 m) to reduce the depth of the deck. Due to the curvature
of the route and its length, it was not possible to use the usual expansion joints to
adopt a continuous structure on supports. Therefore, the frames are independent of
each other. They consist of a short span of 21 m embedded in the supports, and an
isostatic span of 45 m resting on the previous ones by means of spherical supports
that allow longitudinal displacements and rotations. There are no track expansion
joints.
Further Reading
de la Savoureuse, V. (2005). France. In: The Architectural Review.
Yazbeck, N. (2010). LGV Rhin-Rhône, Branche Est - Les études du viaduc de la
Savoureuse. In: Travaux, n. 870.
B.2 France 289
B.3 Spain
Spanish high-speed long bridges are characterised by the use of continuous solutions
with a fixed point usually at an abutment or at one or more intermediate points in
the case of extra-long bridges.
This generally leads to the need to place a track expansion joint in the movable
abutment or two track expansion joints in abutments when the fixed point is in the
centre of the bridge.
The Spanish orography has given rise to a series of long-span bridges in which
structural rigour has been combined with visual appraisal and landscaping consid-
erations.
292 Appendix B Singular Bridges for High-Speed Railway Lines
B.3.1.2 Description
The bridge has a total length of 546 m with a span distribution of 18 + 6 × 24 + 60 +
120 + 2 × 60 + 42 m. It is a novel deck in which an attempt has been made to adapt
the concept of the metal trusses used in large railway bridges to the structural and
constructive domain of prestressed concrete bridges.
The deck is a box girder with lightened webs with circular holes and joined at the
top with ribs. Structurally, it is a vierendeel beam. It has a total depth of 9.15 m. The
webs are lightened with circular holes of 3.80 m. in diameter every 6.00 m. The deck
has a set of transverse beams cross beams with a circular profile depth separated by
3.0 m. with a trapezoidal section with a variable width from 0.50 to 0.60 m.
Further Reading
Manterola, J. and Martínez-Cutillas, A. (2003). The Ebro River Bridge, a new concrete
bridge for railways. In: Presented at: IABSE Symposium: Structures for High-Speed
Railway Transportation, Antwerp, Belgium.
Manterola, J., Martínez-Cutillas, A., and López, J.L. (2003). High-speed railway bridge
over the Ebro River, Spain. Structural Engineering International 13 (3): 174–176.
B.3 Spain 293
Figure B.25 Osera Bridge, interior view (Courtesy of Carlos Fernández Casado CFC & ADIF).
Figure B.26 Osera Bridge (Courtesy of Carlos Fernández Casado CFC & ADIF).
294 Appendix B Singular Bridges for High-Speed Railway Lines
B.3.2.2 Description
The 574 m long bridge is comprised of two parts: a 307 m long steel–concrete struc-
ture crossing the busy highway AP-7 and a continuous prestressed concrete box
girder viaduct over the Mogent River with a longest span of 48 m. The steel bridge is
a continuous structure with five spans measuring 45 + 71 + 75 + 71 + 45 m, respec-
tively. The location of the piers was controlled by the highly skewed angle of the
highway crossing and by the launching process used to erect the bridge. The bridge
is designed as a king post truss with curved diagonals, which supports a composite
steel–concrete deck. The 14 m wide deck accommodates two ballasted tracks, for a
total bridge width of 17 m. The composite deck rests on 1 m deep transverse steel
plate I-beams positioned 3.55 m apart. These beams are connected to the bottom
chord, which is a 1.6 m wide box section with depth varying between 3.5 and 6 m.
Further Reading
Sobrino, J.A. (2010). Two steel bridges for the high-speed railway line
Madrid-Barcelona-French border. In: Presented at: IABSE Symposium: Improving
Infrastructure Worldwide, Weimar, Germany.
Sobrino, J.A. (2010). Realizaciones Españolas. Diez años de ingeniería estructural.
Asociación Científico-técnica del Hormigón Estructural (Cinter Divulgación
Técnica).
B.3 Spain 295
Figure B.28 Llinars del Vallès Viaduct (Courtesy of PEDELTA, photo: Alex Bo).
