‘IWWWWWQ

AN,‘
Bridge Launching

- Marco Resignoli
Parma, Italy

~ '1' . ThomasTelford
 Quay, London E14 4JD.
Published by Thomas Telford Publishing, Thomas Telford Ltd, 1 Heron
URL: www.thom aste1ford.c om

Distributors for Thomas Telford books are

USA: ASCE Press, 1801 Alexander Bell Drive, Reston, VA 20191-440
0, USA Contents
Chuo-ku, Tokyo 103
Japan: Maruzen Co. Ltd, Book Department, 3—10 Nihonbashi 2—chome,
Books and Journals, 648 Whitehors e Road, Mitcham 3132, Victoria
Australia: DA

First published 2002

Also available from Thomas Telford Books

Klein. 2nd edition, 1999. ISBN
Cable stayed bridges. R. Walther, B. Houriet, W. Isler, P. Moia and J-E;
0 7277 2773 7.
2. ICE, Highways Agency and
Current and future trends in bridge design, cbnstruction and maintenance
6.
International Association for Bridge Maintenance and Safety, 2002. ISBN 0 7277 3091
G. L. England, N. C. M. Tsang and D. I. Bush, 2000. ISBN 0 7277 2845 8.
Integral bridges.
Management of highway structures. P. C. Das, 1999. ISBN 0 7277 2775 3. iX
(eds), 2000. ISBN 0 7277 2774 5. Preface
Manual of bridge engineering. M. J. Ryall, G. A. R. Parke and J. E. Harding
Routes et Autoroutes,
Post-tensioned concrete bridges. Highways Agency, Service D’Etudes Techniques des
Research Laboratory and Laboratoir e Central Des Ponts et Chaussées, 1999. ISBN 0 7277 2760 5.
Transport
Prototype bridge structures. M. Y. H. Bangash, 1999. ISBN 0 7277 2778 8.
Xi
0 7277 2881 4.
List of symbols
Steel bridge strengthening. Highways Agency and W. S. Atkins, 2001. ISBN

1. Introduction
Cover photo: the Steine Bridge (courtesy Greisch) 1.1. Introduction to incremental launching 10
1.2. Incremental launching of PC bridges 13

. 11.4
15

y.»
A catalogue record for this book is available from the British Library 1.3. Incremental launching of composite bridges

. h.
;
J :W we gains/«,2:
_‘_,___,.__
1.4. Incremental launching of prestressed composite bridges 16

.4.__
.3,
ISBN: 0 7277 3146 7

'-——--—-.—.__.*‘_,____,_,——————
« g
, \ .._ «r . p - ~ . -. 4 -‘-. Av?
Incremental launching of bridges 19
2.
© Thomas Telford Limited 2002 2.1. Structural system of the superstructure 19
27

1-_
2.2. Static system and seismic design

a4».-..i~.»-..c-:r:-
.g;
Designs and Patents Act 29

‘=-.'-‘
All rights. including translation, reserved. Except as permitted by the Copyright, . 2.3.. _ Geometric- constraints... ., , .. .
in any foi'm or

. m; m,<‘;_. final”...
1988, no part of this publication may be reproduced, stored in a retrieval system or transmitted 32

.: 1;
2.4. Launching techniques

-u.ml_o———-_H___._._—m—
permission of the
by any means, electronic, mechanical, photocopying or otherwise, without the prior written

«1' . -., \-‘(

Publishing Director, Thomas Telford Publishing, Thomas Telford Ltd, 1 Heron
Quay, London E14 4JD. 2.4.1. Launching of light superstructures 33

u fl 3,.“
Friction launching of heavy superstructures 43

. -‘ _’u Ww‘u
2.4.2.

. j, . ‘ «v wnr.
This book is published on the understanding that the author is solely responsible
for the statements made and
2.5. Launch bearings and guide devices 53

': :4:;-:,.-m»:~:'vr':.:1w-iNew,
that such statements and/or 55
opinions expressed in it and that its publication does not necessarily imply

m..." 4....p...—.a..._,,.....—— -——h-»~_a_m__.n_¢—n—H——»—a-—4———

has been made to ensure that 2.6. Launch and lock forces
opinions are or reflect the Views or opinions of the publishers. While every effort 58
the statements made and the opinions expressed in this publication provide a safe and accurate guide, no 2.7. Correction of launch stresses

-
liability or responsibility can be accepted in this respect by the author or publishers. 64

n“: . y :.
2.7.1. Launching nose
88

'
2.7.2. Stayed front system

_
K
94
.
Typeset by Keyword Publishing Services 2.7.3. Temporary piers
4 -..M...»
Printed and bound Vin-Great Britain by MPG Books, Bodmin 97
. ‘ ‘
'
2.7.4. Deck launching onto arches
2.8. Launch stresses in the piers 103
2.9. RTM method for analysis of the continuous beam 110
2.10. Shear and bending moment envelopes 118
2.11. Verification of design assumptions by monitoring 120
2.11.1. Elastic modulus of concrete 121
.

. 2.11.2. Thermal correction of results 123

. ‘ ‘\
. *fiififii'v'lflLJE‘TKf- ~~ .
 WWWGIW"WWWWWWW
vi BRIDGE LAUNCHING CONTENTS. vii

2.11.3. Coefficient of thermal expansion of concrete 123 4.6. Incremental launching of the concrete slab onto the steel beam 5; 257
2.11.4. Shrinkage 123 4.6.1. Technological features 258
4.6.2. Advantages with respect to in situ casting 265
3. PC bridges 125 4.6.3. Stress analysis 267
3.1. Presizing of the superstructure 125
3.2. Deck segmentation and yard organization 128 5. Prestressed composite bridges 271
3.3. Casting phases 131 5.1. Introduction 271
3.3.1. Monolithic casting 132 5.1.1. High performance concrete 272
3.3.2. Two—phase casting in a single formwork 141 5.1.2. Lightweight concrete 275
3.3.3. Two-phase casting in a double fofmwork 142 5.1.3. Introduction to prestressed composite bridges 276
3.3.4. Segment extraction 146 5.1.4. Cross-sectional efliciency 279
3.3.5. Curing yard 149 5.2. PCS bridges with stiflened-plate webs 281
3.4. Assembly and launching of precast segments 151 5.2.1. Web-slab lower node 283
3.5. Launch bearings and guide devices 153 5.2.2. Web instability 284
3.5.1. Pier-cap arrangement 155 5.2.3. Launching 286
3.5.2. EiTects of vertical misalignment of launch bearings 156 5.3. PCS bridges with corrugated webs 288
3.5.3. Local stresses above launching bearings 165 5.3.1. State of stress in the corrugated webs 291
3.6. Eifects of time-dependent behaviour of concrete 170 5.3.2. Cross-sectional deformability 294
3.6.1. Shrinkage 171 5.3.3. Connection between webs and slabs 296
3.6.2. Creep 172 5.3.4. 7 Web instability 299
3.6.3. Prestressing steel relaxation 173 5.3.5. Fabrication of corrugated plate girders 305
3.6.4. Eflects on bridge launching 174 5.3.6. Field activities 307
3.7. Prestressing 178 5.3.7. Launching 308
3.7.1. Launch prestressing 179 5.3.8. Case studies 311
3.7.2. Dimensioning of launch prestressing 190
3.7.3. Service prestressing 194 References and bibliography 319
3.8. Reinforcement 206
Index 333
Composite bridges 209
4.1. Conceptual design and deck presizing 210
4.1.1. Multigirder systems 211
4.1.2. Two-girder systems 212
4.1.3. Composite box girders 217
4.2. Segmentation of the steel girder 221

.__ ‘ ummmmfiv—flmflMmflfi
4.3. Assembly yard organization 225
4.4. Launching of the steel girder 229
4.4.1. Launching bearings 231
4.4.2. Launching nose 235
4.4.3. General launch stresses and instability 236
=~ -‘ -,:i:: _.‘ w

4.4.4. Local launch stresses and instability 239

‘51::-

4.5. In situ casting of the concrete slab

is‘w's-

246
awn—n
1:

4.5.1. Casting methods and cracking control 247

. .
Mama

4.5.2. Longitudinal prestressing of the concrete slab 254

' 1‘:f"“1~'
. .‘ .;

4.5.3. Transverse"prestressing of’ the concrete slab ' 257

.,_. 3
W
,.
4%“;
:»:.ii:..,..,..i2!.i;._-cn
 Preface

*‘v- -
‘ "S'w
” n-—‘—._—_.M’“’
‘
__ 3-
M....l.c_
In recent decades, the availability of new means of calculating and of new materials
has permitted a rapid evolution of bridge design and construction methods, under
pressure from the competitive requirements of the international construction

._,_
industry.
The increase in labour costs has played a fundamental role in the indus-
trialization of bridge construction methods. However, industrialization requires

. -
investments in terms both of research and of specialized construction equipment,
and these investments must be amortized. This requirement has resulted in a
marked specializatiOn of operators, and the advent of bridge subcontractors
equipped for just one or two construction methods, and often for just one structural
material (either prestressed concrete or steel).
Contemporaneously, a changing emphasis on the environment has modified
demand and led designers to plan infrastructures more and more carefully inserted
_ into,...their.-sites...,lhe,ruse..-.o,f.,iunneiasigrfiI&dy,reduces the impact of new
infrastructures on the environment; however, costs apart, this design philosophy
has to face the difficulty of passing under existing obstacles. Rivers, channels and,
to a lesser extent, existing infrastructures, require construction of bridges.
Infrastructures deeply inserted into the environment result in short bridges. In
hilly areas, bridges are often directly inserted between tunnel inlets to span the
obstacle. In spite of their modest length, however, modern bridges must fulfil
several requirements. They must have low maintenance costs and minimum effects
on the environment and the obstacle to be overpassed. They must also be easy to
build and competitive in terms of construction costs.
The length of a bridge has a strong impact on its construction cost. In indus-
trialized countries, labour costs can easily exceed half of the total construction cost
of the bridge. Reduction of labour costs requires industrialization, with related
investments that have to be amortized. If the bridge is short, amortization of
equipment costs (not only the initial investment, but also transport, assembly,
and final dismantling) can be prohibitive. On the other hand, labour costs resulting
from a poorly industrialized yard can be prohibitive as well. The solution is not
easy.
x BRIDGE LAUNCHING

These new demands stimulated research on new structural systems and bridge
construction methods that combine low labour costs and low investments for
adaptable equipment that can be amortized in several jobs. Finally, a better
evaluation of the total cost of a bridge (construction and maintenance during Symbols
its service life) singled out the intrinsic quality of the bridge and its good
adaptability as primary objectives. These issues produced substantial innovations
in medium-length bridges.
For contractors who operate in the international market, innovation and export
capability are two basic and tightly interconnected requirements. The ability to
export depends on the ability to propoSe competitive alternatives, which in turn
derives from cultivating research and innovation. In recent years, the contraction of
the public works market and the resulting tighter competition has led contractors to
reduce overhead expenses, and with them research, and to seek a systematic reuse of
available equipment. This, in turn, is prevented by innovative ideas requiring new The symbols listed below have been used in the text. Some passages have required
investments. . further symbols, described in the respective sections.
In spite of that, however, innovation is an inevitable requirement for the
international activity of high-tech countries, which cannot Compete with emerging a distance between web centres on the top slab axis
countries in labour costs and must therefore exercise their know-how to af base of the fold in a folded plate web
propose advanced alternatives. Some emerging countries have already reached ag peak ground acceleration (PGA)
high technological levels, so, to preserve their export capacity, the most evolved ag,C peak ground acceleration during construction
nations cannot but continue the research of new ideas. Incremental launching of A cross-sectional area
bridges is a logical and effective response to most of these requirements. Ab area of the bottom slab or flange, area of the braces of a temporary pier
The aim of this book is to merge the structural, organizational and economic c cross-sectional area of the concrete slab
aspects of the incremental launching of bridges. These aspects are tightly related f net area of each flange
and must be considered as a whole in the design of such structures. cross-sectional area of the nosedeck joint
The author is a freelance bridge consultant who has spent many years as project area of each pillar of a temporary pier
manager for bridge contractors, leading workgroups composed of designers and S cross-sectional area of the steel girder _
field engineers, and verifying each time that a broad view can result in a successful W c‘ragslgeerisfiai“* area of a web
approach to high-tech structures. This book is addressed to bridge contractors, distance between web centres on the bottom slab axis, breadth of
technical and design-oflice engineers, and owners, to help them to widen their concrete in cross-section, flange width, flange segment length
partial knowledge of the subject. bb breadth of the bottom flange
Chapter 1 introduces the incremental launching construction method for PC, bf step of the folds in a folded plate web
composite and prestressed composite bridges. Chapter 2 deals with the techno- bnf breadth of the neo-flon plate
logical and design aspects common to these three different types of launched bt breadth of the top flange
bridges. Chapter 3 is devoted to prestressed conCrete bridges, while Chapter 4 B total breadth of the top slab
deals with conventional composite bridges, whose launching is similar to the launch 6 web sloped depth
of steel bridges. Finally, Chapter 5 is devoted to prestressed composite bridges. Cd cross-sectional property related to warping (inertial term in the BEF
equation)
Cf friction coeflicient
CR, breakaway friction coeflflcient at bearing i
Cf; friction coeflicient of the contact plates of the friction launcher
C, constant ‘
Ct torsional stiffness
xii BRIDGE LAUNCHING SYMBOLS xiii

theoretical breadth (from web axis) of the side cantilever of the cross— prestressing force in the top slab

SPUEE
section thrust force
distance of the bottom slab axis from the cross-sectional centroid upper prestressing force in the nosedeck joint section
distance of the top slab axis from the cross-sectional centroid effective shear modulus of steel

wQBQCD
(n
longitudinal flexural stiffness of a cross-sectional plate average percent gradient of the launch surface
transverse flexural stiffness of a unit length of web Shear modulus of steel

U)
eccentricity theoretical depth of the cross—section (distance between the centre lines of
eccentricity of the additional prestressing in the front zone of the slabs or flanges)
superstructure depth of the cross—sectional central core

w
o
downward eccentricity of the resulting force hcon average thickness of concrete
efficiency of material depth of the fold
upward eccentricity of the resulting force height of the pier ‘
elastic modulus height of the tower in the cable-stayed front system
elastic modulus of the braces of a temporary pier net height of web plate (measured clear between flanges)
mean elastic modulus of concrete during the heating period total depth of the cross-seCtion (distance between the extreme edges)
mean elastic modulus of concrete during the cooling period total depth of the steel girders (distance between the extreme edges)
elastic modulus of the launching nose moment of inertia about the centroid of the cross-section
elastic modulus of steel moment of inertia of the nosedeck joint cross—section
axial deformation modulus of a folded steel plate moment of inertia of the launching nose
flexural stiffness of the superstructure BEF modulus of foundation, coeflicient for seismic demand during
flexural stiffness of the launching nose construction, generic coefflcient or constant
ultimate stress constants or coeflicients
cube compressive strength of concrete at i days of curing length of the span
28-day characteristic compressive strength of concrete length of the bank span of the continuous beam, length of patch loading
tensile strength of concrete along the bottom flange edge
effective tensile strength of concrete critical length of the front cantilever
ultimate strength of steel length of the launching nose
characteristic yield strength of steel ien'gth“of“'the neo-flon plate
general force " effective length, length of the equivalent hinged rod
additional prestressing in the front zone of the superstructure length of patch loading along the Web edge
prestressing force in the bottom slab inverse of Poisson’s ratio (l/v)
theoretical friction force bending moment
friction force opposed by the formwork negative moment in the first support section at the end of launch
design friction force design bending moment
horizontal force due to the longitudinal gradient of the launch surface moment in the cross-section above support I
hemisymmetrical force moment in the cross-section above support I due to cantilevered masses
breakaway force at support i ultimate cross—sectional flexural strength
lower prestressing force in the nosedeck joint section axial force in a member, compression force
lock force minimum prestressing force applied to the anchor blocks
maximum prestressing force uniformly distributed service load
minimum launch prestressing probability of exceedance of the design seismic event during service life
parabolic prestressing force p€,C probability of exceedance of the design seismic event during construction
force in the stays of the cable—stayed front scheme, longitudinal forces in hemisymmetrical component of the unit load
.. the slabs 7 Ppar uniformly distributed load produced by parabolic prestressing
 BRIDGE LAUNCHING SYMBOLS xv
xiv

uniformly distributed superimposed dead load distance of the centroid from the upper edge
Euler’s critical load angle, coeflicient of thermal expansion, dimensionless progression of
nominal static ultimate shear strength of headed stud connector launch
Critical dimensionless progression of launch
nominal static ultimate shear strength of headed stud connector in the
angle, restraint coeflicient
presence of tension
difference between two values
uniformly distributed dead load
loss of prestress at time t
uniformly distributed dead load of the launching nose
orthogonal strain
unit weight of the steel structure of a composite bridge
creep strain
dead load of the superstructure
shrinkage strain
radius of gyration of the cross-section about the centroidal axis, radius
friction angle, creep coeflicient
of curvature, uniform friction load
rotation
local buckling strength
specific weight
design support reaction
friction force amplification factor for design of launch devices
hemisymmetrical component of the support reaction
design safety factor for loads
reaction at support I
design safety factor for materials
ultimate support reaction applied through bottom flange plate
design safety factor from uncontrolled sliding
symmetrical component of the support reaction
vertical deflection, imperfection constant
radius of curvature of the equivalent tendon
distortional component of the cross—section deformation
support reaction of bearing i
torsional component of the cross-section deformation
support reaction at the friction launcher
BEF characteristic length, slenderness ratio
local yielding strength
BEF characteristic length for the calculation of distortional edge stresses
section modulus
development ratio of the folded web
modified section modulus
Poisson’s ratio
state array in section j
torsional rotation
reduced state array of support section J
flexural efliciency of the cross-section
time, thickness
structural efficiency of design
thickness of the bottom slab or flange _ L1 1 L .. .

construction duration
Uly LllUgUIldl SLI‘CSS

compressive stress (positive)

return period of the design seismic event
critical buckling stress
return period of the design seismic event during construction
critical vertical stress
thickness of the top slab or flange
ideal biaxial stress
web thickness
tensile stress in a prestressing tendon
torsional moment, temperature
initial tension in a tendon
transfer matrix between the cross—sections j and k
allowable contact pressure on PTFE
reduced transfer matrix between the support sections J and K
residual tensile stress (negative)
ultimate tension force in headed stud connector
' " ' tensile stress (negative)
shear force
ultimate compressive stress
vertical misplacement of the launching bearings
total length of the neo-flon plates on a launching bearing
hemisymmetrical component of vertical misplacement of launch
' tangential stress
bearings
critical tangential stress
symmetrical component of vertical misplacement of launch bearings
critical tangential stress for global buckling
total plastic work for a complete corrugation
critical tangential stress for interactive buckling
distance of the centroid from the lower edge
critical tangential stress for local buckling
1, xvi BRIDGE LAUNCHING

75d design ultimate shear stress

TTim Timoshenko’s critical tangential stress
1.0 contemporaneity factor for launching bearing breakaway force
|. Introduction

The first prestressed concrete bridge was built in 1941. From the end of World War
11 onwards prestressed concrete bridge construction developed quickly, thanks to
the pioneers of the new technology: Guyon, Freyssinet, Leonhardt, Magnel,
Morandi, Morsch and Ross, among others.
In the past, the design of almost all these new bridges was limited to quasi-
isostatic systems, since analysis of statically indeterminate structures was not
compatible with the means available for calculation. For many years, the basic
criterion in the design of prestressed concrete (PC) bridges has been to assure the
possibility of simple static analysis. So, multispan bridges built on ,falsework used
the simple support scheme, the first balanced cantilever bridges in the 19503 were
hinged at mid-span, and the first superstructures built with movable shuttering
systems in the 19608 were simply supported at the piers or articulated at the
counterfiexure points.
From“*tlrer‘i96fisr‘onwards;atheeextraordinarfyepregress in ccmputing techniques,
the technological advances in prestressing, and the improved knowledge of materi-
als extended the possibilities in the design and, above all, in the analysis of PC.
These new possibilities stimulated technological innovation and the' definition of
new construction methods. .
In the creative thinking of these 60 years, many construction methods were
developed to guarantee adequate margins of competitiveness of PC over steel
constructions. At the same time, however, the cost reduction of steel plates, the
better organization of workshops, new splice techniques, and new field assembly
methods reduced the lower threshold of competitiveness of steel bridges to spans
that can be easily reached with PC as well. Finally, new combinations of reinforced
concrete slabs and steel girders led to composite sections explicitly addressed to
these borderline spans. This further stimulated research in the concrete and steel—
work industries, and created new connections between these different technologies.
New construction methods took advantage of the most recent technological
advances both in the design of new equipment and in widening the field of applica-
tion of already familiar techniques. So, although the idea of bridge launching
 2 BRIDGE LAUNCHING 3
INTRODUCTION

derives from Palaeolithic technology (think of a tree—trunk), its application to PC These obstacles were gradually overcome. The advances in prestressing tech-
bridges has been made possible by the availability of launching techniques already ‘ nology lightened the superstructure, made it elastic and less subject to cracking,
tested in the field of steel bridges, and of new materials such as Teflon. permitted joining subsequent segments, and allowed introduction or removal of
For many decades, the light weight of steel structures had permitted their launch tendons according to need. The commercial availability of finite—element software
by means of winches and lubricated wooden slides or steel rollers, with a friction for personal computers eased continuous beam analysis in the subsequent support
that, although considerable, did not cause inadmissible stresses in the piers or configurations assumed during launching. The introduction of the steel—Teflon
require excessively expensive launching equipment. The high flexural efiiciency of contact along sliding surfaces permitted reduction in launch friction. Finally,
the steel girders and the ability of the material to work both in tension and in construction equipment reached the technological leVel necessary to move huge
compression eased resistance to the temporary launch stresses and avoided the masses with the required precision.
need for overdesign with respect to service requirements. Finally, launching of the None of this was available in 1959, the year that saw a first attempt to launch
steel girder had many advantages over conventional assembly, as fitting and joining precast concrete segments for the bridge over the Ager River [1]. The superstructure
of the steel elements was performed on the ground instead of in the air, without was built on a timber falsework, which supported precast segments built on one
affecting the areas below the bridge, and with greater safety and simplicity (Fig. 1.1). bank and launched one by one into their final position (Fig. 1.2). The joints
However, compared with traditional hoisting techniques for steel girders, costs between adjacent segments were cast in situ, and prestressing was obtained by
were higher as a result of the operational slow-down caused by the availability of means of long external tendons placed within the box cell and tensioned from the
only one working yard. As a consequence, the adoption of incremental launching two ends of the superstructure. The bridge segments were moved by sliding them
was initially limited to bridges high above the ground or in inaccessible areas. along a wooden rail by means of wooden skids lubricated with engine oil. The
In the eyes of the PC bridge pioneers, some of these aspects were less critical, friction coefiicient was high, but without negative effects as the segments were
others even promising. In fact, construction duration and yard organization for PC moved one by one.
bridges are different from those for steel bridges, and the cost of labour and This first trial demonstrated how it would have been simpler to avoid the false-
equipment is so high that every possible alternative must be carefully examined. work and to launch the whole superstructure directly on to the piers by taking
Front launching of a superstructure built on the ground promised savings in both
advantage of its fiexural stiffness. This concept was applied not much later for
equipment and labour, but in practice this was limited by the much greater weight the Rio Caroni Bridge [2,3]. The falsework was eliminated and the whole length
of concrete and its low tensile strength.

. Fig- .1-1- Incrrmeml launchingeitheired girder of ,rh,e,,Schroteral,Bridge (author) , Fig. 1.2. “Precast segments of the Ager Bridge (courtesy Leonhardz‘)
 4 BRIDGE LAUNCHING INTRODUCTION 5

of the superstructure was built behind an abutment by joining precast segments

built in a casting yard and advanced to the assembly position by sliding along a
wooden rail, with joints cast in situ.
On completion of segment assembly (Fig. 1.3), the superstructure was prestressed
with external tendons distributed within the box cell so as to obtain a centroidal
force. Their anchorages were arranged in the front end—diaphragm of the super-
structure, and at the opposite end they were deviated around a cylindrical concrete
block, whose jacking from the rear end of the superstruCture produced the required
prestressing force.
The superstructure was launched with the help of a front steel-truss nose. One
temporary pier was built in each span to halve the launching span (Fig. 1.4). The
bridge was launched above movable bearings, and at the end of each launch stroke
it had to be lifted to relocate the bearings (Fig. 1.5).
On completion of launch, axial prestressing was not sufiicient for the design
loads, so the external tendons were made eccentric by vertically shifting their
mid-span and support deviation points. At the same time, the rear deviation

lstphase: construction 0? segments Fig. 1.4. Launch of the Rio Caroni Bridge (courtesy Leonhardt)

construction , , , construction 0F
0? segments QOSI'IZIOHIHQ Final piers
cylinder was released. Tendons were finally drawn close to the webs, fastened to
ELI-um them, and protected by means of prepack concrete (Fig. 1.6).
Although conceptually ingenious, the shifting of tensioned tendons was a very
W complex and expensive operation. This stimulated the idea of introducing
undulated tendons ”on completion of launch to correct a permanent axial launch
prestressing,and this concept has characterized most launched bridges since then.
2““ phase: axial prestressing 0? the deck Also the need for complete assembly of the superstructure behind the launch
abutment was quite limiting, and this stimulated the idea of match-casting longer
launching nose construction 0? deck segments directly behind the abutment and joining launch tendons every two
temporary piers
or three segments. Launching of the superstructure thus became incremental.
That was how, in 1965, the incremental launching method (construction of long
deck segments behind an abutment alternating with the launch of the whole deck
section) was applied for the first time in the construction of a bridge near Kufstein,
Germany. Once the new segment has been joined (if precast) or has reached the
necessary strength (if cast in situ) and launch prestressing has been introduced, the
whole superstructure is launched forwards by a length equal to the segment. This
gives free access to the assembly yard or to the formwork, another segment can be
built in contact with the rear one, and so on until completion of the superstructure
(Fig. 1.7). From one point of View, this construction method simulates a horizontal
‘slip form’, with the difference that the form is fixed and the superstructure itself
moves with respect to it, in a process that is somehow similar to a cadenced
_ Fig.1..3. Construction, schemeoftheRio Caroni Bridge extrusion.
 6 BRHXHELAUNCHING INTRODLKHTON _ 7

i— Final bearing GUIDE DEVICE

“III- IIIII

LAUNCHING steel plates

rm

spacer
-------c-- ............
......

- prestresgng Mres
llllllllllll Illlllllllll or strands

l Inl
27i LIFTING ”'Iili i l l concrete web

FINAL ARRANGEMENT -
Asllfl_ IIIII

reMForcement
BEARING RETURN
Mre net

MJecfion pme

KIIII- lllll

LUVERING a; I-
a" V- _._-_ -
I

;;_,gfl—pre-pack concrete

mortar
Fig. 1.5. Discontinuous launch with mobile bearings
anchor bott

In this first experience of incremental launching, the use of a specific launch bush-hammered surFace
prestressing was avoided by reducing launching spans with two temporary
piers per span and by accepting edge tensile stresses, temporarily controlled by
reinforcement and permanently avoided by parabolic prestressing (Fig. 1.8). concrete web

In subsequent years, there have been many improvements in this construction

Fig. 1.6. Final arrangement ofpresiressing tendons in the Rio Caroni Bridge
technique. Sliding is facilitated by neoprene-Teflon (neo-fion) plates inserted
between the superstructure and fixed launch bearings, thus eliminating the need
for lifting. The front cantilever support devices range from steel or concrete launch— beam solver algorithms permit easy evaluation of the stresses and deflections of the
ing noses to cable-stayed systems. Thrust devices have evolved towards compact superstructure in every launching phase as well as of their envelopes.
friction units that guarantee large safety margins and smooth travel, thanks to the Although the incremental launching construction method was originally con-
combined use of high—pressure hydraulic and digital controls. Fast continuous ,..ceived_ for bridges some 100 m long, the search for the greatest cost reduction
8 ” BRIDGE LAUNCHING INTRODUCTION 9

Fig. 1.8. Launch of the third box girder 0f the Kufstein Bridge: the rectangular piers
are temporary (courtesy Leonhardt)

rotation have won a role in critical applications and in many cases they represent
the least expensive and most elegant solutions [4].
In the last two decades, there have been many ingenious applications of launch
techniques for PC bridges. These range from beams supported by arches, curved
superstructures, cable—stayed bridges launched symmetrically from the two banks,
or built along one bank and then rotated into their final position, or built by
launching the superstructure above temporary piers and then suspending it once
Fig. 1.7. Incremental launching it has reached the final position. .
There has also been research and development in the field of steel constructions.
A better knowledge of instability has led to lighter structures, often characterized
extended its use both to short bridges and to very long bridges. This construction
by thin web plates. This required improvements in the launch bearings, which must
method is now applied to small bridges, where labour cost can be lowered without
follow the rotations of the superstructure with uniform support reactions. This also
major investments, and to large bridges, where greater labour cost savings
resulted in lighter noses and specialized systems for transfer of launch thrust.
are achieved by massive field industrialization. In both cases, investments for
The evolution of launch bearings was particularly significant in conventional
specialized construction equipment can be adapted to the specific demand and
composite bridges, where the steel girder is launched first and the deck slab- is
can be amortized in more bridges, as equipment is very adaptable [4].
cast in situ on completion of girder launching. The large deflections and rotations
The technological advances of incremental launching permitted experimentation
of these flexible girders prior to cross-section completion result in several specific
with new types of monolithic handling of the superstructure. So, though incremen-
requirements fOr launch devices.
tal launching still represents the most common handling technique and its evolution
In spite of these advances, the diflerent cost evolution of concrete and steel
has influenced the development of others, both the symmetrical, approach and
plates still suggests the use of reinforced concrete for the compression areas of
 10 BRIDGE LAUNCHING INTRODUCTION 11

medium-span steel bridges. A concrete deck slab is particularly suitable for simply resist both tensile and compressive stresses, but this simplicity is only apparent, as
supported spans; however, the continuous beam offers several advantages, and compressive stresses cause instability. In both cases, therefore, a structural design
control of the upper edge tensile stresses in the support areas is therefore an issue for the permanent actions may be inadequate for self—resisting construction if the
common to most composite bridges. temporary stresses are very different from the final ones. On the other hand, if
The control of longitudinal tensile stresses in the concrete slab requires specific construction stresses are so significant as to require design adjustments, this
segmental casting sequences for the slab, which cause high costs and long construc— might result in a structure that is overdesigned with respect to service requirements.
tion times. This suggested incremental launching of the concrete slab onto the steel Design for construction stresses depends on the level and distribution of
girder. By using the same thrust devices first for the steel girder and then for the the temporary internal actions, which are governed by the dead load of the
concrete slab, it is possible to build the composite bridge in a very restricted yard superstructure and the temporary structural systems used for construction.
and to avoid any interference with the areas below. Experience gained in the launch When construction takes place with the same static system of the permanent
of PC bridges was invaluable for this new step forward. structure, a load increase after completion of construction (for instance, application
Research in PC also led to external prestressing. When external tendons are of superimposed dead load) results in a proportional increase in internal actions.
used, the web thickness of the box girder is no longer restrained by the need to Other types of load (live load, wind) or actions (temperature gradients, imposed
contain and to deviate tendons. This permits lightening the webs and enhancing displacements, settlements) may result in non—proportional increases or sign
cross-Sectional eificiency. However, shear forces diCtate the minimum thickness inversion of internal actions. However, this second family of effects is always
of concrete webs, and steel plate webs are much more efficient for this purpose. superimposed on the basic effects of dead load, and the resulting variations in
This led to the concept of prestressed composite bridges. In these new box internal actions are relatively small. .
girder typologies, two concrete slabs resist the flexural stresses thanks to Variations in the static system have a more profound influence, especially when
prestressing, and tendon deviation reduces the shear stresses in the webs down they result in sign inversion of bending moment in areas of the structure that are
to levels that can be resisted by thinnest steel plates. Compared with the PC box loaded in a stable manner on completion of construction. A mid-span section
girders, these prestressed composite sections are much more efficient, with crossed by prestressing tendons nearby its lower edge is hardly compatible with
related savings in weight, concrete, reinforcement and prestressing. These new negative moments. A support section crossed by prestressing tendons near its upper
sections require Only about a quarter of the steelwork of conventional composite edge will hardly resist positive moments. As any operation on tensioned prestres-
sections. sing tendons is expensive, less eccentric tendons must be used to enhance the
Advance in the launch techniques for PC and steel bridges led to the launch structure adaptability to these temporary load conditions, and this results in a
of prestressed composite bridges. Incremental launching is probably the most less efficient final prestressing scheme. Similar conclusions can be drawn for the
competitive construction method for medium-span box girder bridges. flange plates of the steel girders of conventional composite sections.
As cOnstruCtion "stresses" depend on the deadload of the superstructure and the
|.|. Introduction to incremental launching static system, both of them must often be adjusted. Dead load should be as low as
Construction methods that take advantage of the self-resistance of structures are possible during launching. In composite bridges, the concrete slab is a prime
based on a rather intuitive principle. As several structural components are designed component of the section weight and it is therefore added only after completion
for permanent stress levels that are higher than those reached during conventional of the launch of the steel girder. In prestressed composite bridges and conventional
construction, this temporary overstrength can be exploited to simplify construction PC box girders this is usually impossible. However, appendages can be applied after
and to reduce the carrying demand on construction equipment (and the related launching, temporary support devices such as launching noses and stayed systems
costs). can be used to adjust the internal stresses in the front zone of the superstructure,
The use of structural strength during construction results in significant economic and temporary piers may diffusely decrease launch stresses in the most delicate
advantages for both PC and steel bridges. HoWeVer, several aspects should be taken cases.
into account to correctly estimate the economic implications of self—resisting The economic advantages of self-resisting construction methods often exceed the
construction methods. costs resulting from the design requirements of construction phases. However, a
In bridges made of concrete, the structural material is unable to resist significant broad vision of the overall design economy is necessary with these types of
tensile stresses. The use of prestressing, limits or avoids tensile stresses; however, structures.
this stress adjustment is obtained with tendons whose alignment depends on the Segmental bridges are usually erected [5,6] in one of five ways: span-by-span
distribution of the internal actions to be corrected, and can be modified only by construction with truss, falsework, or movable «shuttering system; progressive
-- expensive operations. In steel-bridges, the structural material is-inherently- able to placement; span—by-span lifting; balanced cantilever construction; or incremental
 12 BRIDGE LAUNCHING INTRODUCTION 13

launching and derived methods. The first three methods do not involve Since construction stresses and permanent stresses are similar, this construction
self-resistance of the structure during construction. method does not require particularly refined prestressing schemes. Therefore,
although it dates back to the beginning of PC bridges, the advances in prestressing
in recent decades have not produced substantial changes in balanced cantilever
a With span-by-span construction, the PC superstructure is cast in situ in a
construCtion, whose optimum field of use is still limited to rather long spans.
movable shuttering system that proceeds from one abutment towards the
The use of the balanced cantilever method has now widened to include the
opposite one. Construction joints are usually placed at the counterflexure
stayed-cantilever construction of arch bridges and the progressive erection of
points. The form carrier provides factory—type operations transplanted to the
cable—stayed bridges. The conceptual solution is the same [5], with temporary
job site. It maybe supported on the piers, from the edge of the previously
stays in arch construction and permanent stays in the cable-stayed bridges.
completed construction or from the pier. The same principle has been applied
in conjunction with precast segments that may be assembled on a steel truss to
make a complete span. Prestressing tendons ensure the assembly of the various
l.2. Incremental launching of PC bridges
segments in one span while maintaining full continuity with the preceding span.
Compared with traditional construction techniques, the inCremental launching of a
' This construction method is rarely applied to steel girders.
PC bridge cast in situ is revolutionary [7] from many points of view:
o In principle, progressive placement is similar to the span-by-span process,
because the construction starts at one end of the structure and proceeds
continuously'to the other end. Its origin, however, is derived from the cantilever 0 The standard equipment is simple, being limited to a formwork supported on
concept. Precast segments are placed in successive cantilevers on the same side the ground, a thrust system and special devices (a launching nose that may be
of the same pier rather than by balanced cantilevers on both sides of the pier. integrated with temporary piers or a stayed front system) that reduce the
Temporary stays deviated by a steel tower placed at the pier limit the cantilever launching stresses.
stresses and deflections. This method is seldom applied to superstructures cast in 0 The formwork, the batching plant (if any) and the steel cage assembly template
situ or steel girders. . V are located in a small yard, so the stocking and handling of loose materials by
o Span-by-span lifting is mainly used for the steel girders of composite sections to means of a tower crane are considerably simplified.
be completed with a concrete slab cast in situ. It is also used for PC l-girders, o The superstructure is built in a fixed, sheltered location. Each operation, from
while the dead load of PC box girders is excessive for this construction method. the steel cage assembly to concrete casting or prestressing, is much simpler and
0 The balanced cantilever proCess is primarily suitable fOr long spans where safer than on a falsework or at the tip of a cantilever, and may be organized in
construction activities for the superstructure can be planned at the deck level parallel rather than in series.
and without the use of extensive falsework. Long viaduct structures with 0 A highmlevelof qualitycofintrol is indispensable at each stage of construction,
relatively short spans are better processed using either span-by—span methods since mistakes or irregularities cause launch difficulties. The construction of
or inCremental launching. This procedure cantilevers Segments from a pier in a well-built and durable structures is therefore in the best interest of the
balanced fashion on each side until mid-span is reached and a closure is made contractor. '
with a previous half-span cantilever from the preceding pier. The same erection o The construction of the superstructure does not require placement of equipment
process is repeated until the structure is completed. For construction cast in situ, such as footbridges or falsework between the piers, with clear advantages when
the movable formwork is supported from the previously erected segment or the bridge spans a river (Fig. 1.9), highway, railway or inaccessible place, or
from a movable form carrier while the new segment is being formed, cast presents reduced clearances.
and stressed. For precast segmental construction of box girders made of 0 Unlike balanced cantilever construction, it is possible and convenient to
steel, reinforced concrete or composite materials, the new segment is usually build long superstructure segments, even more than 30 m long. This reduces
supported from a movable winch system anchored to the preceding segment. the number of construction joints, weak points in any structure. Compared
to balanced cantilever construction, launching results in faster construction
Balanced cantilever construction strictly follows the self-resistance criteria. In a PC and requires only one casting bed instead of two or more pairs of form
bridge, prestressing tendons in the deck slab and compression forces in the bottom carriers.
slab resist the negative moment generated by the dead load of the cantilever and the
weight of the form carrier. The permanent effects of dead load and prestressing are Although the incremental launching construction method is generally associated
similar to construction stresses, and the effects of superimposed dead and service with bridges cast in situ, it is perfectly suitable for precast segmental construction as
......loads...are mostly of thesame sign.» . . . Well. Bridges composed of precast segments arranged along temporary support
 14 BRIDGE LAUNCHING INTRODUCTION 15

Launching of precast segmental bridges [72] can be a very competitive solution

too. Bridges whose length does not permit amortization of the costs of a launching
truss for segmental construction can be optimum targets, especially when it is
necessary to allow unrestricted use of the areas below. Tens of modular highway
overpasses may be built with a precasting plant and few sets of inexpensive launch
equipment (support/launch rails, a launching nose and a thrust device), with all
the advantages of the continuous box girder structural solution. A precasting plant
can simultaneously feed production lines with truss assembly and production lines
with incremental launching, thus further enhancing amortization conditions and
planning flexibility of the project. All the conventional launch techniques
(incremental launching from one abutment towards the opposite one, monolithic
launching from the opposite abutments with mid-span closure, rotation about the
piers adjacent the obstacle to overpass and mid-span closure, launching followed by
transverse shifting) may be used as well.

|.3. Incremental launching of composite bridges

Fig. 1.9. Skye Bridge (reprinted with permission, Dywidag Systems International) In the last decades, many factors increased the competitiveness of composite
bridges and reduced their lower use threshold to spans of about fifty metres,
which previously were the dominion of PC box girders [4,12]. Progress in iron
rails, connected by closure joints or epoxy, and incrementally launched, ofler metallurgy led to rolled steel of high and reliable mechanical properties. The
several advantages. development of design codes based on the strength of materials led to a better
evaluation of safety. The progress of assembly techniques replaced riveted splices
0 Construction of most of the superstructure is independent of pier construction,
with bolted or welded ones. In the last years, this evolution has been further
as precast segments can be stored elsewhere.
0 The use of labour and casting equipment is optimized in a series of highly accelerated by the cost stagnation of steel plates and the parallel general increase
in labour cost.
repetitive operations with the best quality control and amortization of invest—
The qualitative advantages of composite bridges are numerous. The high tensile
ments (precast segments for launched bridges are absolutely conventional in
and shear”’e‘r”ficieney of the steel plates combines with the good compressive
dimensions and handling).
0 Assembly and launching of the superstructure are fast operations that require efficiency of concrete. Rapidity of construction and the possibility of building
most of the structure in a workshop improve planning and risk management.
little specialized equipment (support/launch rails, a launching nose and a
The long durability resulting from different and renewable protective treatments
thrust system). Equipment is easily reusable and inexpensive in terms of initial
and the possibility to modify the structure with time to adapt it to new service
investment and assembly and dismantling costs.
0 Yard operations are limited to the segment arrangement along the support/ conditions further improve the flexibility of design. Finally, the presence of few
structural elements Whose function is clearly recognisable enhances the architectural
launch rails, the joint sealing and the introduction of prestressing in the new
section of the superstructure. The construction duration can be half of that quality of the structure.
The optimum field of utilization of a composite section is the simply supported
required for casting in situ, although with additional labour costs because of
beam. In the negative bending regions of a continuous beam, the bottom flange
these additional activities. ' *
compressive stresses require thick plates or closely spaced stifleners to prevent
So far, the incremental launching construction method has been mainly used for instability, while the tensile stresses in the deck slab require longitudinal rein-
bridges cast in situ, as this permits combining the advantages of the industrialized forcement to control cracking. The presence of a concrete bottom slab in the
and repetitive work processes with those deriving from a small number of construc- negative bending regions for double composite action is competitive only for
tion joints with continuous reinforcement. In addition, segment length and yard long spans. In most cases, therefore, the cross-section is open and only aesthetic
organization can be easily adapted to the specific requirements, thus enhancing the reasons or particular design requirements justify, in straight bridges, the higher cost
.. overall competitivenessof the construction method. Of closing the girder with a steel bottom plate.
 l6 BRIDGE LAUNCHING INTRODUCTION l7

These limitations also influence the launch of composite bridges, as the dispersion composite action. Of course, construction of a concrete bottom slab would be much
of support reactions within the webs and the negative bending stresses affect every simpler in a fixed formwork supported on the ground.
cross-section of the superstructure during launching. Since the dead load of the This led to the concept of box girders composed of two concrete slabs and two
concrete slab is 75 to 85% of the total weight of the superstructure, the steel girder steel webs, to be launched with the completed section. The presence of two concrete
is launched first and the deck slab is placed on to the steel girder only on completion slabs requires the massive utilization of prestressing, and concrete slabs, light steel
of its launch. This solution reduces the launch stresses in the steel girder and avoids webs and prestressing tendons are combined in new typologies of prestressed
tensile stresses in the concrete slab. On the other hand, however, the deflections of composite sections.
the steel girder are large during launching because the flexural stiffness is much In these sections, longitudinal bending moments are resisted by axial forces in the
smaller than that of the final composite section. This requires precambering and concrete slabs and the shear forces, reduced by tendon deviation, are mainly
results in specific launch requirements in terms of both support adaptability and resisted by the steel webs. Structural behaviour depends on the shear force
control of local stresses. distribution between the steel webs and the prestressing tendons. The different
The concrete slab may be either cast in situ or in turn launched onto the steel material efiiciency suggests shaping the tendons so as to balance the shear forces
girder. Manual form handling for slabs cast in situ is expensive, risky and time produced by the permanent loads and one-half of the live loads, so that the steel
consuming, and the final quality is quite poor. Therefore, casting in situ takes place webs resist only the shear fluctuations due to the presence or absence of live loads.
either with form carriers (movable shuttering systems that shift along the steel It is thus possible to obtain slender cross-sections that guarantee the highest
girder to reach the different casting positions) or on precast planks. The use of efficiency in service. On the other hand, any operations in which the dead load
form carriers is preferred in most cases although their cost can be adequately shear forces are not reduced by shaped tendons are difficult and delicate.
amortized only in long bridges. In the case of short or medium-length bridges, Prestressed composite sections actually involve several launch challenges.
precast concrete planks are often used as left—in-place formworks. However, con—
crete planks usually do not contribute to the carrying capacity of the composite
section and their weight therefore increases the weight of the steel girder. Precast
planks are expensive and are to be placed onto the steel girder, with risks for
workers and the areas below and additional costs.
As an alternative to casting in situ, a continuous concrete slab may be cast in a
fixed form placed behind an abutment. The slab is composed of segments match-
cast against each other with continuous longitudinal reinforcement. Segment
casting is followed by the launch of the whole slab section onto the steel girder,
and this sequence (construction of a new slab segment, and launch of the whole
strip) is repeated until completion of the slab, which is finally joined to the steel
girder to achieve composite action.
Incremental slab launching is an interesting construction method. The form
system is much less expensive than a movable shuttering system. Slab construction
is a continuous process, as the discontinuous in-situ casting sequences aimed at
avoiding tensile stresses in the negative bending regions are no longer necessary.
Steel cage assembly and concrete handling are extremely simplified. Slab launching
also avoids interference with the areas below, reduces shrinkage and creep cracking,
eases introduction of longitudinal prestressing, simplifies structural analysis,
improves durability and aesthetic quality, and reduces risks for workers.

l.4. Incremental launching of prestressed composite bridges

In spite of its numerous advantages, a conventional composite section has some
limits when used in a continuous beam. Its construction by incremental launching
requires launching of the steel girder alone and placement of the deck slab later, and
. the difficult casting. conditions of. a.» bottom slab discourage the use of double
 2. Incremental launching of bridges

2.l. Structural system of the superstructure

With the progress of materials technology and calculation methods, the continuous
beam structural system has been shown to have several advantages over statically
determinate schemes. Continuous beams require less structural materials than
either the simply supported beam or Gerber schemes, and guarantee a better
control of deflections and fatigue phenomena. Avoidance of internal hinges that
interrupt beam continuity in the horizontal plane improves the seismic response of
the bridge, and a smaller number of bearings and intermediate joints decreases
maintenance costs. Finally, secondary stresses produced by diiferential settlement
of bearings (if any) may be avoided by shimming, and are definitely reduced by
creep of concrete. For all of these reasons, the continuous beam scheme has been
more and more frequently adopted in recent generations of bridges.
The superstructure of a launched bridge is always built as a continuous beam.
This makes the bending moments in the mid-span regions and in the support
sections Simii'ar""ifivatue“Haring the launch, although opposite in sign. Centroidal
launch prestressing of PC and prestressed composite bridges controls the flexural
tensile stresses alternately at both edges without excessive compressive stresses at
the opposite edges, and similar considerations apply to the compression flanges of
composite bridges.
The advantages of a continuous beam suggest permanently maintaining the static
system used during launching, at least when the bridge length is compatible with the
expansion and contraction produced by thermal and time-dependent phenomena.
In this case, on completion of launch it is only necessary to replace the launching
saddles with permanent bearings and to complete prestressing. Two longitudinal
static systems of the superstructure may be achieved this way:

o Fixed bearing at one end and sliding bearings at all the other supports. This
scheme permits concentration of the horizontal, forces at one abutment, but
requires longer sliding bearings and a large expansion joint at the mobile end.
0 Fixed bearing at an internal pier and sliding bearings at all the other supports.
This scheme is applicableOnly in the presence of a short and stiff central pier
”W
 2O BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 21

that can resist the whole horizontal action. Compared with the previous scheme, dictated design of two continuous beams fixed, at their respective abutments, with a
a central restraint reduces the movement of bearings and expansion joints. central expansion joint. Since the dead load component along the launch plane was
In a case of flexible piers, two or three central piers may be hinged to the greater than the friction force, launching was obtained with a braked downwards
superstructure to better distribute the horizontal forces. sliding. A braking force of 6 MN was sufficient, compared with the l4-MN thrust
force required for launching the superstructure upwards. The same technique
The static system used during launching can be modified once the superstructure
was used to open the central expansion joint. After fixing the upper beam to its
has reached its final position.
abutment, the lower beam was allowed to slide doaards by releasing the
0 Some design requirements (very long bridges, high temperature differentials, ‘ temporary prestressing tendons of the central joint.
slender piers, etc.) can dictate division of the superstructure into shorter The Sinntal Bridge for the high-speed railway ICE in Germany, composed of
consecutive Continuous beams, or even simply supported spans. eight 42.5-m spans, a 51.5-m span and a 30.8-m span, was built by launching as a
9 In a case of variable curvature in plan or profile that makes launching from only continuous beam a sequence of single PC spans temporarily coupled to each other
One abutment impossible, two sections of the superstructure can be launched by joint prestressing. On completion of launch, the spans were disconnected and
from the opposite abutments. As an alternative, if curvature varies in a bridge detached by jacking the joints until obtaining the final design scheme of 10 simply
region not much distant from the ground, the casting yard may be arranged on supported beams. This solution resulted in a significant cost reduction compared to
a temporary embankment raised in this region, so that two deck sections can be the conventional span-by-span construction with a movable shuttering system. A
launched one after the other in opposite directions. In both cases, the central similar application of temporary prestressing of expansion joints permitted the
closure between the two deck sections increases the final degree of redundancy execution of the Basrah Bridge in Iraq [13], in which the central span of a movable
with respect to that used during launching. bridge was launched together with the access spans as a single continuous beam.
o The superstructure 'may be permanently framed to the launch supports to The launch static system may be also modified by joining two sections of the
obtain indeterminate arches or portal schemes. This solution is often used superstructure into a single continuous beam. This solution has been often used
with steel and composite bridges. for bridges with a straight-curved or curved-countercurved plan alignment, which
o The superstructure, launched over temporary piers, may be suspended to a can be obtained by launching two bridge sections with constant curvature from the
tower to obtain a final cable—stayed system. opposite abutments, and then connecting them by means of a central closure joint
[14]. In the final phases of approach, the launching nose is assembled in a raised
In the 49.16 Lot of the high-speed railway ‘TGV Atlantique’ in France, a PC position to slide above the front-end diaphragm of the superstructure already in
superstructure composed of five consecutive three—bay continuous beams, each place (Fig. 2.1). A temporary pier below the closure joint improves the casting
with spans of 29.2, 30.6 and 29.2 m, was launched as a single 445-m continuous conditions, but similar results may be obtained with crossed joint bars like those
beam. After launching, the temporary prestressing of the permanent joints was used for mid-spaniconnection of bridges built with balanced cantilever constrUction.
released to separate the superstructure into the five final beams, which were The .Schnaittach Bridge was built in Germany by launching a 424-m bridge sectiOn
moved towards their final position by jacking the internal expansion joints. with a constant radius of plan curvature of 1000 m downhill from one abutment, and
The same technique was adopted in South Africa for the Olifant’s River Bridge, an 864-m section composed of a front clothoid and'a rear tangent section uphill from
composed of 23 45—m PC spans for an overall length of 1035 m. The final scheme Of the second abutment. The front clothoid required the progressive rotation of the
this railway bridge consists of two ll-span continuous beams fixed at the respective casting cell during the first segment extractions and the use of shifting launch
abutments and a simply supported central span working as an expansion joint. This bearings (Fig. 2.2) placed on temporary extensions of the final piers. On completion
scheme transfers the braking forces directly to the abutments and permits the use of of launch, the two bridge sections were connected by casting the joint segment in situ.
slender piers. The 23 spans were built as a single continuous beam. On completion A similar solution was adopted for the Schrotetal Bridge in Germany. This 495-m
of 1aunch,the superstructure was permanently fixed to the launch abutment. After composite box girder bridge (Fig. 1.1) was built by launching a 404—m section of the
relieving of the launch prestressing of the first internal joint, the free section of U-girder from an abutment and a 5 1-m thinner section from the opposite abutment,
the superstructure (a 12~span continuous beam) was launched forwards until the directly above the railway. A central 40—m varying-depth Span was assembled on the
central span was placed in its final position. After opening of the second joint, this ground and lifted by jacking until closing the steel girder. Then, the concrete slab
operation was repeated on the remaining 11-span front section, fixing it to the was cast in situ with a movable shuttering system. Incremental launching on 68.1—m
second abutment. . spans required the use of a 24.1—m trussed launching nose. '
A similar solution was adopted in France for the highway Oli Bridge, composed Mid-span closure of two bridge sections launched from the opposite abutments
. . .of..1.5141—mPCspanswitha longitudinal5.4%grade.-The.,presence of piers 60 mtall also permits. overtaking a much longer central span. Temporary stays deviated by
22 . BRIDGE LAUNCHING INCREMENTHAL LAUNCHING OF BRIDGES 23

rear extension of
the iounchlng nose steel towers placed onto the superstructure have been used to support the two
cantilevers of the central span of the Mainflingen Bridge [15]. As an alternative,
the superstructure might be directly designed with extradosal prestressing. In this
case, some permanent prestressing tendons deviated by the concrete towers might
support the front cantilever. Concrete towers are integral with the superstructure
and arrive above the piers of the main spans on completion of launch. In a case of a
long central span, this construction method might be more competitive than
balanced cantilever construction; however, no application of this technique has
been reported as yet.
Wide superstructures may be attained by joining parallel beams. The l9-m wide,
twin box-girder superstructure of the Tanaro River Bridge in Italy was obtained by
joining two box girders built in the Same casting yard one after the other. On
completion of the incremental launching of the first half of the superstructure
(Fig. 2.3), this was shifted rightwards by 9 m (Fig. 2.4) to clear the yard alignment
for the construction of the second half. Both deck sections were cast on a SO-m
continuous support through lO—m segments, and launched in four SO-m strokes.
temporary pier The opposite 5.3% grade at the abutments and a l900-m radius of vertical
curvature required the use of temporary locks to avoid uncontrolled sliding.
Fig. 2.]. During the final approach, the launching nose is raised to slide over the Temporary piers aligned with the casting yard halved the launch span (from
superstructure already in place 50 m to 25 m) to further decrease launch prestressing. The total investment for
specialized construction equipment was less than €120 000.

Fig. 2.3. Launch completion of the first half of the superstructure: the rounded end of
Fig 2 2 Launchingbearingsmovable-inmthetransverse-direction {authorj-‘A - ~ the left cantilever is the left edge of the deck slab (author)
 24 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 25

Fig. 2.5. Launching of the superstructure above the arch of Neckarburg Bridge
Fig. 2.4. Transverse shifting of the first half of the superstructure clears the temporary (courtesy Leonhardt)
pier alignment for the launch of the second half (author)

Continuous beams launched onto arches represent a brilliant solution to the

requirements for safety of workers, high structure quality, short construction
duration and cost savings. The two parallel arch bridges” over the Neckar River
in Germany for highway A81 have been built by incrementally launching two 15—m
wide. 365—m long continuous PC box-girder superstructures on to two central
parallel arches and related access spans (Fig. 2.5). Maximum pier height is 95 m.
The central arches have a span of 154.4 m and a rise of 49.9 m.
A similar solution has been adopted in France for the arch bridge over the Isére
River for roadway CD 22 [16]. A 8.6-m wide, 234—m long continuous PC ribbed-
slab has been incrementally launChed on to a central arch and related access spans.
Maximum pier height is 30 m. The central arch has a span of 134.0 m and arise of
24.0 m. ’ i
The arch must have enough flexural stilfness to resist the asymmetrical loads and
the horizontal forces produced by launching of the superstructure from one
abutment towards the other. This'often requires the use of temporary stays that
balance launching loads and brace the tallest spandrel columns. Load asymmetry
might be avoided by launching two symmetrical sup’erstructures at the same
time from the opposite abutments towards mid-span. However, the presence of
symmetrical loads on only some of the spandrel columns induces bending stresses
in the arch, and the cost of two Casting yards rarely makes this Scheme competitive.
In the Veitshochheim Bridge for the high-speed railway ICE in Germany
_._(F..ig. 2.6), temporary- stays were integratedwith counterweights- to balance load Fig. 2.6. Veitsho'chheim Railway Bridge (courtesy Leonhardt)
26 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 27

asymmetry on the arch. The PC box—girder superstructure, l260—m long and been proposed for the central cable-stayed span of the Worli-Ba-ndra Sea-link
weighing 425 MN, is one of the heaviest beams ever launched [17—21]. The central Project in India.
polygonal arch spans 162.0 m with a rise of 24.0 m, and the access spans have a The tramway Palizzi Bridge (Fig. 2.8) is a l56—m-long superstructure with overall
constant length of 53.5 m. depth of 1.0 m, twin box-girder cross—section, and a main stayed span of 66 m
So far, no rigid-frame PC bridge has been built by incremental launching, [24,25]. It was incrementally launched onto six tracks of the FS railways in
although this structural scheme would permit overtaking spans fitted to arch Milan, Italy, with the use of temporary piers that reduced the launch spans down
bridges with lower costs. Many steel portal bridges have been built by launching to about 22 m. Site restrictions required the use of hinged-pendulum temporary
the steel girder on to inclined legs, as with the Piave River Bridge and the Cadore piers that were stayed to external restraints. Refined geometry control procedures
Bridge in Italy [22]. have been adopted to compensate for the settlements produced by the continuous
Compared to the conventional balanced cantilever construction of cable-stayed passage of trains. [26,27].
bridges, bridges launched on temporary piers and then cable—stayed to a tower are
often less expensive. The 2l.8—m wide, 527-m long PC box-girder of the Wandre 2.2. Static system and seismic design _
Bridge in Belgium (Fig. 2.7) was launched onto temporary piers placed in the On completion of launch, the superstructure is a continuous beam supported on
Meuse River and a parallel channel. With lS-m-long deck segments, the launch low—friction bearings and transversely restrained by the lateral launch guides. This
of the 118-‘MN superstructure required a 35-m-long launching nose. On completion temporary support configuration has to be modified to achieve the permanent static
of launch, the superstructure was stayed to a 95.5-m-tall A—shaped tower to attain system of the structure, which will govern its seismic response in the longitudinal
the final structural system with two stayed spans of 144 and 168 m respectively [23]. and transverse direction.
With this construction method, the tower can be erected without any restrictions Several solutions are possible in the longitudinal direction.
during the incremental launching of the superstructure. Compared with conven—
0 The launch bearings may be replaced with PTFE sliders that transfer small
tional stayed—cantilever segmental construction cast in situ, the stays are introduced
horizontal frictional forces to the piers. This solution simplifies pier cap
on completion of launch, with smaller time-dependent effects in both the tower and
details and avoids the need for anchor pins between the bearings and the
the superstructure, easier geometry control, and a much smaller number of
construction joints in the superstructure. Compared with precast segmental
construction of the superstructure, continuous reinforcement through the joints
reduces deck prestressing, which can be attained with long external tendons
tensioned before removal of the temporary piers. This construction method [has

Fig. 2.8. Launch of the first segment of the twin-box-girder superstructure 0f Palizzi
Fzg2...7._ Incremental launchingof the, Wandre Bridge..(c0urtesy;Greisch) .. Bridge .betWeen the vertical towers (author)
28 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 29

superstructure, with related geometric tolerances that are hard to respect with dampers have been used for the twin launched superstructures of the
a launched bridge. In this case, the longitudinal forces resulting from the Megalorema Bridge in Greece. ' '
seismic response of the superstructure are mainly transferred to the
abutments. Sometimes one abutment is able to resist the whole force; in In the transverse direction, it is almost always necessary to anchor the super-
most cases, however, high peak ground accelerations (PGA) and/or long structure to the piers. When the standard guides of PTFE sliders are insufficient
and heavy PC superstructures require the contribution of both abutments. to resist transverse response forces, sliding shear keys may be anchored to the
The seismic demand on abutments may be reduced with dampers activated bottom slab of the superstructure. These shear keys slide into longitudinal guides
by sacrificial shear keys. In this case, the significant displacements of the embedded into the pier caps, so that their transverse action does not restrain
superstructure require large expansion joints, since sacrificial joints might longitudinal displacements. These devices have been used along with longitudinal
complicate the transit of emergency services after a major earthquake. As LUD in the Greveniotikos Bridge in Greece [31]. As an alternative, seismic
an alternative, both abutments may resist longitudinal seismic forces rigidly. isolation of the superstructure works equally well in all directions; in this case,
This behaviour can be attained by creating an expansion joint in the middle of however, the piers are almost always to be designed for an elastic response to the
the bridge or by adding lock—up devices (LUD) to the abutment sliders. LUD DLE demand.
are active at the high velocities of the dynamic response of the bridge (not As regards the analysis methods, a non-linear time-history dynamic analysis
only to seismic demand, but also to braking forces and wind) and practically is often necessary with isolated bridges [28]. When inelastic structure response is
inert at the lower velocities of thermal expansion. Compared with the use of achieved with plastic hinges at the pier bases (and at the pier tops in the case of
abutment dampers, longitudinal seismic displacements of the superstructure continuity between piers and superstructure), the use of the force reduction factors
are much smaller, but the reduction in the seismic demand produced » by provided by codes is often sufficient. However, the adoption of a single-degree—of—
increased damping is lost. freedom model may be excessively approximate for these structures [31], especially
Hinge bearings may be inserted at some piers. This complicates the anchor in a case of moderate seismicity. In these cases, a non—linear pushover analysis [32]
devices of bearings but permits a better distribution of the seismic response is strongly advisable as it permits:
forces between the piers and the abutments. As for the abutments, hinge
bearings can be obtained by adding LUD to conventional PTFE sliders so 0 analysis of the non—linear response of the bridge from the first pier hinging to
that longitudinal displacements are restrained only during the dynamic response the dynamic equilibrium condition during the DLE and then up to the ultimate
to impulse actions. limit state ‘ *
With the opposite intention, the seismic demand on a hinged superstructure can 0 evaluation of the progressive period lengthening and damping increase
be reduced by locking the PTFE sliders at some piers with sacrificial shear keys produced by structural deterioration and a method for taking these parameters
and by equipping them wrth dampers. Breakage of the shear keys lengthens the into accounrindetermining the seismic: demand on the bridge
vibration periods and the intervention of dampers increases the equivalent 0 determination of the onset sequence and progression of plastic hinges and the
system damping, with synergistic effects on the seismic demand. In this case, related pier stiffness distribution to be used for. spectral analysis
the shear keys resist static and dynamic service loads. 0 determination of the force reduction factors to be used in spectral analysis
The PC superstructure may be rigidly joined to one or more piers. Vertical 0 analysis of the non—linear loading c0nditions of plastic hinges and assessment of
reinforcing bars screwed to bar couplers embedded into the pier cap can be their rotation capacity and shear strength
embedded in support diaphragms cast in situ that fill the solution of continuity 0 determination Of the overall safety factor of the structure from progressive
between the pier head and the superstructure. Vertical prestressing bars or collapse condition.
tendons may clamp these horizontal joints to locally increase the axial force.
This eases, control of the hinging sequence of the structureduring the design 2.3. Geometric constraints
level earthquake (DLE) and decreases the ductility demand at these joints. Of During launching, the superstructure is a continuous beam supported on launching
course these interventions are much simpler in the case of composite bridges bearings and transversely restrained by lateral: guides that prevent drifting
with steel piers. movements. Any constraint eccentricity (Vertical misplacement of launching
Seismic isolation of the superstructure [28—30] may be attained by replacing the bearings or transverse misplacement of lateral guides) with respect to the theoretical
launch bearings with low—damping rubber laminated bearings, lead-plug launch alignment causes secondary stresses. These can be tolerated only if small,
bearings, high-damping rubber systems, friction pendulum systems and similar and often With overdesign. Constraint ecCentricity also causes launch problems such
devices. [Low-[damping rubber laminated bearings “integrated with abutment as excessiVe Wear of neo-flonpads and frequent jacking for removal of damaged
 30 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 3l

bearings, with related higher labour costs. Therefore, it is necessary that the launch of the superstructure. These aspects should not be disregarded when defining the
bearings and the lateral guides be perfectly aligned with the surfaces of the super- plan layout of the piers and when positioning the launch bearings and the lateral
structure they come in contact with during launching. This can be guaranteed in guides, especially in a case of significant curvature [33]. A last case is (5) a circular
each launch position only in the case of a common geometry. helix, very rare and substantially comparable with the above even when the radii of
For one solid to slide rigidly inside another, it must be superimposable on itself curvature are small.
by rotation, translation, or rotation-translation. Therefore, its alignment must be: These geometric constraints and the uniform distribution of launch stresses
(1) tangent in plan and tangent or circular in profile; (2) circular in plan and generally suggest the launch of constant-depth superstructures. Minor adjust-
horizontal in profile (no launch gradient); (3) circular in a plane inclined with ments in the vertical alignment of the launch surfaces have often been attained
respect to the horizontal plane; or (4) curvilinear both in plan and in profile by temporarily attaching shims (hard wood or precast concrete planks) under the
(Fig. 2.9). In cases (3) and (4), the projections of the superstructure on to the bottom surface of webs of PC box girders. This solution generally aims to cover
horizontal plane are not arcs of a circle but arcs of an ellipse. In case (4), this small differences in elevation due to transverse rotations of the bottom slab.
distortion also takes place in the vertical cylindrical surface that contains the axis Steel plate or trussed extensions have been applied under the bottom flanges
of variable-depth steel girders. The adoption of these solutions with PC super-
structures is complicated by the higher dead load during launching. Temporary
piers are often necessary to decrease the launch stresses in the final mid-span
sections, and the adjustable casting cell is more expensive. The cost of the
precast concrete shims and of their anchoring and final removal further limits
the overall advantages of the solution with respect to conventional balanced
cantilever construCtion. ' '
In bridges with shorter bank spans and progressively longer spans towards the
centre, varying-depth superstructures have been attained by using a vertical
curvature radius for the bottom slab different than that of the deck slab. The Ile
Falcon Bridge [34] in Switzerland was launched with a convex curvature radius of
24 900 m for the top slab and 60 000 m for the bottom slab. Over the total 720-m
length of the bridge, this permitted increasing the total depth of the PC box girder
from 2.1 m at the abutments to 3.7 m in the middle of the central 73-m span.
Although some types of launching bearings can be adjusted in the transverse
direction “and "adapted "to varying-width cross-sections or Varying pian curvatures,
when designing a bridge (and a launched bridge in particular) it is always
advisable to place the varying curvature zones Outside the bridge itself. If a
transition curve affects the end section of the superstructure, the lower corners
of the cross-section (the only zones of the superstructure thatcome in contact
with launch restraints) can be designed with a constant curvature. Then, the final
alignment of the deck slab can be attained by locally adapting the lateral
cantilevers, since transition curves begin with radius variations so small that
these corrections are generally acceptable. Symmetrical launching from the
opposite abutments with central closure permits covering major differences in
curvature. As an alternative, the casting yard may sometimes be placed in a
central position to launch one section of the superstruCture on one side, and
eventually a second section on the oppOsite side with different plan curvatures,
with optimum amortization of the casting installation.
In fact, incremental launching involves some constraints that the designer must
Fig. 2.9. Incremental launching of the Val Resiel Bridge with a 150-m radius ofplan take into account from the early stages of design. On the other hand, technological
. .carrature (courtesyGeCa)..-.--.- W. -. --. . - progress has permitted realizations that were unimaginable just a few years ago,
 32 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 33

and the technological implications of these high-tech construction methods will 2.4. I . Launching of I ight superstructures
continue to influence the design aspects more and more. The launch of light superstructures (short PC bridges, medium—length prestressed
composite bridges and even long steel girders for non-prestressed composite
2.4. Launching techniques bridges) is often characterized by the use of inexpensive equipment, because of
The handling of a PC superstructure involves enormous forces and requires the the low launch forces and the necessity of assembling and dismantling the yard
guidance and control of large volumes. Although lighter, a steel girder basically rapidly and with low costs.
involves the same launch problems, and lower thrust forces are in this case Drawing by means of prestressing bars or strands is the most common solution,
complicated by much larger deflections and rotations of the support sections. although it presents some disadvantages, especially when launching along inclined
The force necessary to launch the superstructure is proportional to its weight, as surfaces. In fact, the ordinary longitudinal grade of highway bridges falls within the
both friction resistance and the longitudinal force produced by the gradient of the range of variation of the steel—Teflon friction coefficient, 1—40/0. Therefore, when
launching surface are a function of it. In the longest PC bridges this force can easily the launch occurs with new or well-greased neo-flon pads, it can be necessary to
exceed 10 MN, although in most cases it is only a few MN. This is nearly the same brake the superstructure during launching or to lock it during the construction of
strength as prestressing tendons. Therefore, in the first PC launched bridges it was the new segments to prevent uncontrolled sliding.
natural to use the preStressing materials and equipment already available in the As an alternative to towing, thrust devices are made up of hydraulic pistons that
yard. FrOm here came the development of drawing devices composed of one or act directly against the rear end of the superstructure by self—clamping to
more prestressing jacks anchored to a foundation and acting on strands or bars longitudinal kerbs or by locking in racks anchored to similar kerbs. In the case
anchored to the superstructure. of PC and prestressed composite bridges, these kerbs support the superstructure
With time, some inconveniences of the drawing systems stimulated the develop- during segment casting in situ or precast segment assembly. In the case of steel
ment of thrust devices, some applicable to relatively modest loads, some suitable girders, these kerbs are often necessary only for launching, so this solution is
for higher loads. As a consequence, the launch equipment currently used has expensive and the kerbs are replaced by bars or strands anchored to the abutment,
mechanical characteristics, power and cost so varying as to substantially define resulting again in a tow system.
its field of utilization [35]. The use of winches is limited to the lightest loads Back thrust systems have the same weak points as tow systems when launching
(short steel structures) and is progressively being abandoned. The less expensive along inclined planes. None of these devices permit moving the superstructure
hydraulic devices apply a tow force by means of strands or bars and are suitable f0r backwards in a case of excessive advance or launch problems, nor its braking
light superstructures (short PC bridges, steel girders with orthotropic—plate deck or when launching downhill. Therefore, these launch systems should be used with
for composite superstructures, concrete deck slabs launched onto previously prudence.
launched steel girders). Intermediate launcher types apply the thrust force by
self-clamping to longitudinal continuous restraints and are suitable for prestressed 2.4.1’ . 1' . Towsystems
composite bridges and medium—length PC superstructures. Finally, the most Coupled prestressing bars and hollow hydraulic jacks anchored to the super—
expensive devices transfer the launch force by friction and are generally reserved structure (Fig. 2.10) or to the abutment (Fig. 2.11), have been used many times
for the movement of large masses in short times, or to solve particular control with PC and steel girders to reach launch forces of 0.8 to about 1.5 MN.
requirements of the launch forces. Since friction launchers are extremely adaptable, The short stroke of conventional jacks for prestressing bars, usually about
extended amortization conditions may be sought in short PC bridges as well. 200 mm, suggests placing two jacks on each bar (Fig. 2.10) so that they alternate.
Second—hand launch equipment of all types is easily available in the EC. During the stroke of the first jack, the second one returns to idle to be ready to pull
Independent of the transmission modality of the launching force to the super- the bar when the first jack reaches the end of its stroke, avoiding bar relieving. This
structure, every type of launch equipment requires the presence of an anchor scheme requires the use of special guillotine anchorages, and the bar couplers must
element restrained to the ground. The-foundation of the abutment—is often suitable pass through the central hole of the thrust jacks.
for this use, since it is usually dimensioned to resist most of the longitudinal service The use of single-thrust jacks with bar relieving at the end of each launch stroke
loads. When the launching force exceeds the capacity of the abutment (a condition should be limited to almost horizontal launch surfaces. The stroke necessary for the
not infrequent in the final phases of uphill launching of PC bridges), overdesign new bar stressing is lost in each cycle and this lengthens launch duration. If the
may be avoided by joining the foundation of the casting yard with the abutment to casting/assembly yard is long, bar elongation at tensioning may result in excessively
create a large concrete bed that cooperates by friction. In case of need, however, short effective strokes of the thrust jacks. In this case, the jacks may be designed
specific foundation blocks for the launching jacks can avoid the application of any for launching, with an effective stroke of about 1 m (Fig. 2.11). Their cost is
. horizontal forces t,9,,,the,,abutment. , , . reasonable, although amortization is necessarily limited to launch operations.
 34 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 35

Fig. 2.10. The twin—jack back-thrust system for the steel U-gircler of the Schrotetal
Bridge was designedfor a totalforce of I .3 MN and 210-mm launch strokes (author)

In the case of launching uphill, the superstructure muSt be locked during each Fig. 2.1]. Long-stroke bar jack anchored to the abutment (reprinted with permission,
return stroke of the tow jacks to avoid uncontrolled backwards sliding. The modest Dywiclag Systems International)
dead load of the superstructure usually allows utilization of the drawbars
themselves for this purpose. The use of two jacks on the same bar (Fig. 2.10) avoids
any problem, as drawbars are not relieved. In the case of single jacks, the drawbars are individually pretensioned before the beginning of launch (isotensioning by
have sometimes been di'ided into coupled segments as long as the jack stroke. ‘cornbing’). However, when pulling launch tendons with conventional short—stroke
When the jack reaches the end of its stroke, the bar is locked to an anchorage, prestressing jacks, tens of repetitions of the tensioning cycle compromise the initial
the jackis released, its rod is reentered, the superfluous bar element is eliminated isotensioning and require corrections of the force in the strands during launching.
and the new bar head is connected again to the jack. This system is effective only in Although these operations are generally carried out at the anchor end and do not
the case of long-stroke launch jacks and requires a lot of bar couplers. When require dismantling of the launch jacks, they cause significant time losses and
conventional prestressing jacks are used, the less expensive solution is to locate involve some risks in the case of steep launch grades. The progressive loss of strand
them in cylindrical saddles that permit bar anchoring at their front end and isotensioning also suggests adoption of high safety factors when designing launch
avoidance of bar relieving during jack re-entry. tendons.
Drawbars are placed under (Fig. 2.11) or next to the superstructure (Fig. 2.10). There are numerous possible schemes for applying the thrust force to the super—
In the case of high launching forces, drawbars are substituted with tendons of structure, and physical principles for load transfer. The use of frictiOn between the
strands. In principle, by using strands the tow force is limited only by the capacity superstructure and the anchor plates of draw tendons (mobilized by hydraulic local
of the jacks, the addition of further strands always being possible. However, it is compression as in Fig. 2.12, or by prestressing bars crossing the superstructure as in
generally convenient not to exceed the force of 2.5 MN in each tendon in order to Fig. 2.13) is limited to short PC bridges launched along horizontal planes, which
use ordinary 19T15 prestressing jacks with adequate safety factors. If needed, it is require minimum draw forces.
better to increase the number of tendons, although this causes problems of load Higher thrust forces or steel girders require mechanical transmission: transoms
distribution that require electronic controls or powerful hydraulic pumps for anchored to, the rear end of the superstructure (Fig. 2.10) or through pins
wing—"parallel. jackfeeding with reasonablelaunch velocities. Strands in each tendon (Fig. 2.14). In the case of a PC bridge, these load-transfer devices induce high
 36 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 37

anchor plate

neoprene contact surFace .-, ,,,

///////////////////////
zzzzzzzzzzzzzzzzzzzzzI
{III ‘
1/
1/

jack
I
I

)}f

We'll/Illa ' n Will/[M [v] \YV 37 \y \y /

v] V/ V/ \y \y I'I

."'// VI/ W/ "//"’I7 "I/ W/ '47 ‘

l

return stroke roUer

Fig. 2.12. Friction transfer of the thrust force by means of hydraulic compression removable launching pin

' IWIIIIA”
concentrated stresses in regions of the superstructure that often are already loaded
by the anchorages of prestressing tendons. Therefore, transfer of launch forces
requires careful . design and proper local reinforcement. At the end Of each
construction cycle, the transmission devices must be relocated to the rear end of
the superstructure and the draw tendons must be brought back to their initial fi'Ilk'lk/Aé'fil WwI”;
length. Their weight makes this operation difficult, and in the case of long bridges 1'V' I

it may justify the adoption of electrical systems with service trolleys [36].
I draw tendon
The problem of superstructure locking in uphill launching can be solved, as with
drawbars, by locking strands during the return stroke Of the launch jacks. Although Fig. 2.14. Through pin
few prestressing jacks have two opposite strand clamps to avoid tendon relieving, the

launch jacks Can be equipped with conventional anchor plates to lock strands at the
beam hangers
end of each tensioning cycle. This solution involves minimum costs but does not
permit controlled backwards sliding of the superstructure in case of need. As an
alternative, it is possible to use two oversized draw cables and to take advantage of
the difference between launch and locking forces, by relieving the first tendon while
llllllllllllllllllll
IIIIIIIIIIIIIIIIIIIIIII
keeping the second one tensioned. In other terms, two tendons are necessary for
launching, but they are designed so that. one tendon is sufficient to prevent back-
rrrrrrrrrr
I I

wards sliding with the necessary safety factor.

’1’: l/:
In the heaviest bridges it is preferable to avoid anchoring the superstructure by
:3 7 prestresslng bars
means of the draw tendons and to adept specific friction-lock devices. These
k

\
\

guarantee a higher safety when it is necessary to replace some strands or to re-

equilibrate the draw tendons because of strand slippage in the jacks. Finally,
friction locks are always necessary during construction of the new segment when
rear transoms are used to launch the superstructure. 7
launching beam draw tendon Friction-lock devices usually involve hoisting of the superstructure at the
abutment by means of jacks equipped with knurled contact plates able to mobilize
. ”Fig.2.13. Friction transfer of the thrust force brmeans of,prestressing bars a high “friction coefficient against the bottom surface of the superstructure. This
 38 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 39

locking system has a varying effectiveness, locking force being proportional not (Fig. 2.16). Therefore, tendon deviators must be progressively disassembled to per-
only to the friction coeflicient, but also to the continuous beam support reaction mit the jack passing. Even so, the tow force applied along the subsequent chords
at the abutment. This remains practically constant during most of launching but tends to cause drift phenomena in the superstructure, which require the presence of
decreases considerably in the final phases, just when the weight in movement is at many lateral guides (Fig. 2.15) in the formwork area. Finally, the drawbars or cables
its highest and, with it, the sliding force along the launch direction — that is, the cannot be placed alongside the superstructure (a less expensive location because of
force to balance. its good accessibility) as they touch the bridge along the external side of the curve.
In some heavy PC superstructures, friction anchoring has been attained with To conclude, in spite of the special attention required in particular cases (steep
horizontal jacks pressing against the opposite sides of the bottom slab, so that grade, curved alignment) and a certain operative slowness, tow systems remain the
the transverse locking force does not depend on the support reaction of the super- least expensive and most efficient solution for the launch of short PC bridges and
structure. In this case, lock safety depends on contact pressure, i.e. on an artificial even long steel girders. They require simple and reusable equipment and are easy to
force that may vary over time. Therefore, jacks should always be equipped with assemble and dismantle.
safety ring nuts and the time-dependent transverse shortening of the bottom slab
should be accounted for when determining the minimum transverse force necessary
to create a safe lock action. 2.4. I .2. Self-clamping systems
In the case of steep slopes, it can be convenient to launch downhill maintaining As an alternative to draw systems, hydraulic pistons can be anchored directly to the
the superstructure braked, as the braking force is much smaller than the draw force rear end of the superstructure to transfer the thrust force by acting against rigid
required for uphill launching, friction working in favour of equilibrium. In reality, restraints.
however, the possibility of pulling the superstructure backwards should be taken Rear thrust requires anchoring the launch pistons at many locations along a
into account when sizing braking tendons, since this facilitates both the release of continuous restraint. The longitudinal kerbs that support the forming system of
the temporary locks of the superstructure and the correction of an accidental bridges cast in situ or the precast segments of segmental bridges allow resisting
excessive advance. the thrust reaction by means of friction mobilized by hydraulic compression
The draw systems present some limits when the bridge has a curved plan align-
ment, as the launch force is applied along the chord subtended by the jacks and the
draw tendon anchorages, and not along the local tangent at the bridge alignment.
This produces angular discontinuities in the draw strands both at their entry into the
jacks and at the anchorages, which must be reduced by deviation devices (Fig. 2.15).
In this case, the strands cannot slide inside deviators (friction would be excessive)
and it is necessary to plaCe the launch jacks at the rear movable anchorage

8"“ phase Formwor‘k launchlng bearlng

1St phase Formwor‘k , _ q - abutment

.'.‘_-......

\ _ ._ draw tendon
, lateral guide tendon deviator
continuous Foundatlon draw Jack
Fig. 2.16. The rear launch system for the Tiziano Bridge was designed for a 3.2-MN
Fig. 2.15. Deviators of draw tendons and guide devices are necessary in the case of thrust force attained with two launch tendons, one of which was suflicient to avoid
. curvedplan alignment , uncontrolled backward sliding (author)
 4O BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 4l

(Fig. 2.17). As an alternative, a continuous mechanical contrast (Fig. 2.18) is

more expensive but more reliable in a" case of uphill launching.
The use of self-clamping systems is generally limited to light structures launched
along a horizontal alignment, as the high thrust forces required for heavy super-
I I, A-\

, :K~, . \ .\

/~
structures or steep grades cause several difficulties: 2 Ln.“ n
——I——_——_
\'\
0 Pulling of the superstructure backwards is practically impossible (as with the I 'I/ 'I/ r
"/ "/ 7/ 7/

tow systems) as this would require bidirectional anchor devices between the
thrust pistons and the superstructure in every construction joint. Therefore,
recovery of the elastic launchwards deflection of a pier caused by the seizing
of a neo—flon pad is an extremely delicate operation, and the wrong insertion of
a neo—flon pad requires hoisting of the superstructure. Both these operations
increase the launch duration and immobilize labour on the piers.
o The presence of rear pistons slows down the formwork set-up and interferes
with the distribution of the joint reinforcement and the prestressing tendons.

thrust piston

, IWilly;
t
'i.
E
n
i- ”vow-7.”:-
m r
Fig. 2.18. Rcar thrust pistons With mechanical contrast in a continuous rack
—_ A

"
I
__ __—'
r _
In:
’armfianwf

The flexural stresses produced by the inclination of the thrust pistons suggest
launching the whole cross—section, and the use of paired forms (bottom
slab and webs in the rear formwork and deck slab, in the front one) is
discouraged.
The casting/assembly yard has to be located nearby the abutment to transfer the
thrust reaction directly to its foundation. This causes large rotations in the joint
launch .
sections between the segments of the superstructure, whose reduction often
requires the insertion of a temporary pier in the bank span, with additional
costs.
release, Forward stroke
At the beginning of each launch, a high thrust force (static friction) is applied
far from the abutment. In the foundation system of the casting/assembly yard,
hinged at the abutment, the moment produced by this long lever arm may
produce rotations that jeopardize the vertical alignment of the whole casting
_..,Fig. 2.17. Rear thrust pistonswith hydraulic clamps _ ...yard,. unless expensive. indirect foundations are used.
 42 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 43

In the case of light structures launched along a horizontal alignment, however, the the superstructure by two temporary tendons, has. been studied for a downhill
use of rear thrusters offers some advantages over tow systems. Launching is much launch on an 8% slope in Brazil.
faster: this is a basic issue because two or three people are necessary at every On the Whole, the investment required by these thrust devices, certainly
support point during the launch, and labour saving can soon cover the higher functional for light or medium-weight bridges, increases quickly with the weight
cost of launch equipment. Launch is easier in the case of curved plan alignment of the superstructure and soon becomes uneconomical. Their optimum field of
as well. Finally, the use of an anchor rack permits hydraulic braking of the super- utilization is limited to medium-length PC bridges, long U-girders for composite
structure during downhill launching. The braking system of Fig. 2.19, anchored to bridges and prestressed composite bridges. In long PC bridges the rear thrusters
may be combined with strand—drawing systems sized for kinetic frictiOn, so that
the rear pistons cooperate only in overcoming the static friction at the launch
beginning. However, long and heavy PC superstructures suggest adoption of launch
devices based on transmission of the thrust force by friction.

2.4.2. Friction Iaunching of heavy superstructures

In the heaviest bridges, the thrust force is transferred by friction using one or more
pairs of “launchers” placed under the webs of the superstructure. In the simplest
version, a launcher is composed of a vertical jack pushed along a lubricated surface
by a horizontal piston. The piston acts against a foundation block that supports the
superstructure in the pauses between two subsequent launches (Fig. 2.20).
In the most advanced version, a friction launcher is a monolithic device
composed of a sledge containing the vertical jacks. The sledge slides between

- - —Eig.—-- 2.19. - Rea-r braking sys tem for» steep—slope do wnhlillrllaunching vFig... 2.20. Basic configuration of a friction launcher .(author) .
 44 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 45

longitudinal guides along a steel—Teflon base under the thrust of one or two pistons I'll/II

pivoted to a rear steel block (Fig. 2.21) (notice the level of surface finishing that can
be easily reached with incremental launching [37,38]).
Further details of this axial launcher can be found in Fig. 2.22, which illustrates
the standard working sequence of friction launchers:
LIF TING
0 Using the vertical jacks, the superstructure is hoisted a few millimetres from the
support block of the launcher (or placed near the launcher). As a consequence,
the support reaction of the continuous beam is transferred from the support
block to the movable launcher sledge.
0 Using the thrust pistons, the sledge advances by sliding on a low friction
surface. The thrust force is transferred to the superstructure by friction, by
means of the support reaction of the continuous beam on the sledge and the
high friction coefficient between the bottom surface of the superstructure and
knurled steel plates placed above the jacks.
c When the piston limit stop is reached (after anefi‘ective stroke that varies from THRUST
0.25 to 1.00 In), the superstructure is lowered on to the support block by I I
u' '

relieving the hoisting jacks.

III}[I’ll/IIIIIIIIIIIIIIII/l II fl
IIll-IIII/J‘l/l/lll/Il/lIll/I’llIII!
I III! [I’ll I I

o The sledge, unloaded, returns idle to the initial position to start this cycle
agaln.

LUVERING

000.000.!
0 c *—

RETURN

Fig. 2.2]. Front view of an axial launcher for fish-belly cross sections. Raising
--------------6apa6i-ty--is--8.4v-vNand rtherthrust forcerrris 6.4»MN (author) . , ~ 77 . w
46 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 47

For the superstructure to be trailed by the launcher sledges it is necessary that the
ratio of the thrust force FT necessary to move the superstructure to the vertical
force RV; transferred between the superstructure and the launchers be smaller than
the friction coefficient Cf,L between the launcher sledges and the bottom surface
of the superstructure. Otherwise, the hoisting sledges would slip under the super—
structure, damaging its surface, but without succeeding in moving it.

Fr
5 Cr (2-1)
RV,L ’L

The vertical force RV; is the continuous beam support reaction at the launchers. It
depends on the static system of the superstructure, i.e. on the position of the
launchers with respect to the adjacent supports. In very long PC bridges, therefore,
high thrust forces require a significant distance between the abutment equipped
with launchers and the front support of the curing area. This requirement restrains
the distribution of the temporary supports in the curing area and the sequence of
introduction of the launch prestressing.
RV,L is generally calculated by considering aligned supports, as the launchers
Fig. 2.23. Initial launch phases of the Aroncle Bridge. These friction launchers have
hoist the superstructure only a few millimetres. If it is insufficient, forcing the
been designedfor a total raising capacity of 16.8 MN and a thrust force of 12.8 MN
superstructure upwards can increase this reaction. The related stresses must be
(author)
checked for the conditions of partial launch prestressing in these segments, and
partial curing in the case of casting in situ.
Hoisting jacks are generally oversized with respect to the highest theoretical
reaction on the launcher of Fig. 2.21, as recorded during the final launch phases
values of RV’L to be able tohoist the superstructure, even when support mis-
of the Serio River Bridge. The continuous line describes the theoretical
placements or thermal gradients increase RV,L‘ Thrust pistons are designed for
evolution of the continuous beam support reaction, black dots have been
the highest launch thrusts considering the friction in the casting/assembly bed, in
recorded at the launcher, the support reaction needed for launching with the
the launching bearings and in the launchers themselves, in addition to the average
recorded friction (CfaL : 0.029) is dotted, and the effects of overhoisting
longitudinal gradient of the launCh surface.
(forcing) between 13 m and 4 m from the end of launch are evident. When
The use of a pair of friction launchers anchored to an abutment involves some
the lack of vertical force affects long launch phases (uphill launching along a
specific considerations in the initial and final phases of launch.
steep gradient), a second pair of friction launchers may be placed on a more
0 At the beginning of the launch of a PC bridge, the curing area separates the advanced support (generally a short pier). However, synchronization of two or
formwork from the abutment equipped with launchers. The need for a perfect more pairs of launchers is a rather delicate task, especially when their distance is
connection between the launching nose and the superstructure is generally significant. When the lack of vertical force affects only the last launch phases, it
solved by match-casting, i.e. by assembling the launching nose against the is normally simpler to use a steel rear-nose integral with the superstructure and
front head of the formwork. This requires the launching nose and the front connected to the launchers by draw strands (Fig. 2.25). In this case, the draw
section of the superstructure to be trailed on to the launchers until a support tendons are designed only for a share of the final thrust force, as the continuous
reaction sufficient to continue the launch by friction is obtained. For this beam has a support reaction to exploit by friction (Fig. 2.24). In addition, when
purpose, strands anchored to the launching nose and to the friction launchers the rear end of the superstructure approaches the abutment, launching is no
permit using the thrust pistons of the latter for these first launch phases as well longer affected by the extraction friction from the formwork and the friction at
(Fig. 2.23). the temporary bearings of the curing area (which together represent a significant
o In the final phases of launch, RV; decreases until the reaction of the continuous part of the total friction). As an alternative to the rear-nose, front drawing with
beam end support is reached. This vertical force is generally insufficient to strands and prestressing jacks permits solving the problem in a less expensive
satisfy the highest thrust force demand (2.1) by friction, especially when way (Fig. 2.26), although with some interference with the bracing system of the
.launchingtakesplaceuphill. Figure2.24 describes the evolution ofthe support . _ launching nose.
 48 . BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 49

support reaction at the launcher (MN)

0
—40 —20 0
progression toward the end of launch (m)

Fig. 2.24. Decrease in the abutment support reaction in the last launch phases of the
Serio River Bridge

Fig. 2.26. Final front pulling of the Palizzi Bridge (author)

The use of friction launchers (single as in Fig. 2.21, paired as in Fig. 2.23 or
distributed in more pairs) controlled by numeric—control boards (Fig. 2.27) may
result in enormous synchronized thrust forces. This permits moving huge masses, Compared with draw and back-thrust systems, the use of friction launchers offers
such as those of the longest railway PC bridges. many advantages:

0 In draw systems, the breaking of any component of the launch system (bars or
strands with related anchorages and couplers, anchorages of the thrust transoms
to the superstructure, thrust jacks, hydraulic plant) leaves the superstructure
free on low—friction supports. This requires adoptiOn of high safety factors that
involve oversizing of equipment and operative slowness. Friction launchers, on
the contrary, guarantee absolute intrinsic safety, as the worst consequence of
mechanical or hydraulic faults is halt of the launch and descent of the super-
structure on to the temporary supports near the launchers, which are designed
as friction constraints. Therefore, equipment can be designed for the launch
loads and can be overloaded without excessive worry in case of unforeseen
requirements.
o The electronic control of the hydraulic plant permits synchronization of the
action of launchers and recording in real time of the forces produced. By
monitoring RVL it is possible to verify the accuracy of superstructure, geometry
(some types of geometric errors modify the support reactions of the continuous
beam, as discussed in Section 3.5.2) and to correct these errors prior to their
accumulation. By relating Rm, to the raising of the superstructure, it is possible
to estimate indirectly the elastic modulus of the young concrete. The possibility
Fig. 2.25. Rear nose and integrative strand pulling in the last launch phases of the of setting an upper limit for the thrust force permits avoidance of pier overloads
. .Ser.io...Ri.izer Bridge. (author)......,.., ., produced, for instance, by the upside-down insertion of a neo—flon pad. This
50 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES.

o The lateral guides of the launcher sledges also act, by friction, as transverse
guides of the superstructure, and avoid the need for further local devices and
the related labour demand. As an alternative, however, the hoisting jacks can
be left free to shift transversally (Fig. 2.28). As with the draw or back-thrust
systems, it is thus possible to launch by friction superstructures with varying
width.

In order to transfer the thrust force Without stress concentration and surface
damage, the contact plates of the friction launchers are articulated so as to adapt
to the superstructure. This is usually achieved by connecting the jacks of the hoist-

--— launching dlr'ectlon

corrugated lH’ ting

steel plate Jack
.................................. ...........

IIIIIIIII IIIIII

lllllllllll
IIIIIIIIIII

Fig. 2.27. The numeric-control board of the friction launcher of the Serio River Bridge IIIIIIIIIIII IIIIIIIIIII

allowed pre-setting and real-time monitoring of launch parameters (forces and dis- //////////////////////

placements) (author)
'____________
////////////

"5.1: 33%

$mlllllllllllfiii=gffl “a?
’.
I block
I
’
.'““‘“**wwm;_
’WI, IIIII I I
a ' 1"],
............. I I I I I I I
I I I I I I I I I I I I I I I - I I IIIIII
I ’ I I I I I I I I I l / I I I IIIIIIII I
IIIIIIII I I I I I I I I/I/I/I/I/I/I/I/I/I/I I I z I I I I I I I

mistake may seem ridiculous but in fact it is rather frequent and potentially very
dangerous, as the friction coefficient increases dramatically when passing from a mortar shiFting cast—In
. ..-....1.:.},¢ttnaler- . . fixing rail.
greased steel—Teflon contact to a steel-neoprene one. In addition, pier workers

-
are often so hopeful of successfully removing the seized pad (in most cases this is
practically impossible) that launching is stopped when the pier is so significantly
deflected launchwards that it begins vibrating. Without friction launchers, llf-‘tlng
pulling the superstructure backwards is impossible. Therefore, recovery of the Jack
launchwards pier deflection requires local raising, and the horizontal force that
deflects the pier is transferred to the hoisting jacks. These emergency operations
require extremely careful procedures and controls, possibly planned from the
very beginning.
0 Launch cycle automation by means of electronic , , end-of-stroke switches
_./ wing mm: For
facilitates the operations even more, which can be supervised by one technician lateral move
only, and increases the launch speed, which can easily reach 10 m/h. -m- . M}
0 Labour costs are definitely lower because of the higher launch Speed (launching
I II I IIIIIIII
III I IIIIII
l/Ill I b IIIIII
IIII I [IO U men IIIIII

immobilizes two to three people at each support) and the possibility of avoiding
IIIII III Ill/II 11111
III III/ I/I/I/I/I/I/I/I/I/ IIIIII IIIII

tow tendon repositioning and anchoring of the superstructure during these mortar steel plate
phases. Moreover, after launching, the casting/assembly yard is immediately
_ readyfor, the construction of the next segment. V . Fig.__ 2.28. “Shifting launchers
 52 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 53

ing sledge in series and by equipping them with ball—and-socket caps. In the case of phases. This system is perfectly suited for horizontal launching; completed with
PC superstructure, the contact plates should also be relatively ‘soft’ to compensate synchronized lock jacks, it can be used in the case of inclined launching as well
for surface irregularities of the bridge. Their surface is generally engraved in the (both downhill and uphill).
opposite direction of the launch (Fig. 2.29), so that the thin plates thus obtained Friction launchers with hydraulic clamps are often used for the launch of steel
can transfer thrust by bending. The opposite arrangement might easily break the girders [39], although their transfer devices often damage paint. Flat hydraulic
concrete surface because of stress concentration. Similar design concepts can be pliers clamp the bottom flange of the steel girder and the launch cycle is similar
used for the launch of steel girders; of course, in this case, the bottom flanges of to that used for the Olifant’s River Bridge. Also in this case, launching along
the steel girder will be free from splice plates and bolts in the working areas of the inclined planes requires prudence and experience.
friction launchers.
The original vertical—load-based friction launchers have evolved into several 2.5. Launch bearings and guide devices
different schemes. The launch system used for the Olifant’s River Bridge was The launch of PC, prestressed composite and conventional composite bridges
composed of longitudinal thrust pistons acting on a movable transom placed requires diiferent types of launch bearings that comply with the specific charac-
under the superstructure. The transverse jacks of the movable transom clamped teristics of the webs of the superstructure. Launch bearings are described in specific
the bottom slab of the PC superstructure and created the contact pressure chapters related to the different types of bridges.
necessary for friction transmission. The work cycle, similar to that of a Lateral guides are also necessary during launching, to maintain the correct plan
vertical-load launcher, involved clamping of the superstructure, thrust, release, alignment of the superstructure and to resist the transverse forces produced by
and transom return to the initial position. Not depending on the continuous wind and the seismic response during construction. The prevailing function of the
beam support reaction, this scheme has the advantage of not restraining the lateral guides is related to the depth and mass of the superstructure as well as to
design of the casting/assembly yard and not breaking down in the final launch the plan alignment of the bridge. Large masses such as those of PC bridges
require strong lateral guides during launching, especially when it takes place
along a curve. Strong guides are also necessary to resist the seismic response of
the structure during the design-level earthquake (DLE) for construction phases;
lounchlng direction the effects of transverse wind are generally modest. Conversely, design of the
lateral guides for the lighter steel girders of composite bridges is often governed
by transverse wind. ‘
All the piers of a tangent bridge are equipped with lateral guides during launch—
ing although the correction of alignment of the superstructure occurs in most cases
by acting'at three points only: ’ ’ ' "" "
0 near the front end of the casting/assembly yard to correct the position of the
construction joint with the next segment
0 at the launch abutment (especially With draw systems)
0 at the front—most pier, to adjust the direction of the launching nose at its contact
with the next pier.
The design load of lateral guides is often empirical, being based on the force
necessary to move the superstructure by acting at the most appropriate points
during each launch stage. To account for the need for adjustments in the bridge
alignment during launching, some standards for PC bridges (AASHTO, ZTV-K88)
require that the calculated horizontal forces acting on the lateral guides be increased
by 1% of the local support reaction of the continuous beam. At least 3% of the
maximum support reaction should be used when launching the steel girder for a
composite bridge [39]. -
When horizontal or uphill launching takes place along an alignment curved in
_.. _.Eig. 2.29, Contact, plates.f0r,,,,PC,superstructures- . . plan,...the. external lateral guides are also subjected to radial forces resulting from the
 54 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 55

local reaction to the thrust force. This system of forces (of elastic nature and period of the DLE for construction phases is 20 years and age varies between
therefore proportional to the transverse flexural stifihess of the piers) can be 24% and 39% of ag. In regions of high seismicity, these levels of seismic demand
analysed by using a model of the continuous beam placed on elastic supports. may be significant and have different effects in the longitudinal and transverse
The forces thus obtained are added to those estimated for the alignment adjustment direction.
of the superstructure. When launching downhill, the superstructure is braked, the In the longitudinal direction, the superstructure is supported on low-friction
axial launch force in the superstructure turns from compression to traction, the launch bearings. Neglecting friction of the neo-fion pads, the Whole longitudinal
direction of the guide forces is reversed, and lateral guides located on the interior response of the superstructure must be resisted with specific restraints. If friction
face of the girder are required. When the gradient of downhill launching is launchers are used, they also act as lock-up devices and the problem is solved. If
intermediate between breakaway and kinetic friction, the superstructure must be launching takes place by means of tow systems or back pushers, bidirectional
pushed until detaching, and braked during sliding. Therefore, direction of the radial restraints must be used to lock the superstructure even in the case of a horizontal
guide forces inverts during launching [34]. launch surface. In both cases, the most logical location of these restraints is at
The design of lateral guides for Wind action is simpler. The design standards the launch abutment, although this stiff restraint results in short vibration periods
provide criteria for evaluation of the transverse load due to wind. This load of the structure and increased spectral demand. Seismic restraints located at some
is applied to the superstructure with the most unfavourable distribution to piers might decrease the system stiffness by taking advantage of pier flexibility;
determine the transverse forces to be resisted at the piers. Transverse wind generally \ however, locking operations are difficult at the top of the piers and may result in
governs the design of lateral guides for steel girders, but it may often be neglected pier overdesign in the case of isolated bridges.
when launching much heavier PC superstructures. Alignment forces are added to In the transverse direction, on completion of launch, the superstructure will be
Wind reactions determined for the maximum wind speed allowed during launching. connected to the piers by either transverse shear keys or a seismic isolation system.
Exceptional wind conditions might require interruption of the launch of steel In the first case, the piers are designed for the full transverse response to ag with the
girders. In this case, they may be assessed without alignment forces. additional mass of the superimposed dead load and characteristic live load; there-
The design of the lateral guides for the seismic response of the structure during fore, these piers are rarely designed for the seismic demand during construction. In
the DLE for construction phases is more complex. Several seismic design codes the case of isolated bridges, pier design for service conditions depends on the period
determine the seismic demand in terms of peak ground acceleration (PGA) and lengthening and damping increase produced by the isolation system. Since the
spectral modification factors that depend on the importance of the structure and isolation devices are inserted only on completion of launch, the superstructure is
on the site classification. The PGA, ag, for the seismic design of the permanent not protected during construction phases. Also in this case, however, age is much
structure often reflects a seismic design event with a return period [no of approxi— smaller than ag and this is usually Sufficient to avoid overdesign of the piers due to
mately 475 years. Accordingly, the design event has a probability of exceedance Pe construction phase requirements. In both cases, lateral guides are designed accord-
of about 10% and 19% for a design life of the structure of 50 and 100 years, ing to the Conventional CapaCity designprinciples ’[28], with adequate protection
respectively. factors.
Evaluation of the seismic demand for construction phases requires determination
of the probability of exceedance during construction, Pep, and of the construction 2.6. Launch and lock forces
duration to. EC8 [40] suggests that pa,C be not greater than 5%. The return period of The horizontal force necessary to move or to lock the superstructure depends on the
the DLE for construction phases, Inc, can be determined as [40] longitudinal gradient of the launching surface and on the friction forces that oppose
imposed movements (launching) and spontaneous movements (uncontrolled
zno — 1
1 _(1_pe,c)l/rc (2 2) launchwards or backwards sliding).
The effect of the longitudinal gradient is easy to evaluate. By assuming that the
and the related PGA during construCtion, agjc,'is [40]
self—weight of the superstructure Q is concentrated at its centroid, the longitudinal
grade of the vertical profile of the bridge at this location defines the average percent
t k launch gradient Gm. The horizontal force F6 to be applied to the superstructure
erg,C 2 ag (L9) (2.3) to launch it or to lock it along a frictionless launch surface is
tr,0

The coefficient k varies between 0.30 and 0.45 according to the reliability of the m=mg on
seismological data. For instance, assuming a construction duration of 1 year (for Particular care must be paid to the determination of the average gradient of the
._.the.. superstructure) and-..accepting-a 5% probability of exceedance,.-the return , superstructure during the individual launch phases. When the vertical curvature
 56 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 57

radius is small, the launch gradient can be very different from the average grade inserting timber plates instead of neo-fion pads at some bearings; however, these
between the opposite abutments. In the Tanaro River Bridge (Fig. 2.3) the local solutions should be limited to emergency interventions.
gradient is + 5.3% at the launch abutment and —5.3% at the opposite abutment, so o A contemporaneity factor W1 < 1 should be used to decrease the breakaway
that the average grade of the bridge is zero. During the launch, however, Gm was forces at the rear launch bearings, since the elastic shortening of the super-
+ 7.2% at the beginning of the first 50-m launch (with the superstructure entirely structure under the axial compression produced by the thrust force longitudin-
behind the launch abutment), + 5.8% at the beginning of the second launch (lOO-m ally staggers the bearing breakaway. This reduction is influential only in long
superstructure), + 4.4% at the beginning of the third launch (150—m superstructure) bridges.
and + 3.1% at the beginning of the final stroke. A similar situation occurred in the o The friction force opposed by the formwork, FfW, or the foundation kerbs of the
stayed Palizzi Bridge [26,27] (Fig. 2.8). Checking Gm in the individual launch phases assembly area for precast segments is high because of the low bearing stress and
is necessary for a safe design of thelaunch devices and the anchor systems that lock the dirt and wear caused by grout spillage. The breakaway friction coefficient of
' the superstructure during construction of new segments, since uncontrolled sliding the casting cell may be about 10% in the case of a continuous stainless steel—
Can occur backwards as well. Tefion contact, and much higher in a case of different sliding surfaces. ExcessiVe
Assuming that the deck is horizontal, in a launch bearing 1', the minimum tan- tolerances in the vertical alignment of the support rails further increase extrac-
gential force 17,0 necessary to produce sliding depends on the local support reaction tion friction. However, this high friction coefiicient is applied only to the dead
RVJ of the continuous beam and on the breakaway (static) friction coefiicient CR», of load of the deck segment to be extracted from the casting cell.
the launch bearing. 0 The friction component of the transverse forces applied by the lateral guides is
often negligible (even in the case of a curved launch alignment) when using
F? = CRiRV- - (2.5) neo-flon pads between the lateral guides and the superstructure. The use of
rolls, rollers, or other sliding materials (filled PTFE, CM1, CM2, bronze,
In the case of a PC deck, launching occurs on steel—Teflon contaCts. CR, depends on
Ferrozell), frequent when launching steel girders, may increase guide friction
many factors, since the behaviour of PTFE in contact with a polished stainless steel
substantially, although the dead load of the superstructure is in this case much
surface is rather variable:
smaller.
0 CR, decreases when the average contact stress increases. Recommended design
As a consequence, the total friction force Ff does not increase linearly with the
coefficients of friction for bearings with stainless steel sliding on pure PTFE
number n of launch bearings affected by the superstructure, and follows a law of the
continuously lubricated are 8% for a bearing stress of 5 MPa, 6% for 10 MPa,
type
4% for 20 MPa and 3% for 30 MPa and over [41].
0 Temperature: increases in temperature produce effects similar to increases in 11

load.
F. : F..+ Twila-0 , ,,- , (2-6)

-
a Lubrication and surface cleanliness: an imperfect cleaning or lubrication of
neo-flon pads and/or of the sliding surfaces, or their excessive wear, increase When sizing the thrust devices it is necessary to apply a load factor )4: to the friction
friction substantially. In the absence of test data, for design purposes, the force (2.6) to account for events such as insufiicient lubrication of neo-flon pads,
coefficient of friction for pure unlubricated PTFE on stainless steel may be wear of sliding surfaces, and tendency of PTFE to seize to the launching bearings
taken as twice the values previously given before for lubricated contacts [41]. during the intervals between two subsequent launches. The final friction component
Similar conclusions can be drawn for the rollers or sliding devices used to F1: of the design thrust force is therefore
launch the steel girders.
FF = 3433 / (2-7)
The total friction force Ff. does not depend only on the (breakaway condition of and the total thrust force FT is
launch bearings:
FT 2 FG + FF (2.8)
o The breakaway forces (2.5) soon decrease towards the values resulting from the
coefficient of kinetic friction, Cf’i, which is lower. Therefore, the piers should be On the other hand, no resisting friction should be taken into account when designu
designed for a breakaway friction of at least 5% (7% is more prudent in the case ing the lock or brake devices, and uncontrolled sliding will be checked for a force
of tall piers), but friction must be neglected when determining the braking or
FL = VSFG (2-9)
locking forces (kinetic friction may be very low), as required by AASHTO [42].
Incase of. need, the. braking devices for downhill launching may. be helped by , wherethe load factor for uncontrolled sliding ys will be at least 1.3 [42].
58 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 59

To conclude:

0 During launching it is necessary to apply a system of forces able to produce

(thrust) or to control (brake) the movement of the superstructure.
0 During the pauses of the launch it may be necessary to apply a different system
of forces to avoid uncontrolled sliding.

In both cases, great care must be taken in the evaluation of these systems of forces.
In several bridges, the gradient of the launch surface and the average coefficient of
launch friction are similar. When FG and FF have the same sign in (2.8) (uphill
launching), the sign inversion of FF during the stops of launching may cause back-
wards sliding. When FG and FF have opposite sign (downhill launching), great care
must also be taken in the determination of the expected range of variation of FF.
After breakaway, new and well-lubricated neo-flon pads may result in friction
coefficients much smaller than 2%. In this case, downhill launching may require,
at the same time, pushing the superstructure until the breakaway and braking it
immediately after.
Bridge launching affects huge masses, and no intervention is possible if sliding is Fig. 2.30. Limit support conditions for the current zone of the superstructure
uncontrolled. Therefore, launching along inclined surfaces requires prudence and
experience, and often justifies an independent design check by specialists.