296 Appendix B Singular Bridges for High-Speed Railway Lines
B.3.3.2 Description
Situated only one hundred metres from Delicias Station, the overpass spans eight
railway tracks, including the Madrid-Barcelona high-speed railway as well as
regional and commuter train tracks. The deck rigid stays consist of welded steel
box sections. The pier shafts are rigidly connected to the deck. The reduction of the
height of the cross section towards the top of the pier contributes to reducing
the bending moments in these members. The bridge only has support bearings on
the abutments. Track has no expansion joint.
Further Reading
Tanner, P. and Bellod, J.L. (2006). Salto del Carnero Railway Bridge Saragossa Spain.
IABSE Structural Engineering International Journal 16 (3): 200–203.
Tanner, P. and Bellod, J.L. (2011). Concepción y proyecto del puente ferroviario
extradosado Salto del Carnero de Zaragoza- Delicias. Hormigón y Acero 62 (259).
B.3 Spain 297
Figure B.29 Salto del Carnero railway bridge (Courtesy of CESMA Ingenieros).
Figure B.30 Salto del Carnero railway bridge, view from below (Courtesy of CESMA
Ingenieros).
298 Appendix B Singular Bridges for High-Speed Railway Lines
B.3.4.2 Description
High-speed railway line between Barcelona and the French border crosses the AP-7
highway in Riudellots de la Selva (Girona). The structure has two spans of 53.00 m
and a big skew (34∘ ). A truss structure with variable depth has been used due to a
wide range of facts such as high visibility from the highway, strong stiffness require-
ments, and span length. This structural type combines stiffness and lightness.
The structure has two lateral beam trusses with a parabolic shape. Beams are
entirely closed in central area to have enough bending resistance and to transmit
the shear forces through the central support. The deck has transversal beams con-
necting the trusses without any diaphragms between piers, which creates a clean
bottom view from the highway. A concrete slab was constructed over transversal
beams. This slab directly resists railways loads and gives transversal stiffness to the
structure. The bridge is seen by a thousand people daily, and the design aim is the
combination of robustness and aesthetic appearance.
Further Reading
Romo, J. (2010). Riudellots High Speed Line Bridge. In: Presented at IABSE Symposium:
Large Structures and Infrastructures for Environmentally Constrained and Urbanised
Areas, Venice, Italy.
Romo, J. (2011). Puente de la línea de alta velocidad Barcelona-Frontera francesa sobre
la AP-7 en Riudellots de la Selva. Revista Hormigón y Acero 62 (261).
B.3 Spain 299
Figure B.31 Viaduct over AP7 Riudellots de la Selva, aerial view (Courtesy of Talyr).
B.3.5.2 Description
The arch, with a span of 261 m, was divided into six parts by vertical columns.
The arch follows a polygonal guideline. The antifunicular of the arch is perfect in
this way, reducing the deflections that would exist in the area between the vertical
columns if the arch were perfectly curved.
It is a reinforced concrete arch bridge with a continuous prestressed concrete box
deck and two access viaducts. The span of the arch is 261 m and a rise at the centre
is 36.95 m, and therefore a rise -to-span ratio of 1/6.77.
Further Reading
Manterola, J., Martínez, A., Navarro, J.A., et al. (2008): Puente de ferrocarril de alta
velocidad sobre el embalse de Contreras. Presented at: IV Congreso ACHE, Valencia.
Manterola, J., Martínez, A., Navarro, J.A., and Martín, B. (2012). Puente arco de
ferrocarril sobre el embalse de Contreras en la línea de alta velocidad
Madrid-Levante. Revista Hormigón y Acero 63 (264).
B.3 Spain 301
Figure B.33 Contreras Bridge (Courtesy of Carlos Fernández Casado CFC & ADIF).
Figure B.34 Contreras Bridge (Courtesy of Carlos Fernández Casado CFC & ADIF).