With regard to the cross-sections temporarily above the piers, in position A, the
launch stresses are only a fraction of the permanent stresses, and the deck can be
2.7. Correction of launch stresses
sized for the latter. Longitudinal prestressing is still needed to cover the upper-edge
The most characteristic aspect in designing launched bridges is the need to absorb
tensile stress. Web reinforcement is designed for the operational shear stresses and
the stresses due to the temporary support configurations assumed by the deck
assessed for the launch shear stresses without the relieving effect of the parabolic
during launch. In fact, every cross—section of the superstructure passes cyclically
tendons (Whose sloping sections reduce the shear force in the webs) since these are
in mid-span and above the piers, and is therefore subjected to the maximum
installed only on completion of launch.
positive moment, the maximum negative moment, and the maximum shear.
On thebon'trary, in pOSi’tio‘n B, the cross—sectlons that on completion of launch
Thus, although supporting only its own weight, each cross—section has to Withstand
will be in mid-span to resist positive moments and slight shear forces are subjected
transitory stresses that are not only very heavy, but also significantly different from
to the maximum negative moment and shear. They often need to be adapted to
the service stresses. Thermal stresses and differential support settings worsen the
these transitory stresses, and therefore tend to be oversized with respect to the
situation.
service requirements.
The situation evolves between the two limit conditions of Fig. 2.30. Considering
During launching, every cross-section of the deck passes above the piers, and the
two successive piers and a typical span of the continuous beam, the first condition is
need to resist the same transitory stresses requires that both the cross-sectional
the final position A, with the support diaphragms of the deck located above
moment of inertia and the web thickness be constant along the superstructure.
the piers. The second condition is position B, with the deck advanced by half a
This prevents lightening mid-span sections (an operation that with other
span and supported on cross—sections that, on completed launching, will be
construction methods is more and more effective as span increases, especially in
mid-span sections. I V I j j I,
constant-depth bridges) and extends oversizing to the whole deck, causing
In a PC bridge, in both positions, mid-span sections are rarely overloaded. Shear
structural burdening [43,44]. In addition, the cyclic changes of sign in bending
force is low, and the bending moment is lower than the values reached under live
moment and shear ferce prevent the launch prestressing from undulating. This
loads. In addition, the box girder cross—section is well suited to positive moment,
maintains an axial layout and limits its contribution to the uniform compression
since the Wide top slab provides a large compressed area that draws the centroid
of all cross-sections. On completion of launch, this axial compression is insufficient
upwards. Therefore, longitudinal prestressing reducing tensile stress at the bottom
to absorb the stresses due to live loads, and it has to be integrated with more
edge is generally sufficient.
efficient undulated tendons. .It is detrimental as Well, because it increases the stresses
 60 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 61

in areas already compressed by the bending moment diagram of the continuous deck slab. Launching the whole cross-section instead of only the U-girder results in
beam [45,46]. To conclude, launch stresses require numerous checks and specific launch stresses five times greater, and the related costs (oversizing of the steel girder,
types of prestressing that optimize the bridge performance during launch, reducing expensive launch devices, etc.) are higher than those resulting from casting of the
the need for additional material at those points where oversizing, in service deck slab in situ on completion of launch of the steel girder. Therefore, the steel
conditions, would be pointless or even damaging. girder can be presized for conventional design conditions (dead load of the concrete
In a prestressed composite bridge the situation is similar. In these structures, slab and related casting equipment prior to the onset of composite action, and
however, the cross—sectional behaviour under bending moments and shear forces composite resistance to service loading), and minor design adjustments resulting
is more specialized and this results in additional launch requirements. Compared from launch requirements are related to only its own dead load.
with a PC superstructure, the higher cross-sectional efliciency and the lower weight Following the previous comparison, the dead load of the U-girder during
reduce the flexural launch stresses. In spite of this, however, launch prestressing is launching is 25 kN/m compared with 195 kN/m for a PC box girder. Support
always necessary. This is obtained with straight tendons mostly contained into the reactions are much smaller during launching, but the open steel girder is very
concrete slabs. Reaching a correct position of the prestressing force is diflicult with flexible and the large rotations of its support sections require rocking launch
only internal tendons and some external tendons are often necessary. This results in bearings. The cross-sectional centroid is often located below the middle of the
more efficient final prestressing schemes as the external tendons can easily be girder depth, and this may result in oversized top flanges. Finally, the support
relieved and repositioned on completion of launch. diaphragms of the steel girder are much heavier than the standard section and
As regards shear stresses, the external polygonal tendons of permanent the envelope of the launch moments is disturbed by the migration of these
prestressing are usually designed so as to balance the shear forces produced by concentrated loads.
the permanent loads and 50% of live loads with tendon deviation forces. Structures that present so many and diflerent load conditions require a careful
Therefore, the thin steel webs only resist the fluctuations in the shear forces presizing, to avoid excessive stresses in one of the many launch or service stages, but
due to the presence or absence of live loading. During the launch, however, without excessive prudence, which would result in an oversized structure. These
prestressing must be uniform to accommodate the changes in sign of bending aspects are discussed for the different types of launched bridges in the relevant
moment, and the steel webs have to resist full shear forces. This may result in chapters; however, the considerations that follow are common to all these types
locally Oversized details, especially when the bridge is launched without temporary of structures.
piers that halve the launch spans. Most of the launch stresses depend linearly on the distributed dead load q of the
The launch of a conventional (non-prestressed) composite bridge is relatively superstructure, so its reduction is of fundamental importance. If p is the distributed
simpler as these structures are built by launching the steel girder first and then service load, the bridge cost depends on p + q and the overall structure efliciency
building the concrete slab and the weight of the steel girder is relatively small. can be expressed [46] as
Just for a first comparison, let us assume we have to launch a bridge with 50-m P
spans and a 13-m-wide deck slab. A PC superstructure with internal prestressing ps p+q (2-10)
may have an average concrete thickness of, say, 0.60 m, which results in a dead load
of 195 kN/m. A prestressed composite bridge may have an average concrete thick- The reduction of dead load limits the influence of launching stresses and increases
ness of 0.35 m, which results in a partial dead load of 114 kN/m; adding the weight the general efficiency of the bridge by creating a reserve available for service loads.
of two corrugated plate webs, say, 6 kN/m, the overall dead load is 120 kN/m and Since the cost of both materials and construction equipment depends on dead
the weight saving is 38%. Finally, the average thickness of the concrete slab of a load (that is, on the area A of the cross—section), and the moment of inertia I is the
composite box girder bridge may be 0.28 m, which results in a partial dead load of main parameter in estimating carrying capacity, the efliciency of the cross—section
91 kN/m. Adding the weight of a U-shaped steel girder, about 25 kN/m, the total can be evaluated in terms of moment of inertia reached, with self-weight being
dead load is 116 kN/m and weight saving is 41%. With these spans and deck , equal; that is, radius of gyration [45]
dimensions, a composite box girder is not much lighter than a prestressed com—
posite bridge. The narrow bottom slab of the latter does not increase the total
r z— (2.11)
concrete quantity much, and the high efiiciency of prestressing strand significantly
With area being equal, in a section of depth H whose centroid is distance 211 and 21
decreases the total weight 0f the steel webs. These economic aspects will be explored
respectively from the upper and lower edge, the highest radius of gyration is
in Chapter 4.
obtained by concentrating masses at the edges:
At the moment, it is important to note that 22% of the dead load of this
_ composite box girderis dilate,theU-girder and ,theremaining, 78% to the concrete, rrznax : ZuZl , (2.12)
 62 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 63

and the cross-sectional efficiency can be expressed in relative terms as flexural beyond that critical cantilever Lcr for which the negative moment in the cross-
efficiency, pf: section above the first support is equal to the design negative moment (perfectly
fixed beam) of the rear zone:
r
2 I
pf : — — (2.13)
2 __ —
qLé _ 4L2 (2.14)
2 ‘“ 12
rmax ZuZlA

The closer the actual value for pf is to the ideal value of l, the better the fiexural
and which is LCI = 0.4lL. There are three possible solutions:
efficiency. Generally, PC box girders have flexural efficiency of about 0.55, and
prestressed composite bridges can reach 0.70. 0 Limiting the temporary stresses by reducing the spans (i.e. increasing the
Considering a continuous beam composed of many bays of the same span and number of supports by inserting temporary piers between the final piers).
subjected to a constant dead load, far from the ends, the cross-sections above the 0 Limiting the difference between the temporary stresses in the front and rear
piers remain vertical, and the static scheme of each span is that of the perfectly fixed parts of the deck by supporting the overhang with a cable-stayed scheme.
beam. On the contrary, during launching, just before reaching the next pier, the 0 Reducing the cantilever weight by using a lighter steel launching nose.
front zone of the deck overhangs the entire span (Fig. 2.31). The negative moment
In the early applications of incremental launching, the final spans were reduced with
in the deck cross-section passing above the first pier is six times greater than that at
temporary piers distributed along the entire length of the bridge. In fact, it was
the rear supports and the shear doubles.
observed that the use of just one temporary pier located below the front overhang
The launch bearings transfer significant forces (the support reactions of the
involved significant time and costs, acceptable only for spans so long as to make
superstructure), and their most logical location is therefOre directly. below the
launch-stress reduction advantageous in the rear section of the superstructure also.
webs of the cross-section. On the other hand, the launch surface must be regular,
It was also observed that the use of many temporary piers in each span did not
and this requires a cylindrical geometry of the solid to be launched (in most cases, a
actually produce the dramatic stress reduction that might be predicted at first sight,
constant depth of the cross—section). With these technological restraints and the
owing to the different elastic and thermal settlings of the supports, the construction
stress differences of Fig. 2.31, a sizing carried out for the temporary stresses of
tolerances and the fiexural stiffness of the superstructure. This resulted in the
the front section would burden the entire deck design, whereasa sizing optimized
general use of a single temporary pier per span.
for the rear section would be inadequate in the front one.
Nowadays, the cost of the temporary piers for PC bridges (even if prefabricated
This second alternative is clearly less expensive. Therefore, it is necessary to
and reusable), their foundations, and the additional labour they require is justified
introduce devices that control the stresses in the front section of the superstructure, only in the case of very long or variable spans, where savings in launch pre—
stressing (unless made unnecessary by a higher parabolic prestressing) and
reinforcement" may be competitive. The use of temporary piers is also infrequent
—25o.o in launching the steel girders of composite bridges, although in some cases it has
—2oo.o — permitted the launch of the composite section complete with the concrete slab.
, \
These limitations narrow the field of application of the principle of generalized
bending moment (MNm)

_150.0 _ If, I,»

launching stress reduction. Instead, the stress differences along the superstructure
/ r {I [I [11/ /1\
are normally limited to make it possible to overcome the full span with a
_100.0 _
/
cantilever.
-50.0 —
- N 0O
In the first of these techniques, the cantilever is supported by means of a stayed
«onoouuouuuum “onuunmuu .01 oum"WN l I!
.«Mwo ouonuuwow h.“ «ouuouomo»
v.‘:¢:¢:¢:¢:¢3¢30203€03939:¢‘0‘9‘0‘9’0‘09/4‘:*3’3‘3'3‘3'3‘3’3‘o’o'o’o’o‘o’o’otoz02¢”. .&&t
Oouonu’uuuuuw «unom-«n
om.,..§¢§ §oo9
,.:.:¢,
W”.
m’o ’o‘\
to
. _ _. .§.§.:.:\\\
scheme comprising a tower integral with the superstructure and strand tendons
AAAAAAAA ,—-!=:\,‘ , -, v
anchored symmetrically to the tip of the cantilever and two spans behind, thereby
"
a.” "‘§‘¥ v '
".
000000 \
oooooo
'C‘ '
we \
§t“9§§\
\\““.‘}§
2‘

creating a sort of external prestressing with great eccentricity. In relation to the

position reached by the tower during launch, this scheme requires continuous
llllllllilllllii
adjustment of the force in the stays because of the concentrated action exerted
r"
"
—
—

*
—

—
—
1700
1800
1900
2000
210.0
2200
2300
2400
2500

OOOOOOOOOOOOOOOO
FNC‘DVLOCDNQCDOFNCOVLOCO
1.
by the tower on the deck. Just before the contact with the pier the tower passes
nearby the front support and the maximum pull in the stays supports the cantilever
Figure 2.3]. Bending moment envelope in a bridge with 45.5-m spans and 0.207 and reduces its elastic deflection. However, when the new pier has been reached the
. _MN/m ._dead_,_.loed,, without correction devices, inethe front span _ pull must be relieved so as not to increase the positive moment in the front span
64 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 65

with the concentrated action of the tower. Resort to this technique is discouraged
not only by the longer time required for the continuous adjustments of the stay
force and the deflection of the superstructure, but also by its intrinsic delicacy and
complexity, presenting considerable stress concentrations and serious risks in case
of errors.
The second technique provides, instead of a support for the deck end, a limitation
of its cantilever weight by means of a lighter extension — the launching nose — that
anticipates the contact of the deck with the pier. The launching nose is a ‘passive’
solution, perhaps less elegant than the stayed scheme but much simpler; it is safe,
fast and economical, so its adoption has become virtually standard in full-span
incremental launching of PC and prestressed composite bridges. Even with the
use of temporary piers the reduction of the cantilever weight is advantageous,
and it is essential in the longest spans. The stayed scheme also takes advantage
of it. The launching nose thus characterizes practically all the applications of
incremental launching.

2.7.]. Launching nose

The role of the launching nose becomes more and more important as the
self—weight of the superstructure increases. Composite bridges are generally built Fig. 2.32. Tmssed launching nose (author)
by launching the steel girder and then casting the concrete slab. As the self-
weight of the steel girder is only one-fifth of the final cross-sectional weight, the
stiffness-to-weight ratio is favourable during launching and the use of a launching 0 the nose length compared with the span to be overcome, Ln/L
nose may sometimes be superfluous. When necessary, short, light trusses are used, o the unit weight of the launching nose compared with the unit weight of the deck
like that in Fig. 1.1. The heavier prestressed composite bridges are in an front zone, qn/q
intermediate situation: their stiffness-to-weight ratio is higher than that of a o the flexural stiflhess of the nose compared with the one of the deck front zone,
conventional PC superstructure, and relatively light noses can be used in these EnIn/EI.
cases as well. The most demanding launching noses are those for conventional
PC superstructures. The model developed in the following chapter makes it possible to illustrate the
.
In the 30 years during which the incremental launching construction method influence of these three parameters and to optimize rapidly the nose—deck inter-
has been used, launching noses of all types have been used. They have been more action [47], avoiding the trial and error use of sophisticated calculation methods.
or less rigid trusses (Fig. 2.32) or plate girders, made of steel or concrete, with As this model neglects shear strains, results may be approximate in the case of
front realignment sledges or hydraulic systems to recover the elastic deflection. prestressed composite bridges, which are generally affected by significant shear
These launching noses have sometimes been designed according to customs and deformations.
experience, but often they were already available and therefore ready to be
reused. _ 2.7. I . I. Nose—deck interaction model
The most common type of launchingnose for PC bridges consists of two steel In a PC bridge, a correct presizing of the nose—deck system can start from a nose
plate girders connected to the front end of the superstructure with their bottom length Ln equivalent to about two-thirds of the typical span:
flanges aligned with the lower edge of the deck. Acting as an extension of the
superstructure perfectly integral at the joint, the launching nose affects the stresses Ln g 0.65L (2.15)
in the superstructure (or, more exactly, in the continuous beam comprising the deck
and the launching nose itself) and permits their homogenizing. and from a first attempt unit dead load qn obtained from the following equation:
The behaviour of the nose—deck elastic system is governed by three dimensionless
parameters that describe its geometrical and mechanical characteristics: q, = kLfi (2.16)
66 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 67

where, in kN and metres, it is 0.012 5 k 5 0.020 for highway PC bridges and At the start of launch, for or = 0, the bending moment in the support section B is
0.018 5 k g 0.030 for railway bridges, with progressively higher values in the that of the cantilever nose, which in a dimensionless form (the moments due t
case of wide and heavy superstructures. These approximate values for LH and cantilever masses will be marked with an asterisk) is '
qn can be used as a starting point for the optimization of the nose—deck
interaction. ME 1g, L 2
In order to develop a theoretical model of the nose—deck interaction, the follow-
ing assumptions are made [48,49]:
217:2 = ‘2? (In) (2'18)
Once launch starts, this increases (in absolute value) with the dimensionless law
0 The nose and the superstructure are of constant stifiness and weight: since
they actually vary both in the launching nose and in the deck front zone, it is M; a2 a,( a)
correct to consider their weighted average. The concentrated load of the end q 2 g L (2.19)
diaphragm of the superstructure, the temperature gradients and the mis-
placement of launching bearings, have a significant influence on the launch The moment MC at the third support C is
stresses. These load components have not been included not to excessively
complicate the model; however, practical applications need to account for
(2.20)
these aspects, which can be easily introduced through the related support 9L2— k3+qL2 k3+qL2
rotations.
o The number of equal spans behind support C (Fig. 2.33) is assumed to be so with the coefificients k, of Fig. 2.34. Finally, the support reaction at the front
large as to assimilate the superstructure to a continuous beam composed of support is L
infinite bays of constant span L.
0 Launch prestressing is centroidal also in the deck front zone, so as not to fi_fi_{g (H002 +———
an 1+0, +———
a (2.21)
introduce additional redundant unknowns. qL—qL2 2 qL 2L

Defining the progression of launch with the distance x of the nose—deck joint Once support A has been reached, the nose tip is hoisted until it realigns with the
section from support B (Fig. 2.33), the cantilever configurations assumed before theoretical profile of the deck in order to permit the subsequent sliding on launching
the nose reaches support A vary from the starting position, x :— 0, to the contact bearings. The recovery of the elastic deflection creates a positive moment that
one, x = L — Ln, and the dimensionless launching progress reduces MB and the second launch stage starts (Fig. 2.35), in which the nose slides
on support A until the concrete deck arrives. This stage is defined by
05 . —_ ’— 0 (2.17)
1 T /.T...,, “Abuz 1. .
i —1_1n/LJ < u: 2 1.

L L L qL3 qL3
k1: —k2=—k3:—_k4:— 5:—~—-———
3E1 6E1 NEE] 24EI 24¢§E1

k6 ml +0 I) ”in MM” 3)
L E] 3 L 21 OZ
:_1 — 1— =— —_"‘

+EI l ozz+oa3
E1111, 6 2 3

qL 3 qn 2 01 2 or 3 oz 3 05 4
k=—~— —l— —-———
8 2EI{ “(2 0‘)+q( ”K2 3) — —-—
(3 4)i+

-
qL3 EI
[4030 ~— 04)3 +g-él-(4015 —— 15054 + 20013 — 10052 +1)]
24E] 15,],

Fig.-. . 2:3 .3 NQse—djeck, system in the first phase, of{launch .Ez'g....2.34.. Elastic coefficients for nose—deck interaction
 68 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 69

—0.16

moment at support B (M/q1A2)

0.025
\
Ln _o.12 /0.’65A \ \

—0--
//0'1 x

“0‘08
—1/12 4%“
/// I /
' _

—0.04 /

bending
Fig. 2.35. Nose—deck system in the second launch stage
0.00
With the coefficients of Fig. 2.34 one obtains the reaction RA at the front 0.00 0.25 0.50 0.75 on 1.00
support:
Fig. 2.36. Progression of MB with launch for Ln/L = 0.80 and qn/q = 0.10 in relation
[(5 +k4 ) 1 [(7 gm (Ln )2 to the relative flexural stiffness, EnIn/EI
—k -—k —k __._+__ —+ot--l
RA:(k3+k1 2 4 8 q q
c713 k1+k6 —k§/(k3 He) flexural stiffness and tends to an end-of—launch (EOL) value that, setting a = l in
2
(2.23), is
+1+%§(a+1fi) (2.22)
MEOL

and the value of MB in the second launch stage —0134qn

_. g L“ L 2 0106
. .
(224)
61L2
MB _ ME RA
q_q qL
(2.23) As in the cantilever configuration, for o: < 0.20, in this final stage of launch, with
a > 0.90, MB does not depend on E1111, /EI. Therefore the flexural stiffness of the
where M; is the cantilever moment (2.19) and the reaction RA intervenes only for nose can be used only to prevent the launching moments after contact with support
l—LMLsagL A and until stabilization of the EOL value (for 0.20 < a < 0.90) from exceeding the
Equation (2.23) shows that, to be effective, a launching nose must at the same moments at contact and at EOL, since deck oversizing is generally more expensive
time reduce the cantilever moment 1%E and ensure high values of RA. Therefore it than stiffeningthe 'lauHChing” nose. ' '
must satisfy two a priori contradictory requirements: , Optimum stiffness, in this case, furns out to be around En]11 /E1 = 0.200 (thicker
line). Greater stiffness is pointless, since it cannot prevent reaching the EOL
0 On the one hand it must be light, to reduce ME, and sufficiently long that at the
moment, and less stiffness increases the negative design moment of the super-
contact with the pier the portion of cantilever concrete deck, which is much
structure. Moreover, the maximum cantilever moment is much lower than the
heavier, is as short as possible.
EOL one: contact with the pier is uselessly anticipated; i.e. the launching nose is
0 On the other hand it must be stiff, so that recovery of the elastic deflection at
excessively long, and it is convenient to shorten it.
support A produces an initially high reaction RA that immediately increases, as
On the other hand, in the case of a short nose (equal relative weight but
launch proceeds, to balance the increase of ME caused by the progressive
Ln/L = 0.50), the progression of MB is as given in Fig. 2.37. (Note there that,
lengthening of the concrete segment affecting the first span.
with a short nose, the maximum cantilever moment is higher than the EOL one,
For a long launching nose, Ln/L : 0.80, of relative weight on /q = 0.10, the pro- which again is higher than the support moment of the perfectly fixed beam.) Until
gression of MB with launch in relation to the flexural stiffness is described in the contact with support A it is favourably affected by the lower weight of the nose
Fig. 2.36. (Note there that, with a long nose, the maximum cantileVer moment is but unfavourably by the greater length of the cantilever concrete segment. On
lower than the end-of-launch one, which in turn is higher than the support moment reaching A ME is reduced more effectively, and at the EOL it stabilizes on the
of the perfectly fixed beam.) Until contact with support A, for at < 0.20, MB grows value given by (2.24). The effects of the flexural stiffness are comparable as well,
with (2.19). On reaching the support, the positive moment generated by recovering and also in this case optimum stiffness should be around En]I1 /EI = 0.200 (thicker
.. the elastic. .deflectionrreducesME,and as launchcontinues MB depends on the _,line_).....However, keeping MBbelow the EOL value during the intermediate launch
7O BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 71

—0.16
The diagrams examined so far analyse the behaviour of the nose—deck elastic

moment at support B (M/q1A2)

/0\05‘0.025 system for a given relative weight; Fig. 2.39 compares the diagrams of MB calcu-
—0.12 //‘ 0 '\
1 \\ lated for the optimum parameters obtained for qn/q = 0.10 as a function of this
' 02”,;— ratio. MB depends significantly on the relative weight of the nose only during the
—1/12
first phase of launch. In the second phase it is limited to a family of curves
—0.08 / contained in 15—20% of the EOL value, which twist because of M21.
The difierent relationship between MB and qn /q in the two phases of launch
means that, once the system has been optimized for a given relative weight, an
—0.04 / increase in qn/q modifies the diagram of MB towards the type shown in Fig. 2.37
bending (‘short’ nose) and vice versa its reduction modifies the diagram towards the type
shown in Fig. 2.36 (“long’ nose). To force the value of M; on contact t0wards the
0.00
0.00 0.25 0.50 0.75 or 1.00 EOL moment it is then necessary to modify the nose length, shortening the canti-
lever concrete segment (i.e. lengthening the nose) as the relative weight of the latter
Fig. 2.37. Progression of MB with launch for Ln/L = 0.50 and qn/q = 0.10 in relation increases. I
to the relative flexural stiflness, EnIn/EI Figure 2.40 compares the diagrams of MB that restore this condition: starting
from qn/q : 0.10, as the relative weight increases equalization is obtained with
longer noses and for progressively lower values of MB, which tend to the perfectly
stages loses significance since this threshold is substantially exceeded in the previous fixed beam moment of the rear zone of the superstructure.
cantilever configuration. Proceeding by trial and error, the curve of Fig. 2.41 is obtained, which defines, as
In contrast to what occurs for the long nose, in the case of a short nose the an /q varies and for the optimum fiexural stiffness, the values of Ln /L that equalize
maximum cantilever moment is greater than the EOL one, and it is therefore the the negative moment on contact with the EOL one. This diagram makes it possible
first phase of launch to affect the design of the superstructure. Contact with the pier to determine easily the optimum length of the launching nose in relation to its
occurs too late; i.e. the nose is too short and should be extended. relative weight.
Proceeding by trial and error, the optimum nose length is obtained in Fig. 2.38: In conclusion, a correct design of the nose—deck system can be arrived at starting
with EnIn/EI = 0.200, the cantilever moment at contact is equal to the EOL one, from first-trial parameters obtained from (2.15) and (2.16). The values of LH and am
and in the intermediate stages it is lower. so defined will be refined until the equalization of the contact moment with the EOL
one is obtained. Then, the minimum fiexural stiffness necessary to contain the

-0.16 —0.16
0.025 /
0.25
moment at support B (M/q1A2)

moment at support B (M/q1A2)

—0.12 ///
/ \\*
o».05—\ 0.2

//g-;r-\§\
,,//,/°,;j/
—o.12 /// //

‘—1/12 // I '

i)
—1/12
—0.08 / ,, /

\\
\
\
\
/
_o.04 /
bending

bending
0.00 0.00
0.00 0.25 0.50 0.75 0L 1.00 0.00 0.25 0.50 0.75 Ct 1.00

Fig. 2.38. With qn/q = 0.10, the cantilever moment at contact is equal to the EOL one Fig. 2.39. Progression of MB for Ln/L : 0.65 and EnIn/EI : 0.200 in relation to the
foa/L2065 ,, . . ., . relativeweight qn/q. The thicker line is the optimum diagram offigure 2.38
 72 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 73

—o.12
In Fig. 2.42, the progression of MC (2.20) shows that, independent of the nose

0.06
0.12
bending moment at support B (M/q1A2)

0.1
0.14
0.16
ix
*7 length, the recovery of elastic deflection forces MC towards a single curve that
0

/l
AR stabilizes for a = 0.60 on a slightly lower value than the perfectly fixed beam
——-1/12
_008 // l / V/i/l/l/ Il/Il/
one. From this point onward, except for irrelevant additional fluctuations on

2
contact with the following piers, the support moment is no longer affected by the
front elastic system.
In short, with regards to the negative moment, it may be concluded that by

%
adopting launching noses of suflicient flexural stiffness and adequate length for
the relative weight, the deck zone affected by the nose—deck interaction can be
limited to a length of about 1.5L, increasing to 2L in the case of very deformable
0.00
noses.
0.00 0.25 0.50 0.75 Oi. 1.00 The maximum positive moment in the front span is
Fig. 2.40. Obtaining equalisation of bending moment on contact and at the EOL by Mm
AEB=—RA 1—a—@——‘l
L +_1 ._A
R 2 +_@
1 L 2 519...
-
adjusting the nose length as qn/q changes qL qL
.3 (2.25)
q L 2 qL 2q ~L q
and Fig. 2.43 illustrates its evolution in the case of a long launching nose.
moment in the intermediate launch stages under this common threshold may be In contrast to what happens to MB, the flexural stiffness of the launching nose
defined. For this purpose, it is necessary to remind that the diagrams in Figs 2.36— aflects only the progression of M21333 towards the maximum value, which is in any
2.43 are only aimed at illustrating the influence of different parameters 011 the event reached for a 2 0.90. It then starts to decrease because of the eflect of M111. By
behaviour of the nose—deck elastic system. When presizing a launching nose for a shortening the nose, the progression of M23 is substantially similar. The maximum
practical application, the concentrated load of the front diaphragm of the super- value is obtained a little further on in the launch and is slightly higher because of the
structure, the expected misplacement of launch bearings, and the temperature smaller contribution of M21. Therefore, the flexural stiffness of the launching nose
gradients stated by the design code, should always be taken into account. This does not govern the positive moment behaviour of the system, which depends
can be easily achieved by inserting additional components of the rotations of the essentially on the continuous beam static scheme of the superstructure and can be
support sections in Equations (2.20) and (2.22). modified only by means of ME. A synergistic action of increasing the length and the
weight of the nose can reduce at the same time the negative equalized moment and
the maximumpositive moment- Conseqnentlynthechoice among available noses can
1.000 be made by also taking these savings into account.

-0.120

moment at support 0 (M/q1A2)

relative nose length

0.750 3’ N§ -1/12

M//
‘0-080 \ , \a—k‘n
~—%§
0.6x

bending
0.020 0.060 0.100 0.140 0.180
relative nose weight
-0.040 '
0.00 0.25 0.50 0.75 o. 1.00
Fig. 2.41. Evolution of the relative length Ln/L that equalises the negative moment as
, .. .. . (In/é] changes- -v--Fig.-2.42. Evolution of MC for a’iflerent values of Ln/L
74 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 75

0.08
2.7.1.2. Sizing ofthe launching nose

f The optimum stiffness and dimensions of a launching nose depend on many

moment (M/q1A2)
factors, economic considerations included. The convenience of amortizing the
investment over several jobs suggests designing launching noses to be as versatile
as possible. Since the incremental launching construction method for PC bridges
0.04 reaches spans of about 60 m, it may be convenient to design the launching nose

A-B span bending

directly for the longest spans even in the case of initial use on shorter spans, and
to adapt it to the lower depth of the superstructure. The probable need to reduce
the nose length can be addressed by subdividing it into spliced elements, while its
greater stiffness and weight further reduce the launch stresses in the deck front
zone. The role of the launching nose, originally aimed at controlling the negative
0.00
0.40 0.60 0.80 Ct 1.00 launch moments, can thus extend to the control of the positive moment in the
front span. The convenience of the higher investment required for this purpose
Fig. 2.43 . Evolution of the maximum positive moment in the front span for Ln/L = can be verified based on the amortization expectations and the savings in the
0.80 and qn/q = 0.10 in relation to EnIn/EI front zone of the superstructure. -
In most cases, a launching nose for a PC bridge is composed of two braced steel
plate girders. Launching noses made of PC have been adopted in the case of
Finally, it should be noted that the consideration that the first bay of the impossible reuse or difficult transport, to be demolished on completion of launch.
continuous beam is in any event subject to higher positive moments is deceiving, Prestressed composite noses (steel plate webs and PC flanges) have been used for
since correct bridge design involves shorter end spans just to compensate these the launch of extremely heavy and stiff superstructures.
increases, while during launch the front section of the deck has to overcome all If there are no limitations in the depth, truss noses are very efiicient. However,
the longer intermediate spans. the bottom chords of these noses have to withstand the local stresses generated by
For the B—C span one obtains the migration of the support reactions and are therefore designed for high shear
forces and local moments. Together with a great quantity of manual welding, this
MBLEC
max
_ i (MC MB MB
1) 2 gig (2.26) makes the use of truss noses infrequent for the launch of PC bridges, and the
C] E_q_1? 2 general trend is towards the use of plate girders composed of transportable elements
connected by vertical splices with high—grip bolts (Fig. 2.44). On the contrary, truss
and the relative diagram, compared with the progression of MC , confirms that noses are often the best solution rortrieiaufieh of the steel girders of composite
the rear zone of the deck, unaffected by the nose—deck interaction, begins with bridges.
reasonable approximation at a distance of 1.5L from the front end. Preloaded pins (see Fig. 2.46) are frequently used to join the flanges since con-
To draw some conclusions, the design of a PC bridge or of a prestressed compo- ventional splice plates with high—grip bolts require hoisting of the superstructure (to
site bridge built by incremental launching depends on the technique adopted to avoid interference with the launch bearings) and consequently the discontinuous
control launching stresses, a technique that in all cases involves the use of a launch procedures of Fig. 1.5. The use of preloaded pins specializes the action of
launching nose. The optimization of the nose—deck interaction produces savings connection devices: the bolts along the gusset plate mainly resist shear forces while
in structural materials. It may be easily carried out by means of the described model the pins, located in widened holes to avoid lateral contact, resist only axial forces.
(modified to include the effects of the front end diaphragm, temperature gradients, Horizontal splices in webs are necessary only in the case of deep girders for railway
bearing settlement, and the cable—stayed front system, if present) and an ordinary bridges or long spans (Fig. 2.45) and are generally designed for high-grip bolts.
spreadsheet. Adopting the most common scheme, which provides for using just a The bracing system of the launching nose is almost always designed specifically
launching nose, the evolution of the negative moment in the superstructure may be for the bridge, since it is convenient to align the nose girders with the bottom
adjusted by acting on the parameters governing the nose—deck interaction. Among corners of the deck cross—section to use the same launch bearings. The use of
several combinations equally suited to the negative moment, some of them improve wide launching bearings might permit the reuse of a bracing system designed for
the behaviour at the positive moment as well, with a higher cost of the nose more closely spaced girders; however, the high stresses generated in the front
balanced by lower costs of structural materials. diaphragm of the superstructure by the transverse eccentricity of the forces and
the moments transferred by the nose girders make this solution infrequent. When
76 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 77

PLAN

lap plate —hor‘izontal bracing

segment 1 segment 2 segment 3

‘4’...
H}...—

.4»...
H}..—

SIDE VIEW

j lap plate tiFting points Jack

LM

Fig. 2.45. Horizontal field splices facilitate transport of deep girders. The bottom
flange is strengthened with continuous box stifleners (author)

connection system approach device

necessary only in the central launch phases. The front section can therefore be
Fig. 2.44. Typical segmentation of a launching nose lightened, by decreasing both the web thickness and the girder depth, with savings
in costs and cantilever masses.
The design stresses should be kept reasonably low to accommodate load
buying a second-hand nose, therefore, the cost of new horizontal braces and cross— diflerences between the two girders due to construction and assembly tolerances,
frames, should always be taken into account. ‘ flatness defects in the bottom flanges, and vertical misalignment of launching bear-
The external edge of the bottom flange of the launching nose may be aligned ings. Load diflerences can be significant in launching noses with closely spaced
with the external corner of the bottom slab (PC and prestressed composite bridges) girders, and diaphragms and cross-frames are generally too flexible to eflectively
or of the bottom flange (steel girders) of the superstructure. This allows use of the control them. Consequently, each nose girder should prudently be designed for
same lateral guides for the nose and the superstructure. This solution is particularly 75—100% of the whole support reaction transferred by the nose. This criterion
advantageous in the launch of steel girders since the lateral guides can be small low- should also apply for the local prestressing systems that join the nose girders to
friction plates acting directly against the external flange edge. the superstructure.
The guide action on the launching nose should be aimed only at centring the The design of a launching nose starts from the envelopes of the bending moments
support reaction along the nose webs, since alignment corrections of the super- and the shear forces resulting from analysis of the continuous beam. The load
structure by acting on the nose would cause excessive stresses in its bracing-system. envelopes are determined in each nose girder, and the depth of the rear nose section
Should major alignment adjustments be necessary, they can be generated by is taken equal to the overall depth of the superstructure. .
shimming one of the lateral guides at the most advanced pier reached by the super- The transverse Width of the bottom flange may be chosen so as to use the same
structure. ’ neo—flon pads used for the superstructure. Defining the transverse width of a
When designed specifically for the bridge, the highest depth of the launching nose neo-flon pad with bnf, its launchwards length with Lnf, and the total length of
is generally equal to the depth of the superstructure and is maintained Only for few the neo-flon pads on a launching bearing with ELnf (see Fig. 3.23), the calculation
metres from the joint section. In fact, the evolution of the negative moment in the of the contact stress is immediate. The transverse bending moment generated by the
superstructure. is such that the, full . ,flexural stiffness _. of the launching nose , is i 7 contact stress governs the thickness tb of the bottom flange. The transverse flexural
-
 78 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 79

stress generated by the support reaction RA is calculated by assuming that

the neo—flon pads are centred along the web and the support reaction is equally
distributed between the two nose girders
_ 3RAbnf
__ 2.27
at 8% 21:1,, ( )
Then, this transverse stress is increased to account for the eccentricity in the support
reaction and in its local distributiOn, and it is finally compared with the axial stress
allowed by code.
Once the bottom flange area has been determined, the top flange area is calcu-
lated with a trial—and-error process based on the expected web thickness, the result-
ing location of the cross-sectional centroid, and the comparison of the edge stresses
under the maximum and minimum moments with the allowable stress. The use of
diflerent types of steel for the opposite flanges of the nose girders is infrequent. The
final dimensions of the top flange plate are determined based on the influence of the
width—to—thickness ratio on instability of the compression flange. Additional infor-
mation can be found in Chapter 4.
The web thickness, tw, is determined from the envelope of the shear forces trans— Fig. 2.46. Vertical flange stiffeners combined with pre-loaded pins (author)
ferred by the nose girders. The design tangential stress is kept low to account for the
uncertainties about the distribution of support reactions between the two nose
girders, and to decrease the number of vertical web stiffeners, whose spacing can 0 Vertical compressive stresses are generated by dispersion of the support reaction
thus be determined based on the bracing requirements of the nose girders. Based on into the webs.
the final web stiflener geometry, the code equations for web panel buckling define the 0 Local strut—and—tie force systems arise from the application point of the support
minimum thickness of web plates, which should not be less than 10 or 12 mm to ease reaction to the shear keys welded on to the contact plate with the superstruc-
plate handling in the workshop and to reach adequate overall stiffness. Finally, local ture. When the shear keys are located at the lower edge of the superstructure to
buckling of web panels under the thrust of compression flange is assessed. transfer shear forces directly to the concrete webs (Fig. 2.47), transmission of
By assuming a 450 dispersion of the support reaction into the bottom flange and, shear forces is concentrated in the lower region of the nose girders.
again, a perfect load distribution between the two no e girders, the vertical com— a High local stresses are generated by the anchor systems of the prestressing bars
pressive stress at the bottom edge of the web panel above the launching bearing is that clamp the nose against the front end of the superstructure. These anchor
systems also concentrate most of the longitudinal flexural stresses produced by

OZ "— 2tW(2Lnf + 2th)

(2.28) the support reactions.

Detailed criteria for assessment of the web panels are provided in Chapter 4. If the The principal stresses are sometimes calculated in these regions of the nose girders
transverse stresses in the bottom flange (2.27) or the vertical stresses in the web by assuming that the stress-state is a biaxial one. In the most delicate cases, a 3D
panels (2.28) are excessive, closely spaced flange stiffeners can create additional finite—element analysis of the whole joint region (rear portion of the nose girders,
support lines in the flange and improve dispersion of the support reaction into front diaphragm of the superstructure and front portion of the box girder section) is
the webs (Fig. 2.46). The triangular box stiffeners of Fig. 2.45 are less expensiVe necessary.
in terms of hand welding but complicate the use Of preloaded pins for the vertical Clearance requirements of the tensioning jacks of the anchor system between the
field splices due to jacking clearances. launching nose and the superstructure rarely permit adequate distribution Of
In proximity of the rear nose end, the stresses in the lower region of the webs are bottom flange stiffeners (Fig. 2.46) or adoption of the box stiffeners of Fig. 2.45.
high because of the Concomitance of several stress states. In addition, it is often necessary to use four clamping bars on each level (see bar
holes in Fig. 2.47) to increase the vertical lever arm of joint prestressing. In this
0 Longitudinal flexural stresses are generated by the support reaction of the con- case, anchor boxes welded to the nose girders are often necessary, as in Fig. 2.48. A
tinuous beam and by the front nose portion overhanging beyond the support 3D finite-element analysis facilitates nose detailing and the design of transverse
.. point. -...pres.tressing or reinforcement of the front diaphragm of the superstructure,
 80 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 81

Fig. 2.47. Gusset plate and concentrated shear key of the launching nose ofFigure 2.48 Fig. 2.48. Anchor boxes for the prestressing bars of the joint system (author)
(author)

especially when the concrete webs are inclined. Similar considerations govern the
design of a launching nose for the steel girder of a composite bridge. In this case, the
rear end of the trussed nose is often a plate diaphragm directly welded to the steel
girder as in Fig. 2.55.
Nose girders are connected to each other by cross-frames and horizontal bracing.
Cross-frames reduce the load diflerences between the two girders due to eccentric
support reactions; when the nose is twisted, they also oppose distortion. Cross-
frames are designed as crosses or transverse frames with stifl joints between the
vertical web stifleners and horizontal cross-beams placed in the plane of the
horizontal bracing (Fig. 2.49), sometimes supplemented by diagonals that further
improve stability of the compression flange. In this case, it is advisable to design the
cross-beams so that the distance between their bottom flange and the bottom flange
of the main girders is at least one-half of the transverse width bf of the latter [11].
The walkway to the front work platform for nose realignment at contact with the
pier (see Fig. 2.23) can be placed directly on the top flange of cross-beams. In
narrow noses (Fig. 2.48) the longitudinal spacing of cross—frames is often such
that the panels of the horizontal bracing are square; the rectangular scheme is
more common in wider noses, as in Fig. 2.49.
The role of the horizontal bracing is to stiffen the nose girders in resisting
the wind loads and the transverse forces produced by the lateral guides that centre
the nose. webson. the launching.,bearings- Horizontal bracing is designedwith cross Fig. 2.49. Framed crossediaphragm (author)
 -
Mimi: .
82 BRIDGE LAUNCHING

iii-h: .
INCREMENTAL LAUNCHING OF BRIDGES 83

»
or K schemes, where the verticals are the cross-beams and the chords are the main most of the clamping prestressing is applied at the bottom flange level. In this node,
girders. the distribution of the clamping bars is governed by technological restraints such
as the nose geometry, the diameter and length of the stressing jack and the diameter
of its contact plate alongside the anchor nuts of the tensioned bars. The most
2.7. I .3. Coupling to the superstructure practical solution for a satisfactory transmission of longitudinal prestressing
The conneCtion system between the launching nose and the superstructure must between the nose anchor system and the prestressing tendons of the superstructure
guarantee the transfer of the bending moments and the shear forces by providing consists in modifying the layout of the latter so as to avoid the bar anchorages and
perfect structural continuity. On completion of launch, the connection system must to introduce the tendons deeply between the bundles of bars to make the stress
also permit an easy disassembly of the launching nose. transmission as direct as possible. Although tendon curvature causes prestress
In PC bridges, the most effective solution for the transfer of the positive and losses, with this scheme the prestressing forces migrate from the clamping bars
negative moments is a local prestressing system composed of alloy steel bars. Strand towards the prestressing tendons through longitudinal compressive stresses. A
tendons are less frequently used because it is more diflicult to relieve them for scheme with bundles of clamping bars and parallel prestressing tendons would
maintenance or monitoring purposes. The bar portion embedded into the super- involve transfer of prestressing forces through shear stresses and would require
structure is generally l.5—2 times as long as the deck depth to obtain an anchorage the use of transverse prestressing to avoid cracking.
as long as the whole deck depth both for the clamping bars and for the launch Splicing of joint prestressing can be complex also in the upper node, especially
prestressing tendons of the superstructure (Fig. 2.50). when inclined webs move the upper cross—sectional node of the superstructure away
The top flanges of the launching nose are sometimes anchoredto temporary from the top flange alignment of the nose, whose webs are vertical (Fig. 2.51). This
reinforced concrete blocks cast directly on to the deck slab. This solution may be can require a mild transverse prestressing in the top slab that absorbs the principal
necessary in the presence of many anchorages of launch prestressing in the top slab tensile stresses resulting from the shear forces caused by this eccentricity. However,
region affected by the nose, but inclined deck webs or available second-hand noses most of the joint prestressing is concentrated at the bottom flange level and the
deeper than the superstructure are also good reasons for this solution. splicing forces in the top slab are rarely critical.
The positive moment transferred through the nose—deck joint is generally much Starting from the maximum positive and negative moments in the joint and the
higher than the negative moment due to the cantilever nose. As a consequence, distance between the gravity axis of the superstructure and the centroid of the upper
and lower bundles of bars (eu and el respectively), the calculation of the two
prestressing forces (the lower force Fl and the upper one Fu) is immediate. The
l (1.5%2)H clamping bars should be oversized with respect to these minimum requirements to
T HE
resist the load diflerences between the nose girders without joint decompression and
related bar overloading. When presizing the clamping bars, ailowance should be

/ / / / / / / / / / /
launching nose outlin -

\ n’"_“m*___.__

shear keys

\\
temporary concrete segment

Fig- 250. Side. View. Of the anctzar devices of the launching nose, .. . Fig... 2.5]. Front View of the anchor systems of Fig. 2.50
‘-
-:
‘:};s:i:_f
 84 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 85

made for re-entry of anchorages, friction losses, elastic shortening of the compres- are respectively compared with the allowable compressive stress and the least joint
sion concrete at tensioning and prestress losses due to time-dependent phenomena compression stated by the design code. Similar verifications are carried out for the
(launching of long bridges takes several months). Launch procedures and checklists negative moment generated by the cantilever nose and the downwards eccentricity
should also include periodic checking of bar tensioning and retensioning when a of the clamping prestressing. This last load condition sometimes prevails, as in
predetermined threshold is exceeded. The use of one or two hydraulic load» cells is Fig. 2.53.
advisable in the case of long bridges for earlier detection of excessive prestress losses In PC bridges, the design shear force in the joint, V, is transferred through single
in the joint system. shear blocks (Fig. 2.47) or more shear keys distributed along the contact surface.
When the depth of the superstructure allows comfortable work within the box Figures 2.50 and 2.51 illustrate a solution that avoids carving the front end of the
cell, the clamping bars can be tensioned from the rear anchorage. This eases bar superstructure with shear keys. In spite of these shear transfer devices, it is advi—
retensioning in case of need; bar geometry is less restrained, and the bundles of bars sable that the joint prestressing force guarantees joint stability by friction. With a
may be closer to the nose webs because there is no jacking clearance on the nose coefficient of friction of 0.4 for a steel-concrete contact and a safety factor of 2, we
side. This results in compact joints (Fig. 2.52) and no risks for workers during the obtain
operations of bar check and retensioning.
The joint surface, of net area Aj (after subtraction of bar holes) and moment of __I_ < 0.2 (2.30)
inertia I], is designed by assuming a linear distribution of the contact stress along
A+E‘
the joint surface. The edge stresses for the maximum positive moment When the launching nose is deeper than the superstructure and/or the deck webs are
inclined, the nose can be anchored by means of steel blocks (Fig. 2.54) or concrete
2Fu+Fl+M$ax+Fueu_FlelZ blocks directly cast on to the deck slab. In the case of concrete blocks, the solution
Cu
Aj [j u (2.29)
_Fu+Fl Mgax‘i—Fueu—Flelz 12.0
01

-—d|--—
*—_ 1

«F
I.
Ar

It:
’33,

1 14.9

-4>-—
T

Fig. 2.52. Bar tensioning from the rear anchorage results in compact gusset plates Fig. 2.53. Contact stresses (N/mm‘?) in the joint section of the launching nose of
._ (author) 0 , 4 ,, , , , , ,... w. Fzg248
86 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 87

with vertical reinforcement through the joint requires demolishing the block, Similar considerations govern the design of the joint between the launching nose
cutting the joint reinforcement, and patching the deck surface on completion of and the steel girder of a composite bridge. When the bridge has a composite box
launch. In both cases, therefore, the anchor blocks are joined to the superstructure girder section, the nose is complete with a plate diaphragm that joins the nose
by friction. The horizontal force transferred by the nose to the anchor block acts as trusses with the inclined webs of the U-girder. The nose—deck joint may be designed
a shear force Vh in the joint between the block and the deck slab. The minimum as a conventional grip-bolt field splice, although a welded joint on a shortest
vertical force Nmm to be applied to a concrete block by means of vertical pre- sacriflcal front overhang (Fig. 2.55) is often the most eflicient solution.
stressing bars is determined by applying a safety factor of 2— to a friction coeflicient
between two concrete surfaces of about 0.6: 2.7. I .4. Recovery ofelastic deflection
When approaching the pier, the nose tip is generally distant from the theoretical
Vh vertical profile of the launch surface because of the elastic deflection of the nose*
Nmin E E (2.31)
deck system.
The coeflicient of friction can be increased with distributed shear keys obtained by In the case of a regular sequence of spans and of piers with similar vertical
pressing corrugated sheet metal on the fresh concrete of the deck slab. In the case of stiffness, the nose tip deflection at contact is always downward. It can easily exceed
a steel anchor block, we have again 10 cm in the case of a PC bridge, and several tens of centimetres in the case of the
steel girder of a composite bridge (Fig. 2.32). The need to recover this deflection to
Vh
N ~ >m permit the continuation of launch might be solved with a hoisting wedge; i.e. by
H1111 — 0.2 (2‘32)

inclining the bottom flanges of the launching nose so that its advance on the pier
The use of some shear keys in the joint surface is always advisable, as in Fig. 2.54. forces its progressive realignment. In reality, this realignment method is used rarely
In the particular case of this bridge, the whole vertical shear force between the nose and only for steel girders, as the high horizontal forces generated by the wedge
and the superstructure has been resisted with vertical bars. Therefore, the front action can overload the piers. In addition, the launch bearings must be anchored to
bundle of vertical bars is particularly powerful. the pier cap to avoid accidental displacements or overturning, and must be articu—
lated to accommodate the large rotations of the support section.

Fig. 2.55. Joint diaphragm between the truss nose and the U—girder of a composite
,.,,,,bridge... (author)
 88 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 89

In most cases, therefore, realignment is obtained with hydraulic devices. For a PC

bridge, the simplest solution consists of placing on the pier cap, immediately before
the launching bearings, a pair of hydraulic jacks equipped with rounded top plates.
When the launching nose has arrived above these jacks, neo-flon pads are inserted
between them and the nose. Then, jacking realigns the nose and the neo-flon pads

Inn-—
permit the launch to continue until the nose tip arrives on the launching bearings.
This scheme requires jack transfer from pier to pier; in addition, the effective stroke
of flat jacks is quite short and the complete recovery of nose deflection can require
several hoisting cycles, with significant dead times. As a lot of workers are present
at the piers, these dead times are expensive. These weak points often make» it
convenient to use steel brackets hinged to the nose tip and moved by hydraulic
pistons (Fig. 2.23): when they reach the launching bearings, these brackets are
lowered onto neo-flon pads and forced until the elastic deflection has been
-—- deviation tower
completely recovered. When launching a steel girder, these pivoted arms are longer
and moved by less powerful long-stroke pistons (Fig. 2.32). removable pin

2.7.2. Stayed front system

A second solution to the need to reduce the transitory stresses deriving from
cantilever configurations consists in supporting the front end of the superstructure.
This is attained with a cable-stayed device composed of a tower hinged to the
superstructure and strand tendons fixed at the tower head and fan-anchored
symmetrically with respect to it, to obtain an adjustable external prestressing system
with high eccentricity (Fig. 2.56).
In the initial phases of construction, when the first two spans of the super-
structure are still in the casting yard (Fig. 2.57), a short launching nose and a
steel tower are anchored to the superstructure. The tower is positioned so that
when the nose tip reaches a pier, the tower is some metres behind the previous
pier. Transversely, the two legs of the tower are supported on to the deck webs by
means of long-stroke hydraulic jacks that control the vertical force transferred to
the superstructure and permit simultaneous adjustment of the pull of stays.
For this device to be effective in all launch phases, it is necessary to modify the
pull of stays continuously because of the concentrated-load that the toWer transfers concrete seat

to the superstructure. When the nose up approaches the pier the force is the highest
lllllllllll
IIIIIIIIII
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
IIIIIIIIIIIIIIIIIIIIIIIIIIIII
llllllllllllllllllllllllllllll

and the system is effective (Fig. 2.58), but when the tower overtakes the pier and
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
IIIIIIIIIIIIIIIIIIIIIIIIIIIIII
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
111111111111111111111111111111
IIIIIIIIIIIIIIIIIIIIIIIIIIIIII
ttttttttttttttttttttttttttttttt
IIIIIIIIIIIIIIIIIIIIIIIIIIIIII

advances to mid—span the same force would increase the positive moment and must
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
rrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

be reduced by lOosening stays. " ” " ' ' ' "" ' ' ' '
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

In practice, many factors limit theuse of this support scheme to the longest spans
Fig. 2.56. Cable—stayedfrom system
of PC bridges, and always in combination with a launching nose to reduce the
cantilever weight (Fig. 2.59), and with it the required capacity of the stays, as
. much as‘possible.