302 Appendix B Singular Bridges for High-Speed Railway Lines
B.3.6.2 Description
The viaduct over the river Ulla where it flows into the Arosa Estuary. The design
minimises the number of piers on the course of the river and tries to seek the max-
imum transparency and the minimum visual impact possible on the surrounding
landscape. The bridge is a composite variable depth structure with three large cen-
tral spans over the water measuring 225 + 240 + 225 m and 120 m access spans. The
deck has a variable depth with 17.90 m on the piers section and 9.15 m at the cen-
tre of the span. This depth remains constant on the access spans. The four main
piers were reinforced to resist deck rotation and control the level flexion transmit-
ted to the foundations through the frame effect, thus preventing over-sizing. For this
reason, the main piers located at the outer edges of the 225 m spans were designed
with two free-standing partitions driving into the foundations and pier capitals. The
remaining piers are conventional. The deck support over these piers is free spherical
lengthwise bearings. The track has expansion joints coinciding with the expansion
structural joint in the abutments.
Further Reading
Millanes, F., Ortega, M., and Matute, L. (2014). Viaduct over river Ulla: an outstanding
composite (steel and concrete) high-speed railway viaduct. Structural Engineering
International 24.
Millanes, F., Ortega, M., and Estévez, R. (2015). Viaduct over Ulla River in the Atlantic
high-speed railway line: A composite (steel–concrete) truss world record. (ACHE,
ELSERVIER, Hormigón y Acero 66(277).
B.3 Spain 303
B.3.7.2 Description
The bridge has a large concrete arch with an upper deck spanning 384 m over the
Alcantara reservoir. This large arch is the main element of a 996 m long viaduct,
consisting of 12 approach spans with 45 m spans, and two additional spans of 36 m
at the ends.
The deck has a prestressed concrete box section with a constant 3.1 m depth. The
viaduct piers have a maximum height of 65.3 m. Both those found at ground and
those supported on the arch have a variable octagonal cross-section, the aerodynam-
ics of which are beneficial for the arch span, given its large span. Almonte’s main
mechanism for taking the longitudinal forces from trains braking is the fixed point
located at the apex of the arch. At the centre of the bridge there is a 42 m long fixed
point connecting the arch and the deck. Horizontal braking loads are transferred
through the fixed point from the deck to the arch and then into the abutments. All
the columns are connected to the deck with bearings, allowing the deck to move
longitudinally with respect to the columns.
Further Reading
Arenas, J.J., Capellán, G., Martínez, J. et al. (2016). Viaduct over River Almonte. Design
and Analysis. In: Presented at: IABSE Symposium: Challenges in Design and
Construction of an Innovative and Sustainable Built Environment, Stockholm,
Sweden.
Capellán, M. (2015). Puente arco de alta velocidad sobre el río Almonte. ROP Revista de
Obras Públicas n∘ 3562 Madrid. Spain.
B.3 Spain 305
B.3.8.2 Description
Located on the Madrid-Extremadura line, the viaduct has a total length of 1488 m.
The distribution of the spans of the viaduct is influenced by the Tagus River jump,
which is made by means of an arch of 324 m span, with the deck being divided into
six 54 m spans. The access spans are 60 m, between them are two 57 m transition
spans, one on each side of the springs of the arch. The arch is a concrete hollow
box with a variable edge between 3.50 and 4.00 m and a width between 12.00 m at
the and 6.00 m at the keystone. And the deck is supported by bearings on the piers
and abutments. The keystone of the arch is therefore the point at which the deck is
fixed longitudinally. There are two structural and track expansion joints at the two
abutments.
Further Reading
Manterola, J., Astiz, M.A., and Martínez, A. (1999). Puentes de Ferrocarril de Alta
Velocidad Revista de Obras Públicas n∘ 3386 Madrid.
Manterola, J., Martínez, A., and Martín, B. (2015). Viaducto sobre el río Tajo en el
embalse de Alcántara para ferrocarril de alta velocidad. Revista de Obras públicas
n∘ 3562 Madrid.
B.3 Spain 307
Figure B.39 Alcántara Bridge (Courtesy of Carlos Fernández Casado CFC & ADIF).
Figure B.40 Alcántara Bridge (Courtesy of Carlos Fernández Casado CFC & ADIF).
B.4 Japan 309
B.4 Japan
The first high-speed lines were built in Japan. The population density throughout
Japan has meant that most of the lines have been built on viaducts. The typologies of
bridges for the Shinkansen are varied, although the most singular are the extradosed
bridges, a typology that originated in Japan itself.