‘. 3 The need for a long casting yard (about two spans): before launching on the
bank....span,__the.superstructure must be. long.enough_to-,anchor the. stays and, to
90 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 91

/l\ Fig. 258. Position of maximum pull in the stays of the Boivre Bridge (courtesy Spie

I I I W Baligmllesl
M B I I

lh‘. lulklill‘l'tlll

Fig. 2.57. Launch phases with the stayed front system. In case of a short nose, each
launch phase is as long as the whole span

guarantee adequate safety against tilting, unless a temporary pier is added in the
bank span.
0 The complexity of calculations: the tension in the stays depends on the elastic
deflections of the stayed syStem and of the superstructure, on pier settling, on
bearing misplacement, and on the thermal strains both in the deck and in the
stays. In addition, non-linear stay behaviour reflected in the Ernst modulus
[5,27] depends on the stress level. Although the corrections deriving from the
apparent'variation in the elastic modulus of- stays are usually necessary only in" -
the analysis of large cable—stayed bridges, these corrections may be significant
in the front—stayed system because of the low tensionin the stays in the first
phases of launching. ‘
o The need to control and to continuously adjust the stresses in the tower and in
the stays. 7 p _
o The stress concentration in the deck region under the tower, which in spite of . 4 Fig. 2.59. Launch of the Charix Viaduct on 64-m spans (reproduced with permission,
-,stifl76_ning__isoften. afiectedby cracking,problems.. , ,v . . , , ,, , , ,, , , ,,,Scetar0ute~—Berenguier)
92 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 93

o The intrinsic delicateness of the solution, the risks in case of error, and the need At the beginning of launch the stays are unstressed, and they are not necessary
to reach the front pier at the end of each launch to avoid long stops with until a critical advance LCr that generates a negative moment in the front support
stressed stays. In the case of short launching noses, this, last requirement adds section equal to MEOL is reached. In dimensionless form (2.17)
further restraints on the length of the casting yard.
0 The irreversible deck deformations generated by creep of concrete in the case of
ac, L...
L q L+\/(q
=—=———
(1.. Ln
L) —1.2 6 8—~—
—-—— g qr Ln 2 an — e .2
L 2 +012
L.. (35)
2.
launch steps as long as the whole span. In this case, it is advisable to end each
launch with a front cantilever able to reduce the positive moment in the first
Beyond 0501., the action of stays should limit the support moment ME below or at the
span, or to adopt a relatively long launching nose that permits halving the
EOL moment until the contact with the pier; i.e. for over 5 OZ < 1 —— L1.l/L. By
length of the launch steps.
absorbing the excess of ME over MEOL with stays inclined at an angle )8 with respect
0 A longer front deck section requires higher axial prestressing than a conven-
to the horizontal, one obtains the following equation for the adjustment of the pull
tional use of the launching nose.
in the stays FS:
For all these reasons, the combined use of a relatively short launching nose
fizz; [3+@52+0°634@(fi)2_ 0106] (2.36)
L qL sm )8 2 q L 05 g L or
0.4 5 13 5 0.6 (2.33)
This value is only preliminary because the concentrated action of the tower can
and of a cable-stayed system with a tower of height ht generate excessive stresses in the second span [50]. In every launch phase, the pull in
the stays must be higher than a minimum value that depends on the negative
moment capacity of the front support section, but lower than a maximum value
0.40 5 L htL 5 0.44 (2.34) that depends on the positive moment capacity of the mid—Span sections in the first
_ n
two spans. Consequently, it is necessary to determine the diagram of the minimum
is generally limited to wide and heavy superstructures with 60—70—m spans. In this and maximum pull in the stays as the deck position changes (Fig. 2.60 shows the
case, it permits the use of launching noses designed for spans lO40% shorter. Over case of a very short launching nose) and the actual procedure of pull adjustment
these spans, the use of temporary piers is probably the best solution. must reside between these two curves. If they are excessively close or if they
The use of temporary stays is particularly advantageous in the case of a long intersect it is necessary to increase the axial prestressing in the. front zone of the
central span obtained by launching two bridge sections symmetrically from the superstructure, which controls negative and positive moment capacity.
opposite abutments with mid-span closure. The cable-stayed support of the two The tensioning diagram is discontinuous since pull adjustment occurs during the
cantilevers permits overtaking central Spans much longer than 100 m. If the central laumh stops.The calculation offitheeverticalload-onthe: tower is immediate, and
AAV‘.

span is designed with extradosal prestressing, the deviation towers of the external this value is applied with the hoisting jacks. In the case of a narrow pull“ adjustment
tendons may be used during launching to deviate the temporary stays.
Compared with the use of a launching nose only, the action of stays modifies the
evolution of the launch stresses in the superstructure. Before the contact of ’the nose
tip with the pier, the stays absorb most of the excess negative moment resulting
from the use of a short launching nose (Fig. 2.37). Once the pier has been reached,
the stays are relieved and launch continues with the nose only. The action of the

pull in the stays

stays and that of the launching nose are staggered over time and are not super-
imposed on each other. Therefore, the launching nose and the stayed system can be
designed according to independent criteria.
The length and the unit weight of the nose are determined first. Then, the
optimum flexural stiffness is defined according to the criteria examined in Section
2.7.1.1. As already discussed, the launching nose cannot control the end-of—launch
moment, so it is convenient to design the superstructure for the EOL moment
progress of launch
and the action of stays so as not to exceed this design value in the cantilever
configuration. ' I . Fig. 2.60. Pull adjustment in the stays (thicker Zine)
 94 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 95

range (i.e. very delicate operations), the jacks can be placed on dynamometers as The same reasons discourage the use of many temporary piers in each span. In
shown in Fig. 256. fact, the relative settlements and the construction tolerances in the piers and the
Once the pier has been reached, the elastic deflection of the nose tip is recovered superstructure reduce the actual possibility of obtaining satisfactory support
with jacks or hinged brackets with hydraulic pistons to avoid excessive stresses in reactions at all the piers. Without expensive jack-bearings, the stress reduction
the stays and in the tower. that is theoretically predictable with many temporary piers in each span is illusory
in reality. On the contrary, the use of a temporary pier in each span is not affected
2.7.3. Temporary piers by many uncertainties and results in a shorter and lighter launching nose and
The first launched bridges were built by halving the launch spans with temporary savings both in reinforcement and in the launch prestressing (partially balanced
piers. Since then, this solution has been adopted more and more rarely, and only in by more expensive parabolic tendons). These savings often cover the cost of the
bridges made up of very long or varying spans. Only in these cases, in fact, can the temporary piers, their foundations, and the additional labour they require for
economies obtained by reducing the launch stresses, and with them the launch assembly and dismantling, during launching, and to relieve the support reactions
prestressing, cover the cost of the temporary piers, their foundations and the on completion of launch. "
additional manpower they require. The first temporary piers were made of concrete, some cast in situ like the
The temporary piers are generally much more slender than the permanent piers permanent piers (Figs 1.4 and 1.8), others obtained by filling steel caissons or
because of the lower design load and the temporary nature. Therefore, under the tubes and were demolished on completion of launch. More recently, modular
same launch loads, the vertical settlement of the temporary piers, in, is larger than precast towers have been used, composed of reinforced concrete pillars or plates
that of the permanent piers, np. In the presence of a temporary pier in each span, (Fig. 2.61) positioned under the deck webs and braced transversely by steel
the relative settlement members. Concrete modular elements are reusable, less deformable than steel
towers, less sensitive to temperature variations, and generally less expensive
77 = or — np , (237) under 10-m height. On the other hand, steel towers are lighter and can be easily
generates a positive moment in the deck section above the pier and a negative
moment in the support sections above the adjacent permanent piers. In a bridge
with piers of similar height and final spans L , these moments are both

M235 24E]
L2 2? (2.38)
The related edge tensile stresses
24E17
0'u,1 E — L2
2,1,1 (2.39)

depend linearly on the deck depth and are unaffected by the cross-sectional moment
of inertia.
Since the superstructure is designed for permanent spans that are twice as long as
the launch spans, the high depth-to-span ratio during launching results in a stiff
bridge response to differential settlement, with high edge tensile stresses that require
parallel increases in launch prestressing. The launch bearings on the temporary
piers are sometimes shimmed after loading to recover. the differential settlement
and to reduce the higher demand of launch prestressing. In the most delicate cases
the launch bearings are directly placed on to jacks for real—time hydraulic adjust—
ment of the support reactions. Jack-bearings (Fig. 2.65) are also very effective in
controlling the torsionaldistortional effects of bearing Vmisplacement and construc-
tion tolerances (these effects are also amplified by the stiff response of the bridge).
However, jack—bearings are more expensive than conventional launch bearings and Fig. 2.6]. Temporary piers composed of reinforced concrete modular plates (courtesy
,require,.ski_1lssi_erews- , . , Leon/71411741)
 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 97
96

transferred from one launch alignment to the other [51] in the case of parallel
bridges (Fig. 2.62).
Steel towers are so flexible that the absorption of the horizontal launch forces
with tension rods anchored to the adjacent permanent piers (Fig. 2.62) or stays
anchored to the ground is often unavoidable. In this last case, the pull in the stays is
often controlled by jacks equipped with dynamometers or inclinometers to avoid
applying backwards forces in kinetic friction conditions, once the static friction of
the launch bearings has been overcome. In the particular case of Fig. 2.63, the static
scheme of the temporary pier was that of a vertical rod hinged at the base and
symmetrically stayed at the cap to avoid flexural stresses in the foundation [26,27].
The strand stays have been inserted in self-supporting steel tubes to avoid the risk
of a broken strand falling on the electricity p0wer lines. ConventiOnal rubber
laminated bearings permitted rotation of the pier base.
The vertical settlement of a steel temporary pier can be quite large. Therefore, the
best position of tall temporary piers is not in mid-span but slightly before it
(Fig. 2.62), so that the increase in the negative moment in the deck section above
the permanent pier generated by the settlement of the temporary pier is reduced by
a shorter span. This position also allows loading the four pier legs at the same time
during nose realignment, thus taking advantage of the deck flexural stiffness in Fig. 2.63. Hinged temporary pier for the Milan-Palizzi Bridges (author)
restraining the pier cap against rotation to improve pier stability. Finally, steel
piers are much thinner in cross—section than concrete ones and the resulting
lower thermal inertia causes sudden differential thermal settlements. For this rea- To conclude, temporary piers can be used successfully only in particular cases:
son, when weight reduction is not a prime objective, the steel tubes of the temporary low height of the piers, direct foundations, accessible areas, spans longer than 60 m
piers are often filled with concrete. or very variable, parallel bridges built subsequently so as to transfer either the
piers or the whole superstructure from one side to the other. In cases other than
these, on long spans it may be convenient to reduce the deck slenderness (for
instance, 12 g H/L 5 15) and to use a long launching nose or a short nose together
with a cable—stayed system, in both cases without additional supports.
Finally, the use of temporary piers may solve specific launch problems. In the
presence of a narrOw axial pier, temporary“ columns placed on both sides of the
permanent pier and supported on its footing may support the deck webs during
launching. In superstructures with varying width or radius of curvature, temporary
columns adjacent to the permanent piers may support wide launch bearings able to
accommodate the transverse shifting of the support points. A temporary pier may
also guarantee adequate curing of the central closure segment of superstructures
launched symmetrically from both abutments.

2.7.4. Deck launching onto arches

The incremental launching of a continuous superstructure over an arch is a brilliant
solution to the requirements of safety of workers, structural efliciency, high struc-
ture quality, short construction duration and cost reduction.
- ..-.~ . . , f.¢%.,,_._._\ _._.\ .-.-:w. \

Fig. 2.62. The temporary piers of the Steine Bridge have been shiftedfrom one lane to The deck of a reinforced concrete arch bridge is generally either a grillwork
"the,other-onen-(courtesy Greischiilr .. w . of simply supported I-beams completed with a concrete slab cast in situ, or a
98 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 99

continuous PC box girder. The higher flexural and torsional efficiency of the box
girder, the much smaller number of bearings, and the absence of expansion joints
(abutments apart), suggest adoption of this last scheme. A continuous box-girder
deck also improves the response of the deck—arch system to live loads and the seismic
demand by providing additional stiffness at the deck level. This makes the system
behaviour similar to that of a deck-stiffened arch and reduces live load and seismic
moments at the arch springings, with enhanced slenderness of the arch ribs [12].
A continuous PC box girder can be built over an arch with three construction
methods:
0 incremental launching from one abutment towards the opposite one
o span-by-span casting in situ with a movable shuttering system (MSS)
o segmental precasting with launching truss site assembly.

Incremental deck launching over the arch takes place with conventional construc—
tion procedures for PC launched bridges. The deck is built in a fixed casting yard Fig. 2.64. Arch stabilisation by means of water filling
placed behind an abutment, in segments as long as half a span to increase the level
of repetition in the casting procedures and to better amortize the investments for
the forming system.
Deck construction by incremental launching offers a great deal of advantages. along the arch instead of being concentrated at the spandrel columns). Then, the
One of the main design advantages is the possibility of increasing the spacing of deck is launched until it reaches the second spandrel column while water is emptied
spandrel columns and of the piers of the access bays without the restraints deriving out the arch sector between these two columns, so that the dead load of the deck
from the increasing costs of an MSS or a launching truss for precast segments as the replaces that of the water. This process (incremental deck launching on to the
span increases. This permits optimizing the deck—arch interaction and eases the spandrel columns and incremental emptying of the arch sectors between the
choice between a curved or polygonal design of the arch rib. spandrel columns) is repeated launchwards until the deck launch is complete.
Load asymmetry on the arch during deck construction might be avoided by This construction method reduces the flexural stresses in the arch and in the
launching two symmetrical superstructures from the opposite abutments towards superstructure, does not require suspended counterweights, does not generate
mid-span. However, the presence of symmetrical loads on only a part of the dynamic effects and involves minimum costs.
spandrel columns causes bending stresses in the arch and the cost of two casting Incremental- deck-launching on to the arch generates smaller vertical ioad
yards rarely makes this scheme competitive. Therefore, launching usually takes imbalances than alternative construction methods, as no construction equipment
place from one abutment towards the other (like the span—by-span construction adds to the self-weight of the superstructure. On the contrary, the horizontal
with an MSS or a launching truss) and the arch has enough flexural stiffness to forces are larger due to launch friction; Launch forces and load asymmetry
resist the asymmetrical loads. may be balanced with temporary stays that brace the tallest spandrel columns
Load asymmetry on the arch may be balanced with temporary stays or counter- and the arch ribs, possibly integrated with counterweights. The stay systems are
weights, or by filling the arch ribs with water. Long—span arches always have box usually anchored to the piers without slippage at the deviation points, so that the
ribs with diaphragms at the base of spandrel columns: if these diaphragms fully forces on the opposite sides of the piers may be different. This permits the
close the cell (or if they can be temporarily closed), filling the rib sectors with water application of horizontal forces or the balance of deviation forces. Temporary
may create a distributed balancing load. The construction sequence studied for the staying does not affect construction cost much since the same strands and
pursuit of the 300-m arch of the Hoover Dam Bridge Bypass Project (US anchorages used during Cantilever construction of the arch are reused during
Department of Transportation, FHWA) is described in Fig. 2.64. When the launching of the superstructure. Since the construction equipment for a launched
nose—deck joint reaches the first spandrel column, the arch ribs are filled with bridge is much less expensive than an MSS or a launching truss, the cost of
water by pumping in all arch sectors so that the loading scheme is that of a temporary stays does not affect the overall competitiveness of the incremental
uniformly distributed load that progressively increases. When the arch is full with launching construction method. In the most delicate cases, the use .of launch
water, the stresses generated by the water are similar to those caused by the super- reactors (able to avoid application of any horizontal forces to the spandrel
. structure .(waterfilling was 10% heavier than the deck but uniformly distributed 7 columnS) mayfurther reduce temporary stays.
100 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 101

The launched deck spans both the arch and the access bays, so its length may be and its introduction does not affect the segmental construction time-schedule,
considerable. This suggests adoption of a highly industrialized casting yard, the occurs after long segment curing, and is not subjected to the requirements of
organization of which is the same as that for PC bridges launched onto ground- span—by-span construction. Therefore, the polygonal tendons can be as long
based piers. All the casting schemes discussed in Chapter 3 (monolithic casting of as several spans to reduce the cost of anchorages and tensioning and grouting
long or short segments, two-phase casting in one or two adjacent forms) are operations.
applicable. Launch bearings are conventional both in the casting yard and on the During launching, the time-dependent strains of concrete under axial prestressing
piers of the access spans, and adjustable bearings (Fig. 2.65) may be used at the cause uniform shortening of the deck, which is compensated by slightly lengthening
spandrel columns to better control the flexural stresses caused by the elastic arch the segments. The creep rotations that arise during the launch stops for segment
deflections. construction do not superimpose on each other as the deck support positions are
Construction joints are weak points of any PC structure. In launched bridges staggered by half a span. Finally, the eccentric tendons that correct axial launch
with segments as long as half the span, construction joints are located at the span prestressing are introduced after several weeks of curing. The upwards cambers in
quarters and are therefore affected by the lowest long-term flexural stresses. the concrete deck due to the time-dependent behaviour of concrete are therefore
Construction joints are treated to enhance adhesion before match casting the minimized.
new segment, and surface bonding is further improved by launch prestressing. Launching occurs on specific temporary bearings. The permanent bearings are
Construction joints are crossed by continuous through reinforcement that further inserted under the superstructure on completion of launch and are therefore
enhances contrOl of edge stresses, and they are definitely compressed by launch unaffected by deck construction (grout spillage, accidental movements during
prestressing. In bridges launched onto ground-based piers, the axial launch pre— concrete pouring and vibration, deck shortening at tendon tensioning, etc.).
stressing is sometimes partially relieved on completion of launch. In the case of an Bearing presetting is much easier too, because it is unaffected by prestressing
arch deck, the axial launch prestressing controls the edge stresses produced by arch operations.
flexibility and provides a constant long-term protection for structure durability: The standard construction equipment is the simplest (a formwork, a thrust
therefore, launch tendons are generally permanent. device and a launching nose) and relatively inexpensive. In spite of the low
The polygonal (external) or parabolic (internal) tendons are introduced on com- cost, construction equipment can be easily amortized over several projects
pletion of launch to adjust the axial launch prestressing according to service owing to its adaptability, and involves minimum assembly and dismantling
requirements. This integrative prestressing is applied to a self-supporting structure costs. In 35 years, the incremental launching construction method has been
adapted to a great deal of different situations and this has generated several
alternative solutions for the casting yard. Correct yard industrialization results
in optimized casting sequences, reduced construction duration through parallel
rather than serial " processes, impr'ovedriabour’"rotation and easier iorm main-
tenance. The formwork, batching plant (if any), and steel cage assembly yard
are adjacent, and storage and handling of materials with a tower crane are
simplified as much as possible.
Compared with deck casting in situ with an MSS, the incremental launching
construction method has extremely significant advantages in terms of safety of
workers, interference with the adjacent areas, structure quality, availability of
alternative casting processes and construction costs. In addition, the use of an
MSS has some specific disadvantages: an MSS for PC decks with 40—50-m spans
is an extremely heavy piece of equipment with a weight that can easily exceed
1000 tonnes. The use Of this type of equipment attracts severe penalties: very
high investment; little chance of direct reuse in different projects, resulting in
prohibitive amortization conditions; lengthy assembly and dismantling operations;
and high transport, assembly and dismantling costs.
An MSS is rather flexible. Form precambering can control geometry of the final
structure; however, the unavoidable form deflections under the casting loads
Fig. "2.65. , Adjustable jack—bearing .. . . _,,,.9¥¢§1I9.-.-917.3913 in the (concrete and irregularities at the construction joints.
102 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES _ 103

Additional defects may result from the deflections and vibrations of the arch—deck and the environment and a poor aesthetic quality for the structure. Finally, a basic
system. advantage of segmental precasting is the short construction duration. When the
During span-by-span construction with an MSS, the parabolic tendons are ten- deck is over an arch, however, this advantage is only apparent as segment assembly
sioned as early as possible since concrete curing immobilizes equipment and labour, is delayed until completion of most of the piers and spandrel columns not to
and the joint section between two spans (which contains the tendon anchorages) is interrupt the work of the launching truss. Deck launching can start much earlier
covered by match-casting the next span. The introduction of full parabolic prestres- if pier erection proceeds in the same direction, and a longer construction duration
sing after a short curing time increases prestress losses, and the creep rotations may thus be avoided.
caused by prestressing forces designed for full service loads (and therefore excessive The advantages of the incremental launching construction of the PC deck of
for self-weight only) generate irreversible cambers that are aesthetically harmful an arch further increase when the deck is a composite box girder. The vertical
and disturb drivers. Finally, the permanent bearings are affected by deck construc— load on the arch is smaller, load imbalances and launch forces during deck
tion, prestressing shortening included. construction are much smaller, and the overall structural efficiency increases.
As with the incremental launching, the deck is cast on to the arch proceeding Incremental launching can actually open new perspectives in these types of
from one abutment towards the opposite one. Therefore, the self-weight of the MSS structures.
(not much lighter than a whole PC span) increases the load imbalance on the arch
and generates additional flexural stresses both in the arch and in the completed
portion of the deck. In addition, all materials (steel cage, prestressing tendons, 2.8. Launch stresses in the piers
concrete, etc.) must be transferred into the MSS via the deck or with cable cranes, The application of the launching techniques to PC bridges has been permitted by
which increases the risks for workers and the environment and lengthens construc— the commercial availability of Teflon (PTFE), a self-lubricating material able to
tion duration. Launching operations of the MSS are risky too, and cause debris reduce sliding friction at low cost. Previously, the self-weight of PC structures
dispersion. prevented the use of the launch equipment commonly used for the much lighter
Compared with the incremental launching construction method, the only advan— steel girders. The longest rollers would have been expensive and hard to amortize;
tage of deck casting in situ with an MSS is the smaller number of construction joints greased wooden sledges would have transferred excessive horizontal forces to the
(one per span). The number of construction joints is, in turn, one of the main weak piers and would have required extremely powerful thrust systems.
points of the use of precast segments for the PC deck. Segmental precasting has the The development of sliding bearings based on the steel—Teflon contact has
additional limitations described in what follows. permitted the launch of longer and longer PC bridges. The Veitshochheim Bridge
A launching truss for heavy precast segments and 40—50-m spans can weigh for the German High Speed Railway [17—21], 1260 m long and weighing 433 MN,
more than 500 tonnes. Like an MSS, a launching truss involves high investments has been launched, with a total thrust force of 18.2 MN. The incremental launching
and expensive transport, assembly and dismantling operations. Also in this case of the 498-m long Sart Canal Bridge [52], currently under construction nearby La
deck construction proceeds from one abutment towards the opposite one, and the Louviere, Belgium, will increase the launched weight record to 650 MN. This
truss weight increases the transitory flexural stresses in the structure. The main bridge required a specific design of four synchronized thrust units with 5 MN thrust
weak points, however, are the number and the characteristics of the construction capacity each and 2 m effective stroke.
joints. Construction joints are weak points in any concrete structure, especially The force system that the Superstructure transfers to a pier during launching has
when they are not crossed by through reinforcement. These aspects are even more three vector components:
significant when the deck is over an arch, since arch deflections modify the stress
distribution in the structure and shift the counterflexure points, Which may o A vertical component, RV, which is the support reaction of the continuous beam
decompress the deck edges. The incremental launching construction method offers at the pier: it can be eccentric from the pier axis both in the longitudinal
a lot of advantages: one-tenth the number of construction joints, continuous direction and in the transverse plane.
reinforcement through the joints, bush-hammered joints definitely match-cast a A longitudinal component, proportional to RV, generated by the sliding frictiOn
and unaffected by form oiling, axial launch prestressing that prevents or controls in the two launching bearings on the pier and by the local grade of the launch
edge decompression, and a casting yard at the same quality level as a precasting surface.
plant. - . 0 A transverse horizontal component generated by the guide system that prevents
Segmental precasting of a continuous deck requires that some segments are cast transverse drifting of the superstructure: its value does not depend on RV but on
in situ to prevent accumulation of the geometry errors deriving from short-line the driving force and, when launching along a curve, on the local radial
“match-vcasting—"Over4an. arch, segment casting. in situ. involves risks for workers _ component of .the thrust reaction.
104 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 105

The vertical component is much larger than the longitudinal component, which
is in turn generally much larger than the‘ transverse component. Therefore, the
force system transferred to the pier mainly lies in the vertical surface that
contains the longitudinal axis of the bridge. The transverse component and the
effects of the transverse eccentricity of the support reaction can be easily super-
imposed.
Starting from the breakaway friction coefficient C? between the neo-flon pads
and the stainless steel sheets that cover the launching bearings, the static friction
angle qb is such that

tgqb = C? (2.40)
Letting or be the local longitudinal gradient of the launch surface at the pier, Figs
2.66 and 2.67 define all the possible vectorial combinations for or < qb and or > qb
respectively.
Defining the vertical component of the load transferred to the pier by the vector
W, its orthogonal component with respect to the launch surface is
O_N = Woosa (2.41)

and the tangential component is

I a<<I>
if = W sin or (2.42) R2

On‘ applying the thrust force the resultant begins to rotate, and launch starts only V R1
when its inclination exceeds the equilibrium. angle R20R1 = 2gb. Therefore, the limit
equilibrium resultants are Fig. 2.66. Pier cap forces for a local gradient of the launch surface smaller than the
__ __....._ _.cosa
OR 2 OR = 0V (2.43) static friction (0: < ()5)
1 2 cosqb

When launching uphill, to rotate the resultant from 577 to 0R2 it is necessary to
apply a force F equal to cases (i should be directly set to 0. In both cases, the horizontal force transferred to
F = T—T2 = 0T2 _ W = ——D_I7(cos or-tgqb + sin oz) (2-44) the pier is

and the horizontal force transferred to the; pier is COSOJ . .

HZOHIZW
COS sm(¢ — a) (2.47)
C OS CU (b
H = 0—H2 = :67 cosqb sin(¢ + or) (245)
To conclude, when launching uphill the slope adds to the friction, the thrust force
In the case of downhill launching on the {other hand, if a thrustiorce has to be and the horizontal forces applied to the piers are higher, but the movements of
applied (i.e. for or < ch as in Fig. 2.66), the resultant rotates from 70V to 0R1 under the superstructure are better controlled. When launching downhill the forces are
a force F: lower but controlling them is more difficult and delicate, as the reduction in the
coefficient of friction after breakaway may change the direction of the force to be
F = ”T—Tl = 0T1 ---0_T = fiflcosa tgqb — sina) (2.46)
applied.
If the local gradient is higher than friction (i.e. for a > qb as in Fig. 2.67), the In practice, during the stops for construction of a new segment during a downhill
superstructure slides without any intervention, and to brake it the force F must launch the deck is anchored to the launch abutment, and thermal contraction (and,
be directed uphill. In this case, a minimum value of the friction angle must be used to a lesser extent, creep and shrinkage of concrete) cause the superstructure to move
rfor-rrcalculations-instead- of rthewstaticfriction angle (2.40), and. in the most delicate. ,.,slightly.. uphill at the pier bearings. Since the slope and the friction act in opposite
 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 107
106

Once the system of forces applied to the pier cap has been determined, the
eccentricity of its application point with respect to the pier axis can be evaluated.
In theory, the neo-flon pads should assure a uniform distribution of the support
reaction along the contact surface, so that RV can be assumed as centred on the
launching bearings. In this case, the launching bearings can be shifted backwards
from the pier axis to create a load eccentricity, e, able to reduce the moment
generated in the pier—base section by the horizontal component H of the force
applied to the cap. Setting hp as the pier height, the backwards eccentricity suflicient
to cancel the pier base moment is
H ,
e _ RV hp (2.48)

In practice, the inevitable tolerances and the imperfect parallelism of the launch
bearings and the deck surface make the vertical force transferred to the pier
eccentric. This eccentricity is often modest; however, its negligibility must be veri-
fied case by case on the basis of the length of the contact surface of the launching
bearings compared with the longitudinal pier dimension. Finally, the pier itself has
a tolerance in verticality (especially if built with slip forms) and the creep of con-
crete increases over time the launchwards deflection generated by the horizontal
force applied to the pier cap.
The vertical load transferred to the pier is also eccentric in the transverse
direction as a result of level tolerances in the launching bearings and geometric
imperfections in the superstructure, which together generate load diflerences
between the two bearings. Transverse eccentricity of the support reaction is rarely
critical in the permanent piers, which are designed for full service loads; in the
than the temporary piers, when necessary, it may be controlled with jack—bearings able to
Fig. 2.67. Pier cap forces for a local gradient of the launch surface greater
static friction (or > qb) adjust the torsion in the superstructure and the resulting transverse bending in the
pier (Fig. 2.65)."- V, 7
For long bridges or steep launch gradients it may be necessary to place friction
thrusters at some piers to better distribute the thrust action (Fig. 2.68). Front
if any. staying of the permanent piers equipped with thrusters is generally unavoidable.
directions during launching, the launchwards pier deflections are very small,
cause
Therefore, deck contraction may easily recover these residual deflections and The experience gained with friction thrusters placed on stayed piers led to the
the friction forces and the slope com- development of launch reactors. These hydraulic devices, schematically similar to
a situation of ‘negative friction’ in which
necessar y to pull. the deck small friction launchers, apply to the superstructure a thrust force equal to their
ponent act in the same direction. Sometimes it is also
backwards (uphill) to recover launch errors. For all of these reasons, the pier should own internal friction combined with the local grade, so that no resulting
always be designed for the slope added to the friction, independent of the planned horizontal force is transferred to the pier. These devices are directly controlled
launch direction. ' ' " ' ' by pier incl’inometers to make their response instantaneous and free from the
may acciden tally exceed the thrust pulses of the main launchers. Synchronization of the launch reactors
The horizontal force transferred to the pier
infreque nt case is with the main thrust systems is delicate, and several launch interruptions are
theoretical value based on friction and slope. A typical and not
super-
the insertion of a neo—flon pad upside down, with Teflon in contact with the necessary because the reactors do not reach their end of stroke simultaneously.
piers,
structure and neoprene in contact with the launch bearing. In the. case of tall For all of these reasons, these devices are quite expensive and should be reserved
(inclinom eters) and emergen cy for the most critical cases.
it is prudent to use real-time monitoring systems
stops, when the pier is not directly supported with horizon tal braces or inclined As regards stress verification, the pier is a cantilever partially fixed at the base and
’ " Stays. ---v---is----therefore very sensitive- --to- the— couples and the horizontal forces applied to the
 ' BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 109
108

_
. / / / / /
, /
/
I / ' /
very high it is suflicient to calculate the stresses neglecting the pier deflections and to
/
W/ / / / // / / ///
/'
//’/////:////’/
—A/ check the critical cross-sections for combined compressive and bending stresses with
the amplification factors prescribed by codes, which are generally based on Euler’s
critical load:

PE = —2 (2.50)

Euler’s critical load is calculated with a reduced elastic modulus to account for the
creep deflection of the pier generated by the constant application of a horizontal
force to its top. Finally, if the pier is slender and deformable, the stresses must be
determined by accounting for the launchwards deflection, i.e. with an analysis
/ / / -/ / including non-linear P-delta effects.
/ // // I / / // // I / / / / / Similar verifications must be carried Out in the transverse direction as well. In this
/ /
/
/ / / // /
/ / / / /
case, however, the low guide forces applied to the superstructure rarely require non-
linear analyses and the torsional stiffness of the superstructure often results in
A
V
negligible rotations of the pier cap. Therefore, the effective pier length mainly
depends on the transverse mobility of the superstructure and is

hi): L0 5 hP (2.51)
i.e. from one—half to one—quarter of the longitudinal one in the case of a perfectly
fixed footing. In practice, the lateral guides of the temporary piers, instead of
guiding the superstructure, are mainly aimed at stabilizing the pier by restraining
its cap in the transverse direction. In the case of tall temporary piers, this principle
may be used in the longitudinal direction as well. By placing four launching bear-
ings on each pier instead of two (Fig. 2.62) it is possible to take advantage of the
flexural stiffness of the superstructure to fix the pier cap against rotations.
@

HQ The temporary piers are often stabilized with longitudinal stays anchored to the
WE -
(IIIIllr/dfllllWMW/iW
footings of the adjacent permanent piers. Monolithic temporary piers generally
have suflicient flexural strength not to require staying. In the case of two pillars
AV . or RC plates placed under the deck webs, the stays are contained in the longitudinal
bridge plane. In fact, the large transverse spacing of the pillars, the low transverse
Fig. 2.68. Pier friction thrasters and rigid stays are often necessary for steep—grade forces and the shorter effective length (2.51) generally avoid any need for transverse
uphill launching stays. Consequently, the steel cross-braces connecting the pillars (Fig. 2.63) must
resist the transverse forces applied to the pier cap.
Elastic stability of the temporary pier is not much affected by the type of pillar
top. Consequently, pier instability must be assessed during launch as well. The (precast RC plates, steel tubes or composite tubes) sinCe in any case the pillars and
slenderness ratio of the pier, A, is their bracing system form a spatial truss to be assessed by taking the influence of the
shear stresses into account. If the X-braces of the truss are inclined angle a from the
A 2’ E (2.49) horizontal plane, the critical load is [53]
l"

PE
where r is the cross-sectional radius of gyration and L0 is the eflective "length Pcr:
P
(2.52)
(double the pier height in the case of a perfectly fixed footing, and progresswely 1+ .13 2
._ . _ ZEbAb_31na_.co_s_._.. a ._
” ”longer as" thedegree'of base constraint decreases).-When the'slenderness ratio is not
 ' BRIDGE LAUNCH’ING INCREMENTAL LAUNCHING OF BRIDGES lll
llO

where E, is the elastic modulus of the braces and Ab their cross-section. This bridges. In its original formulation, however, the progress of launch increases the
equation is based on the assumption of instability occurring in the elastic field, number of redundants and penalizes the immediacy and simplicity of the method.
so this condition must be assessed by comparing the critical stress (Ap is the The reduced transfer matrix (RTM) method [57—59] overcomes these limitations
cross—section of each pillar) and obtains an exact, simple and economical algorithm:
0 It is exact, because it derives directly from the integration of the elastic beam
Per (2.53)
UCIZZA
differential equations without any interpolation or iterative procedure.
P
0 It is simple, because the procedure is based on the repetitive manipulation of
with the yield stress. small square matrices that takes full advantage of the repetitiveness of the
Local instability must be assessed as well. If the pillars are thin RC plates segments that compose launched bridges.
(Fig. 2.61) of width [9, thickness t, elastic modulus EC, and Poisson’s ratio 1), the o It is economical, because it is carried out by multiplying matrices of constant
plate panels might buckle. By assuming that cross-bracing is distributed so as to dimensions, and repeating hundreds of cycles of the algorithm is rapid, even
hinge them along the vertical edges, the critical load is [54] with small computers.
7T2ECI3 (2 54)
Consider the case of a continuous beam contained in the x — y plane (a local frame
Per:
3(1—-v2)b2 ' ' of reference with the x-axis on the gravity axis of the beam and orthogonal y—axis),
This equation is valid when the height of the temporary pier is not less than three any constraints, constant or variable cross—section with a principal axis of inertia
times the plate width and instability occurs in the elastic field. Finally, temporary contained in the plane.
For each cross-section, the state of stress is defined by three forces N, Vy, and V2,
piers composed of steel and/or composite tubes will be assessed according to simIlar
and three moments MX, My, and M2, and the state of deformation by the three
criteria. Detailed information about the strength of composite tubes can be found
in [55]. components of the displacement n and rotation go.
Assuming the only presence of external forces orthogonal to the x — )2 plane and
2.9. RTM method for analysis of the continuous beam external couples about in-plane axes, the six stress components reduce to three in
The design of a launched bridge requires a considerable amount of calculation each cross-section: orthogonal shear force V : V2, bending moment M = My and
because the highest launching stresses must be found within an infinite number torsional moment T = M2. Their internal work results from the orthogonal out-of-
plane displacement n :. 172, the flexural rotation go 2 gay and the torsional rotation
of support configurations.
In practice, this research is done by discretizing the deck movement. Bending 19 = 90x of the cross—section.
moment and shear force diagrams are calculated in a sequence of closely spaced Since highway and railway bridges are, in most cases, tangent or moderately
support configurations to simulate the course of launching with adequate reliability. curvilinear, and the adoption of box girders (whose'high torsional stiffness limits
Then, the highest values are obtained by finding the envelope. tangential stresses, even when the torsional moment is relatively high) is the norm,
Bridges that are hundreds of metres long require the envelopment of hundreds of straight beam theory is generally sufficient. Then, assuming external forces are
moment and shear force diagrams, each calculated for the worst combination of applied along the x-axis, unknown deformations are only 17 and go, and unknown
thermal gradients, differential bearing settlements and permanent deformations due load effects are only M and V. -
to creep. Such an onerous task requires the use of special computer software that In PC bridges, the action of prestressing should be considered. This may be
simulates the development of launch, and iteratively recalls a continuous beam evaluated by applying a system of equivalent forces composed of distributed
solver subroutine from the subsequent support configurations. orthogonal loads along the curved zones of the tendons, and concentrated loads
It is well known that there are several methods for calculating a hyperstatic at their anchorages [60].
system. In most cases (apart from'artifices of symmetry or shrewd choice of redun— The anchorage loads produce shear forces and axial forces. Axial components are
dants), these methods are substantially equivalent since they are based on the same balanced by the same components at the opposite tendon anchorages and by the
differential equations and require the same number of integrations. axial resultants of distributed friction. These systems of forces (a self-balanced
The practical interest of a method therefore depends on the use of the same type system for each tendon) do not introduce indeterminate effects, as launching
of integrals, or in the possibility of using an agile and repetitive algorithm that bearings (neglecting friction) do not prevent axial deformations in the deck.
limits the risk of mistakes, and can easily be implemented in a computer program. Therefore, the action of prestressing may be computed by superimposing the
The transfer matrix method [56] meets both of these requisites, and its advantages uniform compression, directly deducible from the axial force diagram, on the
' ' ”are "increased-“by"the"repetitiveness? of~ thesegments that-usually compose launched ._ flexural and shear effects of. the couples and orthogonal forces of the equivalent
-
-
 112 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 113

system. The couples are produced by the eccentricity of the tendon anchorages from dgo __ M dn _ 2 63
the x—y plane. The orthogonal forces are the components of concentrated and dx _ EI dx — g0 ( ° )
distributed prestressing loads. These equations are solved through four subsequent integrations. Integrals depend
So, axial force apart, the elastic system is described by a state-array {S} in each on four constants (which, if the state-array of the origin section 0 is defined, are
cross-section of the beam: 770, (00, M0, and V0), on the mechaniCal properties of the beam segment, and on the
77 loads applied to it:

{S} = W
(2.55) ’ L2 L3
M 771 ’70 + 900 2EI 0 6E] V0 +f1(p, L, E, I)
V L . L2
4
901
z
900 __M
E1 0 2E] V0 +f2(P,L,E,1)
__
(2.64 )
and the linear relation between the state-arrays of two subsequent sections by
a matrix [T] which, transferring the definition of the state—array, is called transfer M1: M0 +LV0 —f3(l?a L)
matrix.
. ' Between the Origin Section 0 and Section 1 of the beam (the generic section of the In matrix form, these relationships may be expressed as
7 beam Will be defined by a number, and the support section by a capital letter), we
- -1
have ' L2 L3
l L .
{S}1= [T15 >< {S15 (2.56)

Hfififis

HVE‘Ss
L L2
Between the two ends of the subsequent beam segment, we have again —E _E f2(P,L,E,1) X,

COCO
{Sh = *

ll
{S}2 = [T15 >< {Sh (2.57)

O
0 1 —f4(p,L)

O
that is J 1
0 0 1

O
{Sb = [T15 X [T15 x {Sb (2.58) = [T15 >< {S}o (2.65)
and in a beam composed of n subsequent segments, the relation between the
state-arrays of the two end sections is obtained by progressively multiplying in which the fifth row has been introduced to keep the transfer matrix square. All
the transfer matrices of the single segments: the terms of the first four columns depend only on the mechanical and geometrical
properties of the'beam segment,and the terms of the fifth column depend on
{S}... = [T151 >< [213:5 >< ...... X [T15 X [T15 x {515 (2.59) external loads, as well as the mechanical and geometrical properties of the beam
segment. , .
and can be summarized in one transfer matrix:
In PC bridges, the construction and service stages are characterized by few load
{SL5 = . [T13 >< {S}o (2.60) types, and a complete library of cases is easy to create:
obtained by multiplying o For a distributed load p applied uniformly to the whole 0 —> 1 element, by
[T13 = [T]3__1 X [T13 >< ...... x [T]? x [T15 (2.61) integrating
4 3 2
The terms of the transfer matrix [T](1) between the end sections of the oriented 0 —> 1 TL pL pL
T15 24E] 25 6E] 35 2 45 P (2 66)
beam segment, Of length L and conStant fie'Xural stiffness E], are obtained from the
equilibrium difierential equations: o For a distributed load whose density varies linearly from p0 in section 0 to p1 in
dV _ dM section 1:
2V (2.62)
dx _ 1? dx
2&5: T :im L3 T35 = _ 2190 6+191 L2
and from the compatibility ones, ([7 positive downwards and neglecting shear strain 15 120E] 25 24E] (2.67)
—— a hypothesis generally acceptable for PC bridges but often approximate for T45 : _P0:P1L
{prestressed...co.mp.os.i.te..bridges)whichware--. .
-
114 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 115

and the generalization to any type of distributed load is immediate by observing A17

OOOHO

oer—too

CHOOO
OOOOH
that the terms TiS are the subsequent integrals of the —p(x) function divided by —E1 A90
right _
in the first two rows. [Theft _
AM (2.72)
The case of distributed loads applied only to portions of the beam segment or of AV
variable E] can always be solved by integration. In practice, it is generally expedient
1
to divide the beam segment into sub-segments fully affected by the distributed load
or of approximately constant EI, and obtain the transfer matrix by multiplying (Lmnz (Lmn3
(L-X)

)_n
them.

CD
2E1 6E1
For a prestressing force N whose eccentricity varies linearly from 60 in section
1 c—n (w
0 to 61 in section 1 (parabolic prestressing or straight prestressing but varying
[Tliight = (2-73)

HOOO
COO
E1 2E1
position of the centroid), the beam segment is loaded by a moment that varies 0 l (L — x)
from M0 2 N60 to M1 2 Ne}, and it is
0 0 1
T15 : _2MgE-;M1L2 T25 : _M02;IM1
L T35 Z 0 T45 = 0 (2.68) 0 0 0 0

If the concentrated action comes from an elastic support (piers with different height
For a linear thermal gradient AT between a warmer upper surface and a colder
or temporary piers alternated with the stiffer final ones) and therefore is
lower surface, the constant value of the moment (where H is the deck depth and or is
the coefficient of thermal expansion)
AT (2.69) the left —> right transfer matrix becomes
M0 = M1 : EIE‘H—
1 O 0 0 0
must be introduced in Equation (2.68). Thermal gradients may have significant
0 1 0
effects during launching since the deck is not protected by the permanent paving.

b—‘OOO
cor—o
For instance, for an unprotected superstructure DIN 1072 [61] requires considera- [THEE 0 0 0 (2.75)
tion of linear thermal differences of + 150C and —50C for steel bridges, + 80C and k 0 1
—7°C for composite bridges, and + 100C and —-3.50C for PC bridges. BS 5400 [62] 0 0 0
requires the use of more refined non-linear gradients.
Usually, the continuous beam is solved by assuming support reactions as redun-
In the case of concentrated actions or distortions applied to a section at. distance
dants (Fig. 2.69). if
x from the local origin 0, an infinitesimal sub—segment will be introduced. The O -—>
Starting from the free back end 0 of the deck, once it is defined by subsequent
1 segment is divided into three sub-segments 0 —> left, left —> right and right —> 1,
multiplication the transfer matrix [T]’0A between the ends of the first 0 —> A beam
with the central sub-segment limited by the cross—sections immediately on the left
segment is
and on the right of the section interested by the concentrated actions. The transfer
matrix [TE is always obtained by multiplying {mi=[Trso 07o
-wn=[ftxw%=[flamxurE%4Tfitxwm arm In the same way, once the transfer matrix [TLE between the ends of the first span is
defined, it is
with
{sa=trfixini ova
2E1 6E1
x x
infi= QJD
beer—x

_ ff _2EI
i—QOO
booo

l x
0 l
.,.0.. .0. Fzg-269-Sra-tic- system waft-he continuous beam
116 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 117

in which [Tl]: includes the unknown support reaction RA between the external load m = 7.7] = constant 27K 2 7.7K = constant (2.83)
as well. Consequently
By extracting VJ from the first equation of (2.82) and substituting it in the other
{S}B = [T12 X [T]? x {510 (2.78) three, a direct relation between rotations and bending moments in the two support
Proceeding in the same way, one oversteps the last support N and reaches the front sections may be obtained:
free end of the launching nose, section 1. The relationship between the state—arrays
7.7K = constant
of the end sections is
VK = ——
T14 GDJ ——M
T14 J + —(
T14 77K — 77.7 — T15 ) T45
T12 T13 l 0 0
+
{S}1= [T1111 >< [T1§_1>< ...... X [T11 X [T13 x {510 (2.79)
in which each span-by—span transfer matrix except [Tk? contains a redundant
unknown. It can be summarized in T T T . .
50K = (1— T24—_12)€01 + (T23 — T24 —13)MJ +““—24 (77K — ’71 — T15) + T25
T14 T14 T14
{S}1= [T15 >< {Sio (2.80) T T T o .
Without these redundants, Equation (2.80) would be four linear equations with MK = ‘T34i90J +(1“ T34£)MJ +351“ (77K — ’7J — T15) + T35 (2-84)
T14 T14 T14
eight unknowns, and four additional equations would need to be obtained by
Therefore, once defined the reduced state—array {SR} of the support section
imposing the boundary conditions in the two end section state-arrays.
The presence of the unknown support reactions requires defining an equal num- 90
ber of additional equations. These may be obtained by calculating the support {SR} 2 M (2.85)
section state-arrays and by imposing zero deflection (or a definite displacement in 1
case of settlement). For this reason, the number of equations progressively increases
during launch owing to the new supports afiected by the beam, and the solution and the relative reduced-transfer matrix (RTM) from the third and fourth equation
requires an iterative algorithm, generally slower than the direct solution. Most of (2.84)
important, however, the method as a whole loses brevity and repetitivity. T24T12 T24T13 —-—-—(
T24 ° ' _
Since additional unknowns are concentrated in the support section state-arrays l —mT——
T14 23 T14 T14 71K ’71 —T)
15 + T25

-
(where they constitute the only possible discontinuity), the transfer matrix method T T T T T o 0
K— —fl I‘mfii 34 ("K—"J“T15)+T35 (2.86)
can be improved to operate only on the continuous terms of these arrays (the
[TRh — T14 T14 T14
rotation (p and the bending moment M) to calculate the remaining ones in a second
time.
Consider two subsequent support sections J and K of the continuous beam that 0 0 1
limit the generic span J —> K. We have in a continuous beam with n rigid supports the relation between the reduced
state—arrays of the end supports is
15hr: [T11< >< {5}] (2.81)
{SR}N= [TR]§_1X ...... >< [TR],EX{SR}A: [TR]? >< {SR}A (2.87)
that is
so we have two equations with four unknowns.
77x I m + T1291 + T13MJ + T14VJ + T15 Since the bending moments MA and MN are immediately calculable, the two
99K = 991 + T23MJ + T24VJ + T25
(2.82) rotations goA and goN are the only unknown terms, and Equation (2.87) can be
MK :MJ+T34VJ+T35 immediately solved in closed form:
VK 2 VJ + T45 1
90A = T—(MN — TR22MA — TR23) _ (2-88)
where coefficients are the terms of the [TEi transfer matrix. J and K have a vertical R21
deflection of zero. In practice, it is better to assume a constant displacement to take Knowing 90A and therefore, the reduced state-array {SR}A, rotation and bending
forced movements (raising needed to replace the neo—flon pads, to remove bearings moment in each support section may be defined by Equation (2.87), and the relative
and to analyse the effects of vertical misplacement of launching bearings) into shear force VI may be defined by using the second equation of (2.84). Once defined
account...Therefo.re,.we.set. , ., . . ....the. support section state-arrays (731 is assumed constant), shear force, bending
-
118 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 119

—0.15
moment, deflection and rotation may be obtained in every cross-section of the deck
by means of the transfer matrix between the section at issue and the nearest support
“0.10
one.

moment (M/pLA2)
In bridges built by incremental launching, the RTM method allows one to solve
-0.05
the continuous beam in an exact, simple and economic way. The procedure is based
on the repetitive manipulation of square matrices composed of only ten varying
0.00
terms, six Of which are always expressed by the same equations, and four that can
be collected in libraries integrated in closed form, even for the strangest load types.

bending
0.05
The procedure takes full advantage of the repetitive nature of the segments that
make up launched bridges. Since it is carried out by multiplying small matrices of
0.10
constant dimensions, the repetition of hundreds of cycles of the algorithm is rapid,
even with small computers.
0.15
The calculation of stress and strains in each cross—section is very fast as well. It is
sufficient to memorize the terms of the transfer matrix between the nearest support
section and the section at issue to reduce the problem to a simple matrix Fig. 2.70. Dimensionless bending moment envelope in the 'threefronr spans oft/1e deck
multiplication. Practical applications of the RTM can be found in [57], [58] and [59].

2.I0. Shear and bending moment envelopes for both the positive and the negative moment. Since the additional weight of the
The use of the envelopes of bending moment diagrams (and from these of the support diaphragms may increase launch prestressing, it is sometimes expedient to
envelopes of the edge stresses), instead of hundreds of pages of printout, is definitely cast them in situ on completion of launch.
advantageous. The envelopes of shear forces are basic design tools for the launch of The launch prestressing force is almost always balanced on the centroidal axis of
prestressed composite bridges and steel girders, but shear stresses are less critical in the mid-span sections. Therefore, it becomes eccentric in the stiffened support
PC bridges and are often assessed directly for the worst-case condition. For both sections due to a definitely lower location of the cross-sectional centroid. This
bending and shear, the envelopes resume the evolution of launch stresses and generates a positive moment peak that may be significant even with low eccentri—
highlight the deck regions that require strengthening. cities, as in Fig. 2.70. These moment peaks, however, do not immediately increase
The bending moment envelope during launching looks like Fig. 2.70. This curve the edge stresses because of the much higher moment of inertia of these crOss-
takes both dead load and launch prestressing into account, since the undulations in sections. Finally, 'the'neg'a'tive moment envelope 1n the front span highlights the
the deck gravity axis (in relation to the subsequent cross-sections) generate eflects of the use of a ‘long’ launching nose, as in Fig. 2.36.
prestressing eccentricities that influence the edge stress diagrams. The bending and shear envelopes rarely include the effects of thermal gradients
In the rear deck section (launching proceeds to the right of the figure), the and differential bearing settlements, since the first may be immediately superim—
migration of the support diaphragms generates regular undulations in the envelope posed and the second generate a huge number of load combinations. Therefore,
with a period equal to the span length, since the highest moments are reached they are often calculated for some typical situations and then added. In the presence
when these heavier sections are in mid-span. Consequently, the envelope of the of temporary piers, the envelopes are also affected by the different deformability of
positive moment has its maximum at the support sections and the envelope of supports, which is calculated by taking the low thermal inertia of the steel piers into
the negative moment has its maximum at the mid—span sections, where the deck account.
is supported when diaphragms are in mid—span. ., . Irregular sequences of spans cause anomalies in the envelopes. The presence of
The undulation generated by the support diaphragms in the moment envelope a longer span in a regular spacing influences the envelopes of the whole deCk
depends on the concentrated load AQ corresponding to the weight increase of the section that passes above that span. If the longer span is close to the launch
diaphragm with respect to the standard cross—section of the deck. The amplitude of abutment,this may cause overdeSign of the whole superstructure. In the case
these undulations is approximately of irregular spans, therefore, the use of homogenizing temporary piers with
the temporary elimination of the support on some of the final piers may be
L
AM 2 iAQ§ (2.89) expedient.
120 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 121

2.I l. Verification of design assumptions by monitoring stresses should be expected when the instrumented section passes on to the launch—
The transitory stresses that arise during incremental launching are so high as to ing bearings. A preliminary finite—element analysis of the support region of the deck
influence the bridge dimensioning. Consequently, the design assumptions and the may define the theoretical stress distribution and permit a better comprehension of
technical specifications about bearing positional tolerances, thermal gradients, the measured data.
time-dependent effects and so on must be carefully evaluated, since they directly
affect the competitiveness of the project.
In the most delicate cases, deck monitoring during launching permits checking of 2.I 1.]. Elastic modulus ofconcrete
the design assumptions. It may also provide information about the elastic modulus Knowledge of the elastic modulus of concrete is of fundamental importance for the
of concrete, the coefficient of thermal expansion, the influence of topographic and calculation of stresses.
geometrical defects, the effects of shrinkage and creep, and the stress distribution The statistical code equations that relate the elastic modulus to the compressive
when passing on to the launching bearings. strength of concrete (e.g. [64]) are generally too approximate to analyse the beha-
It is expedient to choose the section to instrument among the most stressed ones viour of instrumented sections, and direct determination is necessary. Because of
that is, in the front section of the superstructure. This also permits early retrieval of the great influence of this parameter on the results, it is convenient to use several
the first data. The areas monitored by the strain gauges should overlap (Fig. 2.71), methods at the same time.
to eliminate incongruous data. Embedded strain gauges at the lower edge of the One method consists of measuring the strains of cylindrical specimens subjected
webs should be aligned in more directions (e. g. horizontal, vertical and inclined by to monoaxial compression. However, the small dimensions of specimens and the
450) to measure longitudinal and vertical compressive stresses as well as tangential different curing conditions provide only approximate data and do not lead to
stresses when passing on to the launching bearings. Strain gauges sunk into con- reliable results. ‘
crete may be complemented with external extensometers that measure the distance A second method is based on tensioning of prestressing tendons and correlates
between two points of reference fixed to the deck surface. This allows monitoring a the average shortening recorded by many extensometers distributed along the sec—
PC section even after its casting. The use of glued strain gauges is also common tion with the total force applied by tendons. Strains must be read immediately
practice when monitoring the incremental launching of steel girders, especially before and after each tendon tensioning in order not to superpose thermal strains
when using trusses [63]. and the first creep strains on the instantaneous elastic ones. The mean of read
As well as the investigation of launch stresses, deck monitoring can be aimed at values, weighted according to the surface of influence of the extensometers, deter-
analysing the time-dependent behaviour of concrete. In this case, the section mines the average deformation plane of the section. The individual strains generally
to instrument is often chosen from the most stressed ones in the final support deviate from this average plane. The differences are higher when recording the
configuration of the superstructure. strains due to tensioning of launch tendons (when the concrete is relatively
To determine stresses from the recorded strains it is necessary to know the eiastic young) than when tensioning the parabolic tendons on completion of launch.
modulus and the coefficient of thermal expansion of concrete. Sunk strain gauges Although these deviations do not affect the calculation of the elastic modulus,
are generally equipped with thermocouples to identify the thermal strains. It is also they indicate that this measuring method tends to become imprecise as the
convenient to apply different methods for measuring the same parameters so as to number of tendons increases. In fact, this method adds very small strains that
compare the results and eliminate most of the uncertainties. Lastly, significant local are affected by measurement errors, and even the assumption of conservation of
plane sections is questionable in these tensioning sequences. On the other hand,
operating simply with the difference between the final average strain and the
5 thermocouples initial one would introduce strain components due to creep, thus underestimating
______ \mm,_ __,....___ the elastic modulus.
A third method is based on measuring the front support reaction RA in two
longitudinal instrumented block
strain meters
subsequent launch stages (Fig. 2.72). It leads to a reliable evaluation of the flexural
\ straln meteraFI/
stiffness E1, and on a cloudy day (without significant thermal gradients) it is only
affected by the geometric tolerances in the superstructure and the manometer
—————o—-{: —o——
tolerances of the weighting jacks. After reading the front support reactions RAG)
3 strain meters
placed at 45° 8 thermocouples and RMD in the two subsequent launch positions, in the instrumented section S it is

Fig-2.71- .Sample.._..0f an instrumented section __MS(1):RA(1.)X1_Pd _ _, __ . , (2-90)

122 BRIDGE LAUNCHING INCREMENTAL LAUNCHING OF BRIDGES 123

Only specimens, although less representative, permit measuring the elastic mod—
ulus directly. In all other cases, indirect determination of the elastic modulus is
affected by the uncertainty of the actual cross-section of the superstructure.