310 Appendix B Singular Bridges for High-Speed Railway Lines
B.4.1.2 Description
The bridge consists of several sections, including a continuous extradosed section
with spans of 65 + 105 + 105 + 105 + 65 m. The deck is a prestressed concrete box
with a depth of 2.50 m and is connected monolithically to the shaft of the piers,
which have a height of 12 m above the deck. This bridge is the first application of
the extradosed bridge typology on a high-speed railway line.
Further Reading
Yuyama, K.Y. and Watanabe, J. (1998). Innovative new type of cable-stayed bridge -
Yashiro bridges of Hokuriku Shinkansen line it. In: Prestressed Concrete in Japan.
1998. XIII FIP Congress National Report, Amsterdam, Holland.
https://structurae.net/en/structures/yashiro-bridge.
B.4 Japan 311
B.4.2.2 Description
It is a continuous monolithic bridge with foundations by means of caissons
of 11.00–12.5 m in diameter and 10–16 m in vertical dimension. The deck is a
pre-stressed concrete box with a variable depth that is connected monolithically to
the piers. The track is ballastless.
B.4 Japan 313
Figure B.44 Kumagawa Bridge, view from below (Courtesy of the Zenitaka Corporation).
314 Appendix B Singular Bridges for High-Speed Railway Lines
B.4.3.2 Description
The Bridge on the Tohoku Shinkansen extension was constructed to cross a broad
highway, river, and reservoir near the famous Sannai-Maruyama archaeological
site (Figure 2). It consists of two 75 m spans and two 150 m spans of extradosed
bridges. The bridge with 4 continuous spans has a pneumatic caisson foundation.
The total length is 450 m with spans: 74.18 + 150 + 150 + 74.18 m. The height of
Pylon is 17.5 m, and the material of the stay cable is Epoxy-coated strand. The
bridge features saddles in the pylon in which the strands run through a curved steel
pipe The installation of sliding elastic bearings on four piers except the centre pier,
which is the stationary point of this bridge, reduces the transfer of displacements
into the piers and girder due to expansion and contraction of the girder caused by
seasonal thermal fluctuation.
Further Reading
DYWIDAG Systems International (ed.) (2009–2010). Moderne Technik und Tradition:
Extradosed-Brücke in Japan. In: DSI Info, vol. 17.
Tamai, S. and Shimizu, K. (2011). The long spanned bridge for deflection-restricted
high-speed rail Sannai-Maruyama Bridge. In: Presented in 9th World Congress on
Railway Research Lille France.
https://structurae.net/en/structures/sannai-maruyama-bridge.
B.4 Japan 315
B.5 China
The long viaducts currently being built in China are isostatic sections that are pre-
fabricated in their entirety and assembled directly from previously built sections.
In addition, the orography of mountainous areas and the width of riverbeds have
led to the construction of long-span bridges: arches, cable-stayed bridges and even
the first suspension bridge are undoubtedly a sign of China’s determination to build
a high-speed network that does not shy away from any technical challenge.
318 Appendix B Singular Bridges for High-Speed Railway Lines
B.5.1.2 Description
The bridge deck has two levels: the upper one for a motorway with 3 lanes in each
direction and the lower one for 4 high-speed tracks. The deck consists of three truss
planes, braced together, which are braced by three cable planes to the two main
pylons. The main lateral offset spans have a short span adjacent to them to reduce
rotations at the expansion joints of the structure. The railway tracks are ballasted.
Further Reading
Qin, S. and Gao, Z. (2017). Developments and prospects of long-span high-speed
railway bridge technologies in China. Engineering 3 (6): https://doi.org/10.1016/j.eng
.2017.11.001.
B.5 China 319
B.5.2.2 Description
The main bridge has a total length of 1.615 m and is the longest truss span built
to date. Transversally, the bridge is made up of three trussed beams, braced against
each other, which accommodate the six railway tracks that run along the bridge. The
hangers are rigid steel tubes with a hollow rectangular cross-section of 1.00 × 1.40 m.