2.1 1.2. Thermal correction of results

Thermal correction of the results gained from deck monitoring is a problem that is
neither simple nor to be underestimated, since the evolution of temperature in the
superstructure is sharp and can affect the measured strains. It can be approached in
two ways.

r "Fix 0 Measuring strains frequently during operations that take a long time (e.g.
tendon tensioning) and interpolating the data so obtained. Once the elastic

W1)
deformation has been estimated, it is possible to cancel the effects of all
time-dependent phenomena.
0 Measuring temperatures in the instrumented section and calculating the mean
Fig. 2.72. Evaluation of the elastic modulus of concrete by measuring the support expansion 80 and the rotation go to determine from these the self-balanced
reaction in two subsequent launch stages thermal stresses, and then subtracting the related strains. This method assumes
a constant distribution of temperature along the superstructure, and a knowl-
edge of the coefiicient of thermal expansion and the elastic modulus of concrete.
in the first launch position, and
2.1 1.3. Coefficient of therma! expansion of concrete
MS(2) = RA(2)x2 — Pd (2.91)
Knowledge of the coeflicient of thermal expansion of concrete is fundamental when
in the second, from which taking readings over an extended period of time. It is also necessary for determining
the strains due to creep and shrinkage when the instantaneous strain is known [65].
AM 2 RA(2)x2 + RA(1))C1 (2.92)
The coefficient of thermal expansion must be defined case by case, since it
In a fibre of the instrumented section at distance 2 from the cross-sectional centroid, depends on many parameters that are hard to define, such as the mineralogical
the difference in the flexural curvature is proportional to the strains measured in the nature of aggregates and the humidity and temperature of the concrete. The coefifi-
two support positions: cient is evaluatedby recording the average strains of the instrumented section when
temperature changes. This requires a sunny day with high thermal differentials.
A99 2 m (2.93) Since the coefficient of thermal expansion governs both the mean expansion 80
z
and the rotation go of the cross-section, the average Strain of the centroidal fibres
from which is often used to reduce the influence of gradients as much as possible.

E]
_ AM
_ A70 (2.94) 2.11.4. Shrinkage
The time-dependent strains recorded by extensometers combine the effects of
In practice, the front support reaction should be measured in several launch phases
shrinkage and creep, and the component of strain due to shrinkage can be evaluated
(e. g. during the launch stops for repositioning of the thrust devices) to obtain an
only by means of specific measuring techniques.
average value of E1 less affected by the geometry tolerances of the superstructure.
Other similar solutions are possible. On a cloudy day, the elastic deflection of 0 A first method consists of building a sample segment of the superstructure,
the launching nose may be surveyed in subsequent launch phases. When using adequately instrumented and placed near the bridge to guarantee similar
friction launchers, deck raising at the launcher may be correlated with the vertical hygrometric conditions.
force transferred by the lifting jacks, without any need for instrumenting a section. 0 A second method consists of instrumenting a concrete block placed inside the
When these methods are combined, nose surveying provides information on the box cell (Fig. 2.71) with an immersed strain gauge. In the first curing stages,
variation of the elastic modulus of concrete over time, while the friction launchers shrinkage is proportional to the ratio of the drying perimeter to the cross-
provide . average. data about young concrete- sectional area... Therefore, it is necessary to dimension the concrete block with
124 BRIDGE LAUNCHING

the same proportions as the superstructure, and to seal its ends to reproduce the
hygrometric conditions. In spite of this, the concrete block is protected on each
side while the superstructure has protected internal surfaces but external
sides exposed to wind and drying. Therefore, the ”measured data can be 3. PC bridges
approximated.
o A third method consists of sinking a strain gauge mechanically isolated from the
rest of the section into the deck concrete, to obtain readings independent of the
creep strain.

Each measuring method confirms that shrinkage strains have a markedly seasonal
evolution, with shortening in dry periods and lengthening in the humid ones [65].
This contradicts laboratory data and often produces a total shrinkage substantially
less than expected.
3. I . Presizing of the superstructure
The most characteristic aspect in designing launched bridges is the need to absorb
the stresses due to the temporary support configurations assumed by the deck
during launch. Structures that present so many and different load conditions
require careful presizing, to avoid excessive stresses in one of the many launch or
service stages, but without excessive prudence that would result in an oversized
structure.
The flexural capacity of a PC section is governed by the depth kc of the central
core, which depends on the flexural efficiency pf (2.13) and on the total depth H of
the cross—section:

I
h = H: H .

The highest fluctuation of the bending moment during launching is

AMmaX = kyAL2 (3.2)

where y is the specific weight of concrete, A is the cross-sectional area and L is the
span. Neglecting thermal gradients and differential bearing settlement, with an
optimized launching nose it is k = 0.16 +0.22 in the deck front zone and k z
0.12 +0.15 in the rear zone.
Launch prestressing can avoid edge decompression only if the moment
fluctuation (3.2) maintains its resulting force within the central core. Therefore,
the minimum axial prestressing Fmin is

AMmaX _ kyAL2
FH1111
- : he p (3.3)

For the same span, Fmin depends linearly on dead load and inversely on the depth of
the cross-section and on its flexural eflicienCy. Consequently, a higher efficiency can
also permit savings in, prestressing.
i
.i

is
_
.
126 BRIDGE LAUNCHING PC BRIDGES 127

The role of area is emphasized by expressing the minimum prestressing as By increasing the bridge depth it is also possible to reduce the thickness of the top
2L2 slab, governed only by punching and transverse strength (in the areas unaffected by
Fmin = 1521121? [H (3.4) prestressing tendons). For the top slab of highway bridges, it is (again in metres,
and using the symbols of Fig. 3.1)
Since Fmin is a quadratic function of the area, it is imperative to reduce the latter as Cl
much as possible. t:——- (3.7)
t 25+30
In order not to penalize the flexural efficiency (2.13), the reduction in area should
be lower than the reduction in the moment of inertia, so that the radius of gyration with minimum values of 0.22—0.26 m depending on the distance between the webs
(2.11) can increase on the Whole. Since static and technological constraints prevent and the presence or absence of transverse prestressing. The bottom slab thickness is
excessive cross-section lightening, it is almost always necessary to increase the more conventional, due to the convenience of lowering the centroid as much as
moment of inertia at the same time. possible:
As discussed in Section 3.7.2.1, the cross-sectional centroid should be lowered as
much as possible to enhance resistance to negative moments. This might be 3, = (b — W)(0.07 — ——L_—) (3.8)
3700
achieved by thickening the bottom slab, but then the cross-sectional area would
increase and the structural efiiciency (2.19) would decrease. Therefore, it is generally with a minimum value of 0.18 m, and 0.24 min the regions affected by longitudinal
convenient to take advantage of the wide top slab and to increase the moment of prestressing tendons. This value slightly increases in the support sections to better
inertia by increasing the deck depth. This intervention marginally afiects the cross- resist the negative moment. In railway bridges, the minimum thickness for both
sectional area; the moment of inertia increases with the quadratic ratio, and the slabs is about 0.30 m.
radius of gyration linearly. The structural efiiciency increases as a result of the The web thickness of mid-span sections is usually governed by the launch shear
higher carrying capacity, and the flexural efficiency increases as well. stresses. When a cross-section passes above a pier, it resists tangential stresses due
A larger deck depth thus becomes the aesthetic characteristic of PC bridges built to dead load. In the case of a final span section, it resists much higher stresses than
by full span incremental launching. The bridges tend to lose the typical slenderness those reached in service in its permanent position. Therefore, its webs are designed
of continuous beam decks, especially for the longest spans, and assume the common for these temporary stresses and their thickness increases with the span, contrary to
look of simply supported box girders. In case of need, however, the use of bridges built with other techniques. In the case of highway bridges with internal
temporary piers may avoid these aesthetic consequences. prestressing, the thickness of each web may be estimated (in metres) as
The dimensional implications of the full span incremental launching method have
L
been studied in some statistic analyses [43,45,46]. The linear regression of the cross— z... = 36% + 0.02) f f (3.9)
sectional depth for highway bridges is (in metres) \ f

L
H = 09.4 +227
— (35)
-
compared to
L
H _ 0041—21? (3.6)

for PC bridges built by other methods [44]. In railway bridges it is often

12 g H/L 5 14. . v . .
The aesthetic efi"ects of a higher cross-section can be reduced by the use of sloping
webs to avoid marked cantilevers, and by modifying the surface texture of the
concrete. This is easy to do in a fixed formwork placed on the ground. Cross—section r
asupp---------------ana-np- -p----~--------
par-1.." d Arp'-"l".'.'. '-.' o n a . gunn- I'll- -,,_ A

transverse dimensions confirm this trend. In launched decks, the ratio of the canti-
lever length—to-top slab width is equal to 0.24 as compared to 0.27 for bridges built
with other techniques, but the ratio of the bottom-to—top slab width is always equal
t0 0:44, independent of span andtechnology. . . . , I Fig 31.. Cross-Section symbols
128 BRIDGE LAUNCHING PC BRIDGES 129

and with external prestressing joints and to allow reuse of the same internal formwork. As the span increases, the
I _ BL — 500 + 0.30 (3.10)
segments become shorter than the typical span to limit the cost of the installation.
In this case, the segments may be one-half or one-third of the span. In short bridges
W * 2000
or in the case of external prestressing, for any given span length, it may be
with a minimum value of 0.30 m. Web stiffening at the permanent support sections convenient to use much shorter segments, building more segments on a continuous
is rarely necessary since the parabolic tendons (tensioned on completion of launch) support and launching them as a whOle.
reduce the shear force in the webs. When the number of segments composing the span is defined, the position of the
The web thickness of mid-span sections of railway bridges is about 0.50 m for construction joints should be verified from the structural point of View (in relation
conventional spans of 40—50 m, and in the support sections it reaches 0.90—1.00 m. to the counterflexure points in the final support configuration) and from the con—
A consequence of a higher deck depth and thicker webs is increased consumption struction point of View. In fact, the internal formwork of the box girder is
of materials. An upper-bound value for the concrete quantity hcon (ratio of the dismantled by backwards extraction through the joint section. If the diaphragms
overall volume of concrete, including stiffening and support diaphragms, to the are cast together with the rest of the section (Fig. 3.2), the joint might be placed
top slab surface) in internally prestressed highway bridges launched on the full immediately in front of the diaphragm so that most of the formwork can be
span may be obtained (in metres) from recovered Without obstruction.
L If the segment is built in two casting phases (bottom slab and webs first, and, the
11,0, = 0.25 + TE (3.11) top slab later), the recovery by backwards extraction affects only the form table for
while in railway bridges it increases to the top slab and the joint can be placed anywhere, provided that the support
diaphragms are tapered. As an alternative, diaphragms can be cast in a secOnd
L stage, even after completion of launching, by means of casting pipes through the
km, z 040+
. —
100 ( 3.12 )
top slab. Finally, when launch-prestressing tendons are coupled instead ofjoined by
Prestressing steel (almost always strand as it is easy to introduce into empty ducts) overlapping, the area of the couplers decreases the moment of inertia of the joint
is 40—60 kg/m3 in highway bridges and 60—70 kg/m3 in railway bridges. section, which then should be placed at the span—quarter locations.
Reinforcement is 90—150 kg/m3 in highWay bridges and 140—150 kg/m3 in railway In most launched bridges, each segment is match-cast in-place against the pre-
bridges. vious one, and organization of the whole yard therefore hinges on the formwork.
Whatever the subdivision in segments and casting phases may be, to build a seg—
3.2. Deck segmentation and yard organization ment in the shortest time it is necessary to avoid concentrating in the formwork
Planning of deck construction (choice of a casting sequence and of a prestressing activities that might take place elsewhere. Therefore, production is organized into
scheme, design of the casting yard and of the iaunch equipment, etc.) depends on parallel rather than serial processestintegr-‘Jeufirpartiai prefabrication of steel cages,
length of the bridge, the time available for its construction and the length of the continuous rather than single inserts, precast anchor blocks for prestressing
deck segments compared with the typical span. tendons, etc). A correct handling scheme is essential, and the use of a tower
The bridge length influences the level of industrialization of the casting yard. crane is decidedly advantageous for this purpose. ‘
Long bridges are composed of a large number of segments and permit amortization
of the investments necessary to industrialize the production cycle to reduce labour
costs. Shorter bridges require a less expensive organization, since it is impossible to construction Joint —\ launching __
reach an adequate level of repetitiveness [12].
Deck segmentation depends 0n the bridge length and on the time available for
construction, taking the duration of yardassembly, start—up of production and final
dismantling into account. This is true when, as in most cases, the deck segments are i ' 1/ I I

match-cast against each other in situ; in a case of segmental precasting, time is less A Form extraction ‘
of a constraint as segment assembly is a much faster proceSs.
When the optimum average length of segments is defined, this dimension must be
adapted to the typical span, so that segments correspond to the span or constitute a
whole fraction of it. In bridges composed of a large number of short, equal spans,
segmentsmaybeas long as, thewhole ,Sp.an.,to. reduce. thewnumber of construction : . fig. 3.2. Construction joint location for monolithic segments
130 BRIDGE LAUNCHING PC BRIDGES 131

If the number of segments is high and enough space is available, it is convenient Cage assembly can also be done by transferring preassembled web cages and slab
to create a preassembly area for most of the loose material (bent bars, ducts, panels into the formwork. In this case, the preassembly template is very simple
anchorages, inserts) to transfer them into the formwork as a whole. (Fig. 3.4) and cage segments can be handled with a conventional tower crane.
Steel cage preassembly outside the formwork has many advantages. Cage Yard industrialization reduces the time required for each activity, compared with
assembly is a slow and delicate operation because of the presence of ducts, conventional methods. Moreover, operations are simplified, the number of workers
anchorages and minor forms. Preassembly permits extracting these activities in the narrow formwork area is reduced to a minimum, the full utilization of
from the critical path and confining them to a specific area and organization. personnel is permitted by parallel activities performed during concrete curing,
By separating the casting yard and a preassembly area, it is possible to attain a and formwork can be made cleaner and more durable.
better rotation of labour, especially during concrete curing stages. After segment A high level of industrialization does not require prohibitive investments, as most
extraction, the formwork can be cleaned, realigned and prepared without of the equipment is standard (tower crane, gantry crane, movable coverings, non-
constraints, and cage transfer does not compromise results. Usually, the specialized construction equipment). Only the launchers, the launching nose and
preassembled cage includes the back-joint shutter, complicated by the presence parts of the formwork are specific for this construction method, and they are easily
of tendon anchorages, and after transfer of the cage and insertion of the internal reusable.
form, the segment is ready for casting.
The preassembly template can be adjacent to the formwork or separated from it, 3.3. Casting phases
as in Fig. 3.3. The adjaCent template does not require cage lifting, since it is possible The construction cycle of a segment is the sequence of operations between the
its advancing into the formwork during launching of the superstructure by means of launching of two subsequent segments. Its duration generally varies from 7 to 10
rollers. A separated template requires a specific gantry crane (which can also be calendar days, though it is expedient to reach this lower limit so that the concrete
used as a covering) for cage transfer, but it makes the two activities more inde— can cure over the weekend.
pendent and simplifies bush—hammering of the construction joint and tensioning of The sequence of operations depends on the casting procedure, on the eventual
launch tendons. cage preassembly and on the time required for concrete to reach the strength
necessary to tension launch tendons and resist launch stresses. Natural hardening
of 35-MPa concrete is rarely complete within the duration of a weekend and

Fig. 3 .3. Cage preassemblyfor thefirsl casting phase of-Ifhe Serio River Bridge. In the
bashsrmd: the internalferms are ready for transfer..(aurh0r) , Fig. 3 .4. Breassembledweh cages- (author)
 132 BRIDGE LAUNCHING PC BRIDGES 133

appropriate devices should be used, such as precast anchor blocks, local thermal
cycles with electric wires embedded around tendon anchorages, or steam curing of
the whole segment. Therefore, it is often convenient to directly increase the design stole Form
strength of concrete. A higher specified value allows the necessary strength gain
1
in the scheduled time and at the same time lightens the superstructure and
tunnel Form
reduces the influence of launching stresses, unexpected events and time-dependent
phenomena.
Whatever the chosen solution may be, subdivision of the deck into segments, sub—
segments and casting phases always involves the presence of construction joints
between the elements. The adoption of vertical joints is inevitable, and the adoption
of horizontal joints has several advantages. Horizontal j Oints restrict the quantity of
concrete processed daily and allow the execution of subsequent operations without
excessive interference, the dismantling of the internal form (webs and diaphragms)
with a tower crane, and the easy removal of rubble. If the horizontal joints are at
the top slab level, they also permit support of the form table for the top slab from
the webs and avoid the horizontal cracks that often affect monolithic casting as a
extractlon roll Foundation
result of the settlement of fresh concrete in the webs.
In the vertical joints, the rear end of the front segment is cast against the shutter
and its surface is smooth and weakened by oil. In the new segment, shrinkage and
imperfect adhesion tend to produce horizontal cracks orthogonal to the joint sur—
face. These cracks can be avoided by bush-hammering or chipping the contact
surface so that roughness acts as distributed shear keys, by wetting the joint
Fig. 3.5. Ribbed Slabs are easily built by monolithic casting
prior to casting the new segment and by placing a vertical wire mesh near the
joint. The use of specific shear keys is unnecessary, since with superficial irregula-
rities of about 10 mm the relative movements necessary to mobilize the distributed
shear keys are smaller than those required by concentrated keys, and the reinforcing Even the construction schedule is rather rigid because of the interference that
bars crossing the joint further increase the shear strength. occurs at the only working location and the dead times for concrete curing.
The difference in temperature between the cured segment and that in the setting Therefore, Cage preassembiy is often unavoidable, and this increases the length of
phase should be reduced until concrete has reached the strength necessary to the casting yard and requires expensive handling equipment.
activate the reinforcement and the distributed shear keys, since in these first phases The use of internal left-in-place forms is economically acceptable only in the
the limited adherence is unable to control crack width. In particularly severe shortest bridges. Therefore, the internal formwork is almost always self-extracting
climates it can be necessary to maintain the temperature in the cured segment or through the support diaphragm and adjustable by means of hydraulic pistons
to heat its rear end to 15—250C, and to cool the setting concrete of the new segment. (Fig. 3.6). It advances with the superstructure during launching and is extracted
The need for these measures depends mainly on the heat of hydration of cement and backwards to place it in the new casting position only after the preassembled cage
is highest in the case of particular materials such as high performance concrete [66]. of the next segment has been transferred into the formwork. To avoid geometric
problems that might jeopardize the speed of the operation, the steel cage is
3.3.]. Monolithic casting _ . . V _ . . .. . _ assembled on an internal template that stiffens it during transfer, with additional
In principle, monolithic segment casting is adopted only for ribbed slabs (Fig. 3.5). costs.
In fact, the operations involved in monolithic casting of box girders are rather The need for refined forming systems and heavy handling equipment involves
complex, having to satisfy requirements of cage assembly, concrete placement, significant investments. These can be amortized only in long bridges, and compe-
vibration and internal form dismantling, that often conflict with each other. titive alternatives can be found by subdividing the superstructure into shorter
Therefore, fOr long segments this scheme is so rigid as to limit its adoption almost segments. Independent of the span, and sometimes even of the bridge length, a
exclusively to long span railway bridges, whose significant deck depth eases the more .marked deck segmentation than the conventional one (half of the typical
access 190 theboxcell . ' i _.__._V__S,..pan)_ can reduce the construction cost of the bridge.
134 BRIDGE LAUNCHING PC BRIDGES 135

main thruster

casting cell

curing Pall

Fig. 3 .6. Rear end of a self-extracting internal formwork (author)

Even with the same general principles, it is possible to single out three different
construction schemes that adopt short monolithic segments:

0 construction in a fixed casting cell, with the general launch at the extraction of
each segment
0 construction in a mobile formwork, with launch only after completion of a Fig. 3 .7. Fixed casting cell and general launch by (back thrust
certain number of segments
0 construction in a fixed casting cell, with segment extraction independent from
the general launch of the superstructure. the superstructure in the bank span can be combined with a temporary pier to
reduce the flexural stresses in the non-prestressed portion of the superstructure,
In the first scheme, the formwork is a casting cell similar to those usually adopted as in Fig. 3.7. '
for short—line segmental precasting (Fig. 3.7). It is placed along the bridge align- The advantages of this scheme are a small, automated formwork, repetitive
ment at a certain distance from the abutment. The construction cycle provides for operations, and a more efficient rotation of labour. The disadvantages are in the
the addition of a new segment match-cast against the rear end of the superstructure length of the curing rail, the need for precise regulations and controls (errors
and the general launch, in perfect analogy with the case of longer segments. accumulate with the number of joints), the multiplication of launch operations
At its extraction from the casting cell, the segment, not prestressed, slides along a. that immobilize labour, and the delicate handling required for the back thrust of
curing rail that progressively receives all the segments of a span or of a substantial segments not completely cured and prestressed.
part of it. The presence of a long curing rail is indispensable, as without it the delay A second scheme avoids the frequency of launch operations over relatively short
in clearing formwork produces such long dead times as to make a highly indus— distances, which is the main inconvenience of the first scheme. In this case, the
trialized yard useless. When the planned number of segments is cast, launching curing rail supports a section of superstructure as long as the whole span, and a
prestressing is applied to the whole deck section supported on the rail. To reduce short formwork shifts along the rail to cast the subsequent segments (Fig. 3.8).
the cost of tendon anchorages, this solution requires a long distance between the When the whole deck section is finished, launch prestressing is introduced in one
Casting 0611 and. theabutment AS an alternative, a temporary falsework supporting ._.,..operation.. just before launching. This allows a drastic reduction in the number of
136 BRIDGE LAUNCHING PC BRIDGES 137

mobne Formwork
prestressing, the construction joints between the segments can coincide with the
diaphragms. Diaphragms can be cast together with adjacent segments or together
l '1;-_ {fig i!_uss;ii§§m!llz
fififl“J---I l -.. .

with the previously cast ones, to assemble their steel cages through the solutions of
continuity in the webs.
Even in this case, construction proceeds in an opposite sense to launching. As the
curing rail
last segment is subject to the local stresses of tendon anchorages, it is convenient to
adopt precast anchor blocks (as in Fig. 3.40) or to anticipate its construction. This
last intervention is practically unavoidable in the case of external launch prestres—
sing, although it complicates recovery of the internal forming system of monolithic
segments. By proceeding in the opposite direction (towards launch), one can rely on
cured anchorages and the last segment built is subject to distributed, homogeneous
lllllllllll I ’ I ' ' ‘ u----u .
' I IIEEFBEIIE
compressive stresses; however, recovery of internal forms is more complicated.
Dramatic savings of labour and equipment can result from the combination of
short segment casting with mobile formwork, launching of long superstructure
sections and transverse shifting of the whole superstructure on completion of
launch to clear the casting yard alignment for the construction of a second parallel
_ launching pin superstructure.
The Tiziano Bridge in Alessandria, Italy is a ZOO—m bridge composed of two
parallel box girders joined at the top slab level and spanning four continuous
SO-m bays. It has been built by shifting the 10-111 formwork shown in Fig. 3.9
along a ESO—m curing rail in an opposite direction to launching. In spite of the refined
draw Jack structural solutions (internal launch tendons and external service prestressing),

Fig. 3.8. Shifting formwork and conventional launch

launchroperations and permits movement of the superstructure without risks, since

it is wholly prestressed. This also permits a reduction in bar splicing because
reinforcement of the whole span can be progressively assembled on the curing
rail. The disadvantages are a more complex formwork than that Of the previous
scheme (it has to shift too) and the slightly higher cost of foundations.
Segments can be cast in two phases (bottom slab and webs first, and the top slab
later). In long bridges, whose construction may be more industrialized, monolithic Fig. 3.9. Shifting formwork for the Tiziano Bridge. Reinforcement is assembled in
casting, is-_oft,e_n__convenient- When there are ,manydiaphragms becauseof external parallel....(auth0r).
138 BRIDGE LAUNCHING PC BRIDGES 139

construction of this bridge required minimum investments for specialized construc-

tion equipment.
On completion of launch of the first branch of the superstructure, transverse
shifting took place after tensioning the polygonal tendons of service prestressing
and relieving the residual support reactions at the four temporary piers, one of
which is shown in Fig. 3.10. The bidirectiOnal launching saddle is shown in
Fig. 3.11 above the high-strength-mortar rail for transverse shifting. Laminated
neoprene pads have been inserted between the transverse rails and the bottom
sliding surfaces of the launching saddles to lock the saddles during longitudinal
deck launching. Before transverse shifting, these neoprene pads were replaced with
conventional neo-flon pads with equal thickness.
After completion of the second branch of the superstructure, the two box girders
were joined with a cast-in-place continuous kerb. Finally, the four temporary piers,
placed along the casting yard alignment to halve the longitudinal launch span, were
demolished. '
A third construction scheme combines the advantages of the two previous
schemes and avoids their weak points (Fig. 3.12). The casting cell is fixed and
placed at the rear end of the curing rail. All the segments compOsing a span are
built by match casting and progressively launched along the rail by back thrust as in
Fig. 3.11. Bidirectional launching saddle for transverse shifting (author)
the first scheme. When the last segment of the span has reached the necessary
strength, the whole span is launched by back thrust to the rear end of the super-
structure, to which it is connected by a closure joint. After the introduction of
launch prestreSsing, the whole superstructure. is launched conventionally with
a launch step as long as the whole span. If the distance between the casting cell
and the rear end of the superstructure is longer than the span, during the operations
of closure, prestressing and main launch it is possible to proceed with the construc-
tion of additional segments. In this case, therefore, these additional operations do
not lengthen the constructiOn schedule". '7 W W W :
With this scheme, the casting cell is fixed and potentially automated, foundations
for a mobile formwork are avoided, the general launch operations are still con-
centrated while the single extraction launches are simple and local, and the closure
joint against the rear end of the superstructure avoids accumulation of geometric
imperfections. The disadvantages include the long curing rail, the cost of the
extraction devices and the additional closure joint casting.
Compared with the use of longer segments, all these schemes introduce additional
operations, and as the bridge length increases, this compromises the savings in
equipment with higher labour costs. On the other hand, a common advantage is
the possibility of introducing launch prestressing independent of the production of
segments, and to reduce the cost of tendon anchorages as prestressing affects the
whole span.
With internal prestressing, these construction schemes may be competitive in
relatively short bridges, in which the investment required by a 20—30-m formwork
. cannot make incremental launching competitive compared with other construction
Figr3..],0_”jinansyemgnshiflmgofthefirstbranchofthe. Tiziano Bridgetlauthor). . . . . ., methodsln the case ,of external prestressing, and particularly in the case of
140 BRIDGE LAUNCHING PC BRIDGES 141

auxiliary rear jack curing Pail rear

contflever

worklng draw rear extractlon

platFor‘m Jack support launchlng pln

C
MT
-----------

concrete temporary piers

cast-ln—sl‘tu segment Form lowering Jacks

Fig. 3.13. Aerial casting cell on temporary piers

3.3.2. Two-phase casting in a single formwork

In incremental launching, the most popular casting method is the two-phase scheme
inside the same formwork, with a horizOntal joint at the level of the top slab
(Fig. 3.14).
The external formwork does not presehtparficuiar problems: its design can aim
Fig. 3.12. Fixed casting cell, segment launch by back thrust, and conventional general for operative simplicity by means of sophisticated devices, or for the reuse of
launch ' modular equipment that is easier to amortize.
The most expensive type of external formwork consists of rigid steel frames
hinged at the base (Fig. 3.15). Being designed for a specific cross-section, this is
rarely reusable and its adoption is justified only in long bridges that make the cost
antagonist launch tendons, when construction joints between segments coincide of its whole amortization lower than the labour cost that it avoids.
with deviation diaphragms, the use of short formworks can be convenient even Less expensive formwork consists of modular towers that are dismantled using
in very long bridges. Finally, specific applications permit a solution in the case of their manual adjustment screws, or of towers supported on transverse rails and
settlement or no embankment (Fig.3.l3), since the cost of a small formwork dismantled by side translation. The simplicity of these schemes and the easy amor-
supported on piles or temporary piers is still reasonable. tization of the formwork are balanced by higher labour costs, which limit their use
In spite of the lower equipment costs, construction duration is substantially to short bridges. ‘
the same as for ordinary formworks. A half-span formwork (about 20 m) with a An intermediate type, less expensive and more adaptable than the first, involves
weekly cycle permits daily progress of about 4 m. A 4-m casting cell soon the use of modular towers installed on a lowering platform supported by hydraulic
reaches a daily cycle (except for the more complex pier segments) and an 8—m jacks. Lowering of the platform facilitates segment dismantling, the towers can be
form with 2-day cycle and cage preassembly permits better control of unexpected reused with new cross—sections, and their regulation screws are used only when the
events. . .. _ ,. ,, . ' ' .___f;o,rm geometry has to be adjusted.
-
 142 BRIDGE LAUNCHING PC BRIDGES 143

Fig. 3 .14. Two-phase casting in the same formwork for the Palizzi Bridge (author)

Unlike external formwork, the shape of the internal form depends on the position
of the horizontal joints. When the joints are placed at the upper nodes of the cross-
section, the web form is limited to side panels, or includes a bottom slab
counterform in the case of very sloped webs (left cell in Fig. 3.16).
In short bridges, the top slab formwork is made with materials that are left in
place (precast concrete planks, corrugated sheet metal or plastic plates). In longer
bridges, steel or plywood panels are used, supported on modular tower or on
self—supporting steel frames sliding along extraction rails anchored to the webs.
In a case of joints placed at the lower nodes of the cross-section, internal formwork Fig. 3 .15 . Monolithic steelframe formwork. CZoSe up, the parking support for the top
is similar to that for monolithic casting and is therefore rather expensive. In a case Slab compacting roll (author)
of intermediate joints (e.g. very wide bridges in which the deck slab is stiffened by
transverse ribs), both types of internal formwork can be used. In the case of a single form, steel cage preassembly largely prevents the activity of
Compared with monolithic casting of the whole cross-section, the scheduling of one gang from interfering with or impeding the activity of the other. Even so, there
activities is generally less restrained, and by preassembling the steel cage it is not is some interference: the adjustment and the completion of the steel cage after its
difficult to achieve an eificient construction schedule. Finally, two—phase casting in a transfer into the formwork, the duct connection and so on. In the case of long
single formwork is generally adopted when two parallel bridges are launched atthe segments, the activities of the two gangs can be staggered longitudinally (for
same time (Fig. 3.17), because this scheme produces an optimum rotation of labour instance, starting the concrete placement on one side while cage assembly is still
eVen without steel cage preassembly. going on at the opposite side of the segment), but this is generally a remedy for
inadequate scheduling rather than an efficient organization. In the case of short
3.3.3. Two-phase casting in a double formwork segments, these solutions generally lead to chaotic situations that are in conflict
Progressive labour specialization has led to the subdivision of workers operating in with the cadenced cycles that characterize incremental launching.
the yard into two distinct gangs: the reinforcing bars workers, often working for a As an alternative, a long external formwork can be combined with a much
subcontractor, and the carpenters. shorter_..._intcrnalform. Work can be developed in different points, but costs are
 144 BRIDGE LAUNCHING PC BRIDGES 145

generally high. The forming system for the Charix Bridge (Fig. 3.18) was composed
of an external formwork as long as the whole span and of a shorter shifting internal
form [67].
When the yard is long enough, it can be convenient to divide the production into
two distinct formworks in which different elements of the segment are built in
operational sequences staggered both in time and in space. The advantages of
this solution are many. By operating in two distinct areas, steel cage assembly in
one formwork does not jeopardize the activity of carpenters in the second form-
work and does not require hoisting equipment. The time available for each activity
increases, the critical situations reduce and the concrete surface finish improves as
dismantling may be delayed. The main disadvantage is the higher cost of the form-
work and of any protective structures, which must be double the length. Another
potential disadvantage is the higher shrinkage difierential between the two elements
of the cross-section (AASHTO [42] requires that when the bottom slab and webs
are cast first and the top slab afterwards, the time between concrete placements
must not exceed 3 days). However, creep is so high in these first curing phases that
shrinkage tensile stresses fade quickly.
In the less—expensive configuration, with horizontal joints at the top slab level and
forwards shifting of the rear segment on fixed bearings, the casting yard is divided
Fig. 3.16. Construction joints at the top slab level ease dismantling of internal form into two zones (Fig. 3.19). A rear form, shorter than the completed segment, is
and removal of rubble (author) i . intended for casting the bottom slab and the webs. A front form is intended for
I casting the top slab and the remaining short portion of the first-phase U-segment.

Fig. 3.18. Forming system for Charix Bridge (reproduced with permission,
-Fig. 3.17. Parallel launching-results in optimum labour rotation (author)-- - , -- - _. . Scetaroute.,_.Berenguier)... .
 146 BRIDGE LAUNCHING PC BRIDGES 147

+—— rear Formwork —-{— —'-¢— Front Fornwork ——{— launch pin There are two main problems to be solved:
15" phase casting ~\ 2"“ phase casting _\ /— draw tendons
0 reducing the friction during extraction to prevent the superstructure from
dragging the formwork
....._ ..__.___._._
“m“...gm\\\\\\.““‘\\\‘\am“. mm . mg.
.,,,,,,,._m,.,,,,..,.,,,,,,,,,.”manna.___.m““w mmuum
o reducing longitudinal and transverse flexural stresses in the young concrete.
r
I]
.i.—-m !:,—h|’ , m=mr m=hl .-. . ——»‘.
mmmm
‘— _ ..
Al .’ WM2;W
977W "4! "1/ "I/ "1/ "4r "1/ "(I'V/ 'V/ "I/ "4/ V/WJ_
.7 'I/ 'I/ 'I/ 'I/ 97 '4/ 7/ "I/ "I/ "4' 'W "/1" / Ila”
- . V . _ . V - . .

To reduce friction it is necessary to detach the formwork from the concrete surface.
.

'7
/ _,."

/
/

This can be accomplished either by opening and lowering the formwork, or by

.l

I 1. ll

{If-“ting Jacks completion Fornwork

hoisting the segment.
abutment

continuous Foundation launch bearing draw Jacks

Hoisting of the segment permits insertion of low—friction launch bearings between
Fig. 3.19. Two-phase casting in double formwork and rear segment shifting on fixed the segment soffit and the bottom form surface. These bearings can be mobile or
bearings fixed during the segment extraction, depending on Whether they are integral to the
superstructure or to the casting cell. In the case of mobile bearings, the segment is
raised by four jacks placed at the corners of the formwork. The two front jacks are
The rear U-segment is not match~cast against the front segment, as the presence equipped with launch bearings and remain fixed during segment extraction, while
of a solution of continuity and the adoption of a specific working cycle are the rear jacks are only aimed at inserting two neo-flon pads that support the
necessary, as follows: segment during extraction and slide with it along the formwork. In the case of
fixed bearings, neo-flon pads are inserted between the segment and launching
0 During curing of the rear U—segment, internal web forms are dismantled and the bearings protected during segment casting by precast planks (Fig. 3.20). The
rails for the form table of the top slab are fixed to the concrete webs. position of the rear fixed support is chosen so that the rear cantilevers of the
0 When the necessary curing has been reached, the rear U—segment is raised from superstructure (both at the beginning of segment extraction and when the rear
the fixed bottom formwork to insert neo-flon pads between the segment and deck end leaves the rear support) are compatible with the flexural strength resulting
the extraction bearings. As the rear U—segment is not yet joined to the front from partial curing and prestressing. This support scheme is also adopted in the
segment, raising does not cause stresses in the young concrete.
0 The general launch of the superstructure draws the U-segment forward, sliding
above the extraction bearings. precast concrete plank -—-— bottom f-‘or‘mwor‘k --1
0 Finally, when extraction is completed, the U-segment is lowered onto the
launch bear‘lng — hydraulic Jack —
extraction rails of the front formwork. The form table for the top slab, which
is still inside the previous segment, is drawn backwards and placed above the
U-segment, and the forms for the cantilevers and the short closure segment are
set up. '

This organization scheme allows construction of a 20—30 In segment per week.

Segment extraction on fixed bearings has been gradually replaced by the use
of continuous extraction rails, which, though more expensive, do not require raising
of the rear segment, nor the short closure segment, nor additional labour for ' . ‘iv'
.51.-.- a“?
_ :.v.- g5",
. -' ". 2-,3?-,' 9.;
fiv‘~91“
.2" 51.5:
$7., 4' a 4.4, ‘m_”
. 93ft —‘,‘- -' .9“ '~'1‘- __;.'Ig.
J-'€>.f’}flv’f,h’5’£}_-"Ziv - . *flick»;
_;."_‘_-.. .. ‘ az-T'q
"Ju’gfi, if? 51- o_k.-|
9 .3 . 9,.;-'}. . . r :44", .é.
"I." ¢" a;' .,

insertion of neo-flon pads during launch. ‘

- - m
my,”mun/um,mum,”.1 mmm . 1 m, , . ,

\ zIIIIIIII “WNW“ -§

4\ ‘ mlilllfllllll -" 'mmwmh

l

I \ \ \\\ \ \ \ \\ \ \\\ \ \ .
3.3.4. Segment extraction
Though the cylindrical shape of a launched deck can resemble continuous construc— \\ \ \t \\ ' \ ., . \ , \ ‘ \ \\ \ ‘
\\1’-‘oundatlon
. . the casting
0? . cell \ \ \ \ \ \
tion processes such as extrusion or slip forms, construction of the superstructure is
\ \ ‘\ \ \ ‘\ \ \ \ \ \ \ \ \
cadenced and segment casting is followed by curing inside the formwork. At dis- \ \ \ \ \ \ \ t \ \ \ \ '\ \ \ \ \ \ \ t \ \ \ \ \ x \ \
mantling, therefore, the new segment must be supported by bearings that permit its
Wevxtraction from the castingcell. .. ‘ ,.-_Fig 3’ 20 Segment- extraction. .onfixed bearings
 148 BRIDGE LAUNCHING PC BRIDGES 149

case of aerial casting cells (Fig. 3.13) as it permits concentration of support launch prestressing absorbs the low tensile stresses produced by friction and by the
reactions on to the temporary piers. first bending moments. Neo—flon pads advance with the superstructure and fall on
With both these schemes, the bottom form can be a concrete slab covered the ground in front of the formwork, to be cleaned and rearranged along the
with a plastic sheet, as no vertical movement is required. However, these savings extraction rails for the construction of the next segment.
in equipment and labour can be compromised by the settlement of the slab and the The vertical alignment of the extraction rails must be extremely accurate, as the
stresses caused by segment raising in the young concrete. surface of the superstructure that comes in contact with the launching bearings
Segment extraction by raising has been progressively replaced by the use of during launch is match-cast against their upper surface. Any difference between
continuous low-friction supports that avoid vertical movements in the super- the rail alignment and the- theoretical alignment of the launch surface generates
structure. In this case, two longitudinal extraction rails are placed under the deck secondary stresses in the superstructure during launching. Local defects cause stress
webs and support a central lowering form table, while the external form of the webs accumulation, and diffused defects have consequences similar to the differential
is conventional. This solution is adopted also in the case of shifting formwork settlement of bearings.
(Fig. 3.21). AASHTO [42] requires that the extraction rails be sufficiently stiff that their
In the shortest bridges, in which the frictional forces are less significant, the deflection during casting does not exceed 2 mm. They must also be supported on
segment is often supported on the extraction rails through steel plates lubricated foundations as stiff as possible, and provided with regulation devices able to set
with grease, hydraulic oil, soap or other antifriction materials. In longer bridges, their vertical alignment with tolerances from the theoretical profile not larger than
continuous stainless steel sheets are spot-welded to the upper flanges of the extrac- l—2 mm. It is then necessary to frame the casting yard in a wide closed traverse, to
tion rails, and neo-flon pads are aligned above them to obtain a continuous steel—. use continuous foundations instead of spread footings, and, in the case of settling
Teflon contact. In both cases, the upper surface of these extraction bearings is soils, to use piles and a light prestressing of the foundation kerbs to better control
protected by sheet steel or greased paper that prevents sticking to the fresh concrete, deflections. '
which is match-cast above them. Casting cells equipped with extraction rails are more complex and expensive than
Lowering of the central form table and opening of the lateral forms allow the those based on segment raising. On the other hand, they allow more precise vertical
segment to rest on a low—friction continuous support. Deck launching extracts the regulation and compensation for eventual settlement of the platform, and decrease
segment from the casting cell and trails it on the first curing supports, and the axial the secondary stresses in the young concrete.

3.3.5. Curing yard

Before beingextracted from, the formwork, the segment almost always receives just
a portion of the launch prestressing. Therefore, the casting cell is separated from the
abutment to let concrete cure enough to resist the whole launch prestressing and the
flexural stresses deriving from its launch on to the bank span.
In the area between the formwork and the abutment, the superstructure is
supported on closely spaced bearings that ’r’e’duCe the ,lspans to values consistent
with the grade of curing and prestressing reached by the segments. These supports
are small launching bearings placed on reinforced concrete pillars, and their
number is as small as possible because of their coSt and the labour they immobilize
during launching. ‘
PartiCular care muSt be taken in positioning the firSt temporary support ahead of
the formwork, which is subject to several restraints: ’

0 At the end of each segment extraction, the rear end 0f the superstructure (i.e. the
construction joint against which the next segment will be match-cast) must be
perfectly orthOgonal to the local vertical alignment of. the bridge, since angular
discontinuities would produce secondary stresses in the continuous beam during
W . .. -Fig- 3.21... Segment... extraction with- 5.02m long continuous rails (author) . . . , __-._launch..This requires thatthefirst support be placed at such a distance from the
150 BRIDGE LAUNCHING PC BRIDGES 151

abutment as to fade the fiexural rotation produced by the bending moment in Once these two extreme supports of the curing yard have been positioned, the
the bank span. spacing of any intermediate supports is governed by the hardening rate of concrete,
0 The rear end of the superstructure must also remain firmly supported on the the grade of prestressing applied to the segments and the available space. When the
formwork (raising would create a vertical discontinuity in the deck soffit that number of supports is excessive, the introduction of launch prestressing may be
makes launch difficult) and this requirement in turn affects the distribution of accelerated by adopting shorter and less overlapped tendons, or a temporary pier
support points in the curing yard. may be inserted in the bank span to bring the formwork closer to the abutment
0 The first support also acts as a transverse guide during and at the end of without jeopardizing fading of flexural stresses (Fig. 3.7).
segment extraction. This suggests positioning it as close as possible to the Finally, the spacing of curing supports may be e‘Xcessive in the final launch
formwork to use its foundations as a transverse restraint. phases, when the rear end of the superstructure cantilevers out for as much as
0 An excessive distance also complicates the extraction of the last segment the whole distance between supports. In this case, rather than decreasing these
because of the concentration of the support reaction provided by the formwork, temporary spans, it is better to introduce additional intermediate supports to be
which often breaks the sofiit. This can be avoided by rounding off the front end used only in the final launch phases, disengaging them during the previous stages by
of the extraction rails to progressively decrease the support reaction, or by not inserting neo-fion pads or the launch bearings themselves. As an alternative, a
supporting the superstructure in the last phases of segment extraction with rear steel nose may be used, especially in the case of thrust devices operating by
hydraulic jacks placed on neo—fion pads. The same problem arises when the friction (Fig. 2.25). '
rear end of the superstructure definitely leaves the launching bearings of the
curing yard (Fig. 3.22). 3.4. Assembly and launching of precast segments
The very first applications of the incremental launching construction method were
The distance between the last curing support and the abutment must also be for structures composed of precast segments arranged along temporary support
considered carefully, as it has to guarantee an adequate fading of the bending rails and connected by closure joints (Fig. 1.2).
moment of the continuous beam bank span and a positive reaction at both these Assembly and launching of precast segments has many advantages:
supports. Otherwise, the superstructure would rise from the curing support and the
static system would change. 0 Construction of most of the superstructure is independent from pier erection, as
precast segments can be stored elsewhere.
0 The use of labour and casting equipment is optimized in a series of highly
repetitive operations with the best quality control (especially if segments are
cast in a precasting plant) and the best amortization of investments (precast
segments for launched bridgesvarewabselutely conventional both in dimensions
and in handling).
0 Assembly and launching of the-superstructure are fast operations that require
few workers and little specialized equipment (two‘ support/launch rails, a
launching nose and a thrust device).
0 Launch equipment is easily reusable and relatively inexpensive in terms both of
initial investment and of assembly and dismantling costs.

Precast segments can be coupled by means of dry joints or epoxy, although this
solution has two disadvantages:

o Geometric imperfections of the superstructure depend on the number of joints,

and long bridges can require stringent geometric tolerances during short-line
match-casting of segments. This requirement is easily solved if the assembly
yard behind the abutment can be as long as a span. In this case, precast seg-
ments can be arranged along the support/launch rails and glued to each other
with conventional temporary clamping. Then, the superstructure section so
_._._o.btained..canbe joined.....to the rear end of the launched superstructure with a
152 BRIDGE LAUNCHING PC BRIDGES 153

short segment cast in situ (as in the case of Fig. 3.12), which avoids cumulative 3.5. Launch bearings and guide devices
errors. The technological evolution of incremental launching of PC bridges has led to a
0 The cost of longitudinal prestressing is higher than for cast-in-yard launched clear-cut simplification of support and guide devices of the superstructure, all based
bridges, because of the need for avoiding joint decompression both during on the original principle of steel—Teflon contact.
launching and under infrequent service load conditions. Release and In the very first launched bridges, sliding occurred on bearings integral to the
repositioning of a part of launch tendons on completion of launching can superstructure and supported on stainless steel plates integral to the piers, with
reduce prestressing costs and results in more efficient permanent prestressing the discontinuous launch cycle shown in Fig. 1.5. The length of each launch cycle
schemes, even if the labour cost slightly increases because of these additional was limited by the pier cap width. When limit stops were reached the launch had to
operations. be interrupted and the superstructure had to be lifted, pier by pier, to reposition the
bearings at the beginning of their stroke. The time and costs involved in these
As an alternative, segments can be precast with reinforcement emerging from the operations soon suggested adoption, instead of mobile bearings, of specific launch
joint surfaces and arranged along the support/launch rails at a distance of 0.20— bearings composed of neo-flon pads inserted between the superstructure and fixed
0.60 m to splice the reinforcement and to fill these solutions of continuity with no- launch saddles (Fig. 3.23).
shrink concrete. Shrinkage cracks in the closure joints can be further controlled Neo—flon pads are made up of neoprene sheets alternating with steel sheets and
with low local thermal cycles, or by casting the joint segments in the second half of glued to a lower Teflon plate. The multilayered elastomeric bearing improves the
the night [2] to rely on the thermal expansion of precast segments produced by the uniformity of the contact stress and absorbs the flexural rotations of the super—
daytime temperature during their setting. structure and surface irregularities, Whereas Teflon reduces sliding friction between
In both cases, yard operations are limited to the arrangement of segments along the superstructure and the launching saddles. Some trials have been made with
the support/launch rails, joint sealing and introduction of prestressing in the new through pins that stiffen the elastomeric bearing [68], but wear of Teflon generally
section of the superstructure, and construction speed can double compared with suggests avoidance of these higher investments.
casting in situ.
So far, the incremental launching construction method has been mainly used for
cast—in-yard bridges, as this permits combining the advantages of industrialized and
repetitive work processes with those deriving from a small number of construction neo-Flon plate screw coupling __

joints with through reinforcement. In addition, segment length and yard organiza-
tion can be easily adapted to the specific requirements, thus enhancing the overall ‘
competitiveness of the construction method. ‘
Launching of a precast segmental bridge can be a very competitive solution
too. Bridges whose length does not permit amortization of the investment and
assembly and dismantling costs of a launching truss for segmental construction
can be optimum targets, especially when it is necessary to allow unrestricted use
of the areas below. As an example, in a highway intersection with long and short
bridges, the launching truss might be used for the longest viaducts while the
remaining bridges might be launched, thus multiplying the assembly lines with
savings in both time and costs. In a case of many precast segmental overpasses on
existing highways, a precasting plant might feed several inexpensive, modular-i
launch yards for fast-track projects with all the advantages of the continuous
box girder structural solution. Even if a lot of used launch equipment is available,
8mm pollshed stainless steel sheet
costs are so reasonable that custom-made solutions are often advantageous and
concrete or steel launching saddle—
further ease project optimizatiOn. 1'
20mm neoprene plate
Obviously, a precasting plant can simultaneously feed production lines with
highestrength mortar seat
launching—truss assembly and production lines for launched bridges, thus further
improving amortization conditions and planning flexibility of the project. llFtlng Jack seat

' Fig 3.23.. Launching bearing.