The edges of the rectangle formed by the rigid hangers are chamfered to eliminate
possible aeroelastic vibrations. The tracks are ballasted throughout the structure.
Further Reading
Gao, Z., Yi, L., and Xiao, H. (2009). Dashengguan Bridge – the longest span arch bridge
for high-speed railway. In: Presented at: IABSE Workshop: Recent Major Bridges,
Shanghai, China.
Gao, Z., Yi, L., and Xiao, H. (2010). Dashengguan Bridge—the largest span steel arch
bridge for high-speed railway. Structural Engineering International 20 (3).
https://structurae.net/en/structures/dashengguan-bridge.
B.5 China 321
Figure B.49 Nanjing Dashengguan Yangtze River Bridge (Courtesy of China Railway).
Figure B.50 Nanjing Dashengguan Yangtze River Bridge (Courtesy of China Railway).
322 Appendix B Singular Bridges for High-Speed Railway Lines
B.5.3.2 Description
The bridge deck has two levels: the upper for a motorway with 3 lanes in each direc-
tion and the lower for 4 high-speed tracks. The deck consists of three truss planes,
braced together, which are braced by three cable planes to the two main pylons. The
main lateral offset spans have a short span adjacent to them to reduce rotations at
the expansion joints of the structure. The railway tracks are ballasted.
Further Reading
Qin, S. and Gao, Z. (2017). Developments and prospects of long-span high-speed
railway bridge technologies in China. Engineering 3 (6): https://doi.org/10.1016/j.eng
.2017.11.001.
https://structurae.net/en/structures/tongling-road-rail-bridge.
B.5 China 323
Figure B.51 Tongling Yangtze River Bridge (Courtesy of China Railway, Baidu).
B.5.4.2 Description
The main structure is a concrete arch bridge with a span of 445 m and a rise of 100 m,
giving a rise-to-span ratio of 1/4.45. The arch cross-section is a box containing three
cells, with a constant depth of 9.0 m, but varying in width and wall thickness to
improve the transverse stability for the structure. This is important because the
width of the bridge deck is only 13.4 m resulting in a small width-to-span ratio in
this super-long span arch. In each adjacent pier, the two 65-m girders are rigidly
connected to the pier and form a T-shaped rigid frame structure. The overall super-
structure is in the form of a (2 × 65 m + 8 × 42 m + 2 × 65 m) prestressed concrete
rigid-frame/ continuous girder with a total length of 599.6 m. Such a continuous
structure would effectively improve the vertical and transverse stiffness and help
assure effective operation of the high-speed train. The main arch was constructed
by the embedded Concrete-Filled Steel Tube (CFST) scaffolding method. The
40-segment, 445-m-span steel tube truss arch was first erected, then hoisted by
cable crane, and erected by the cable-stayed cantilever method. Concrete grade of
C80 was injected into the steel tubes after the closure of the steel tube truss arch to
make the truss arch a CFST structure, resulting in significant improvement in the
stiffness and strength of the trussed arch.
Further Reading
Chen, B., Su, J.Z., Lin, S. et al. (2017). Development and application of concrete arch
bridges in China. Journal of Asian Concrete Federation 3 (1): 12–19.
Chen, B., Liu, J., and Tabatabai, H. (2019). Recent research and application of arch
bridges in China. In: 9th International Conference on Arch Bridges. Porto, Portugal.
B.5 China 325
B.5.5.2 Description
The Yachihe Railway Bridge is nearly 300 m over the reservoir and has ac main arch
span of 436 m. The twin track structure is composed of a steel–concrete composite
truss. The approach spans on either side of the main arch are composed of a T-beam
of 2 × 61.75 m.
Further Reading
Chen, B., Liu, J., and Tabatabai, H. (2019). Recent research and application of arch
bridges in China. In: 9th International Conference on Arch Bridges. Porto, Portugal.
B.5 China 327
B.5.6.2 Description
The Wufengshan Yangtze River Bridge is designed for both road and rail use, with
a total length of 6.409 km and a main bridge with a single span of 1092 m. A 4-lane
railway plus a two-way 8-lane expressway. It is the world’s first road-high-speed rail
suspension bridge. The distribution of the spans is: 84 + 84 + 1092 + 84 + 84 m.