'
,
'w-«(Wa—..
.‘i1:1 gr{i .- :
154 BRHXHELAUNCHING PC BRIDGES 155

Launching saddles are rounded reinforced concrete (Fig. 2.45) or steel plates disadvantages of discontinuous launching are also avoided, such as load
covered by a tensioned stainless-steel sheet polished to a roughness of about eccentricity on the piers (which, compounded with horizontal friction forces,
2 ,um. They are usually supported on high-strength mortar or on multilayered often required pier overreinforcing or temporary cable—staying).
elastomeric bearings that further improve the stress distribution in the case of TransVerse guides, necessary to maintain the correct plan alignment of the
excessive irregularities in the superstructure. In some cases, steel launching saddles superstructure during the launch, are placed beside the launching bearings.
have been placed directly on to the permanent bearings to be removed on They produce fixed or adjustable (Fig. 3.25) guide forces through rollers or
completion of launch. steel—Teflon contacts similar to those of the launching bearings. These transverse
In the most delicate cases, launching bearings are supported on jacks fed by a constraints are almost always placed outside the webs. However, special offsets
hydraulic circuit able to control the torsional couple produced by the support can be used in the case of rounded cross-sections, such as the central offset of
(Figs 2.65 and 3.24). This scheme (adjustable closed hydraulic ring) is frequently Fig. 2.21, whose lowered form can be seen in Fig. 3.3 between the two extraction
adopted on temporary piers, as the adjustment of the vertical force transferred by rails [38]. Additional information about the placement and design criteria of
the launching bearings permits a better control of transverse instability of the pier. transverse guides can be found in Section 2.5.
Electric (for remote control) or hydraulic (for local control) load cells can replace
the hydraulic jacks for a less expensive support arrangement; in this case, correction 3.5. I . Pier-cap arrangement
of support reactions takes place by using specific neo—flon pads with different The possibility of inserting hoisting jacks under the webs of the superstructure is
thickness or by shimming standard pads. Finally, superstructures with varying indispensable both during the launch, to replace damaged bearings or incorrectly
plan curvature or varying Width may require a transverse shift of bearings during inserted neo-flon pads, and on completion of launch, to remove the launching
launch (Fig. 2.2): in this case, launch bearings are placed in transverse sledges that bearings and to insert permanent bearings. In both cases, hoisting is limited to a
also contain the service hoisting jacks. few millimetres and the jack stroke requirement is therefore a few centimetres. In
The possibility of inserting neo—flon pads without lifting the superstructure the temporary piers, on the contrary, when launch is finished and the parabolic
makes launch continuous and the total speed is more than double. Many other prestressing tendons are tensioned, it is necessary to eliminate the residual vertical

jock

jock support Frame

Fig. 3.24. The launching nose is approaching a pair of hydraulic launch bearings
(auth,0r)__ _. .. . ., . . .Flg325i Adjustable transverse guide
 156 BRIDGE LAUNCHING PC BRIDGES 157

force transferred by the superstructure before beginning pier removal. In this case, bearings and to install permanent bearings. Althoughthis last operation occurs
the hoisting jacks must guarantee longer synchronized vertical strokes. when the deck is supported on permanent support sections and is often completely
As in friction launchers, hoisting jacks have ball—and-socket caps that permit the prestressed, other situations that require lifting may arise in any launch moment
distribution plates to adapt to the superstructure, which is often irregular or sloped. and in the presence of only the launch prestressing. Therefore, each cross—section of
Any use of wedge shims in these hoisting devices is to be avoided because of the risk the superstructure must be checked for the stresses produced by the expected
of plate expulsion. Jacks must be provided with safety ring nuts " that limit the maximum bearing misalignment and the predicted lifting.
hydraulic support of the superstructure to the phases of lifting and lowering. During calculation one might impose vertical displacements at each launching
It is expedient to place the hoisting jacks on low—friction supports (stainless steel bearing (a reasonable value is :l:5 mm alternated), or a vertical displacement of
sheets over neo-flon pads), in order not to concentrate on the jacks the horizontal :|:10 mm in only one bearing assuming all the others are perfectly aligned. Then,
force deriving from the elastic deflection accumulated by the pier because of the one might add a vertical displacement of $10 mm at the abutment when using
launch friction and the local longitudinal grade of the launch surface. As the steel— friction launchers, and a vertical displacement of 3:5 mm at the front head of the
Teflon contact so obtained generally has a lower friction coefficient than the neo— formwork because of the elastic settlement of this support in the final extraction
flon pads used for launching, often dirty, the pier tends to recover a portion of its phases of the segment.
elastic longitudinal deflection during hoisting. These movements must be accounted As launch prestressing is often designed for the tensile stresses at the mid—span
for when arranging the hoisting equipment on the pier cap, to avoid the risk of the lower edges, the moment variations caused by bearing settlement do not affect it
external jacks falling. much. At the same time, the variation of negative moment in a support section
In many cases, the use of hoisting batteries composed of many small jacks instead subject to a vertical movement, Wm
of a few strong units is expedient, as this facilitates manual jack shifting under the
E1
superstructure and results in a less restrained design of the hoisting scheme even in AM .——.. —K-L—-2—- Wm (3.13)
small pier caps. When pier caps are too small to allow placing jacks under the webs
of the superstructure, it is possible to use temporary cantilever supports anchored with K ranging from 3 to 6, is reduced in quadratic ratio by a generally long span.
to the pier by means of dry friction joints with distributed shear keys and through In both cases, therefore, support settlement is often tolerable without increasing
prestressing bars (Fig. 3.24). prestressing.
The situation is different in the transverse direction, as vertical misalignment of
3.5.2. Effects of vertical misalignment oflaunch bearings two launch bearings on a pier results in transverse eccentricity of the support
Vertical bearing misalignment generates secondary stresses in the continuous beam. reaction. ‘An eccentric load applied to an elastic beam produces torsion and, if
In the long term, differences in level of 10—20 mm are not very significant, as creep warping is restrained, secondary longitudinal stresses. If the cross-section of the
of concrete redistributes the stresses they produce. In the short term, however, these beam is not ri‘gi'd'in“ its plane, an eccentric load also causes distortion.
stresses cannot be neglected, and this is particularly true during launching. In a PC box girder bridge, eccentric live loads generate relatively modest effects,
Vertical misplacement of launch bearings may derive from local geometric imper— i since torsional stiffness is generally so high as to limit tangential stresses even under
fections in the support alignment that can be definitively corrected by shimming, or high torsional moments. The frame stiffness of' the cross-section, along with the
from general geometric imperfections in the superstructure that cannot be re- in-planeflexural stiffness of its plates, rapidly reduces distortion [69]. In most cases,
covered. It may also derive from unforeseen events, (for instance, expulsion of a therefore, important transverse effects can arise only inthe presence of eccentric
neo-flon pad) that can be avoided by using specific launch procedures and well- concentrated loads. _
trained crews. In all cases, the effects of vertical misalignment of launch bearings: The highest concentrated loads that affect the deck are the support reactions, and
can affect any cross-section of the superstructure and not only the final support the vertical misplacement of the bearings of a support section makes the two
sections, Which are generally stiffened and able to mobilize. the high. torsional reactions unbalanced, i.e. their resultant becomes eccentric.
stiffness of the box girder without significant distortional phenomena. This possibility is generally disregarded in design of non-launched bridges, since a
The effects of vertical misplacement of launch bearings are different in the long— bearing failure is improbable, casting of the deck in situ avoids any transverse
itudinal and the transverse direction. In the longitudinal direction, it is always eccentricity in the support reactions and bearing seats are generally cast in situ in
necessary to provide for vertical movements of bearings during bridge design. In the case of precast segmental construction. In addition, the transverse eccentricity
fact, the superstructure must be lifted at any support (while fixed at all the other of the support reaction affects a stiffened cross-section. This makes distortion
supports to avoid prohibitive costs) to remove damaged launching bearings and negligible, and the problem becomes one of torsional overload in a structural
_ ,neorflon .pads,__.On_eempletion 0f 1a,unCh,,,it must also be lifted to remove the launCh schemethat can ;._e_a.si_ly absorb it.
158 BRIDGE LAUNCHING PC BRIDGES 159

In launched bridges, diaphragms and stifienings are generally limited to the These two errors produce rather different effects, since the cross-section is normally
permanent support sections. Therefore, they cannot prevent distortion effectively, not prestressed in the transverse direction [71], and the distance between bearings is
since for this purpose their spacing would have to be reduced so far as to be much smaller; i.e. rotations are amplified. In the limit case of the Serio River Bridge
uneconomical [70]. (Fig. 3.39), which originated the study of the model proposed below, the 42.6-m
But differences in elevation between launching bearings occur in practically every span is about 40 times longer than the transverse distance between launching
case, because of their positional tolerances, their failure, or operator errors. In bearings, 1.1 m. «.
addition, the general geometrical irregularities caused by the formwork positional Large rotations and high torsional stiffness produce high torsional moments at
tolerances can produce the same efiects of bearing misplacements, making even supports. Because of the short transverse distance between launching bearings,
their most careful positioning useless. torsion causes large differences between the two support reactions. As a conse—
In the deck support section of Fig. 3.26, in the presence of a vertical misplace- quence, the overloaded bearing breaks (Fig. 3.27 and 3.28) and the cross—section
ment of the bearings placed under nodes 3 and 4, the distances of the deck support distorts.
points from the theoretical launch surface are W3 and W4, respectively. By imposing strict tolerances in the formwork and. bearing positioning, it is
The bearing misplacement may be divided in a symmetrical error Wm, whose possible to reduce these overloads. Nevertheless, in the most delicate cases jack
effects are only longitudinal and result from Equation (3.13) bearings able to adjust the torsional couple produced by the support may be
necessary at each pier, with high costs. In the presence of slender temporary
m
W_—-—— 2 r 3.14) piers, these problems are made more delicate by pier instability.
Finite element analysis does not allow the evaluation of torsion and distortion
and in a hemisymmetrical transverse error, Wh separately, and requires sophisticated, laborious models and long processing times.
The model developed in [72] ([73] in Italian) permits the analysis and optimization
_W4~—W3 of the interaction between torsion and distortion with a simple spreadsheet, and
W (3.15)
h 2 results in a simpler, faster design.
which locally rotates the bottom slab

¢=_ (3.16)

wallflnana
. vllnn
,
runninlnvavn‘rauwyirnmar
l’lil’ll(It’ll!)
”nonllnllull
‘uv‘uvnnuunarnmvnhrfl’ifi
unall
nnmaa”
m”
nnun (”A .0
(11:11:11, 7

Fig» 3+261. Unbalanced Supportreactions due to. misaligned ”launchingbearings , Fig, 3.27, Collapsed launching bearing of the Serio River Bridge (author)
160 BRIDGE LAUNCHING PC BRIDGES 161

Neglecting shear strain, the distortion nd(x) of the deck is governed by a

differential equation identical in form to that for beams on elastic foundations
(BEF) [74]: '

d477d
dx4 .
where Cd is a cross—sectional property related to warping, k is the deformational
stiffness of a unit length 0f the box cell, and ph is any hemisymmetrical unit load.
The property of the cross—section that corresponds to the BEF moment of inertia
Cd is defined by evaluating the share of hemisymmetrical load resisted by the
in-plane flexural stiffness of the plates. It can be expressed [72,73] as
[9
Cd = Z C8(C2C7 + C3C8 + 2C4C9) (3.19)

The coefficients of Equation (3.19) are summarized in Fig. 3.29 baSed on the cross-
sectional dimensions of Fig. 3.1. y
The vertical shear force V produced in the centre of the bottom slab by the cell
frame action for a unit hemisymmetrical load is
V __ cit/if + (bzc/tivxza + b)
Fig. 3.28. Broken launching bearing of the Palizzi Junction Bridge (author) (3.20)
— (a + era/o3 + (Ia/rot + (2%v + (119 + 292)]
and the cross-sectional property that corresponds to the BEF modulus of founda-
tion is
In a straight single-cell box girder with constant cross-section, the effects of a
difference in height between the two support points of the cross-section deviate k2 2(a+b)E (3 21)
from Bredt’s theory. In fact, none of its assumptions (rigid in-plane rotation, (l — v2){(a4/t§)[a/(a -|- b) — V] + (cbz/rf’w)[(2172/(a + [9) ——~ (2a —|— b) V]} ' .
torsional moment applied through the theoretical distribution of tangential stresses,
and free warping) is really respected.
Therefore, its deformation results from the superimposition of a rigid rotation 522:, + C(a + 2b)rw ’ F(a+ 2d)3 1)
——~—— C t —— _t
c
_ _ 1 __
"
and a distortion. The first effect depends on the torsional stiffness of the girder, (a+2d)3 C2 2 a 6612 1 t 4 b 2( Cifiw
=
1
which limits its capacity to rotate about its axis. Distortion depends on the inter- ft + C(2a ‘i— b)fw
a
action between the frame stiffness of the cross-section and the in-plane flexural
stiffness of its plates. [)2 b 6
C3 = "mi-51b C4 : _Cl:t +6(2 — C1)lw]
If the x-axis corresponds to the deck axis, nt(x) defines the vertical displacement
of the deck lower nodes due to torsion, and nd(x) the same displacement due, to _ (a+2d)3 b c a _-b c a
distortion. They are both hemisymmetrical, i.e. equal and opposite in Sign at the C5‘ a 6ab tt+Zlb+§(1+E)ZW C6—‘C Zrb+‘(2+“)t“’]
c, c, %_§
two nodes 3 and 4 of Fig. 3.26.
In the support section subjected to a misplacement of launching. bearings (the

a
Section 0 assumed as origin of the frame of reference), in order to maintain the

w
Q
x

:2

('3
I|
[\J

+
53

[\3
|
I

|
09
contact with bearings we must have

WhO = 77d(0) + 7M0) (3-17) h

C9=m— 1+(a—b) C a—
All the other support sections of the deck are assumed to be perfectly aligned. c Q—Q—

Fig. 3.29. Coefiicients for determination of Cd

 162 BRIDGE LAUNCHING PC BRIDGES 163

BEF is solved for the infinite beam scheme. In fact, the deck supports cannot R
restrain BEF, since, except when a diaphragm passes above them, the cross-section
2 gl3 iA
l
01 h iA is co
X . l l
——-.-

can distort. As diaphragms are effective only if they are so close as to be excessively 7////// ////////////7//// ////////////////T///////7///////// ///////////////// //////x

a:
.2

i
expensive [70], their presence is neglected from now on. R118 RhA Rh R
In the original section 0, the difference in height between launching bearings L 4, L+ L :lfiAL “8
unbalances the two support reactions R3 and R4 (Fig. 3.26). These may be divided
in a symmetrical component Fig. 3.30. Infinite BEF loaded by hemisymmetrical componentaof support reactions
R30 : (3.22)
2
R )t .
equal to the 50% of the continuous beam support reaction and which produces only nd(x) 2 2h]: e“}‘x(cos Ax + Sin Ax) (3.28)
longitudinal stresses (bending moment and shear force), and a hemisymmetrical
component and its first zero point is at an abscissa of 371/4A, over which deflection is generally
negligible.
R4 — R3
Rho = (3.23) In PC launched bridges distortion fades within distances of about half a span,
2 and the effects of the hemisymmetrical components of the support reactions do not
which is applied to the BEF to analyse distortion, and produces a torsional moment superimpose to each other. In each support section I of the deck, distortion is
T = Rhob caused only by the hemisymmetrical component R111 of its own support reaction.
(3-24) Therefore, from Equation (3.28) we obtain
whose effects in the deck may be studied with the theory of torsion.
tA
If the deck cross-sections were rigid, the torsional moment due to this couple 77d1 — “um—2k (3 . 2 9 )
would be limited to the two spans supported at section 0. In fact, in the case of
constant spans, two balancing couples equal to The transverse rotation of the bottom slab is obtained from Equation (3.16):
T _ RhIA
TA = :lz—Z— (3.25) (3.30)
would arise at the adjacent support sections A (whose bearings are assumed to be and, if bearings are aligned, an equal and opposite rigid rotation of the Whole cross-
aligned). section must occur to maintain the contact with launching bearings:
But these balancing torsional moments are also produced by hemisymmetrical m=+n., m--»m on)
components RhA of the support reaction
For congruence, the difference between the rigid rotations 29A and 2913 of the support
TA = RM!) (3.26) sections A and B is equal to the torsional rotation of the span A—B under the
and therefore these support sections also distort. Rigid rotations occur in sections A torsional moment
to compensate distortion and maintain contact with launching bearings, and a TB 2 Rt (3.32)
portion of torsion migrates towards external spans.
From a theoretical point of view, this migration affects the whole deck. In reality, Imposing this condition results in
high values of Cd and k damp distortion rapidly, and the phenomenon is limited to 2/1216 + (b/Ct)L
: R 3.33
the first two spans on each side of the original cross-section 0. Therefore, BEF is M 42t/bk + (zb/C,)L ho ( )
loaded as in Fig. 3.30. ii )t/bk
Let RhB = Ail/bk + (212/ Ct)L Rho . (3.34)

A_4 k 327 where the torsional stiffness Ct, neglecting restrained warping, is defined by the
”" 4ECd (') second Bredt’s equation (m = 1/1) where v is Poisson’s ratio)
The equation of the right branch of the BEF elastic line produced by a concentrated _ m/[2(m + l)]E[(a + b)h]2
(3.35)
, 193d Rho appliedto its origin 1,5,[75] " 6.1/2; + b/rb_+ Ze/rw
t
 164 BRIDGE LAUNCHING PC BRIDGES 165

Once RM and RhB have been calculated, the distortion is (x 2 0) a couple of unit forces, are obtained by linear proportion with the actual

ndrx) = ndorx) __ ndAtx) — ndgtx) +3—tB

hemisymmetrical component.
(3.36) The reliability of this model has been tested in [72,73] by comparing results with
the finite-element analysis of four continuous beams sized according to the statis-
where the different components (l is the absolute value of x) are
tical criteria of Section 3.1. With web inclinations up to 200 (ie. statistically, in
A . more than 70% of launched bridges) the model closely approximates the results of
77d0(x) = 5%e_m(cos Ax + Sin Ax) (3.37)
finite-element analysis, with an average error of 2.3%. Fmite—element analysis also
A . ' confirmed the damping distance of Wd(x).
ndA(x) : [gig—e—ML‘WCOSML — x] + smAlL — xi) (3.38) Higher slopes produce a progressive deterioration of the results, since restrained
warping generates shear gradients that reduce distortion.
R . The analysis showed that only about half of the launch surface local deformation
ndB(x) = file—“urkcosrlu —— xi + S1n2t|2L .— xl) (3.39) was absorbed by torsion, and therefore could mobilize the high torsional stiffness of
box girder. The remaining share was absorbed by distortion, and produce
and the last addend corrects the frame of reference to take into account the rigid
important longitudinal and transverse stresses. '
rotation in B.
A 10-min vertical misplacement of launching bearings in the transverse direction
Finally, the nodal displacement m(x) due to torsional rotations is (in the two
modified the support reactions calculated with aligned bearings with hemisym—
spam) ' metrical components by about 40%. A difference of 25 mm (easily caused, for

m0""4 (nu—Ci;
—b21L —x) R1.0+ RL
his] c.)
340
instance, by the expulsion of a neo-fion pad) decompressed one launching bearing
and doubled the load on the other one.
These results further confirm the load differences measured in the field.
mA ‘Bm = “fall—ma
192 ~
(3.41) Monitoring of the Amiens Viaduct [76] showed support reaction perturbations of
1; about 20%, and monitoring of the Champigny—sur-Yonne Bridge showed that only
one of the two abutment bearings was actually loaded.
With nd(0) and 77(0), and therefore Who (3.17), defined, for any couple of hemi- These overloads are significant in most launched bridges, and are particularly
symmetrical forces Rho the calculation of the actual hemisymmetrical component of delicate when launching occurs with temporary piers. In this case, in fact, the deck
the support reaction based on the assumed misplacement of launching bearings is is designed for much longer final spans and is therefore particularly stiff on the
immediate, by linear proportion. shorter launch spans. This may cause high transverse stresses both in the deck and
Clearly, in the absence of a bilateral constraint this force cannot exceed the in the temporary piers. ' ’’'
reaction R30 produced by each launching bearing in balanced conditions. When
the 50% of the continuous beam support reaction is reached, only one bearing 3.5.3. Local stresses above launching bearings
carries the whole reaction, and the deck rises from the second one, avoiding further The deck zones that pass above launching bearings are subjected to complex stress
deformations. states. The beam behaviour of the superstructure, the launch prestressing, the
Setting average positional error (3.14) of bearings, the thermal gradients, and the creep
deformations progressively accumulated by the superstructure, generate normal
3k and tangential stresses in the longitudinal plane.
EC l‘WCQ In the presence of transverse differences in the level of launch bearings, torsion
and distortion generate normal and tangential stresses in the longitudinal and
the longitudinal stresses caused by distortion at the web corners are
transverse direction. Normal and tangential transverse stresses are generated by
_ 3Rh0
03 (3.43) thermal gradients in the cross-sectional plates, and vertical compressive stresses
— (1 + C1))l‘wc2 are produced by dispersion of the support reactions within the webs. Also the
0'2 2 C103
sliding friction and the local gradient of the launch surface generate longitudinal
(3.44)
stresses, and since these are local effects, they compromise the conservation of plane
and these stresses superimpose on the longitudinal stresses produced by restrained sections. Since most of these stresses are significant, the local state of stress has to be
, ,Warping:Finellra-th¢transversr frame, stresses, determined from Equation (3 .20) for, checked with adequate.__strength. criteria.
166 BRIDGE LAUNCHING PC BRIDGES 167

3.5.3. I . Vertical compressive stress

In the absence of temporary piers, the support reactions transferred by launching
bearings are not much smaller than the loads transferred by the permanent bearings
on completion of launch. But whereas the permanent support reactions are applied
to cross-sections that are almost always strengthened and stiffened by diaphragms,
the launch reactions are applied to cross-sections devoid of any stiffening. It is
therefore necessary to check their correct distribution in the longitudinal, vertical
Fig. 3.32. Sofit arrangement for the inclined bottom slab of Megalorema Bridge
and transverse directions.
In the transverse direction, the support reaction should be transferred directly to
the web axis, without transverse eccentricity that would generate a local couple.
Rmax = R30 + Rho (3.45)
Avoidance of transverse bending due to eccentric support reactions is the more
important the more slender are the webs and the bottom Slab. Webs 0.50 m thick The maximum compressive stress in the multilayered elastomeric bearings of neo-
hardly suffer from transverse bending, but 0.30-m webs can require lateral widening flon pads should be limited to about 12 MPa, since it is impossible to completely
at the base to increase the available contact surface for neo—flon pads. When the avoid local defects that increase actual compression.
webs are inclined, external vertical edges may be adopted (Fig. 3.31). R X .
It is also advisable to design the zones of the bottom slab soflit that come
in contact with the neo—flon pads as horizontal crosswise. For inclined sections,
As an alternative, the contact surface may be designed for the highest balanced
the soffit scheme of Fig. 3.32 may be adopted to avoid creating transverse drift
support reaction R30 (3.22) and a reduced centact stress of about 8 MPa.
components in the support reactions.
The contact surface of launch bearings (Fig. 3.23) is dimensioned for the highest UPTFE
. _ R30 <
baLnf ____ 8MPa (3.47)
support reaction Rmax, which is determined by adding the maximum expected effect
of bearing misalignment (3.23) to the maximum support reaction resulting from The transverse width bnf of neo—flon pads is generally 30—50 cm. Launch pads are
continuous beam analysis (3.22) often designed l—2 cm narrower than the neo—flon plates used on the form extrac-
tion rails. This permits placing the main launch pads in the prints of the extraction
rails, in contact with a surface of the superstructure as regular as possible. This
solution is particularly advantageous in the case of board—marked finished surfaces.
The external edge of neo-flon pads should be 7—10 cm distant from the external
edge of the deck to avoid breaking and spalling the concrete cOver in the corner.
This sometimes requires the local widening shown in Fig. 3.31. AASHTO [42] also
requires that the post-tensiomng ducts for parabolic prestressing have a concrete
cover of at least 15 cm to the underside, as shown in Fig. 3.33. Similar restrictions
should also apply at the upper edges of the cross-seCtion‘at the support points of the
tower for the temporary front-stayed system.
The total length ELnf of neo-flon pads has a static relevance, as the flexural
rotations of the continuous beam support sections during launching compromise
the uniform distribution of the contact stresses. Since these load anomalies depend
on ELM, the length of launching bearings must be reduced as much as possible. In
the case of long launch bearings, the use of thicker elastomeric bearings in the
neo-flon pads improves load distribution and avoids excessive wearing.
AASHTO [42] requires that the neo-flon pads be able to compensate for local
deviations of the sliding surface of up to 2 mm by elastic deformation. When the
overall thickness of neo—flon pads is insufficient for this purpose, the launch saddles
can in turn be placed on multilayered elastomeric bearings, as shown in Fig. 3.23.
Fig. 3-31. Lowerflede schemes- ,, . _This....scheme..may.. so.lve,..emergency situations due to excessive geometrical tolerances
168 BRIDGE LAUNCHING PC BRIDGES 169

and the vertical component of the contact force (i.e. the support reaction of the
{xx/11y
I / / /
empty ducts continuous beam) is not orthogonal to the gravity axis of the superstructure. In
,,,,,,,,,,, neo-f-‘lon pad turn, as discussed in Section 2.8, sliding friction rotates the resultant and produces
/
additional longitudinal components.
These longitudinal forces are easy to determine at each launching bearing when
x], /
_ / / I
l ,/ ,’//:/ ,/ / l I, / / /’/’,/ l I //,///
mm 15cm , , ,x /
/ 1],] / //,l / //,”/ I
the local gradient of the sliding surface and the distribution of the continuous beam

\
\
I I / j I I f /
_

support reactions are known. The local longitudinal compressive stresses generated
by these forces are high before dispersion within the cross-section, and they are
superimposed on the longitudinal compressive stresses generated by the negative
bending moment of the continuous beam and by launch prestressing.
7 ‘to 10cm
Near the lower edge of the support sections, the diagram of the longitudinal
compressive stress differs from the theoretical Navier’s one and reaches much
Fig. 3.33. Transverse clearance of nee-flan pads higher values. Both deck monitoring and finite—element analysis confirm this phe—
nomenon, which compromises the conservation of plane sections when passing on
in the deck, but requires experience to avoid movements in the launch support launching bearings. These perturbations of the stress diagram dissipate within
systems. short distances, comparable with the deck depth. Also the tangential stress
The length Lnf of the individual neo-flon pads has no static influence and is distribution cannot be evaluated theoretically. Finite-element analysis confirms
chosen according to preference. Short pads are light and easy to handle but require high perturbations, which dissipate within half the deck depth.
frequent insertions, and therefore a higher level of attention from the workers. These aspects become even more significant when using friction launchers, since
Longer pads facilitate accurate positioning (especially when defects in the deck the whole launch force they transfer to the deck is much higher than the launch
surface tend to rotate them) as well as the introduction of shim plates. resistance at the pier supports. Therefore, when designing the friction thrust units,
In the transverse direction, launching bearings are at least 4 cm wider than the hoisting jacks must always be equipped with ball-and—socket caps to improve the
neo-fion pads in order not to damage Teflon in a case of eccentric insertion. In the adhesion of the contact plates to the deck soffit, and the contact surface must be as
longitudinal direction, they have two rounded ends, about 15 cm long, which large as possible to decrease the average contact pressure.
facilitate the insertion and the expulsion of neo-flon pads. Therefore, their total These measures are indispensable when using friction launchers placed on the
length (in metres) is piers, as in Fig. 2.68. In fact, in the case of abutment launchers, the distance
between the launchers and the adjacent supports of the continuous beam is much
L, a 0.30 + 2L,f (3.48)
shorter than the diStance between 'tiie”pierS’, "and the thrust force therefore super—
When launching bearings are placed on the pier cap, adequate clearance for the imposes on smaller support reactions and negative moments.
insertion of hoisting jacks must be maintained both behind and in front of bearings,
since jacks must in turn be aligned under the webs of the superstructure (Fig. 3.28). 3.5.3.3. Stress verification
Long neo-flon pads often cause hoisting difficulties, as they remain trapped between When passing over launching bearings, the lower edge of the deck is subjected to a
the jacks and the superstructure. Shorter pads may be advantageous for this triaxial state of stress.
purpose. With accurate placement of launching bearings and neo—flon pads, in the presence
Finite element analysis confirms the 450 dispersion of the support reaction of low thermal gradients, the transverse stresses in the lower cross—sectional nodes
inside the webs. This implies sharp variations in the vertical compressive stresses are smaller than the stresses in the other two main directions. Therefore, the state of
close to the launching bearings, which must be taken into account in the local stress can be assumed as plane and the principal stresses may be derived from
stress verification. In stress analysis, the presence of non-grouted ducts for Mohr’s circle:
parabolic tendons and any eccentric location of the support reactions should be
'
taken into account. ol}_ax+oz
2 :l: 51 \/(0X — 0,) 2 +4TX,2 (3.49)
02
3.5.3.2. Longitudinal compressive stress
The force transferred to the superstructure by neo-flon pads is not orthogonal to where (TX is the longitudinal normal stress, 02 is the vertical normal stress, and s IS
the contact..surface..-.ln . fact, the. deck. is generally launchedalong an inclined plane, .thetangentialstress..in-..the...plane..- .
 l70 BRIDGE LAUNCHING PC BRIDGES l7l

The principal stresses must be compared to the critical stresses derived from a effects on bridge launching are concerned. Much more detailed information can be
concrete strength criterion based on biaxial compression. For instance, according to found in several international design codes such as MC90 [64].
MC90 [64], by assuming that the biaxial compressive strength is 20% higher than
the uniaxial strength, and setting 3.6. I. Shrinkage
a Shrinkage is a slow, load-independent concrete deformation due to chemical and
a = -—l (3.50) physical reactions related to the cement setting. Its course is rather long, though
02
about one—third of the total strain occurs in the first 15 curing days and in the first
where 01 is the largest principal stress at failure and 02 is the smallest, the strength year it reaches 80%. '
of concrete under biaxial compression may be estimated from the following Shrinkage depends on the mechanical properties of concrete, on the quantity and
equation: quality of cement and admixtures, on the water-to—cement ratio, on the aggregate-
to-cement ratio, on the aggregate rigidity, and on the humidity and temperature of
a2 = —————'—,f.... (3.50 the curing environment [77].
Typical shrinkage values of air—cured concrete at 50% humidity and 210C are
The 28—day mean value of compressive strength fem may be related to the 28—day reported in Fig. 3.34. An increase in the water-to-cement ratio increases shrinkage
characteristic compressive strength fck (in MPa and for cylindrical specimens) with because of the higher quantity of water that can evaporate and of the higher
the following equation [64]: porosity of concrete, which facilitates its evaporation. A reduction in the
aggregate—to-cement ratio increases shrinkage because of the higher volume
fem ck + 8 (3.52)
of cement, which is responsible for this phenomenon, and of the lower volume of
Assuming load lines with the same shape as the ultimate one and, for instance, an aggregates, which oppose the contraction of volume.
elastic limit of 0.6k in axial compression, it is possible to determine the biaxial At any instant I, the shrinkage strain so, measured from a reference instant to can
elastic dominion of concrete in the 01—02 plane. Then, the state of stress in the cross— be expressed by
sectional lower node is easy to check by drawing in this diagram the points
Sosa: r0) : 8cs.ooK1(ta 1‘0)
corresponding to the pairs of principal stresses (3.49). (3-53)

These states of stress, applied many times and for long periods during launching, where the final value 863m and the function K10, to) (which tends to a unit value
might cause irreversible local strains due to creep of concrete. In most cases, for t—> 00, more rapidly for thinner elements) are established by the applicable
however, the curing period of concrete is so long as to limit this risk. standards.
Shrinkage is higher on surfaces exposed to the air and lower inside concrete,
3.6. Effects of time-dependent behaviour of concrete because of both the smaller variations of humidity and the presence of reinforce-
In the design of conventional reinforced concrete structures, concrete is considered ment, which opposes these deformations. Consequently, the conservation of plane
linearly elastic, although in reality it has an inelastic and time-dependent behaviour. sections requires the presence of strain gradients, i.e. of self-balanced stresses.
This model permits adoption of simple design methods but leads to underestima— Although creep rapidly reduces these stresses, in the first curing phases they can
tion or neglect of phenomena that may produce serviceability limit states and rapid exceed the tensile strength ofconcrete and generate distributed superficial cracking,
structure deterioration, and therefore to an incorrect evaluation of the state of which in turn causes deterioration. Similar stresses arise in the Corners of the
stress and of the real safety factors of the structure.
Time-dependent phenomena became evident with the difiusion of PC bridges
W/C
built by balanced cantilever construction with mid-span hinges. In fact, this con— 0.4 0.5 0.6 0.7
struction method requires application of 'prestressing with high eccentricity and
after a short curing period, and mid-span hinges do not allow redistributionrof A/C
3 800 1200
internal stresses. This results in large time-dependent deflections of the opposite 4 550 850 1050
cantilevers. 5 400 600 750 850
6 300 400 550 650
Although it uses the same materials, the incremental launching construction 7 200 300 400 500
method generates much smaller time—dependent effects. These effects are analysed
by superimposing the effects of shrinkage and creep of concrete on those of the Fig. 3.34. Typical Shrinkage values (um/m at Six months) of air-cured concrete as
» ~ ~ relaxation -of—-~pres—tressing steel. These effects {are-idiscusseduhere only as far astheir . . . .. . ..,funcri0nsu 0f the W/C-‘cmd A/C ratios
172 BRIDGE LAUNCHING PC BRIDGES 173

cross-section and, in the case of webs and slabs of very different thickness, they are .1 specimen is subjected to a constant compression 00 from an instant to measured
made worse by differential shrinkage. from its casting, its shrinkage—free deformation at at any instant I is
With the incremental launching construction method, the deck is cast on the 80‘» to) = Seiflo) + 8def(t’ to) + Eccfra Z0) (3-54)
ground and most of the parameters that govern concrete shrinkage can be easily
controlled. The water-to-cement ratio may be reduced by using admixtures, by The instantaneous elastic deformation 8e1 of concrete, fully recovered when
mixing concrete in situ and by placing it with buckets or conveyor—belts instead relieving the specimen, is
of by pumping. 00
86100) = (3.55)
Further local reductions of the water—to-cement ratio may be obtained in the E00)
upper region of the top slab by means of roll-compactors (Fig. 3.35) that expel the The deferred elastic deformation edef, alsO reversible, is about 40% of the instanta—
bleeding water mechanically. This increases the compactness and the tensile neous elastic deformation at 28—day curing, reduced by a coefficient K20, to) that
strength of the concrete cover to top reinforcement and reduces permeability, depends on the load duration, varies between 0.75 and 1.00, and can be assumed
with enhanced protection. Roll compacting also reduces the number of workers equal to 1.00 for periods longer than 3 months: '
required for concrete finishing, and this cost saving soon covers the investment for
00
equipment. t,t ”A: 0.40—K 1,: 3.56
If necessary, forms equipped with draining felts can be used for the whole
external surface of the deck to reduce concrete porosity, to increase the superficial Finally, the irreversible deformation due to creep, sec, can be expressed in the
strength, to protect the surface during the first curing phases and to avoid any use following form:
of oil. 00
z,z =—-—KK t,z‘ 3.57
3.6.2. Creep .
This last strain component is the higher the younger concrete is when loaded. It
Shrinkage is a load-independent phenomenon, whereas creep is the spontaneous
depends on the hygrometric conditions, the thickness of the element, and the qual—
growth over time of the deformation produced by a constant load. If a concrete
ity and compactness of concrete, since the cause of creep is grains sliding caused by
the Viscous behaviour of the cement paste [77]. The coefiicients K3 and K4 are
defined by the applicable standards. On the whole, most of bridge, design codes
lead to expressions for Equation (3.54) like '

80, to) .= .——..—-~. f7° a . (3.58)

till, 1‘0) ,
and the analysis of the total deformation becomes an elastic calculation with a
conventional modulus E*(t, r0). '

3.6.3. Prestressing steel relaxation ,

The time—dependent phenomena in PC are complicated by viscous phenomena
in prestressing Steel known as steel relaxation. Relaxation occurs in a tensioned
tendon of constant length, and therefore it causes loss of prestress.
As steel relaxation develops over a short period of time, tendon retensioning a
few days after the initial tensioning has often been adopted in the past. However,
these operations are expensive, and impossible with tendons coupled or anchored in
construction joints. With the commercial availability of low—relaxation strand,
tendon grouting after the initial tensioning has become a standard practice.
The evolution of steel relaxation can be expressed in the following form:
Aop(t)
Fig. 3.35. Roll—compacting unit movable on longitudinal rails along the whole og—-— 2 K5 + K6 log(z‘) ' (3.59)
- . .. _, . , .0110. _. .
casting‘cell-—---(-au-th0rj------——---—--—--- , ~ ,
 BRIDGE LAUNCHING PC BRIDGES 175
174

where A0p(t) is the loss of prestress at time t, opo is the initial tension in the tendon,
and the coefiicients K5 and K6 depend on the steel properties and are defined
dead load and
experimentally. equivalent system
In practice, the length of the tendons is not constant over time because of the --.-=-.*

time—dependent shortening of concrete. Consequently, the loss of prestress due to

steel relaxation is smaller than the theoretical value (3.59). Some design standards
I |
I |

provide for reduction coefiicients to account for this effect. .. . \ ! mL/cr‘eep def-‘lectlon
,

-
Ii I|
3.6.4. Effects on bridge launching
I I
In a continuous beam supported on sliding bearings, the time-dependent deforma-
tions of concrete cause a general shortening of the beam. Neglecting bearing friction
and losses of prestress, this uniform shortening does not affect the state of stress in
the beam either during launching or in the service stages.
Since the launch tendons generally extend for two or three deck segments, launch
prestressing is introduced gradually in the deck. In most cases, the parabolic 7

tendons of service prestressing are tensioned on completion of launch, i.e. after a

I .....
long curing period. Consequently, the time—dependent component of the longitu—
dinal shortening of the deck is much smaller than in bridges entirely prestressed
after a short curing, and the related prestress losses are smaller as well.
[I creep deflection

In long launched bridges, the front segments reach their final position after many I I
months of curing, while the rear segments are much younger. Both creep and
shrinkage vary along the deck, and the segment length must be corrected over Fig. 3.36. Creep deflections in bridges built by incremental launching (below) and with
time to avoid misalignment of the support sections of the deck with respect to other construction methods (above)
the piers.
When the deck has reached its final position, concrete creep does not modify the
stress distribution. In fact, it produces effects similar to variations in the elastic
load of full prestressing in bridges cast in situ span by span. On the other hand,
modulus (3.58), and if these variations are similar along the deck (the secondary
however, the launch prestressing cannot control the fiexural rotations produced by
effects of the different curing periods of the segments are generally negligible),
dead load bending, and a part of these rotations is made irreversible by creep until
stresses are unaffected.
the parabolic prestressing (introduced only on completion of launch) definitively
However, the support scheme of the continuous beam changes continuously
interrupts the phenomenon.
during launching. Therefore, the effects of creep are not limited to service
In most cases, however, the length of the launch step is half the span. In this case,
stages but also affectthe construction stages, since the deck assumes permanent
the diagrams of flexural rotations superimpose only every two launch steps, and
deformations during each stop between two subsequent launch steps.
during each stop most of the deck is subjected to rotations opposite in sign to those
When the launch step is as long as the bridge span (segments as long as a span,
at the previous stop. Therefore, permanent rotations tend to cancel each other, their
several short segments launched as a whole, etc.), during each launch stop necessary
envelope becomes more chaotic, and the cumulative effect reduces significantly.
for construction of a new segment, the deck is subjected to the same stresses of the
As launch goes on, these permanent deformations disturb the launch stresses.
previous stop. AlthOugh shifted one span forward, the cross—sections of the deck are"
This phenomenon is negligible only in the shortest bridges (whose construction
subjected to the same flexural rotations. As time goes by, concrete creep makes a
duration is too short to cause significant creep effects) with segments of different
part of these rotations irreversible, and the deck undergoes a permanent deforma-
length than the span. In all the other cases, the strain accumulation can be
tion similar to that of a continuous beam cast directly in its final position but
evaluated by assuming that any stress increment A0 generates an irreversible
opposite in sign, since launch prestressing does not produce an equivalent vertical
deformation Aecc that does not depend on the stress increments applied previously.
load.
Additional information about this assumption of ageing linear viscoelastic material
The situation is illustrated in Fig. 3.36. On the one hand, the axial launch
can be found in MC90 [64].___ , _ _ _, __
prestressing-w avoids ~-~-the upwards cambers produced by- the excess of equivalent ,
-
 176 BRIDGE LAUNCHING PC BRIDGES , 177

The analysis is repeated at each launch stop. The permanent deflections devel-
20 A
oped during the stop are added to those accumulated during the previous stops to
obtain the new deformed profile of the continuous beam. The movement of this

lower edge stresses (MPa)

undulated beam on rigid supports permits the calculation of the additional launch
stresses. The calculation is simpler if one assumes the continuous beam as un—
deformed and applies to its support sections the vertical displacements that
correspond to the beam deformation at those points.

O
_|.
On completion of launch, the final permanent deformation is determined by
superimposing the deformation due to parabolic prestressing on the end—of—launch
deflections. This correction is generally modest because of the long curing of
concrete.
The analysis of the secondary launch stresses due to concrete creep requires
o/\ /
laborious calculations. During the preliminary design stages, these analyses may front spans
be avoided by keeping a margin of about 0.5 MPa from the allowable tensile stress
in the worst load condition. CalCulations must be refined during construction with Fig. 3 .38. Lower edge stresses in the three front spans of the Amiens Viaduct
experimental data on the evolution of strength, elastic modulus, and shrinkage
strain of concrete. phenomena. These aspects are even more significant in ribbed-slab bridges, as the
Figures 3.37 and 3.38 show the extreme fibre stresses in the three front spans of higher position of the gravity axis increases the differences between edge stresses.
the Amiens Viaduct during launching. They have been calculated by means of two Much larger irreversible deflections may occur in the case of stoppage of work.
computer programs, one of which takesinto account creep of concrete and the The creep deflections produced by a 21-month stop in the construction of the Serio
instant prestress (thicker line) instead of the loss-free final prestress (thinner line). River Bridge [38,78] required corrective devices such as shimming of the nose
The distance between these curves confirms the importance of the phenomenon. In bottom flanges (Fig. 3.39) and integrative local external prestressing to complete
fact, although concrete creep reaches 85*90% of its final value in 10 years, 30% is
launch safely.
developed in the first month, 50% in 3 months and 75—80% in 1 year. Although
these data are very approximate, they show that the construction duration of a
medium-length launched bridge is such as to mobilize most of the time—dependent

20
upper edge stresses (MPa)

10

3% front spans
Fig. 3.39. Shimming of the nose bottom flanges to compensate creep deflections in the
. ,Fig. 3.3,7....Upper.-.edge.stresses.in the three front spans of the Amiens -Vl'aduct,. . _._..._._.,frontspan(author) _ _ . . .. . _
178 BRIDGE LAUNCHING PC BRIDGES 179

3.7. Prestressing because of the cost of its removal. In fact, it increases the safety factor from
During incremental launching of a short-span PC bridge, the edge tensile stresses decompression in case of unforeseen events, such as support settlement and seismic
due to dead load bending moments can be resisted by reinforcement. In most cases, phenomena. It also improves deck behaviour near the inversion points of
however, it is better to reduce or cancel these stresses by means of an appropriate the bending moment diagram, covering the wide movements of the null points
launch prestressing scheme. produced by the presence or the absence of live loads and thermal gradients.
The cyclic changes in sign of both bending moment and shear force do not allow As this scheme tends to cause higher consumption of prestressing [43,45,46], it
use of the parabolic final tendons for this purpose, since their action would be has been constantly improved to make the centroidal component of the final
inappropriate in temporary support configurations that differ from the final one. prestressing even more effective and less expensive, and several different schemes
Therefore, launch prestressing cannot but be centroidal, and its contribution is have been tested. '
limited to a uniform compression that reduces tensile stresses in the edges decom-
pressed by bending moment (partial prestressing), or that guarantees adequate
margins from their decompression (total prestressing).
The obvious idea might be pretensioning, simple to obtain with straight layouts. 3.7. I. Launch prestressing
This avoids the cost of anchorages, couplers, ducts and grouting, and guarantees Bending stresses in the deck segment supported on the formwork are very low.
better compaction of the concrete and better strand corrosion protection, compared Prestressing introduced before launch is mainly aimed at avoiding shrinkage cracks,
with strands confinement in grouted ducts. absorbing the low tensile stresses due to segment extraction from the casting cell,
In reality, however, even allowing edge tensile stresses (partial prestressing), the and preparing the segment to be launched over the temporary curing supports
average compression to apply to the cross-section is rather high, about 5—7 MPa, placed between the casting cell and the launch abutment, and then over the first
and this requires prestressing forces of tens of meganewtons. The application of span.
such pretensioning would result in very expensive tensioning devices and locking The advances in incremental launching saw the creation and the optimization of
systems for the superstructure, which is supported on low friction bearings and many different schemes for launch prestressing:
movable with much lower forces. Therefore, launch prestressing is always obtained 0 Permanent internal straight tendons joined to each other by couplers in the
with post—tensioned tendons, arranged either internally or externally to the concrete construction joints between the subsequent deck segments.
cross-section. 0 Temporary or permanent internal straight tendons anchored in accessible blis—
This unifOrm compression is particularly ineffective on completion of launch, ters and joined to each other by longitudinal overlapping of their anchor zones.
since it overloads the deck edges already compressed by the bending moment 0 Temporary or permanent external straight tendons together with one of the
diagram. Moreover, the straight layout of tendons cannot reduce the shear force previous types or alone. p , 7 7
near the supports. 0 Temporary or permanent parabolic tendons, either internal or external, along
As the cost of launch prestressing increases with the span, for the longest spans with temporary opposing tendons that make the eccentricity of the resulting
incremental launching, Scan compete with the balanced cantilever construction force constant, and straight external tendons that raise the prestressing force
method only by reducing the influence of temporary stresses; i.e. introducing towards the gravity axis of the superstructure, reducing the overall eccentricity.
temporary piers that halve the average span of the launch static scheme. In all
the other cases, it is better to regularize launching stresses by means of a launching In all these cases, tensioning the launch tendons directly to their final stresses
nose or a stayed front system and to optimize the use of axial prestressing. requires that the fresh concrete behind their anchorages does not collapse.
As an adequate amount of prestressing is available from the beginning of launch, Several solutions are available for this purpose. ,
the first approach is to shift tendons on completion of launch from the centroidal Precast concrete blocks containing tendon anchorages may be adopted
position to, an eccentric one, so as to. use the same strands both for the launch (Fig. 3.40). Since defects or voids behind these blocks can break them during
prestressing and for the final one. The first launched bridge [2] was actually pre- tendon tensioning [l l], the inclination of the joint surfaces and the joint reinforce-
stressed in this way, but savings in material were cancelled by the cost of labour ment must be studied so as to facilitate vibration of concrete of the superstructure.
necessary to shift the tensioned tendons, and this prestressing scheme was never As an alternative, concrete around the anchorages can be locally heated by means
adopted again. of thermal insulation, warm air or immersed electric resistances. Local heating
More specialized schemes were studied [71], composed of permanent launch permits a better continuity of reinforcement than does the use of precast anchor
tendons integrated on completion of launch with parabolic cables. Thus, launch blocks. Finally, higher-strength concrete can be used in the zones around the
prestressingbecamea permanent componentofthe deckprestressmg;and notOIfly , , , _._g anchorages
180 BRIDGE LAUNCHING PC BRIDGES 181

by its own launch prestressing. Tensioning the newly added tendons relieves the
pressure transferred by the anchorages in the previous segment, and the recovery of
the elastic strain of the concrete compresses the joint locally and, for equilibrium,
decompresses the zones further from the anchorages, resulting in tensile cracks.
Therefore, unless launch prestressing is higher than about 4 MPa, the zones of
the joint section distant from the couplers should have a specific reinforcement
[4,79]. Besides, it is expedient to use many tendons to guarantee a good distribution
of the prestressing force in the cross-section, and this complicates the joint geometry
even more.
Because of these problems, the evolution of this scheme of prestressing was fast,
and followed two different directions. In one scheme, tendons are coupled in each
joint but tensioned alternately every two joints, resulting in a more progressive
introduction of prestressing (Fig. 3.41). However, this coupling scheme is expensive
and does not comply with several design standards (e.g. AASHTO [42]) which
require that not more than 50% 0f the longitudinal post-tensioning tendons be
coupled at one seetion.
In the second scheme, couplers are staggered so as to prestress two subsequent
segments with each tendon and halve the number of couplers (Fig. 3.42). When
the casting yard is of sufficient length, several small tendons canbe used, coupled
Fig. 3.40. Precast anchor blocks in the preassembled cage (author)
every three joints to further save couplers and to make introduction of launch
prestressing more progressive. With this scheme, if the strands are inserted in
empty ducts after casting, recesses for the couplers are necessary to anchor the
3.7. I. I. Coupled straight tendons strands. On the other hand, prefabricated tendons emerging from the rear joint
After the initial experiences with shifting tendons, launch prestressing was obtained
complicate the cage assembly in the formwork and make its full preassembly
by straight tendons as long as the deck segments, joined by couplers in each con-
impossible (Fig. 3.43).
struction joint of the superstructure. In effect, coupled joints offer some advantages:
low secondary stresses, low prestress losses due to friction, simple formwork, and
reinforcement. But they have many limitations as well:
launching ollrectlon —--
0 A high prestressing cost because of the short average tendon length, the
additional cost of couplers and the high requirement for skilled labour. ,1
0 The impossibility of removing a portion of prestressing on completion of

I“ if
launch.
_........_t I
o The frequent need to‘increase the slab thickness near the joint to contain the
couplers, and the tendency to extend this thickening to the whole deck, resulting
__'__.._/
in a greater dead load. _—T—-q——i
0 The need to check the joint section accurately, since couplers reduce the moment
f * '* /
of inertia considerably. ‘ ‘
o The sharp introduction of the whole launch prestressing in fresh concrete, even
segments
+—-—\+——4«——+——+
l+8 l+2 [+1 I

though it is centroidal and therefore unable to produce irreversible rotations.

—q— ordinary coupler placed
on a tensioned tendon
This last aspect has often been the cause of distributed cracks in the joint section
because of the elastic strains of concrete near the anchorages. The force transferred -E1— coupler placed on a partially
tensloned tendon
by each anchorage deforms the concrete locally, and with it the construction joint
against which the concrete of the next segment is match-east and then compressed . , , Fig 3.41... .Tendons. coupled at...each.jo.z'nt.but tensioned alternately every two joints
-
-
 182 BRIDGE LAUNCHING PC BRIDGES 183

launching direction —a- o The handling of short duct segments instead of long and heavy prefabricated
tendons speeds up the cage assembly, saves labour and guarantees a better final
I
IL
i ' ’ F, result.
0 Concrete curing time can be used for fabricating the tendons. This reduces the
criticality of the steel cage assembly when it occurs directly in the formwork,
,1—+ ’ 44—!
i d, I

J...
and permits a better rotation of labour.
I + l 0 Because of the absence of tendons emerging from the rear joint, integral cage
i ' I
1’
Ii
preassembly is possible.
1 I 0 The average tendon length increases, and this reduces the cost of anchorages
segmentsJ; l+3 I I+2 J4 i+1 I I
and the prestress losses due to their re-entry.

__V__
0 Friction losses in the longer tendons can be reduced by symmetrical tensioning
4— ordinary coupler from both ends.
0 The costs and implications of couplers are avoided.
Fig. 3.42. Staggered couplers permit prestressing of two segments with ego/1 tendon
Launch tendons are arranged in the slabs, similarly to the mechanically coupled
In spite of these limits, schemes that use couplers are very competitive in slender ones. Their front (launchwards) anchorages always emerge inside the boX cell, and
superstructures where access to the box cell is dificicult. Tendon overlapping may be the rear anchorages can emerge as well, to be always accessible, or can be placed
more advantageous in the deeper sections. within the joint cross—section. In this latter case, the match—casting of the next deck
segment covers them, and the tendons cannot be relieved.
By placing the rear anchorages in the construction joint, the cost of blisters
3.7. I .2. Overlapped straight tendons halves, but special reinforcing schemes are necessary because of creep strains due
The recent advances in placing strands in empty ducts have made joining by over- to local compression at the anchorages. Unlike coupled tendons, the anchorage
lapping an often-adopted and competitive solution. It has several advantages: forces do not reduce with time, and the creep deformation modifies the stress
distribution along the joint, decompressing the zones in front of the anchorages
so much that even an average initial compression of 3 MPa may be unable to
prevent tensile stresses. Joint cracking may be avoided by additional reinforcement.
On the contrary, rear anchorages in blisters (Fig. 3.40) are always accessible and
permit one to relieve and recover . some.of...the.-pre.stress tendons on completion of
launch. This operation, although it reduces the permanent prestressing in the
structure, is generally not very competitive and is adopted only to correct the
undesired effects of launch prestressing. In fact, recovered strands are hard to
reuse, anchorages are lost as well as ducts, which must be grouted to fill holes in
the deck, and the impossibility of grouting temporary. tendons runs the risk of
strand corrosion under load. Since this risk is acceptable only for short periods
of time, the possibility of recovering non-galvanized or PVC-cOated strands is
limited to short bridges.
By inserting strands in empty ducts, tendons can be made very long and can
join several deck segments in order to introduce launch prestressing progres-
sively. The tendon length and the connection scheme characterize the progress
of launch prestress in the superstructure. In turn, the allowable bending moment
in the deck cross—section according to the prestress and curing reached defines
the spacing between the curing supports placed between-the casting cell and the
launch abutment.
WWW W . . Fzg 73.43 _. “Prefabricated-— tendons. emergingfrom the joint (author)
184 BRIDGE LAUNCHING PC BRIDGES 185

3.7. I .3. External straight tendons

The first launched bridge over the Rio Caroni did not use any permanent launch
prestressing, since service prestress was obtained by varying the position of launch
tendons. For this purpose, they had to be external. Later, the scheme composed of
centroidal internal launch tendons corrected on completion of launch by parabolic
and short, straight tendons characterized most launched bridges.
In some bridges, an intermediate solution has been adopted, composed of three
families of tendons as in Fig. 3.44: span “tendons cop tendons parabolic tendons

0 Internal straight final tendons distributed in the slabs so as to produce a

centroidal prestress force. ,l, l
. i ‘L I
External straight temporary tendons, also centroidal, to be relieved on comple—
lllllllllll

0
. I - I I I I I I I I I I / I I / I- l ' I I / I f I f I I I I I f
‘ / I .— ""u""; '”A

,
/ /

’
/' / / / xllulm" L4" ’r ._,_"Wmn_

9"
A

tion of launch to reduce deck compression without introducing secondary

moments. In the case of total prestressing, these additional launch tendons
a I/ V,
can be sized to cover bending moments produced by thermal gradients and l IIIIIII
’éfi’g’i’é’g’g
/ -l/
/ / /// /
walla"; rluuu‘. ' "-uup” dunnuuwul
/ / f
/
/
‘ 'uuuunu ’

bearing settlement, and rarely exceed 25% of the total launch prestressing.
o Parabolic, span and cap tendons introduced and tensioned on completion of
launch (Fig. 3.45). Fig. 3.45. Integrative end-of—laanch prestressing obtained by placing strands within
empty ducts
The use of external tendons for launch prestressing [80] complicates the formwork
and the steel cage, but offers many advantages, such as savings in ducts and grout-
ing, the simplicity of tendon removal, and the improvement of the mechanical
properties of the cross-section owing to the lower number of holes. In order to improve the competitiveness of this scheme, external tendons can be
anchored directly to the support diaphragms, permitting reuse of their anchorages if
additional external prestressing is necessary in the future. In this case, diaphragms
must be built before launch, and the undulation of the bending moment envelope
due to their dead load slightly increases the cost of launch prestressing. Moreover,
diaphragms should be adequately cured before stressing the tendons, and this
increases the length of the curlngzonefi if 7 7
The adoption of a large number of temporary tendons is rarely competitive,
except when the number of superstructures to be built subsequently is such as to
amortize the recovered material, or when this can be reused as parabolic or poly-
gonal final prestressing. On the other hand, external tendons can be useful for
temporary “tendons Flnol ‘tendons absorbing the tensile stress peaks in the front zone of the superstructure, since
they permit sizing of internal prestressing for more uniform stresses.

3.7. I .4. Antagonist tendons

'_/’J’/’/’/’/’/’/'/' 'l'l’l /'l'l'l’l’l'l’lrf‘l’l‘l'll'l’l'l’l’l'l'l’/'/'/'/’/'/’/’/'/’I’I’l’l'f
’/ /’/'/’/'/’/’/’/'/” 'j'/’/’/’/ /’/'l l’/ /’/\'/’/’/"'/’/’l /'/’[’/'/’/'/’/’"/’/’/’/’/'/’/’/’/ 1

‘ E’
Antagonist prestressing has met a certain success in recent years. The principle is to
separate the final prestressing from the temporary one. The final prestressing is the
'l’l’f’l’l'l’l'
’/’/’/’/ / /’/’/
/ l'l’l'l’l’l’lll'l'll'l'l’l'l’l’l’lrl'l’l’I’l'l’l'
l’ll'/’/'/'/’/'/'/
l’l'l'l'l'l’l/l/
/'I’/’/"/’/’/’/’/’/'/’/r/_r/r/'/’/’/’/’/’I'/'/'/ /’/’/ /'/'/’1
typical one for bridges cast in situ, with parabolic internal tendons or polygonal
external ones, sometimes complemented by cap and span tendons. The temporary
prestressing consists of tendons that compensate the eccentricity of the final tendons
Fig. 3.44. Launch prestressing obtained by integrating internal final tendons with during launch, until a centroidal prestress force is obtained.
external temporarytendons .. ,, .
 186 BRIDGE LAUNCHING PC BRIDGES 187

Temporary tendons can be, in very variable proportions: compromise this result with anomalies of the shear force diagram near the deviation
points (especially in the case of internal final tendons, which are parabolic and not
0 Short and rectilinear, placed inside or close to the bottom slab in the support polygonal as are the external antagonist tendons) makes the positioning of
sections and to the top slab in mid-span, and coupled by overlapping in deviation points difiicult and complicates internal formworks.
transverse beams that, in the case of external final prestressing, can be its In spite of these difficulties, these schemes permit one to optimize prestress both
deviation diaphragms. In this case, the optimum length of the deck segments in the construction stages and in the service ones. Even construction times are
is one-third of the current span. shorter, since the end-of—launch prestressing operations involve fewer tendons,
0 Long and parabolic, with polygonal layout symmetrical to the final tendon although the recovery of the antagonist ones is rather difficult.
layout with respect to the gravity axis of the superstructure. In this case, Tendons are generally chosen for strength, in order to limit their number
deviation saddles are necessary along the span (Fig. 3.46). and avoid excessive congestion of the box cell. For a safe reuse of their strands,
especially for the final prestressing, some precautions should be taken during the
The high axial compression produced during launch by final and antagonist
temporary use:
tendons makes it possible to obtain extremely efficient cross-sections, since the
reduction of principal tensile stresses limits the web" thickness. The need not to
0 Temporary launch prestress should be less than 85—90% of the allowable tensile
stress, to reduce risks of local strand plasticization.
detofl 1 0 Strand portions notched by the anchor keys must be eliminated. This require-
ment is almost always met by limiting prestress during the temporary use of
strands, as the higher elongation at final tensioning permits the notched zones to
m
emerge outside the pulling anchorage.
m...—

—'-—-—... J ,_..—--:.:v-—~._ f-«es__>~—_.. firmer—Him“

I . | o The strands should be galvanized and placed in perforated polyethylene pipes

_...........

A B| £ that permit corrosion control in case of damage to galvanization during the

“Fa—u...—
__.4_____

strand insertion. To guarantee the safety of workers in the case of breakage of a

strand or of the whole tendon, these pipes are anchored to the deck by means of
collars.
-— temporary tendons couplers
0 Only one temporary tendon should be dismantled at a time by pulling it back-
wards inside plastic sheaths, which facilitate extraction and protect the tendon
11111
1111111
from dirt, and immediately reused as a final tendon in the same span.

Q I DETAIL 1 I % Antagonist launch tendons were used in the Sathorn Viaducts [81] along with
internal final tendons (Fig. 3.46). Since the cross-sectional centroid was so high
as to penalize the mid—span efiiciency of antagonist tendons contained into the
box cell, they emerged above the top slab and were deviated by steel saddles rested
temporary tendons on the superstructure (Fig. 3.47). In the support zones, they were contained inside
the webs. Both final parabolic tendons and antagonist tendons were anchored in the
steel deviotor‘s
permanent support zones of the superstructure, and therefore each launch step was
Final tendons
as long as the whole span.
An even more radical application of antagonist prestressing was tested in the
A I.
Amiens Viaduct [76]. The launch tendons, entirely external, were located as folloWs:
¢
/ two with a final trapezoidal layout deviated at span quarters, two with a final
triangular layout deviated in mid-span, fOur in antagonist position with a rigor-
ously symmetrical layout, and six straight and close to the top slab to raise the
prestress force to the gravity axis. On completion of launch, span by span, the four
Final tendons temporary tendons
antagonist tendons were relieved and repositioned parallel to the four permanent
.-.Fig. 3.46.. Prestressing scheme for. Sathom Bridge polygonal tendons stressed before launch (Fig. 3.48).
188 BRIDGE LAUNCHING PC BRIDGES 189

3.7. L5. Prestressing oftemporary joints

Most launched bridges keep the continuous beam scheme used during launch per-
manently. In some cases, however, it is necessary to. create intermediate expansion
joints, and the temporary joints are designed so as to guarantee the safe transfer of
shear forces and bending moments during launch, and their easy separation when
the deck has reached its final position.
The use of external tendons or bars for temporary joint prestressing is encour—
aged by the presence of the end diaphragms, which can easily contain their
anchorages. A totally external prestressing also reduces the congestion of internal
tendons in the end support sections of the superstructure. As an alternative, the
joint tendons can be contained in the slabs (Fig. 3.49).
Once the maximum fluctuation of the launch bending moment has been deter-
mined, the joint prestressing is designed with the same criteria used for the current
deck eross-sections, and distributed so as to produce a centroidal force.
The joint compression must guarantee adequate safety from both edge de-
compression and joint sliding. If necessary, the joint can be equipped with shear
» . keys, whose design criteria are those already discussed for the anchor systems of the
Fig. 3.47. Antagonist tendons emerging above the top Slab (courtesy Virlogeux) .1 launching nose.

sectlon A—A

0.000.000. I

shear keys

axial prestr‘esslng Joint prestr‘esslng

A paraboUc tendons
Fig. 3.48. Prestressing operations upon launch completion of the Amiens Viaduct
 190 BRIDGE LAUNCHING PC BRIDGES ‘ 191

3.7.2. Dimensioning oflaunch prestressing uncracked section and a reduced allowable tensile stress. On completion of launch,
In a superstructure of constant dead load, composed of many equal spans, and with the dead load and 10-30% of the live load are carried without edge decompression,
a launching nose of adequate size, the envelope of the launch bending moment while the remaining portion of the live load is carried by partialized sections.
diagrams is like that shown in Fig. 3.50. Edge tensile stress is not limited, and crack width is controlled according to
The superstructure can be divided into two zones: a rear current zone, character- environmental aggressiveness.
ized by a repetitive course of launch stresses and therefore by constant envelopes,
and a front zone characterized by peaks of both bending moments and shear force. 3.7.2. I. Rear zone ofsuperstructure
Usually, launch prestressing is sized for the current zone, and then adapted to the For a beam subjected to a centroidal prestress force F, the effects of a moment M
particular requirements of the front zone. applied to the beam may be evaluated by introducing an eccentricity, e, to the
Applicable standards and design preferences guide in the choice between three centre of F, where
different design criteria for launch prestressing:
PE_F
360
(' )

0 Total prestressing for all load conditions. This criterion is generally adopted

_
__
..~w___wh_._
only for precast segmental bridges with epoxy joints. For instance, AASHTO If exterior loads act on a beam, the resulting bending moment together with a
[42] does not allow tensile stresses in glued joints and requires a minimum uniform compression could be produced in the unloaded beam by a prestressing
compressive stress of 1.4 MPa if external launch tendons are used. tendon with an eccentricity equal to (3.60).
a Total prestressing only for dead load, so that thermal stresses and accidental In the rear zone of the superstructure, each cross-section is subject to bending

MM
events can decompress the edges and produce cracks, whose width is controlled moments that vary from +Jllmax to —Mmin
by reinforcement. These additional stresses may be determined by enveloping AM : Mmax + Mmin (361)
full thermal effects, full bearing misplacement effects, full creep deformation
effects and 50% combinations of these. To annul edge tensile stresses, the minimum axial prestressing, 17min, should be such
0 Partial prestressing, allowing edge tensile stresses independent from the load that the maximum upwards eccentricity of the resulting force
combinations.
e, _ Mmax
Fain (3.62)
Generally, the most competitive solution for cast-in—yard bridges with through
reinforcement in the joints consists in allowing low edge tensile stresses during and the maximum downwards eccentricity
launch (1—2 MPa) to be absorbed by longitudinal reinforcement. Reinforcement M .
ed = “7‘“ (3.63)
is sized for the total tensile force in the concrete computed on the assumption of an
maintain the resulting force inside the central core of the cross—section, whose depth
hC is defined by Equation (3.1)
mm ‘ eu + ed 2 he 2 pSH (3.64)
)1‘111'1.111111%
111’ 11) 111111111; 1', 1‘

n 111111 11
*-

from which
1 ”111111)H
11111111 1W) 111

111111.
moment (MNm)

_30 0 ‘(mmmm’ 11111111111\i 3 1,)

11 11.»
1111111111 11m11 .m. 11mm
111 .1 jgjimj ji1111111’1‘“1111111M 1.11111111111111111;

1111111111 1 1 H 1 18.111111)
11111111111100“ “

{1111111111111111111
"—

11111111
m

' _(111111131111111111111111
---—

«11111111111111
-——

152m 1 i1..."11ii 11111111 ’111111111111 1111111111” 11""11111mm

1‘ ll 111))1 1 1m1m""1“"“1111'
‘hu—e

i'iiiii’ii’i'i’i
jiiw
”(111111 ,111111111111 AM
‘29/1‘330110.t11111111
a

ME»...«I....—

1 1111‘ 1111«1 11 j.
1111.1161111
13mm“)
111111.,‘ 1 11 111111
1 1111 . 111111
—.flm—’—.

1111)
1111111111111.11111.1.1‘. 1111(1'1111111
i 11) I

H‘ r1111
«n;—

1’1 mm...
‘— 4-,. ‘—
M

«1111)
1

1 1
1 111+
1W 1 11 '1$ 11111
1:” “i 1'\i1.
m 11'
W1“
«at—‘—

<-.- 4
W

4——

.—

(1% W1
,~*~%

are 1111#1311111111‘ 111111‘1l11(1611111111

W31 1111.9111i1mm 1 11,11,131 *1iii111“tilii‘iii‘i'"W
1 , 111111111111 MNmmi1i1i11i$£1i1i1{Im“M 'H\
fl»

”is in?)
J-
.5!-

.'
O
O

i-%— -“-—
eh,

\it
y‘x‘»

00
J.

11mm
"'-b
«-
*‘u
‘w

iiiii‘gitgiégiiigil?!11
1'
19“»

‘~
.1».

11111i¢111111111ii11i. ii 1..
Miran 111i O'O'O'IW M 1 «(n 1 - \ IW’ In the case of partial launch prestressing, similar considerations will apply, so that
5’"

11‘ m1l1it; 1.
imimi,l\1si1ii111
\vwvjmmrit'nwtg'g'qga4 i’i‘i’iifl'if"Wjiiii‘i‘jg‘ii‘i
new)
8a.iiol 1 “I
a!
- t 1 111"?" (f 1'! I
1

1 'm. 3*1’ iliiioiiiiifi?”

0111110,)? 1:1
“ha.— 4:a.

iiiiiiiiiw We air 1? .1...girirsaainr

—‘.‘.
bending

van,

"111 W" i
10.0 k oai'u ii'i‘i‘i'i'iti‘ii‘i’iiiiemn
1’ .‘I as “:5. ”a

1W.» iiii‘i’iiiiiiik
O 6
defined edge tensile stresses are not exceeded.
~
4'-

20 . 0 —
-91, #255“ O
In reality, no cross—section can have a “central core perfectly dimensioned for the
gg‘ggéfih
“is.“,ud I; /‘

maximum fluctuation of launch bending moment, and launch prestressing will be

‘afio‘v'o'o 1"

higher than this minimum value.

—
—
—
—

—
—
M

*
“

"“

-
—
—
—
"

By imposing limitation on the allowable edge stresses

250 0
230 0
190 0

210 0

240 0
220.0
200 0
180.0
160.0
170.0
150.0
140.0
120.0
130.0
110.0
100.0

F M (3.66)
~ Fig. 3.50. Typical bending momentrvenvelopew __.__:?7F_SZ+TZH§ “C _ .
192 BRIDGE LAUNCHING PC BRIDGES 193

F M This condition is rarely satisfied, since the wide top slab draws the centroid upward,
7-01 E Z ""' 721 5 0'0 (3.67)
and the minimum launch prestressing will be such as to not exceed the allowable
edge tensile stress in the worst condition:
Where ac is the allowable compressive stress and —at is the allowable tensile stress,
Mmin
the worst conditions are F1 2 (—1 2., — 09/1 (3.75)

1
—0't +
min
Zn 5 (3.68) Mmax
I
M
21 (3.69) Fmin : max(F1, F2) (377)

The launch prestressing tendons are almost always arranged inside the two slabs,
which define the minimum moment of inertia of the cross-section
whose axes are respectively at a distance a?t and db from the section centroid.
1 2 H& ‘ (3.70) Launchprestressing is composed of a top-slab force I7t and a bottom-slab force
0c + 0t Fb, which are determined so that the resulting force is centred on the gravity axis to
avoid secondary moments in the deck:
I > H @33— (3.71)
— 0'c + at
F =
db F (3.78)
If a launched bridge is presized as a slender continuous beam, the position of the t dt'l‘db
gravity axis is generally rather high. During launch, the highest tensile stress is likely
Fb : F _ Ft (3.79)
to be reached in the lower edge in mid-span owing to the small section modulus
I/21, even though the maximum positive moment Adm,1X is lower than the negative The case of external launch prestressing is treated in the same way.
moment at support Mmin. Therefore, the lower edge tensile stress often governs the
axial prestressing. 3.7.2.2. Front zone ofsuperstructure
When the same cross-section passes above a pier, its upper edge is far from In the front deck zone, the maximum negative moment and, in a lesser degree, the
decompression, since axial prestressing is only partially unloaded by the low tensile maximum positive moment, depend on the behaviour of the nose—deck elastic
stress resulting from a high section modulus I/zu. On the contrary, its lower edge is system. Even using an optimized launching nose, the maximum negative moment
overcompressed, as the axial prestressing adds to the high—compression stress gen- at the front support is much higher than at the rear supports, and the back launch
erated by the maximum bending moment and, again, a small section modulus I/21. prestress must be increased. Even mid:span ,, positive moment may require
For the allowable tensile stress to be reached both in the upper edge of the additional prestressing. '
support section and in the lower edge of the mid—span section, we must have The additional prestress, Fa, with eccentricity, 6,, positive if upwards from the
gravity axis, is designed to respect the allowable edge stresses in the worse launch
£_Mminz _£_Mmaxz
(3.72)
A I u_A I conditions. Neglecting its secondary flexural efiects, at both edges of the front
support section we should have
from which
F +Fa +Fa€a “— Mmin
3 = £911 (3.73) A I
Zn 2 _0t (3.80)

Zu Mmax

For the bending moment produced by constant dead load, the optimum section has F :1 Fa_ Faea —M-
[- maIEO,C (381)
the centroid at a distance from its upper edge equal to one—third of its depth, Mmin
being about double Mmax in the rear deck zone. " and in the most loaded span section
In reality, the other launch stresses (thermal gradients, bearing misplacement and
time-dependent effects) affect Illmax more than Mmm, and the centroid should be F+Fa+Faea+MmaxZ a
(3.82)
lower A I u " C
F :1 Fa_ F333': MmalZ_0,t
0.33 5 % 5 0.42 (3.74) (3.83)
194 BRIDGE LAUNCHING PC BRIDGES 195

Once the eccentricity ea has been determined according to the position of the The lifting action of parabolic tendons balances a part of permanent loads (dead
additional tendons, the calculation of F3 and the verification of the edge stresses load, q, and superimposed dead load, ppem), and the ratio of the prestressing
are immediate. equivalent load to the permanent loads is a good indicator of the distribution of
The same problems can affect the rear end of the superstructure, especially when longitudinal stresses in the superstructure. It is convenient [80] to reach at least
using friction launchers. In fact, these devices need a certain distance between the ppar
> 0.60 (3.85)
last curing support and the launcher itself in order to produce a high support q +pperm
reaction at the launcher, and this distance can become critical in the cantilever
configuration on completion of launch. but values of about 0.70—0.80 are preferable, limiting the creep deformations
Since the last segment of the superstructure is never very cured, instead of locally and forcing the stress state in the unloaded structure to approach a centroidal
increasing launch prestress it is preferable to introduce a temporary support compression, ensuring the longest durability.
between the last curing support and the launcher. This support is kept disengaged The additional tendons can be arranged in the cross-section in several ways,
during the previous launch stages by not inserting neo-flon pads or the launching according to the total force needed: in the conventional parabolic layout inside
bearings themselves. the webs, in a polygonal external layout deviated by diaphragms or saddles, or in
Finally, the maximum positive moment in the back end span is lower than the an external parabolic layout in contact with the internal surface of the webs.
moment reached in the front zone of the superstructure, since the end span is
generally shorter than the interior ones. However, the principles for calculating 3.7.3. I. Internal tendons
the additional prestressing of the back end span are the same. The additional internal tendons (Fig. 3.45) can be divided into parabolic tendons
arranged in the webs, cap support tendons placed in the top slab and span tendons
placed in the bottom slab.
3.7.3. Service prestressing Parabolic tendons can extend over several spans, but in this case they suffer high
On completion of launch, the location of lightened and stiffened cross-sections
friction losses. When limited to only one span, they can be anchored at the zero-
finally matches the stress diagrams due to dead load. Launch prestressing tends
moment zones of adjacent spans, so that they overlap in the support sections. This
to be excessive, since the better distribution of dead load reduces moment and shear
layout doubles their effect in terms of both negative moment and, if anchored near
diagrams with respect to their launch envelopes, and the correct distribution of the
the bottom slab, shear force. As an alternative, they can be overlapped at any point
mechanical characteristics of the cross—sections reduces the edge stresses even more.
of the structure by means of special blisters (Fig. 3.51).
In this context, the structure has to be modified to resist service loads. For this
In both cases, overlapping requires crossing of tendons and anchoring beams or
purpose, additional centroidal tendons would not be Very advantageous, and it is
blisters, and this makes parabolic tendons rather cumbersome. In addition, some
preferable to shape the additional tendons so as to obtain the highest eccentricity of
standards require that post-tens’i'oning ducts in the webs have concrete cover of at
the resulting force. The importance and the effectiveness of this correction depends
least 15 cm to the underside (Fig. 3.33) not to disturb the dispersion of support
on the quantity of centroidal prestressing that remains in the superstructure, and
reactions in the webs during launch, and this requirement reduces the lever arm of
therefore on the launching technique. Using temporary piers, launch prestressing
internal tendons. For all these reasons, once the shear stresses have been reduced by
can be relatively low, and the correction to introduce after launch is important and
the inclined portion of the tendons, additional prestressing requirements can be
eflective. In the case of full span launch, on the contrary, launch prestressing is
fulfilled with local straight tendons.
generally so high that only a few tendons can be added, which cannot improve
Cap tendons are difficult to tension. However, they are rarely indispensable, since
substantially the efficiency of the final prestress unless some launch tendons are
the parabolic tendons are added to the centroidal launch prestressing, and this is
relieved.
usually enough to cover the negative moment. In contrast, span tendons placed in
To evaluate the efficiency of parabolic prestressing in resisting permanent loads,
the bottom slab are easy to install and more effective because of the longer lever
the flexural effect it produces can be assimilated to a distributed upWards load equal
’
arm from the gravity axis. They are frequently used in the end spans of the super—
to
structure, especially when it is not possible to reduce the length of these spans with
respect to the internal ones. Their secondary effects can help in reducing the need
_ Fpar (3.84)
Ppar — “Rt for cap tendons.
The internal tendons are tensioned on completion of launch and, to avoid risk of
where Fpar is the parabolic prestressing force and Rt is the curvature radius of the corrosion, their strands are inserted immediately before tensioning. Since their
equivalenttendon{60]. ' ‘ -.-..ducts....ar.e.embedded. in the .deck. during. construction, they must be watertight and
 196 BRIDGE LAUNCHING PC BRIDGES 197

longitudinal section section B-B Concrete in monolithic webs is stronger than concrete containing many ducts,
which reduce the resisting section, disturb stress distribution and complicate
vibration. By eliminating the internal parabolic tendons and their supports, the
density of reinforcement is reduced, bar splicing can be optimized, the duration
of steel cage assembly is shorter, the general economy of construction improves and
the negative influence of poorly skilled workers is limited.
The visual control of the bundles of strands guarantees their optimum alignment
both in the straight zones and inside deviators, and this allows a good control of
tensioning and elongation. Because of the low prestress losses (friction is limited to
deviators, where it can be further reduced), this leads to higher average tensile
stresses in the strands, and therefore higher efficiency of service prestressing. Low
section A-A friction also allows using tendons as long as several spans, and this reduces both the
number of anchorages and the labour costs. In most launched bridges, in fact,
polygonal tendons are introduced on completion of launch, and tendon length
I?!” . does not afiect construction sequences. In addition, the long curing period of
concrete permits the anchorage of very powerful tendons.
External tendons are unaffected by concrete cracking (unlike internal tendons,
Fig. 3 .51. Web blisters for splicing ofparabolic tendons at the quarter-span points
which come in contact with the atmosphere near cracks) and provide a more valid
solution in case of partial prestressing, because of the particular sensitiveness to
stiff. The use of stiffening polyethylene pipes inserted into all the ducts before corrosion of tensioned strand.
casting, and then removed to be reused in the next segment, is always advisable. Maintenance costs are lower, since tendons can be inspected easily (they are the
elements most subject to wearing in a bridge, along with expansion joints) and this
allows easy repair or replacement while maintaining the structure in service.
3.7.3.2. External tendons External tendons pose the problem of resisting local forces and moments near
External prestressing is adopted more and more frequently in several countries, anchorages, which is more delicate than with internal prestressing due to both the
because of bOth the immediate savings deriving from the shorter duration of absence of the force distribution due to friction, and the convenience of using
construction and the long—term savings deriving from higher quality. powerful tendons to avoid excessive congestion of the box cell. The simplest
The web thickness can be limited to that strictly necessary to resist the principal solution is to anchor tendons to massive elements, and therefore to the support
compressive stresses. This reduces the dead load of the superstructure (especially in diaphragms. To obtain a tendon layout suitable for both moment and shear, it is
the longest spans, Where the web thickness is often dictated by the need for tendon then necessary to introduce sOme deviations, with saddles, ribs or diaphragms. The
containment) and increases the cross—sectional efiiciency. web thickness depends on the shear force reduction near the support sections of
Launch prestressing can be obtained with few powerful tendons placed in the the superstructure, and therefore on the longitudinal distribution of the deviators
nodes of the cross-section, which are no longer affected by the parabolic tendons. (Fig. 3.52).
This reduces the slab thickness and allows tendon coupling directly in the nodes, By fan—shaping the tendons from the support section, it is possible to obtain a
Without anchor beams and blisters. This also reduces the congestion of external shear force reduction diagram close to the triangle, and at the same time a resulting
tendons inside the box cell and improves the deck behaviour at the ultimate limit tendon layout close to the funicular diagram of permanent loads. The use of four or
state. If the prestressing force so obtained is insufiicient, some polygonal external six deviators in each span makes it possible to attain extremely thin webs, and since
tendons can be tensioned during construction in their final layout, and other in this case the launch stresses require a high launch prestressing, the presence of
tendons can. be used temporarily in an antagonist layout to be relieved and reposi— many diaphragms can be used to deviate or to anchor antagonist tendons.
tioned on completion of launch. If necessary, few' straight external tendons may As an alternative, all the tendons can have the same parabolic layout instead
raise the prestressing force to the gravity axis. The low residual launch prestressing of being fan—shaped, to reduce the angular deviation in the saddles closer to
deriving from the removal of most of the launch tendons makes it possible to resist the supports without penalizing the average angular deviation, and to avoid
most of the permanent loads with polygonal tendons, which leads to an extremely interference and geometric errors.
-..-efiicient final prestressing.
198 BRIDGE LAUNCHING PC BRIDGES 199

o.-Two ollaphr‘agm tayout

surface of the webs, as in Fig. 1.6. After tensioning, sheet metal forms were hung on
.‘ ‘ “ ‘NXXXNRXNNNNNRRXR\\\E:RXNN‘NXi \\‘\\\“ the strands, filled with fine crushed aggregate and then injeCted with cement grout.
The shear connection to the web surface was obtained with distributed stirrups bent
'\

VI/I/r/I/I
\\
\
\\
\

during casting and then straightened, or with high-strength bolts that guarantee a
\\
\
\ \
\

light prestressing of these kerbs against the webs, limiting the shrinkage cracks of
\ \
\

\\\\\\\\\\\\\\\\\\\N\‘ .\‘.\*\*\‘\~\\\ . ‘\X\

\ \
\

A A the grout.

3.7.3.3. Transverse prestressing

\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\
In superstructures wider than 12 m, transverse prestressing of the top slab begins to
be worthwhile. It is definitely advantageous fOr widths over l7-18 m, since trans—
verse bending generated by the live loads depends on the square of the deck width.
At the same time, prestress losses due to anchorage re—entry are less critical, and the
cost of anchorages less prohibitive.
Above certain span lengths, also the general need for dead-load reduction makes
c-Por‘abollc layout a light transverse prestressing very useful, as it results in a thinner deck slab.
‘ RRXKRXXNX\\ ‘ \1 ‘ XXX\&\ ‘» \ ‘ \ ‘ x ‘1. Finally, slab deformations, cracking and cyclic stress fluctuations in reinforcement
are better controlled.
\ \ ‘

/'/////////’4
\

’/’/‘/’/’/’/
\

The level of prestressing can be chosen freely. Total prestressing is superfluous

\
\
\
\
‘
\
\
\ _ and difiicult to attain, since the live load is much higher than the dead load. Partial
VVV “““ ‘\\\1 ‘\\\\\
\ \ x \ \‘\‘\‘\‘

prestressing is often sufficient, and the transverse tensile stresses due to permanent
\\\\\\\\\\\\\\\\\\ \ \ \\\

loads and, for instance, 30% of the live load can be absorbed with closely spaced
(0.5 to 1.0 111) light tendons (300—600 kN). Their small anchorages do not increase
the slab thickness, and the pulling anchorages can be alternated with the dead ones
(Fig. 3.53).
In the case of partial longitudinal prestressing, the transverse tendons in the
negative bending zones of the continuous beam are adversely affected by slab
cracking. Waterproof ducts such as rigidpoiyethyiene pipes should be used in
these cracked areas. Finally, an important contribution of transverse reinforcement
is always necessary, as the different layers of bars prevent obtaining an adequate
eccentricity of the transverse tendons, especially in the zones distant from the webs.

Shear Force reduction

Fig. 3.52. Shear force reduction in the superstructure

pulling anchorage l

Since the dead load of deviators suggests reduction in their number, the tendon I - ' v '
v v

layout is often trapezoidal, with only two intermediate deviation points placed at
vvvvvvvvvvvvv
a

about one—third of the span for fiexural reasons. In this case it is not possible to
obtain the optimum layout for shear, which requires deviation points at the [dead ancha
quarters of the span, and the webs are thicker.
Very powerful external tendons have been used in long-span bridges. They were
obtained¥by.__.,dcvia_tingthe ,strandswithin, steel saddles. anChqrcd. tethe, internal , ,, , W t353 ...A.lternated.anchorschemefor transverse prestressing-0f the top slab
 200 BRIDGE LAUNCHING PC BRIDGES 201

3.7.3.4. Vertical prestressing 3.7.3.5. Recent trends

The impossibility of undulating the launch prestressing force prevents modification Although the use of external tendons in launched bridges is quite recent, the par—
of the shear force diagram with it, and each cross-section of the deck has to resist ticular requirements of this construction method have already resulted in solutions
full launch shear stresses. Consequently, the web thickness must be such as to so innovative as to represent the most recent developments in bridge prestressing.
maintain the principal compressive stresses below the allowable values, while The first applications developed the idea of mixed prestressing: final straight
shear reinforcement in the webs controls the crack width produced by principal internal tendons in the correct position for the final static scheme, temporary
tensile stresses. ’ straight internal antagonist tendons removed on completion of launch, and poly—
If the principal tensile stress exceeds the allowable value in localized zones of the gonal external tendons assembled and tensioned on completion of launch
deck, rather than increasing the web thickness with expensive adjustments of the (Fig. 3.55). Compared with the conventional tendon layout (launch prestressing
forming system, it can be expedient to use local vertical prestressing. This can be in the slabs and parabolic tendons in the webs) this scheme permits lightening
attained with high-strength bars or U-shaped strands anchored on the top slab both the slabs and the webs.
(Fig. 3.54). ' Subsequent advances saw a final internal centroidal prestressing integrated with
A distributed vertical prestressing can change completely the stress state in the straight external tendons during launch (Fig. 3.56), and corrected on completion of
webs, reducing or eliminating the principal tensile stress, and avoiding cracking. launch with straight cap and span internal tendons and polygonal external tendons.
For vertical prestressing to be effective the web's must be rather deep, and this limits In both cases, launch prestressing is obtained by superimposing several groups of
its adoption to long-span bridges. Moreover, some design specifications require one tendons according to the deck segmentation. This makes it possible to relieve and
, to determine the shear stresses. (and consequently the principal tensile stresses in the remove some tendons on completion of launch. In the second case, these operations
webs) by neglecting a portion or the whole duct diameter. In these cases, the use of are avoided by using temporary external launch tendons, but their removal
vertical prestressing may even be detrimental. may require addition of some final straight tendons, both in the span and at the
Vertical bars or strands are short and expensive to install. Their effect is rather supports.
unpredictable because of the anchorage re-entry, which makes the actual distribu—
tion of vertical compression irregular. For these reasons, it is expedient to use
vertical web prestressing only to resist the peaks of principal tensile stresses, and
during construction
to absorb the remaining share with conventional reinforcement.
As an alternative, the uSe of precast webs vertically prestressed by adherent
strands makes it possible to reach construction and prestressing savings and an 996-966

optimum state of stress at the same time.

In service

0.8 0.8 0.8 0.8 0.8 .

.' , ' / ' , ' x ' x ' I/xfi/IZ/Y/YI/olZ/r/x/xfi/x/x/I/élfl/I/z'fl/I7447/ x/x/I/z/I/IZI'fl/ééxj/rfl/o

\sdw l
/ I
l
‘ , , , Xyncyyyyyxp", , , / , , , , , X/yyyyx/x/yx, ,/,/,/,/,/,7,/,//, , , ,
l

Fig. 3.55. Final straight internal tendons, temporary straight internal antagonist
Fig. 3.54. Vertical ....web..prestressingwith .Ueshaped strandsanchoredattlze deck slab . ,. . _ . tendons. (dotted), and polygonal external tendons
-
202 BRIDGE LAUNCHING PC BRIDGES 203

Final internal tendons tensioned bet-‘ore launching Final tendons streSsed beFore launching

r: i DECIDED
III III/III/IIIII I.
I
/ I , 1 , 1 / ,

I I / I I I I I I I I I I I I I I I I I I I I I I I I I I I l I l I
I / / I I
I I / I I

temporary external tendons tensioned ioeFore launching /

I
I I I I I I/ I I I I I I I I I I I I I I I I I I I I I I I

i ‘l‘

fl fl
l |

——_-
.JLJl...
-——i._‘
_.J_iL...

nun-u—
III-II—

u—m—
l I
I l

temporary tendons stressed ioei-‘ore launching

I I’I’l'4%l/I’/’I'I I’I'l'I’I’I’III’III III I’l’l I I I’/ I l’l’é’l’l’l’l’l I’/ I /

i i
Final Internal tendons tensioned ai-‘ter launching I

z , Xx/pyxyg/xy,’In/Juo/lfl/X, , , z , ,/,’, , , , , ,/‘,’/,y,/,;/,/37,,o ,7,/,/,/,/,',’ , ,,

li l I C] l I I C'/'/'/' (r / 1'/'.r 9’/’/’/'/’/’/ [III/II / l’l’l’fyf/‘ZZIIII/l'l

—
xflyxflZ/r/zlliz'zll'lix' x’x
I

l | |_
r . 'I ' I
z I 2/ I I I I //////////x// I x x z z I r r x r z x I z// / x I z I x// r / x I I / x / I]

f-‘inal external tendons tensioned ai-‘ter launching

1 H H i
Final tendons stressed at the end 0? launch
l ‘I/IV I I
. '
’./m
I

, l , 1 [NZZZZZ’AZ ,/

Fig. 3.56. Final launch tendons integrated with temporary launch tendons and [J ////// if”, 6/9/7’1/{9/1/ / I

corrected with internal tendons and polygonal external tendons

I

.r a l '

I
yvxxaxhxmyyaaiawaxmyz*Ia/ .’/.%/.G{’/c///’/f’ . .’ .9

Other innovative prestressing schemes derive from the evolution of polygonal

antagonist prestressing. In one scheme (Fig. 3.57), final parabolic internal tendons
sized to resist most of the dead load are tensioned before launch." Their eccentricity Fig. 3.57. Final internal tendons compensated daring launch by antagonist polygonal
is compensated by antagonist polygonal tendons and reduced by additional straight external tendons and completed with straight and polygonal external tendons
tendons, all external and temporary. On completion of launch, and after relieving
the temporary launch tendons, prestressing is completed with some polygonal and
some span tendons, all external and sized for service loads. The parabolic internal launch, compensated by antagonist tendons and integrated by additional straight
tendons avoid excessive obstruction of the box cell. tendons, all temporary and external, which are relieved and repositioned in the
This concept was improved in bridges built with totally external prestressing correct final polygonal layout on completion of launch. This scheme may be further
'* [82,83]. A-shareof-----the polygonal tendons, is tensioned in itsfinalposition before. .-.....im.p.r.o_.ve.d..-.with some final. external. straight tendons.
-
 204 BRIDGE LAUNCHING PCBRIDGES 205

f-‘inal Internal tendons stressed beFore launching

As already discussed, it might be preferable to maintain few straight tendons
inside the concrete (placed in the nodes of the cross-section) to reduce the conges-
tion of tendons inside the box cell (Fig. 3.58) and to improve the deck behaviour at
the ultimate limit state. gun‘s-A
2? 4...".

In this case, prestressing can be schematized as in Fig. 3.59:

"hr-0‘ h......””n~”~...--.-..--

0 straight internal tendons in the cross-sectional nodes, tensioned during temporary launch tendons
construction and joined by couplers gag/[I I I I I I
l 1” IT I I I Igfl/I/Iflflflfifl/Ifl/GZ 'lI/J’Z/I/IYIY 2"”;J/I/I/I/I/I/I1/IVI' I'
0 one half of the polygonal external tendons tensioned during construction in the
final layout
0 the other half used temporarily in an antagonist layout to be relieved on I I/ ’ I/I/I/I/I/I/i/I/I/I/I/I7EI I/I/I/I/I/I/I/Ifl/Iflfi/I/II’I/x’I/I/I/I/I I;i’/I/I/,/,/, I I/

\\
completion of launch and repositioned, span by span, in the final layout
0 temporary straight external tendons that lift the prestressing force to the Final external tendons
centroidal level I’ I' I; 'fl/I/I/IYI I I/I,'/I/'"I/I 7/; I/IWIGx/Iflé’l/I/I/IVII’I/I’ I I/I/I/I/I/I I/I/I/I/I/VI’I' I I

0 some internal span tendons in the case of long spans.

I

i ' I l
The low residual centroidal prestressing deriving from the removal of most of I I I’I/I/I/I/I/I/I I/I/I/I/I/I/I/I/I/I/I/I/I/I/I/I/I/I/I/I//I/I/I/I/I/I’ /I/I/I/I/I I I/
the launch tendons permits sizing the polygonal tendons to resist most of the
permanent loads (with the purpose of resisting fully their flexural effect), and this temporary antagonist external tendons
leads to extremely efficient permanent prestressing.
I I I '/f/é'/;/////,I/l,y/Y/yl/l/l I I I I III I I I I I I I I I/l/lflflfi/ljfi/rl/l/lfl/'/l I I
I

| | | l
I I l7/////////// I/I/I/I/I/I I/I/I/I/I/I/I/I/I/I/I/I/I/I/I/I/I/I/I',/I I I/I/I/I/I III I I/

span tendons stressed ai-‘ter launchlng

* 6’" I' I' I ' I- I' III/I7I/Ij’37I/II/I7I7IJ/I7I7I7I: I I/I/I/I/I/If;

IV
r .

I." [fig-nu, p. ‘4‘

tun-'u-r-rul A, ..... .21.. ..........._......... ”Z... .. 22“

external tendons stressed aFter laLTnching

I I I, I/I/I/I/I/I
lI I/I/I/I/Ifl/Ifl/I/I/I/Ifl/I/I/I/I/I/I , I I/I/I/I/I/I I/I/I/I/I/III I. ./
/

' I I I’I/I/I/I/I/I' ,/,/,/,/,/,/,;,/,/,/,/,/,/o/,/,/,/,/,/,/,/,/,/,/,/,/, /I/I/I/I/I I’/I77I7'

f

Fig. 3.59. Few internal straight tendons placed in the cross-sectional nodes reduce
congestion of tendons and improve deck behaviour at the ULS

Fig. 3.58. External launch prestressing in the Charix Viaduct (reproduced with
, permission, .Scetaroute, ..Berenguier)
 206 BRIDGE LAUNCHING PC BRIDGES 207

3.8. Reinforcement slab, which is subject to the highest stresses both during launching and in service
In a box girder built by incremental launching, reinforcement is distributed [79].
according to conventional principles. Adequate stiflness of the steel cage is necessary both for transfer of preassembled
Two orthogonal bar layers are placed at the edges of the top slab to resist cages and during casting and Vibration, to prevent movements that might create
longitudinal and transverse stresses due to the permanent load and the live load, cavities around the bars. These cavities, in fact, permit corrosive agents to come
secondary stresses and distortion. Two similar bar layers are placed in the bottom into contact rapidly with a larger steel surface and accelerate degradation.
slab to resist the same stresses (live load apart) and to distribute the support
reactions towards the webs. For this purpose, the compressed struts must be
checked along with the horizontal reinforcement for the tension rod action
(Fig. 3.60). ’

firm-"mums
Two orthogonal bar layers in the webs resist the principal tensile stresses

._
generated by the launch and the final shear forces, as well as the edge stresses

..-._ .__._-#_mwnmw___P——-4W_q__mm. .
due to transverse bending, thermal gradients and differential shrinkage.
Local reinforcement is finally necessary to resist the thrust force, which depends
on the type of launchers. Using friction systems, it is necessary to resist the super-
ficial stresses produced by the contact plates. When using back transoms or through
pins, it is necessary to resist local concentrated forces and to check the web-slab
nodes for the shear forces produced by the thrust.
Transverse reinforcement is generally composed of straight and bent bars
designed according to the casting phases of the cross-section. Bent bars are more
expensive, but necessary to stiffen the steel cage, acting as spacers of the bar layers.
Finally, it is a good practice to distribute wire mesh on the whole external surface
of the cross-section to further stiffen the steel cage and to limit the concrete cracks
due to shrinkage and thermal gradients. The section of this additional reinforce-
ment may be about 0.1% of the deck cross-section, and about 0.3% in the bottom

/ / I I / / //////’///// /////,///////,/
/

FlgS’ 60 DiSpersion of launch supporrreacrions produces 6.: .Srrut-and-tie,,,actiQn,,

-
4. Composite bridges

In recent decades, many factors have increased the competitiveness of composite

bridges and reduced their lower threshold of utilization to spans of about 50 m,
which were previously the domain of PC box girders.
Progress in iron metallurgy has led to rolled steel of high and reliable mechanical
properties. The development of design codes based on the strength of materials has
led to a better knowledge of structural safety. Progress in the assembly techniques
has replaced riveted joints first with bolted joints, and eventually with welded joints.
The composite action between concrete and steel has been thoroughly investigated
and comprehensive design rules have been codified by standards. Finally, in the last
few years, this evolution has been further accelerated by the rather static cost of
steel plates compared with the general increase in labour costs.
The qualitative advantages of composite bridges are significant. The high tensile
and shear strength of the steel plates combined ,with the low—cost compressive
strength of concrete result in efficient and competitive structures. The rapidity of
construction is a sourCe of financial"savings,”"an’d"the 'pOssibiiity of building most of
the structure in a workshop is a prime advantage in terms of quality. The high
dimensional precision of struCtures enhances reliability of calculations. Extended
durability is ensured by various renewable protective treatments, and by the
possibility of replacing damaged components and modifying the structure to
adapt it to new service conditions. Demolition work is less difiicult than for PC
bridges, and recycling of the steel structure further reduces the total cost. Finally,
the architectural quality of structural elements whose function is clearly recognizable
enhances the aesthetic qualities of the bridge.
The difierent efficiency of PC and composite bridges derives from the different
efi‘iciency of the materials, involved. The efiiciency of a material can be evaluated
from the ratio of the strength to the specific weight [54]:

y ,

45-MPa concrete has a compressive efficiency of eM = 45000/25 % 1800 m and a

much lower tensile. efliciency; consequently, shear efficiency is low as well. The
 210 BRIDGE LAUNCHING COMPOSITE BRIDGES 211

tensile efficiency of a steel plate is much higher: even relating it to yielding, a Fe 510 steel frame is much lighter than the final composite structure, and local stress
EN-10025 steel plate with fy = 355N/mm2 reaches 6M 2 355000/77 2 4600 m concentration further complicates the situation.
and is therefore 2.5 times more efficient, although instability reduces compressive In spite of this, the launch stresses hardly affect presizing of the Steel frame, which
efficiency. Consequently, composite bridges are more eflicient than PC ones, as is almost always carried out according to conventional criteria for composite
their masses are more distant from the gravity aXis and this increases the cross— bridges. Then, launch requirements are met with purpose-designed launch devices
sectional flexural efliciency (2.13), and the material performance is much better. and, if necessary, by adjusting some details of the superstructure.
The optimum field of utilization of a composite section is the simply supported The cross—sectional design depends on the concrete-to-steel ratio of the composite
beam. In a continuous beam, the positive moment regions take full advantage of the section, on the weight-to-cost ratio of the steel components, on the construction
composite action, but the stress states in the negative moment regions are critical both and assembly costs of the steel frame, on the desired slenderness and on the
for the; steel girders and for the concrete slab. The compressive stresses in the bottom requirement for torsional stiffness.
flanges require thick plates or closely spaced stiffeners to prevent instability, and the As the concrete slab affects the total dead load significantly, its thickness is a
cracked slab contributes to the composite strength only through reinforcement. prime design parameter. The slab thickness depends on the deck width, the pre-
In spite of these problems, the many advantages of the continuous beam scheme scribed strength of concrete, and the structural system of the composite deck. If
almost always lead towards this solution. Consequently, the behaviour of shear connectors are provided throughout the length of the bridge, the deck slab
composite sections under negative moments has been thoroughly investigated to cooperates with the steel frame in resisting flexural stresses, and sections in both
reduce the weak points and to ease control of their effects. positive and negative bending are considered composite.
When compressed, the deck slab is often fully cooperating, although some
4.I. Conceptual design and deck presizing standards require the use of a reduced effective width in the case of very thin
As discussed in Chapter 2, the most characteristic design aspect of bridges built slabs to account for shear-lag effects. When calculating stresses from negative
by incremental launching is the need to resist transient launch stresses that are bending moments, the composite section consists of the steel girders plus the
significant and different from the final stress state in the structure. longitudinal slab reinforcement.
The reinforced concrete slab of a composite bridge represents 75—85% of the On completion of launch of the steel girder, the segmental casting sequence of the
total cross-sectional weight. Since launch stresses depend mainly on dead load, the deck slab is reflected in calculating the stresses caused by loads applied separately to
steel frame of a composite bridge is almost always launched alone, and the concrete the steel girder alone, to the short—term composite section, and to the long-term
slab is cast in situ or plaCed onto the steel frame only on completion of its launch. composite section. The cross—sectional properties are evaluated from the modular
Exceptions are possible: a 675-m highway bridge has been built in Germany [84] ratio
by directly launching the final composite deck, but this required the use of tempor—
ary piers, the cost of which is acceptable only in very particular cases for composite rz=—-— ' ' “' (4.2)
bridges. Two parallel 194—m bridges have been recently launched with complete
composite section on the UK’s West Coast Main Line railway, but this required between the elastic modulus of steel and concrete. Time-dependent effects are taken
oversized web plates to avoid buckling caused by migration of support reactions. into consideration by most of current codes [8 5—87] by increasing the modular ratio
When the steel frame is launched alone and the concrete slab is cast later, the steel for short-term actions no
frame must be able to resist the dead load of the deck slab before the onset of
72 = n0[l ‘l— K1K2(t, to” . (4.3)
composite action. Since this distributed load is about four times greater than the
self-weight of the steel frame, huge margins of strength seem available for the where K20, to) is the concrete creep coefficient and K1 depends on the type of action.
launch stresses of the latter. According to EC4 [85], for example, K1 is equal to 0 for short-term actions, 0.55 for
In reality, composite bridges are competitive for spans, beyond 50 m, and-these permanent actiOns, 1.00 for shrinkage effects and 1.60 for slab prestressing through
spans suggest the use of steel girders with constant depth but varying cross-section. imposed deformations. In the French instructions [86] it is [(11903 to) = 2 for all
The flange dimensions and the web thickness vary along the girder according to the permanent actions, but the new Spanish recommendations RPX-95 [87] propose a
flexural and shear strength demand of the composite deck, and to technological and more sophisticated age—adjusted modulus [88].
economic requirements.
Moreover, casting the deck slab raises the gravity aXis of the composite section, 4. l . l. Multigirder systems
which governs the final distribution of flexural stresses. For all these reasons, The cross-section of the first composite bridges contained several steel girders that
,-,-,the.,launch .stressesin the,,steel,,,frarne-.areanomalous.and important even if the supportedthe concrete slab. directly. Nowadays, the use of multigirder systems is
 212 BRIDGE LAUNCHING COMPOSITE BRIDGES 213

rare, as the low coSt of steel plates and the high cost of labour penalize workshop MW/ / x719””/x/j’fl’Z/////V/,Z///W&J
and field assembly activities.
Cost evolution progressively led to a marked simplification of composite bridge
conception. First, trusses were abandoned, to the advantage of steel-plate webs.
Then, multigirder systems were abandoned, to the advantage of two—girder systems, =LL .=L=.
or U—shaped girders for composite box sections in the case of higher torsional
strength demand. Fig. 4.2. Two-girder composite bridge
In a multigirder composite bridge (Fig. 4.1), the field splices in the girders, and
between the girders, the bracing systems and the cross-frames, require expensive
workshop and field operations. In addition, the plate girders are relatively slender, The concrete slab is heavier than with multigirder systems, and this can be a limit
and this increases the density of web stiffeners, with additional welds. on the longest spans. On the other hand, the cross-beams are more spaced in the
In spite of these weak points, the multigirder systems are interesting for short longitudinal direction, about 8 m, and this reduces their cost. "
spans and wide deck slabs, or for very slender bridges (say, with HW % L/30 where Some standards restrict the longitudinal spacing of cross—frames or cross-beams
HW is the total depth 0f the steel girder), or to facilitate crane assembly of paired for curved bridges. For instance, AASHTO [89] requires that the length of unsup-
girders. Since these cases do not involve incremental launching, this structural ported flange between cross—frames or diaphragms be shorter than 25 times the
system will not be deepened further. compression flange width or one—tenth of the radius of curvature of the girder web.
These limitations may be insufficient to guarantee frame stability during launch-
ing, and the longitudinal spacing of transverse bracing systems should be reduced.
4. I .2. Two-girder systems In order to keep the warping stresses due to non-uniform torsion to reasonable
The advances in welding techniques of thick and very thick plates permitted the values, diaphragms and cross-braces should be more closely spaced as curvature.
development of composite bridges based on two parallel I-girders (Fig. 4.2). Two- increases. The following limits are suggested: 4.5 m for plan curvature smaller than
girder systems range from the 50—1 lO-rn spans of conventional launched bridges to 60 m, 5.0 m for 60—150 m curvature, 6.0 m for 150—300 In curvature, and 7.5 m for
g the much longer spans of cable-stayed bridges. Compared with the multigirder, curvature radii longer than 300 m.
systems, the two-girder solution reduces the number of web stiffeners, field splices,
In the open-grid frame of a two-girder composite bridge, the cross-braces are
cross-frames and bracing systems. prime load-carrying members. Their function is to resist torsional and distortional
When the deck is 13—20 m wide, the concrete slab can be supported on the main loads and to provide lateral support to the compression flanges, and therefore they
girders in the longitudinal direction, and on floorbeams spaced about 4 m in the should be attached close to the flanges. This is particularly necessary in curved
transverse direction. This solution usually permits adoption of a constant slab bridges; in straight bridges, cross—beams are frequently joined to the main girders
thickness of 20—24 cm. The slab is joined to the main girders and floorbeams by near the gravity axis of the steel frame (Fig. 4.2).
headed stud connectors, which are welded to the upper flanges with automatic In this case, the vertical web stiffeners are designed to carry this transverse frame
processes. _ action. This results in less expensive field splices and in a vertical clearance between
When the deck isnarrower, the concrete slab can be supported only on the main the bottom surface of the concrete slab and the upper flange of cross—beams that
girders, Which are in this case spaced about a = 0.553 (Fig. 3.1) and joined by
facilitates the use of sliding form tables for in situ casting of the concrete slab.
cross-beams. A reasonable slab thickness (say, 0.25 m in mid-span and 0.30 m Compression flange instability in the negative bending regions is controlled by
above the webs) permits reaching a transverse spacing of the main girders of the flexural stiffness of cross-beams, even when the distance between the bottom
about 7 m, and an overall deck width of about 13 m. In this case, the vertical
flange of the latter and the bottom flange of the main girders is one-half of their
depth of the steel girders may be presized as HW = L/28. transverse width bb.
A solution used sometimes to reduce the transverse span of the concrete slab is
MZ’X/ZZ’Z/ZZ’X/ZZ/%Z’Y/ZWZIZZ’Z/ZW/xflW/xZZ’Z/ZWZ/ZZV/ZWY/ZZ///’//ZZ’//ZZ///%7/Zi the use of a secondary longitudinal I—beam directly supported on cross—beams or
"’ Tu”; *“ LTJ 'T’ ’ ‘T' cross-frames. The field splices between the longitudinal stringer and cross-beams

JILL .L_J. lfl Jfl .U_L.

should preferably be bolted to avoid the fatigue problems of a welded joint due to
the difference in stiffness between the longitudinal stringer and the main girders.
This floor grid system is often combined with the use of precast concrete
, -Fig- 4-]... Multi—girdercomposite bridge. .. _____,_planks_...__that_ _create_._a . left=in=place form . for the concrete slab. In this case, it is
 214 BRIDGE LAUNCHING COMPOSITE BRIDGES 215

prudent not to take this forming system into account in terms both of flexural bars in the slab until reaching the negative moment capacity required by service
strength of the composite section and of local strength of the concrete slab. loads. , .
Conventional thickness of precast planks for such spans (about 3 m) is 4—6 The use of wide, thin flange plates is justified when the flange area is adjusted by
cm: since planks burden the structure significantly, their use is advantageous varying the plate width rather than the thickness, provided that buckling is con-
only on the shortest spans. trolled. However, the use of flange plates of constant width and varying thickness is
When two I-girders support a concrete slab of conventional thickness, the neutral generally preferable, since this simplifies both the form tables for the concrete slab
axis of the composite section under a positive moment is normally just below the and the lateral launch guides.
top flange of the girders, and the entire concrete section is compressed. The incor- To avoid instability, the width-to—thickness ratio of compression flange should
poration of the concrete slab in the beam properties increases the composite not exceed specific code limits that allow for geometric imperfections and residual
moment of inertia, but also increases the distance to the extreme tension fibre, weld stresses. A frequent requirement for straight girders is [l 1]
and the section modulus for the tension steel increases only moderately.
For \a more efficient composite action, the cross—sectional area of bottom flanges b Es
—Z <
_ 0.9 fy
—— (4.4)
should be larger than that of the top compression flanges; this also lowers the
centroid of the composite section towards the gravity axis of the steel girders.
Span lengths suitable for launching always require the use of welded plate girders, where b and 1‘ are the compression flange width and thickness, respectively, ES is the
and therefore there is a wide choice in specifying the size and thickness of the top elastic modulus of steel, and fy is the characteristic yield strength.
and bottom flange. When the steel frame of a composite bridge is assembled directly in its final
In the past, the flexural strength of the steel girders was adapted to the bending position, the geometric restrictions of Equation (4.4) affect the bottom flanges in
moment diagram by superposing cover plates onto the base flanges. Nowadays, the negative moment regions, and the top flanges in the positive moment regions. If
both webs and flanges of welded plate girders are obtained from plates directly the steel frame is launched, these restrictions affect both flanges for the entire bridge
supplied with the desired dimensions. Thick and very thick plates with certified length.
mechanical properties are readily available, and make it possible to reduce the These restrictions apply to straight girders, the flange plates of which are sub-
number of cover plates, or to avoid them at all by directly adopting flange plates jected to a uniform compressive stress across the flange width. In the case of curved
with varying thickness. girders, normal stresses due to non—uniform torsion and related restrained warping
Varying-thickness plates range between 20 and 150 mm, the variation rate of develop, the longitudinal stress varies across the flange width and localized yielding
thickness is usually 5~7 mm/m, and the maximum thickness variation is about near the flange tip causes a sudden decrease in the buckling coefficient. Therefore,
40 mm. The interest in varying-thickness plates had increased considerably in some standards decrease the allowable width—to—thickness ratio, or the allowable
recent years, as they permit savings in material and in welding costs (chamfering, compressive stress in the flanges. AASHTO [89] limits the allowable compreSsive
welding, controls, grounding flat). Moreover, full-penetration butt joint welds can stress based on the length of unsupported flange between cross-frames or dia-
be located in less stressed areas, and both aesthetics and fatigue behaviour improve. phragms, the radius of plan curvature of the girder web, the radius of gyration
Finally, the use of varying-thickness bottom flanges results in a regular launch and the width of compression flange, and the ratio at the diaphragm location of the
surface, as cover plates are avoided. _ flange tip stress caused by non—uniform torsion to the calculated bending stress in
In the support regions of the continuous beam, negative dead and live load the flange.
moments are usually greater than positive moments at mid—span. If the concrete According to Equation (4.4), the width of a compression flange with fy = 355
slab is not compressed artificially (by internal prestressing tendons or by stresses N/mm2 cannot exceed 21.6 times its thickness. Some codes prescribe additional
created adjusting the support levels during construction), longitudinal tensile geometric limitations related to the distance between the dead load counterflexure
stresses exceed the tensile strength of concrete. In this case, only longitudinal slab points of continuous beam, with equations for the effective flange width to be used
reinforcement can be considered to act compositely with the steel girders in for calculations. Finally, the dimensions of flange plates depend on launch require-
calculating the cross-sectional flexural strength, the moment of inertia for deflection ments, on the distribution and the type of shear connectors, and on the segmental
calculation, and the stiffness factors for continuous—beam analysis. casting sequences of the concrete slab. As a result of all these requirements, the width
Again, the additional flange area of the steel girders can be obtained by welding of the top flange usually varies between 400 and 800 mm, and the width of the
cover plates to the beam, top and bottom, or by using thicker or varying-thickness bottom flange varies from 500 to 1200 mm when the design span varies from 30
flange plates. Effective combinations of plates and reinforcing bars may be devel— to 100 m.
Woped,by..._sizing_.th.e.._.to.p..flanges forthelaunch stresses, and, by .addinglongitudinal, .
216 BRIDGE LAUNCHING COMPOSITE BRIDGES 2l7

The girder webs are almost always thicker than the minimum shear strength diaphragms. Web thickness is 10—12 mm. Flanges of transverse diaphragms are
requirements to limit the number of vertical stiffeners. The web thickness, tw, 220—300 mm wide and about 20 mm thick. When floorbeams directly support the
should not be less than 10 m not to aflect web strength with a minimum of concrete slab, their upper flange is about 300 mm wide to facilitate distribution of
corrosion, and to facilitate plate handling during fabrication. In practice, {W is the headed stud connectors. Finally, the total depth of cross-beams is 1/15 to l / 10
generally not less than 12 mm. of the transverse spacing of the main girders.
The current trend is tolimit stiffening of web panels to vertical stifleners and to When presizing a composite bridge, it may be useful to have an idea of the weight
take postbuckling strength into account in design. Therefore, slender cross-sections qs of the steelwork related to the total surface of the concrete slab. For a two—girder
(Class 4 sections according to EC3 [90]) are often adopted in the support regions of composite bridge with vertical girders, the following statistical relationship [1 1] may
medium and long-span continuous bridges, and the bending moment redistribution be used (in kg/mz): '
due to cracking of the concrete slab and to web buckling at the ultimate limit state
should be taken into account. q, 2 0.1051}6 + 100 (4.6)
Both EC4 [85] and BS 5400 [55] provide instructions to account for the first where L (in metres) is the main span of a continuous beam with optimum span
aspect, and the effects of web panel buckling on cross—sectional properties are distribution.
considered for section assessment, but not for structural analysis. In reality, two-
girder composite bridges with Class 4 support sections will likely have their webs in 4. I .3. Composite box girders
the postbuckling field at the ultimate limit state (ULS), and a redistribution of As an alternative to the multigirder and the two-girder systems, a composite box
bending moment towards more rigid regions should be expected. girder can be obtained by placing a concrete slab on a U—shaped steel girder
Moment redistribution depends on the plastic rotation of the intermediate sup- (Fig. 4.3).
port sections, and a certain amount of ductility is therefore needed. As beam The U—girder is particularly effective in resisting the flexural stresses during
ductility increases, so does the moment redistribution that can be achieved, up to bridge erection because of its wide bottom flange and improved lateral stability.
the level of redistributiOn that corresponds to the formation of a plastic mechanism. The low position of the centroid and the wide compression flange improve negative
While the ultimate moment of a cross-section decreases softly when increasing web moment capacity, and these girders are therefore particularly suitable for lOng—span
slenderness, available rotation capacity changes abruptly when the web slenderness launching with short noses.
increases from Class 1 to Class 2, so slender sections often do not have a
The closed shape of the composite section provides considerable rigidity in
satisfactory ductile behaviour. resisting torsional effects, and this can be useful in the case of curved bridges or
The ratio of the net depth of the web plate 11w (clear between flanges) to thickness axial piers, or for very slender superstructures. The cross—sectional resistance to
should be such as to avOid web panel instability under the thrust of compression distortion is higher than in the twojgirder,systems, and more uniform along the
flange. This mode of instability is covered by several standards. According to EC3 member. However, composite box girders achieve torsional and distortional stiff-
[90], it must be ness only after the onset of composite action, and temporary horizontal bracing is
_1/2 necessary at the top flange level to close the torsion flow both during launching of
h A E h E
—W S 0.55 /J__3 1+ W S) (4.5) the steel girder and during construction of the deck slab.
tw Af fy ( 3rfy Additional advantages of the composite box girder are good corrosion resistance
(the external faces are smooth and the internal ones are protected) and a pleasant
where AW and Af are the cross-sectional areas of the web and the compression aesthetic aspect. On the other hand, composite box girders are more expensive than
flange, respectively, and r is the radius of plan curvature of the girder web. In a the two-girder systems owing to a higher quantity of steelwork, more complex
straight girder with fy= 355 N/mm2 and, say, Af : 2AW, the limit ratio is about
224; in practice, however, it rarely exceeds 200. AASHTO [89] sets an upper limitvvof
170 for curved bridges with vertically stiffened web panels, and similar equations / / //x’zz///////,/////////W/
/ / / /I //

//// /////,'/ ///7—1

rule the introduction of longitudinal stiffeners. RPX-95 [87] (always for fy =
355 N/mmz) sets an upper limit of 160 for vertically stiffened webs in the support
regions of continuous beams, and of 240 for mid-span sections. These limits
increase to 250 and 350, respectively, in the presence of longitudinal stiffeners. nnnnn
Cross-beams are often welded I-girders. They are spaced at about 4 m when
rdirectly.supporting.....theconcreteslab, and about8 .m when they. actastransverse , . Fig... 43. Box girder composite bridge
 218 BRIDGE LAUNCHING COMPOSITE BRIDGES 219

workshop and field activities, and higher transport costs. Cost differences can be
reduced by fabricating full-width girder elements in the workshop, so that long—
itudinal field welding at the middle of the box can be avoided for the whole bridge
length. Exceptional road transports can often reach 25—27 m in length, 6 m in width
and 70—80 tonnes in weight, and a 6-m-wide girder may be suflicient for a
l2-m—wide composite bridge.
Unlike I-girders, U—girders have a wide, thin bottom flange, stiffened by an
orthogonal grid of stiffeners necessitating consideration of shear lag and out-of-
plane buckling. Differences in flange configuration and behaviour articulate web
response, and tension field theories developed in the context of plate girders may
not be readily extrapolated to box girders. In addition, there are differences between
the stability of perfect and imperfect plated structures, and classical elastic theory is
not satisfactory in the presence of initial geometric imperfections and locked-in
residual stresses.
Bottom plate thickness generally varies from 12 mm in mid-span to about 50 mm
above supports. The central portion of the bottom plate is often lightened, although
this requires thicker flange plates below webs and two additional longitudinal welds
for the entire bridge length. Savings in weight, reduced shear-lag effects and
enhanced local flange behaviour balance the higher welding costs.
As this stiflened plate is expensive, webs are often inclined to reduce the trans-
verse width of the bottom flange and the number of longitudinal stiffeners
(Fig. 4.4). This solution slightly increases workshop costs, especially in curved
bridges, but improves aesthetics, reduces the shear—lag effects and pier cap dimen-
sions, and facilitates control of local buckling. It also makes it possible to optimize
the transverse distance between top flanges, with enhanced control of distortional
stresses in the cross-frames (with related fatigue effects at the top flange level), and
optimum distribution of transverse flexural stresses in the concrete slab.
Composite action can be used to increase the buckling strength of the bottom
plate in the support regions of the continuous beam and to decrease shear-lag
effects. Shear lag is particularly severe in the regions immediately adjacent to the Fig. 4.4. Internal stiflening of a launched U—girder with inclined webs (author)
application points of concentrated loads, in flange plates that are wide in relation to
their effective length and in flange plates that are longitudinally stiffened. In a
continuous box girder, these circumstances coincide and this can generate excep- costs, with low-cost field activities and minor consequences on the dead load
tionally high shear—lag effects in the bottom flange over intermediate supports. bending moment diagram.
Bridge bearings are often placed directly under the webs to transfer the support Double composite action [91] is very effective in resisting negative moments in the
reactions directly. Dead load counterflexure points of the continuous beam are support regions. Its contribution may be extended to most of dead load effects by
relatively close to the intermediate supports, so that the effective span of the bottom casting the bottom slabs prior to the deck slab, and by precompressing them by
compression-flange is about 40% of the design span and the resulting span-to—width temporarily altering the support level of the superstructure as discussed in Section
ratio is well within the critical range. Finally, the bottom flange is often heavily 4.5.2.
stiflened against buckling under longitudinal compression. The web plates of U—girders are generally more heavily stiffened than those for
A substantial reduction in shear-lag effects due to longitudinal stiflening may be I-girders (Fig. 4.4). This is also necessary to control deformations during welding
obtained by stiffening the critical support regionsof the continuous beam with and assembly. On the other hand, plate thickness can in this case better adhere to
reinforced concrete slabs cast in situ directly on to the bottom plate (Fig. 4.5). structural requirements. In the areas more affected by shear stresses, the thickness
. These concreteslabs stabilize thebOttom flange and .avoid most, of stiffening of web plates may be presized (in mm) by [11]
-
220 BRIDGE LAUNCHING COMPOSITE BRIDGES 221

’ / / / /1 //
// / /’/1:ZZ////’/l///////’//////%Z///’//////k1/,7////77] The reinforced concrete slab is generally supported only on the top flanges
provided that they are not excessively spaced. It may also be supported on floor-
beams to obtain a bidirectional plate action in the deck slab, but transverse support
complicates recovery of the internal form tables. Left-in-place concrete or ribbed-
/ x' /’//////////7/////// sheet planks are used only when the U-girder is so thin as to make recovery of
internal forms excessively expensive.
Fig. 4.5. Concrete bottom slab for double composite action If the concrete slab is very wide, the use of two or several parallel U—girders
decreases the quantity of steel of the bottom plate. This solution is advisable
only for short spans and very slender bridges, especially when incremental
BL launching of the U—girders can be combined with their transverse shifting.
tw _ 10 + 1—66 (4.7) The total depth of the U-girder can be presized by [ll]

where B is the total width of the superstructure and L is the design span (both in (4.8)
metres).