Further Reading
Qin, S. and Gao, Z. (2017). Developments and prospects of long-span high-speed
railway bridge technologies in China. Engineering 3 (6). https://doi.org/10.1016/j.eng
.2017.11.001.
B.5 China 329
Index
a b
abutments, of bridge 90, 108, 112, 113 ballasted track 186, 187
access on 241–242 definitions, functions, and qualities
continuous deck 32
with dampers located on 223 evolution of 36–37
with single fixed point at one longitudinal design 65–66
203–211 optimised 41–43
with expansion joint partial ballast removal 119, 120
only in structure 90 ballastless track 43–45, 66
in structure and track 90–92 bearings, for bridges
fixed loads arrangement continuous decks 97, 99
forces acting on 138–140 friction 124
loads transmitted by deck 137–138 layout 112
transversal cross section of 137 POT bearings 95, 98, 112
fixed point 91–95 replacement 242, 243
accidental situations 118 simply supported bridges 96–98, 224,
aerodynamic actions, from passing trains 226
123 spherical bearing 96, 98
Alcántara Bridge 20, 306, 307 track-structure interaction 189–190
alert limit (AL) 47 Beipanjiang Bridge 18–20, 324, 325
alignment, track geometric quality 47 bi-block sleepers 37, 38
Almonte Bridge 20, 304, 305 braking force 121–122
antimetric vibration modes 169–171 bridge substructure elements, impact on
anti-noise panel, on bridge 14 126
AP-7 Viaduct Almeria, Spain 82 British high-speed bridges 12
arch bridges Built Bridges 66
tied 85, 86
upper deck bridges 83–85 c
Arroyo de las Piedras Viaduct 223 cable-stayed bridges 13, 86–88
asymmetric switch rail profile 197, 198 cable-supported bridges
Avignon viaducts 8, 284, 285 cable-stayed bridges 86–88
track-structure interaction (contd.) Unstruttal Bridge 9, 10, 18, 19, 272, 273
stresses 194–195, 229–234 upper deck arch bridges 83–85
track joints 196 USA high-speed rail bridges 12
vertical displacement 196, 234–235 U-shaped beams 75
vertical loads 192
track-vehicle dynamic interaction 44–46 v
traction and braking forces 121–122 Veitshöchheim Bridge 7, 266, 267
horizontal forces 131–132 verifications, of track-structure
track-structure interaction 191–192, interaction
195 horizontal displacement 195–196,
traffic loads 231–234
application of 126–128 stresses 194–195, 229–234
for fatigue 128 track joints 196
train load 162–164 vertical displacement 196, 234–235
transitional situations 118 vertical acceleration
transversal behaviour, seismic design control of 62
damping devices plus bearings of deck 48–50
104–105 vertical deformation, of deck 51–53
fixed transversal support 101–102 vertical loads 133–134
relative vertical rotation axis 101 dynamic factor 131
transversal damping systems eccentricity 131
102–104 live loads 119–121, 130–131
transversal wind bearings reactions 136 permanent loads 130
transverse deformation and vibration, of POT bearings 136
deck 53–54 track-structure interaction 192
transverse load imbalances, between rails vertical track stiffness 46
121 vertical traffic loads 186–187
traveller’s landscape 13–15 Viaduct Over AP7 Riudellots de la Selva
truss bridges 82, 83 298, 299
21st century, structural engineering of Viaduct Over River Ulla 302, 303
24 vibration modes 161, 162
Vignoles rail 40, 41
u
U cross sections 67–69 w
UIC leaflet 774-3 54 wind speed 132–134
UIC leaflet 776-3 52, 53, 200 Wufengshan Yangtze River Bridge 88,
Ulla Bridge, Spain 21, 24 328, 329
tubular-framed 89
Ultimate Limit State (ULS) 31, 117 y
Universal dynamic train A 167 Yachihe Bridge 326, 327
Universal dynamic train B 167–168 Yashiro Bridge 310, 311
WILEY END USER LICENSE AGREEMENT
Go to www.wiley.com/go/eula to access Wiley’s ebook EULA.