-
The upper portion of webs includes flange plates, the dimensions of which where B is the overall width of the superstructure and L is the design span (both in
depend on the flexural strength required during launching and under the segmental metres). Some codes suggest keeping the ratio of the Overall depth of the composite
casting loads of the concrete slab, and on the type of connection with the concrete section (steel girder plus concrete slab) to the span length above 1/25, and the ratio
slab. Top flange width may vary from 600 mm for short spans to 1500 mm for long of depth of steel girder alone to span length above 1/30.
spans and wide deck slabs. Plate thickness varies with similar proportions, and can Again, in the presizing stage it may be useful to have an idea of the unit weight qS
reach 150~200 mm in exceptional cases. of steelwork referred to the total surface of the superstructure. For composite box
When the compression flange is restrained by a U-girder, the transverse frame girder bridges, the following relationship [ll] may be used (in kg/m2)
stiflness and the transverse moments applied to the shear connectors are more
uniform along the member than in two-girder systems. In spite of this, instability qs = 2.85L -|- 45 (4.9)
of top compression flanges before onset of composite action is controlled as in
I-girders. where L (in metres) is the main span of a” continuous beam with optimum span
When top lateral bracing is prOvided, the section becomes a closed section before distribution.
the onset of composite action, and this is particularly eflective both during launch
and during segmental slab casting. Such closing affords a high torsional rigidity and 4.2. Segmentation of the steelgirder
inhibits the rotational displacement of the box girder. This action further minimizes Field activities for a PC superstructure and for a steel frame to be incrementally
the warping torsional displacement and thus warping stresses. launched on to the piers are completely different. The basic reason for this
Although this problem is most serious for curved bridges, top flange bracing is difference is the material itself: PC versus steel.
usually provided even in straight girders for handling and launching. For curved A PC launched bridge is built on site. In most cases the bridge is cast in situ by
bridges, AASHTO [89] requires a minimum cross-sectional area of the diagonals of handling small, loose components (fluid concrete, reinforcement, strand, etc.) Even
the top flange lateral bracing based on the flange width between webs. AASHTO when the bridge is composed of precast segments, these are built in a casting cell
also determines the stiffness required for the bracing system based on the induced starting again from loose components, without any dimensional restraint. Segment
torque in section, the equivalent plate thickness, the torsional constant of the closed transport may cause some difliculties, but standardized segments (2.5—3.0 m wide)
box, and the angle measured between the longitudinal axisof topflange and the top generally cause only minor problems.
lateral diagonal. The situation is different for the steel frame of a composite bridge, and design is
Cross—frames are used also in U-girders to minimize distortion stresses and, if influenced by the fabrication method from the very beginning. In fact, the basic
curved, warping stresses as well. For curved bridges, AASHTO [89] provides a material is created in a foundry, transported to a workshop where it is modified,
criterion for determining the maximum spacing of cross-frames based on the painted in a paint shop, transported to the field and finally assembled to be
span length and the radius of plan curvature. The minimum area of cross-frame launched. Each of these steps involves dimensional restraints, and this explains
members is determined based on the, minimum width of box, the box depth, the the importance of assembly techniques and their influence on the total cost of the
longitudinal distance between cross-frames, and the web plate thickness. _ budge __ ._
 222 BRIDGE LAUNCHING COMPOSITE BRIDGES 223

Standard dimensions of steel plates vary from country to country. Maximum Finally, piece dimensions and weight and the available space govern the design of
plate width is determined by the roll width of the rolling mill, and in Europe is the assembly yard and the choice of hoisting equipment.
generally about 4.5 m. Maximum plate length and thickness depend on the volume Design characteristics of field splices affect the structure segmentation. If simpli-
of the ingot to be rolled, which in turn depends on the foundry. Maximum com— fication of field activities is a prime objective, high-grip bolt splices may be very
mercial length can exceed 30 m for plates thinner than 20 mm and 1.5—3.8 m wide. useful. Since these joints weaken the cross-section, they are often placed as close as
Maximum thickness is not limited by rolling operations but mainly depends on possible to the dead load counterflexure points of the continuous beam. Launch
welding requirements: 150 mm is generally an upper threshold. Maximum plate requirements impose the requirement that the splice plates of bottom flange are
weight is 15—20 tonnes, and minimum weight is 3—5 tonnes. composed of parallel plates that do not affect the central flange strip below the web,
Steel plates are delivered by truck or rail wagon. Maximum standard dimensions as this flange area is used as sliding surface during launch.
are about 24 m in length and 2.7 m in width, or sometimes as much as 33 and In most cases, however, full—penetration butt weld joints are preferred, since they
2.8 m. Commercial dimensions are strongly influenced by the standards of the offer many advantages in terms of aesthetics and maintenance costs. Since welding
individual countries. does not adversely affect the cross-sectional strength, field splices can also be
Steel plates are processed in the workshop to create the individual bridge com- distributed more freely in the structure.
ponents, the characteristics of which are influenced by the workshop handling Welded field splices are normally located on a single plane, as staggering the butt
facilities. Workshop organization has a significant influence on the fabrication welds of flanges and webs does not improve the performance of the girder. It is also
cost and affects the design itself of the steel structure. The fabrication sequence is easier to prepare the joints and maintain proper fit by flame cutting and levelling
generally as follows: when all splices are located in the same plane. An inherent advantage is to extend
the fillet welds of flanges to the web to the very end of the girder, because this
0 Fabrication preparation (programming of numeric-control equipment, planning
provides better support when the flanges are clamped together for temporary
of assembly sequences, workshop organization).
support during assembly.
0 Working of the individual components (piece marking, scribing of piece profile .
From an organizational point of view, field activities should be reduced to a
and positional references, flame cutting, grinding and edge preparation, drilling
minimum. It is much more diflicult and expensive to carry out segment alignment
0f holes, sandblasting and cleaning).
and joint welding on site than in the workshop environment. Maximum simplifica-
0 Assembly of segments (positioning of the individual components, locking and
tion of field activities suggests a subdivision of the steel frame into elements as long
tack—welding of plates, edge preheating, welding, straightening after welding).
and heavy as possible, and joint welding permits optimized location of construction
0 Segment completion (welding of stiffeners, minor plates and headed stud
joints for assembly purposes.
connectors, machining of surfaces at butted joints).
Consideration of these technological and economic requirements is useful in
0 Final cleaning and painting. GeOmetry and weld controls are distributed along
the theoretical approach to optimum" proportioning of the girders. Considerable
the whole fabrication process.
analytical work has been done in the area of optimization of plate girder design [5].
Bridge segments are usually transported to the assembly field by road or rail; river The usual approach is to relate the section characteristics to a given set of bending
or sea transport is adopted only in exceptional cases. Truck transport is generally moments and shears.
the most flexible solution as it permits reaching most sites. In this case, the most The web depth is assumed constant, and the number of flange cut-off points as
practical weight for bridge elements is 20—30 tonnes, depending on the type of well as the minimum length of each flange plate are decided from practical con-
trailer and the form of the piece. Maximum length is about 22 111; maximum siderations. Each span is divided into segments based on the number of plate cut-off
width is about 3.20 m if loaded horizontally and 4.50 m if loaded vertically. points and the shape of the moment diagram. For a span composed of n girder
However, exceptional transports can exceed these weights and dimensions. When segments, the bending moment distribution is known once the maximum moments
the piece length is critical, the element can reach 33—35 m in length, 3 m in height, Mi have been determined in each segment.
3 m in width, and 70—80 tonnes in weight. When the piece width is critical, the Either the total cost or the total Weight of the girder may be considered as an
element can reach 25—27 m of length, 5—6 m in width, and again 70—80 tonnes in objective for the optimization process. Most investigators chose the weight as the
weight. Railway transport is subject to more stringent limitations, as bridge com- relevant parameter [5], but the same approach can be followed on a cost basis.
ponents must be placed on standard wagons. Maximum length is about 32 m, Assuming a symmetrical I-girder with constant web depth h (distance between
height about 3 m, width about 2.5 m, and weight is 50 tonnes. The usefulness of flange plate centroids) and thickness 1“,, but with variable flange plates, the total
rail transport is also determined by the possibility of unloading wagons in a railyard weight Q for a span L is
,. close to the. site. _
\
 224 BRIDGE LAUNCHING COMPOSITE BRIDGES 225

easily adaptable to asymmetrical I-girders with different flange plates and vertical or
Q g y(htwL + Z 2Afib,) (4.10) inclined webs, as well as to U-girders.
As regards the bending moment to be used for calculation, a composite bridge
where y is the specific weight of steel, Afi is the cross-sectional area of a flange, and built by incremental launching is affected by three states of stress:
b, is the length of the appropriate girder segment. In an optimized proportioning, 0 The launch stresses, which depend on the self-weight of the steel frame, on the
the section modulus SX is related to the allowable edge stress a concentrated weight of support diaphragms, on the movement of the precam-
2: h2 M bered girder on to aligned bearings and on the effects of vertical misalignment of
SXgAfh+L=_ . (4.11) launch bearings in the transverse direction. After enveloping the launch
6 a
moments determined for uniform dead load and aligned bearings, the effects
Extracting Af from Equation (4.11) and combining gives of concentrated loads and precambers may be calculated in some characteristic
locations and superimposed. Transverse eccentricity of support reactions may
§y[htwL+Z(
n
'hoill— h “0]
2M—b- b-
(4.12) be taken into account by increasing the total support moment: a 30% increase is
often sufficient for girder optimization purposes.
0 The slab casting stresses, which depend on the segmental casting procedure.
Because a is constant, for a girder with a chosen web the weight Q is minimum.
They are computed by progressively applying to the steel girder the dead load
when the value of Z |M,b1| (based on the absolute values) approaches a minimum.
of the concrete slab and the concentrated load of the movable shuttering system.
If this criterion is applied to all spans of the continuous beam, it can be shown [5]
Progressive onset of composite action may be disregarded when optimizing the
that the summation Z Z [M1191 [j of the values of Z |Mibi| for all the j—spans must
steel girder.
be a minimum to yield the minimum girder weight.
o The service stresses, which depend on the dead load of the composite super-
Usually, a two-stage iterative procedure is followed to minimize ZZ lMibilj'
structure, the superimposed dead load, and the live load. The composite action
The first Step is to determine the optimum location of the cut-off points based on
is considered in the positive moment regions of the continuous beam, and the
the number chosen. The resulting flange plate lengths are independent of the
contribution of longitudinal slab reinforcement is considered in the negative
optimum web height, which is determined later by considering the final design
moment regions.
moment and shear.
In the first stage of the analysis, the bending moment diagram is determined for These three bending moment diagrams are superimposed at the end of each second
assumed lengths 1),. In the second stage, the values of moments are kept unchanged stage of the iteration process, and the girder segmentation is optimized for the
and the segment lengths are varied to minimize Z Z |Mibi |j with a search process. prevailing state of stress.
The new values of I), are, fed back into the first stage to recalculate the moments. When the final Cut-off point distribution is determined, the total length of
This is followed by. the second-stage computation to evaluate 19, again, and this the girder elements to be delivered to the assembly field is chosen based on the
procedure is repeated until Mi values are compatible with the values of [9,. technological requirements discussed above, including the lifting capacity of
A web choice can be based on the same procedure [5]. Using a modified section the gantry crane for field handling (Fig. 4.6).
modulus § for the continuous beam as a whole instead of that required for the
maximum moment ' 4.3. Assembly yard organization
The field activities for a steel frame to be built by incremental launching may be
g 2 Z: lMibilj (4.13) divided into two groups of operations alternating in time: the assembly of a new
O'ZLJ‘ section of the steel frame behind the abutment, and the launch of the whole girder
section on to the piers.
where Z Li is the summed length of all j-spans, setting dQ/dh = 0 one obtains the
The assembly yard is often twice as long as the end span of the bridge, to control
Web depth h for a chosen thickness tw as
overturning before the contact of the short launching nose with the first pier. When
site restrictions limit the space available, concrete counterweights may be placed on
h : 3—5 (4.14) the rear end of the girder to increase the safety factor against overturning.
(W The assembly yard often contains three girder elements. The length of girder
provided, however, that the depth-to-thickness ratio satisfies the criteria stipulated units should preferably exceed 20 m to reduce the number of field splices, and is
., by the applicable...standards. This optimization method .fo.r,,.symmetrical Lgirders is, ,. kept as constant as possible to avoid shifting the assembly bearings. With a three—
226 BRIDGE LAUNCHING COMPOSITE BRIDGES 227

overall depth. The flexibility of the steel girder eases lowering on to permanent
bearings on completion of launch, and when the sliding surface is higher than the
abutment back-wall, the embankment can be completed before construction of the
steel girder to improve rainfall drainage.
Each girder element is placed on adjustable assembly bearings located at the
quarters of its length to control deflection and rotation of the end sections, and
to ease access to the overhead welding areas of bottom flanges. In the case of two
I-girders, the assembly bearings are placed directly under the webs. When the
U-girder is composed of two L—girders and a central bottom plate, two additional
support lines are necessary, and field welding is not limited to transverse splices but
also includes two longitudinal joints as long as the whole bridge. In this case, the
assembly bearings may be widened transversely to support also the central bottom
plate (Fig. 4.7).
The vertical alignment of assembly bearings is set according to the planned
precambers by means of adjustment screws. In the case of curved bridges, trans-
verse alignment of bearings occurs contemporaneously. Then, the girder elements
are placed onto the bearings according to their moment length (which varies with
the temperature) to keep the butt weld clearance for field splices constant. This also
makes it possible to respect the positional tolerance of the support diaphragms on
Fig. 4.6. Wheeled gantry crane for field handling (author) completion of launch. Minor geometric corrections are made by joint jacking or bar
pulling, and hydraulic clamps finally lock the element edges at the field splices. In
segment yard organization, during positioning and geometry control of the rear
segment, the central segment is welded, and the front segment is painted at field
splices. In this case, three specialized working areas (assembly, welding and paint-
ing) may be organized at fixed locations, the launch steps are as long as a girder
element, and the length of the transportable units governs the assembly and launch
programme. Of course, tack-welding of field splices during assembly must
guarantee adequate stiffness not to jeopardize results during launching of the
assembled units towards the welding area.
As an alternative, as with PC bridges, the launch steps may be as long as several
segments. In this case, the field activities interfere with each other and must be all
completed before launching, but much less time is wasted in launch operations.
When the bridge is short, the steel girder may be entirely assembled behind an
abutment and then launched in just one operation. The temporary supports in the
assembly yard may be designed according to modular criteria so that the cost of a
larger number of assembly bearings can be amortized in more jobs: in this case, the
cost savings in the alignment and launch operations can be significant.
The transverse width of the assembly yard is equal to the width of the steel frame
plus a 3—4-m lane on both sides for access and operations of the gantry crane;
therefore, the total width approximately coincides with that of the completed super-
structure. The sliding plane of the steel girder is often more elevated than the level
of permanent bearings, as the launch bearings must meet technological require-
ments (related to the flexural rotations of the support sections of the continuous Fig. 4.7. Adjustable assembly bearings for L-girders with a central bottom plate
.«beam, and to- the adjustment and distribution of support reactions) thatincreasethe , fl
228 BRIDGE LAUNCHING COMPOSITE" BRIDGES 229

this phase, minor forced adjustments of the cross-sectional geometry are still 0 Complete the welding of the web. For deep webs, the vertical welding is
possible. sometimes divided into two or more sections following a back-step sequence,
The geometric tolerances allowed by standards may be excessive when they affect resulting in amore-uniform transverse shrinkage of the joint.
the bottom flange surface that comes in contact with launch bearings, since defects
of planarity in the sliding surfaCes may cause launch difficulties. For instance, BS These sequences generally permit adequate control of thermal shrinkage of welds.
5400 [92] accepts unintended deviations from planarity due to a plate misalignment Most butt joints: in: field splices of webs are the single V-type. For thick webs
in a welded butt joint when these do not exceed the lesser of 0.15 times the plate (usually above: 14.1mm) jadouble V—joint may be specified to reduce the amount of
thickness or 3 mm. A sharp 3-mm step in the contact surface may cause many welding and tobalanee the welding about both sides to eliminate angular distor-
problems when passing on to launch bearings, especially when these have PTFE tion. Design ofiflangeabutt joints to be field-welded depends on flange thickness. A
sliding surfaces. Therefore, local irregularities in the sliding surfaces should be double V—joint W—it‘li‘} half of the welding on the top of the joint and half on the
carefully deteCted, flattened and repainted before launching. The combined use of . bottom is suitable t‘oinhibit distortion but requires considerable overhead welding.
assembly bearings placed at the element quarters and of hydraulic clamps that align For this reason, a prequalified joint may allow the double V-joint to be prepared so
the element edges before welding is always advisable. that a maximum weld of 3/4 of the flange thickness is on the top and the remaining
High—grip bolt joints are seldom used in the main girders for aesthetic and 1/4 on the bottom: ‘ -
maintenance reasons as well as the design restrictions that affect optimum girder Automatic weldingflsesometimes used for the longitudinal field splice of U-girders
segmentation, and field assembly is almost always by welding. A welding sequence composed of twO Iii—girders. In wide bridges, transport restrictions may require the
for field splices [5] is described in Fig. 4.8. It is based on the following general presence of acentralib‘ottom plate between the two L-girders, and two longitudinal
outline, whereby both the flanges and web are alternately welded to a portion of field welds are necessary for the entire bridge length. Automatic welding often
their thickness after they are secured with tack welds: requires additional field—assembly operations for the support diaphragms and the
transverse bottom flange stiffeners, owing to the clearance requirements of the
o Weld the central portion of the double V—joint of both flanges (1/3 to 1/2 of total welding unit. Onathe other hand, savings in labour costs resulting from automatic
plate thickness) full width. welding may be important.
0 Weld a portion of the thickness of the single V-joint of the web (about 1/2) full Transverse field splices are always welded by hand. Finally, all field welds are
depth. cleaned, flattened and controlled, and paint cycles are applied to the field splices so
0 Complete the welding of the flanges. that the steel girder is launched without any need of additional interventions.
‘ Temporary bracing systems to be removed on completion of launch or after casting
the concrete slab are bolted, and high-grip bolts are also used for field splices
)( subject to fatigue (eg. longitudinal stringers that support the concrete slab in
Flange
mid-span) and for minor facilities (maintenance footpaths, de—watering pipes).
On assembly completion of a new girder element or group of elements, adjustable

s05
launching bearings installed in fixed locations along the assembly yard are jacked
against the main girders (Fig. 4.9). When the steel frame is supported on to the
launch bearings, the assembly supports are lowered and the launch can start.
As an alternative, the rear end of the girder section in the assembly yard can be
supported on a mobile saddle that lodges the thrust jacks [93]. In this case, rein-
forced concrete launch kerbs are necessary for the entire length of the assembly
yard. The rear support saddle is equipped with stainless steel bearings that slide
on neo-flon plates aligned along these concrete rails to ensure a minimal friction
web
coeflicient.
”////&/////
4.4. Launching of the steel girder
X The steel girder of a composite bridge carries its own weight and the dead load of
the concrete slab before the onset of composite action, and the service loads with
Fig. 4.8. Weldingseguence for.field splices . composite action. Although the. launch stresses are anomalous with respect to the
230 BRIDGE LAUNCHING COMPOSITE BRIDGES 2‘31

acting on a curved girder cause primary twisting moments that require an appro-
priate torsional strength for the structure. Therefore, U—girders are generally
preferred to two-I—girder systems properly stiffened for torsional rigidity.
Fabrication and launch of a curved U-girder do not involve particularly complex
operations, provided that the plan curvature is constant and the torsional precam—
ber is attained with deeper external webs rather than by rotating the bottom plate in
the cross plane. Segment geometry is repetitive (flexural precambers apart) and
cross-frames may be kept orthogonal to the main girders in order not to modify
their dimensions and to ease aSsembly. This also facilitates the analysis of launch
stresses.
In practice, however, the steel girder is not circular in plan, as the curvilinear
girder is fabricated in straight units jointed to shape. Since it is expedient to use
girder units as ”long as possible to reduce the number of transverse splices, the
distance between the curve and the local chord can be important. This requires
particular care during launching to keep the support reaction aligned with the web
axis, especially when the plan curvature is smaller than 500 m.
The launching nose is generally short, and stabilization of the front cantilever
may require additional cross—frames and vertical web stiffeners in the girder region
passing above the front support when the nose tip reaches the next pier. Instability
Fig. 4.9. Adjustable launch support for the assembly yard. Oh the left, one of the
in these front-support cross-frames (caused, for instance, by buckling of a compres-
multi-layered elastomeric bearings lodged in the rocking arm below the web (author)
sion member prOduced by transverse unbalance of the support reaction) may cause
collapse of the cantilever. Therefore, flatness and bolt tension of these stabilization
devices should be checked frequently during launching, along with the vertical
final service stresses, the steel girder is almost always designed for the loads applied
alignment of launching bearings. It is also advisable that the lateral guides of
on completion of launch and for composite resistance to live loads, and the launch
launching bearings aCt bidirectionally on the individual bottom flanges to prevent
stresses require only minor design adjustments.
transverse movements in both directions, which suggests adoption of Varying-
When the girder is launched along a straight alignment, lateral—torsional instabil-
thickness flanges with constant width.
ity of compression flanges is a prime concern. This form of instability may aflect the
When the plan curvature varies, "iaun’c’hi‘ng”bearings are designed so as to allow
bottom flanges of I—girders in the negative bending regions, and the upper flanges of
transverse displacements of the support points. The support reactions are much
both I— and U-girders in the positive bending regions. If the steel frame is composed
smaller than for PC bridges, and the shifting deviCes discussed in Chapter 3 are
of two I-girders, horizontal bracing is distributed at the top and bottom flange level,
easily adaptable. Lateral temporary piers are superfluous, as light steel brackets can
and cross-frames join the main vertical web stiffeners at the flange levels. If cross-
be anchored to the pier caps to accommodate large transverse displacements of
beams are used instead of cross—frames, their flexural stiffness must provide
launch bearings.
adequate control of rotation of the I-girders about their longitudinal axes. In the
case of U—girders, horizontal bracing is present only at the upper flange level, and
4.4. I. Launching bearings
cross—frames are always necessary to resist distortion.
A launched deck made of PC is a heavy and stiff structure. High support reactions
The situation is more complex when launch takes place along a curved alignment.
require launching bearings with large sliding surfaces, and, combined with a high
There are two possible solutions:
flexural stiflness of the superstructure, discourage vertical movements of the sup-
0 to use straight girders from one abutment to the opposite one by adjusting the port points. The small flexural rotations of the continuous beam support sections
plan curVature of the concrete slab with varying width cantilevers can be absorbed with neo~flon pads, the thinnest launching bearings are rigidly
o to use curved girders. supported onto the piers, and they can be replaced with permanent bearings on
completion of launch with minor deck hoisting.
The first solution is limited to shortest bridges with minimum plan curvature and The steel girder of a composite bridge is much lighter and more flexible, and webs
,doesnotinvolve greattorsional stiffness demand. In the second cas,e,,ve,rtica1,,loads . ., . .-.ar.e_.ve.ry....thin..- Dispersion o.f...suppor.t.. reactions into the webs requires long, narrow
232 BRIDGE LAUNCHING COMPOSITE BRIDGES 233

sliding surfaces, and the large fleXural rotations of the support sections dictate the economic reasons and in order not to excessively increase the length of the rocking
use of rocking launch bearings. This increases the overall depth of launch bearings, arm, whose elastic deflections adversely affect load distribution between rolls.
and major lowering operations on to permanent bearings are necessary on Lateral guide is usually assured by a roll with vertical axis that acts against the
completion of launch. lateral edge of bottom flange.
Moreover, the steel girder is not rectilinear in the vertical plane. Significant An initial improvement may be achieved with cable bearings. In these devices,
precambers are necessary to compensate the elastic deflections of the steel girder support reaction is distributed by a tensioned ring-cable that directly supports the
under its own weight and the distributed load of the concrete slab before the onset rollssThese bearings are thinner and more stable than roll bearings with rocking
of composite action. Minor precambers are also necessary to compensate the arm, and the descent of the steel girder in case of breakage is limited to few
deflections caused by the superimposed dead load and the time-dependent stress centimetres. Their stability and robustness encourage the use of cable bearings in
redistribution in the composite superstructure. Since the steel girder is light and the most delicate launches. When launch occurs along a curve, cable bearings may
flexible and the concrete slab is about four times heavier, the elastic deflections are be combined with orientation ball-plates and shifting supports, which respectively
significant, and the required precambers are significant as well. permit orientation and transverse alignment of the rolls under the girder webs.
Precambers are attained by assembling straight girder units to the planned curved Cable bearings placed on hydraulic jacks permit measuring of support reactions
shape. Therefore, the bottom flanges are affected by angular discontinuities at most and adjustment of their transverse distribution. This eases the launch of U-girders,
splices between girder elements. Since the girder units are as long as possible to particularly in the case of curved girders with torsional precambers attained by
reduce the number of splices, these angular discontinuities are often significant. rotating the bottom flange plate in the transverse plane. This also eases the control
When the angular discontinuity is a peak (support regions of the continuous of web instability in the support regions of the continuous beam.
beam) and this peak passes above a launch bearing, the support reaction tends to Several launches have showed that fabrication tolerances in the steel girder can
concentrate at the peak. Since the peak is a butt weld joint, it cannot be reinforced disturb the support reaction distribution so much that one launching bearing
with a vertical stiffener, and support reaction concentration may trigger local unloads while the opposite one resists the entire support reaction of the continuous
instability in the bottom flange and the adjacent web region. This is a potentially beam. These issues do not affect the design of launching bearings much, as these
dangerous situation, since instability aflects the compression flange of the devices are often overdesigned. However, transverse eccentricity of the support
continuous beam. reaction increases the launch stresses in one web and generates additional transverse
For all these reasons, launching bearings for the steel girder of a composite and longitudinal stresses that must be assessed in terms of both strength and
bridge are much more complex and refined than those for a PC deck. A launching instability.
saddle for a PC deck is a steel plate covered with a tensioned stainless—steel sheet Sliding bearings with PTFE skids-a can be used as an alternative to roll-based
and placed on a mortar seat, and its total depth rarely exceeds 10 cm (neo-flon pads bearings for low—friction launching. Compared with a PC deck, support reactions
included). A launching bearing for a steel girder is a complex mechanical device are much smaller, sliding friction is not the main problem, and in most cases the
with a depth that can exceed 1 m in the most delicate cases. Consequently, launch bottom flange can directly slide along lubricated PTFE skids without interposition
generally occurs along an alignment more elevated than the final one, and the steel of any stainless steel sheets. The support reaction is often distributed with multi-
girder has to be lowered on to permanent bearings on completion of launch. Large layered elastomeric blocks covered with a dimpled PTFE plate. These blocks are
vertical displacements complicate these lowering operations, which almost always aligned in rocking frames placed on ball—and-socket articulations or transverse pins
require progressive sequential adjustment at all of the piers, whereas just one pier is (Figs 4.10 and 4.11).
generally jacked when placing a PC deck on to permanent bearings. Elastomeric blocks are often thick: this does not greatly increase the total cost of
Launching bearings are much longer than those for PC bridges, to facilitate launch bearings, and improves both the transverse distribution of support reactions
dispersion of support reactions into the webs and to control web panel buckling. between the launch bearings and the longitudinal distribution of the contact stress
Bearing length is determined by taking into account: diflerences in elevation of during the passage of flange splices. In the most delicate cases, elastomeric blocks
launching bearings due to installation tolerances, diflerences in elevation of bottom may in turn be placed onto batteries of spherical springs lodged at the bottom of the
flanges of the steel girder due to fabrication and assembly tolerances, and angles at rocking frame.
transverse flange splices related to precambers. The contact stress of multilayered elastomeric blocks may be higher than in the
One type of launching bearing is composed of rolls assembled on a rocking arm neo-flon pads for launching of PC bridges, as the block thickness eases absorption
that follows the flexural rotations of the continuous beam. The number of rolls of local defects in the sliding surface. Again, the allowable mean stress in the
depends on the vertical load and on the roll diameter, and almost always varies elastomeric blocks governs the design of the sliding bearings, since unconfined
between 2-. and..8.-. ...The-.1toll.number. should be reduced to. a minimumboth for,-,- .. . PTFE can easily carry compressive stresses much higher than 30 MPa.
 234 BRIDGE LAUNCHING COMPOSITE BRIDGES 235

Fig. 4.10. Side View of a sliding bearing with elastomeric blocks lodged in a rocking
frame (author)

In the current state of practice, roll-based bearings (both rocking-frame and cable
bearings) and sliding bearings are complementary. Sliding bearings are suitable for
slow launch of high loads, whereas roll—based bearings are suitable for medium
loads and high launch velocities.
Most of the launch devices for PC bridges can be used to generate the launch
force. However, back—thrust systems require the presence of continuous contrast
kerbs and are therefore rather expensive.
The low dead load of the steel girder often suggests the use of tow systems with Fig. 4.11. Upper view of the sliding bearing of Figure 4.10. The low-friction lateral
bars or strands (Fig. 2.10). As an alternative, a winch placed between the abutment guide is made with Ferrozell (author)
and the rear end of the steel girder can provide the thrust force while a second,
opposite winch avoids wrong displacements and permits backwards pulling. This 4.4.2. Launching nose
solution, along with the use of roll-based launch bearings, results in the fastest The launching nose is frequently used during launch of steel girders to control the
launch operations (0.5—1.0 m/min). However, recovery of the front deflection at launch stresses and to ease recovery of the elastic deflection of the front cantilever at
the contact with the pier and repositioning of pulling and braking ropes slow the contact with the pier.
launching down,- and usually it takes half a day to launch a 50—70—m-bridge- section. Since the steel girder is relatively light and the launching nose must be much
The steel girder can be equipped with a rear nose that lodges the launch winches lighter to control the negative cantilever moment, the use of truss noses is very
and, if necessary, a counterweight that controls overturning in the first launch frequent. In this case, each nose truss can have a V—shaped cross—section with one
phases. bottom chord and two braced upper chords that enhance control of compression
Because of the higher friction at launching bearings, thrust or pulling devices are chord instability. This solution requires the use of a transverse diaphragm at the
often designed for 10% of dead load plus launch gradient. The thrust force should nose—girder joint that distributes the concentrated loads transferred by compression
be continuously monitored to avoid overloading the relatively delicate launch chords towards the upper flanges of the girder. This diaphragm (Fig. 2.55) may be
, bearings. ._.we_ldc_d___to___the_ girder and flame-cut on completion of launch.
236 BRIDGE LAUNCHING COMPOSITE BRIDGES 237

The nose is shorter than that for a PC deck because of the better negative uniformly distributed load both for the interior spans of the continuous beam and
moment behaviour of the steel girder. For a conventional continuous beam with for the front cantilever.
the bank span Lb shorter than the internal Spans, Lb /L = 0.7 + 0.8, a nose with An effective torsional rigidity should also be used to account for the cross-
Ln/L : 0.2 + 0.4 locates the stiffened girder region of the first internal support sectional distortion of the U-girder. Effective cross-sectional properties do not
above the front pier in the most delicate launch phases, thus often avoiding the affect analysis much, since the eflects of the varying stiffness along the continuous
need for additional stiffening of this girder region. When launching along a curve, beam are often secondary. However, the effects onstrength verifications are much
this solution also eases control of torsional and warping eflects in the front more significant.
cantilever. In the most delicate cases (wide girders, curved launch alignment, no horizontal
The hydraulic systems for recovery of front elastic deflection are designed for bracing at the upper flange level, as in Fig. 4.4), a 3D analysis may be necessary. In
large vertical movements, sometimes greater than 1 m. Since reduction of negative this case, the U-girder is modelled by 4-node or 8-node plate elements, and the
cantilever moment is less urgent than for a PC deck, the bottom nose chords are computational effort increases significantly. Therefore, 3D models are often used to
often rounded (Fig. 2.32). In this case, the front hydraulic pistons recover only a evaluate local effects and the effective stiffness, but general launch stresses are
portion of the cantilever deflection, and the remaining portion is recovered by the obtained from the 2D single-line beam analysis.
wedge action of the rounded nose. The longitudinal force generated at the front In the case of a steel frame with two or more I-girders, the launch stresses are
support by friction, grade and wedge action is high, but the support reaction is so analysed with 3D grid models of the entire frame, complete with bracing systems
small that its horizontal component is rarely able to overload the pier. and cross-frames or cross-beams. The resulting forces are enveloped before asses-
On the contrary, the horizontal force is easily able to overturn the deep launch sing the stress distribution and compression member instability. Sign inversion is
bearings, which must be anchored to the pier. Large rotations are requested for taken into account in enveloping results unless the frame is symmetrical with
launching bearings to accommodate this wedge action, and since all bearings are respect to its gravity axis. In this case, any member can be direCtly assessed for
loaded in this way at the nose arrival, their rotation capacity must be much greater the resulting force taken as a compressive one.
than that resulting from the flexural rotations of the support sections of continuous Similar considerations guide the analysis of the launch stresses in the cross-
beam. frames or cross-beams. This analysis is carried out by imposing vertical displace-
The rounded profile of the bottom chords affects the support action of the ments to the support nodes of the frame model equal to the prescribed maximum
launching nose. For the launch stress analysis in the continuous beam, the truss tolerances in the vertical alignment of launch bearings and in the fabrication and
nose is often modelled as a straight girder extension, with flexural and shear field-assembly processes for the steel frame. Larger support displacements are
stiffness determined with a specific 3D model of the cantilever nose. The rounded imposed to analyse the lowering phases for placing the steel frame onto permanent
contact surface is modelled by applying to the front support in the finite—element bearings on completion of launch.
model a downward displacement equal to the distance of the bottom chord from These analyses are particularly compiex in the presence of several I-girders owing
the projection of the bottom girder flange (launch surface). When the envelope of to the higher grade of redundancy of the cross-section, which requires enveloping a
the support reactions applied to the launching nose is determined, local stresses are great quantity of results. Finally, analysis with 3D models is practically unavoidable
analysed by applying these forces to the cantilever truss nose model. in curved bridges, since twisting of the curved girders is resisted by interaction with
adjacent girders. _.
4.4.3. General launch stresses and instability Additional analysis features derive from the irregular profile of the steel girder,
The launch stresses in the steel girder are determined by criteria similar to those precambered to reach the final design geometry after addition of the concrete slab.
examined for PC decks. The use of internal action envelopes instead of hundreds of Since the girder is flexible before achieving the composite action and the concrete
diagrams is still advantageous, although single—line modelling of the continuous slab is heavy, precambers — often of some tens of centimetres — are important.
beam is often insuflicient and the stress envelopes are in this casedetermined by Precambers follow the span sequence of the continuous beam, and if the spans
using 3D models. are not constant, the support points of the cambered girder are not aligned during
A 2D single-line analysis may be sufficient for rectilinear U—girders stiffened launch. The situation is similar to that discussed in Section 3.6.4 in relation to the
during launch by hOrizontal bracing at the upper flange level. Shear—lag effects creep deformations of PC deck. A180 in this case the girder is analysed as a recti-
(non—linear distribution of longitudinal flexural stresses due to in—plane shear flex- linear one, and vertical displacements equal to the local diflerence between the
ibility) can be important in the bottom flange plate. If this is wide, it cannot fully straight girder and the precambered geometry are imposed on the support nodes
contribute to the flexural stiffness of the cross-section, and an effective width must of the finite-element model.
bewused for ..analysis......BS 5400 [9.2] provides effectiveiwidthratio coefficients for . ,
 238 BRIDGE LAUNCHING COMPOSITE BRIDGES 239

Bending and shear envelopes are assessed for both strength and stability. A plate interrupted and the value of this ‘critical stress” was calculated [94].- The present
girder subject to mainly flexural stresses is affected by three modes of instability, all means of testing and calculation confirm that, beyond this critical stress, the growth
of which can arise during launch: rate of transverse deformation does not increase, but tends to decrease until the first
plasticization occurs. Only at this point does the deformation begin to increase
a The lateral-torsional instability of compression flange: a beam loaded in the
more quickly than the load, and the plate collapses. Therefore, the critical stress
vertical web plane suddenly becomes unstable and twists laterally under a
represents a border between two different states of elastic equilibrium, and does not
critical load smaller than the flexural strength.
constitute the limit value of the carrying capacity, which is much higher.
0 The torsional instability of compression flange, partially related to the previous
The ratio of the ultimate stress to the critical stress increases with the plate
mode and controlled by the vertical web stiffeners and the braCing system.
slenderness, generally varies between 3 and 8, and for the usual web panel slender-
0 The local instability of web panels, along horizontal lines in the case of pure
ness of composite bridges is 5—6. However, the web panels are not perfectly flat and
bending, and along inclined lines in the presence of shear forces.
are often aflected by residual welding stresses. Test models showed that, because of
Lateral-torsional instability of the I—girders is controlled by the cross-frames or this, they use a portion of the postcritical dominion even at the serviceability limit
cross-beams, which avoid relative movements between flanges, and by the horizon— state, and the safety factor from the ultimate stress is therefore lower. Moreover,
tal bracing, which resists the torsional forces and distributes the concentrated the postcritical behaviour of a non-stiffened plate is different from that of a stiflened
stiflening action of cross members along the girders. When launching occurs plate because of the postcritical effects of web stiffeners.
along a curve and the I-girders are closely spaced, warping phenomena can also Several standards base the analysis and the ultimate—limit-state assessment of
affect the steel frame as a whole. In this case, the transverse distance between girders steel and composite sections on the concept of effective width. This concept is
should be increased, since when the cross-sectional moment of inertia about the applied not only to webs, but also to flanges and, in the U-girders, to the bottom
horizontal axis is smaller than that about the vertical axis, the cross-section is stable plate zones between longitudinal stiffeners. Some recent codes (EC3 [90], EC4 [85])
against lateral-torsional instability. Since consideration should also be given to ofler significant advances with respect to former texts, and introduce non—linear
possible instability of constituent parts of the cross-section, the adoption of U— instability theory. This approach makes it possible to take into account the post-
girders offers a better control of these phenomena. critical behaviour of slender plates subject to normal and tangential stresses,
All steel beams should be restrained at their supports against rotation about the according to safety factors from the ultimate limit state deriving from the most
longitudinal axis.'Direct restraint with cross members is impossible during launch recent research.
because of the longitudinal migration of the support points, and indirect restraint
must be attained by transversely locking the girder flanges at the launch bearings,
and by enhancing the interaction between the components of the structural system. 4.4.4. Local launch stresses and instability
BS 5400 [92] provides instructions for optimum sizing of horizontal bracing The incremental launching construction method is particularly suitable for two—
(3.1.1 1"

between flanges and of restraining cross-beams. Appendix F of EC3 [90] takes girder or U-girder composite bridges with steel frames that have only two support
into account the imperfections of welded girders and provides complete infor- points at each pier. This eases the launch, reduces the cost of launching bearings
mation about lateral—torsional instability of symmetrical and non-symmetrical and simplifies the analysis of the launch stresses because of the lower level of
sections. transverse redundancy.
Buckling of web panels is covered by all standards. This form of instability does Complex states of stress arise in the girder regions immediately above launching
not represent a potentially dangerous situation like instability of slender compres— bearings. The continuous beam behaviour of the steel frame, the movement of a
sion members, since the panel edges are immobilized by flanges and vertical precambered girder above aligned launch bearings and thermal gradients generate
stiffeners, and the growth of out-of—plane elastic deformations is resisted by normal and tangential stresses in the longitudinal direction. Torsion and distortion
membrane stresses that develop a natural stiffeningaction in the panel. ' resulting from misaligned launch bearings and geometric imperfections in the girder
After buckling, the web plates develop a reserve of strength because of the elastic generate normal and tangential stresses in both the longitudinal and the transverse
work of membrane stresses. In the precritical flexural state, when stresses are low, direction. Local vertical compressive stresses are generated by dispersal of support
the girders deform keeping sections plane. At buckling, axial stresses begin to reactions into webs. Finally, both the launch friction and the gradient of the launch
redistribute towards the stable zones of web panels, the slope of tension diagonals plane generate additional local stresses, which also compromise the conservation of
changes and a new equilibrium arises based on higher stresses. plane sections.
Years ago, instability of web panels was approached with manual methods. Since most of these stresses are significant, the local state of stress must be
....When,-the ,.plate__.___b_e_gan to._assume, a, transverse deformation,therloading ,was, . checked with adequatestrength criteria.
 240 BRIDGE LAUNCHING COMPOSITE BRIDGES 241

4.4.4. I . Vertical compressive stress where LW is the length of patch loading measured on the longitudinal web edge, 1W is
On completion of launch, the steel girder of a composite bridge is supported at its the web plate thickness, kw is the net depth of the web panel measured clear between
final pier sections, where the vertical web stifleners distribute support reactions to flanges, and fy is the nominal yield strength of steel. In practice, however, the
the entire depth of web panels and prevent local instability. A correct distribution vertical cOmpressive stresses due to dispersal of launch support reactions are almost
of stiffened and lightened sections results in optimum bending and shear behaviour always much higher than this threshold.
of the steel girder, which can resist the dead load of the concrete slab before the A non—stiflened web panel subject to a concentrated support reaction applied
onset of composite action. through bottom flange is affected by three collapse modes that depend on load
During launching, the support sections of the continuous beam are devoid of any intensity and on the slenderness of the web panel (Fig. 4.l3):
local stiffening. Although the steel girder is much lighter than the final composite
superstructure, its ability to resist concentrated support reactions applied in the 0 Local yielding is caused by web yielding immediately above the load, followed
plane of webs must be assessed in any cross-section, in terms of both strength by the plastic deformation of the bottom flange.
and stability. 0 Local buckling is triggered by localized buckling in the lower part of the web
The vertical compressive stress in an unstiffened web panel due to a support panel, for a vertical depth about 50 times the plate thickness [1 1], followed by
reaction applied to the bottom flange may be calculated on the assumption that web yielding and the onset of a plastic mechanism in the bottom flange.
0 General buckling of the web panel extended to most of its depth is assessed by
the load is dispersed uniformly. Load dispersion may be assumed to take place at an
angle of 600 from the line of load application through the thickness of the bottom the code equations for panel buckling under shear and bending.
flange, and at an angle of 450 through the web panel itself (Fig. 4.12). In terms of local effects at the bottom flange level, the cross—sectional strength RR
Stress concentration due to angular discontinuities in the bottom flange is con- for a support reaction applied orthogonally to the web plate through bottom flange
trolled by the flexibility of launching bearings and by the local flexural stiffness of is the lower value of the local yielding strength Ry and the local brickling strength
the flange plate. In the most delicate cases (launching bearings with long rocking Rb.
arms, long girder units with precambers applied only through field splices, inclined According to EC3 [90], the local yielding strength of an 1-, H- or U—section is
webs), a 3D finite-element analysis of the interaction between launching bearings
and the steel girder may be necessary.
Ry : (Lb + Ly) tw
4§L_ I (4.16)
BS 5400 [92] permits disregarding the effects of in-plane patch loading on the VMI
longitudinal edge of web when the vertical compressive stress oz in the web plate is
where Lb is the length of rigid patch loading measured on the launch surface (lower
(in MPa)
edge of bottom flange — Fig. 4.12), the partial safety factor from instability is 7M1 :
1.1 and Ly is
1 355
0,;5 3 ——~EL—— /———- 4.15)
r.-—-2z
fyv Lwhw fy
£31 1-— 1:19é 2
(
y—bfw .fy 4.17
()

where UK is the longitudinal normal stress in bottom flange and [9b and tb are
respectively the width and thickness of the bottom flange plate, with the limitation

local yielding local buckling panel buckling

lliliii
m V
IHIIIIII
a—L —o
b Illllliil llllllili

,.,,Fig- 41.2..._._1.2i512_er5ql__oflaad through anunstiffened_web__.. Fig. 4.13. Local buckling,,,modes ave launch bearings
242 BRIDGE LAUNCHING COMPOSITE BRIDGES 243

bb /tb _<_ 25. Always according to EC3 [90], the local buckling strength of web panel the whole section, and add to the longitudinal compressive stresses produced by the
is continuous beam negative moment.
The longitudinal axial stresses in bottom flanges above launching bearings at the
Rb = gag/MW tb/tw + 3(tw/tb)(Lb/hw)] (4.18) ultimate limit state depend also on the type of girder. For an I-girder, the stress-

-.
VMI strain relation is bilinear. In the absence of instability (Class 1 and Class 2 sections

”Hz-«-
where hw is the net depth of web panel measured clear between flanges, with the according to EC3 [90]), collapse is not reached when the first yield occurs at a
limitation Lb/hW 5 0.2. moment My, but when the external moment reaches the plastic moment capacity
The design support reaction RD is determined by taking into account the of the cross-section Mp. The ratio of the plastic moment to the first—yield moment
expected misalignment of launching bearings and the geometric irregularities in depends on the cross—sectional geometry and on the type of steel. For Class 1 and
bottom flanges due to fabrication and assembly tolerances. A 30% increase of Class 2 sections, one may very approximately take Mp E 1.15My. Class 3 sections
the maximum theoretical support reaction Rm21X is sometimes suflicient [39], pro- have MR 2 My since instability prevents from exploiting the postyielding field.
vided that stringent support and geometry tolerances are prescribed and actually Finally, the flexural strength of Class 4 sections is governed by instability.
respected. The behaviour of a U-girder subject to negative moment and shear is much more
complex. Because plane sections do not remain plane, simple beam theory is not
RD 2 1.312max (4.19)
adequate to predict flange stress diStribution. If the webs are assumed stocky and
Based on the design load factor yL prescribed by the applicable standards, we only the bottom flange is slender, first yield occurs in the outer panels of the flange
must have where normal and tangential stresses are maximum due to shear lag. Since this plate
panel and the stiffener bounding may be slender, the applied stresses are magnified
VLRD 5 RR (4-20) at the plate surface due to non-linear elastic out-of—plane buckling. As a conse-
The flexural strength MR of the cross-section is determined as the lower value of the quence, both illy and Mp can be smaller than their theoretical linear-elastic values.
plastic moment and of the critical moment for general web-panel buckling. Then, Two assumptions may be made [5]:
letting MD be the design bending moment of the cross-section, EC3 [90] requires o The load-carrying capacity of the U-girder is exhausted when yielding of the
that an additional stability criterion be satisfied to account for the contempora- outer panel occurs.
neous risk of both buckling modes: 0 The bottom flange has the capacity to redistribute the stresses across its width so
i/LRD + i/LMD : 15 (4.21) as to equalize the available capacity in the lower—Stressed central area.
Rb MR In reality, the load-carrying capacity depends on whether the outer panels continue
and, of course to carry their share of the load as redistribution takes place or they rapidly shed
load.
VLMD 5 MR (4-22)
In order to understand the behaviour of the compression flange of a U—girder
Finally, the state of stress in the webs is assessed according to the conventional above launch bearings, the behaviour of each component must be predicted up to
criteria of biaxial strength, which depend on the influence of the tangential stresses and beyond collapse, including the effect of boundary conditions. This analysis
as well. includes the eflects of shear lag, restrained warping, cross-sectional distortion and
imperfections.
4.4.4.2. Longitudinal compressive stress Shear lag eflects are covered by many standards. Among the parameters con—
Support reactions are not orthogonal to the launch surface. The steel girder is sidered are flange plate orthotropy, ratio of flange to web area, cross-sectional
almost always launched along an inclined plane, and the vertical components—of shape, aspect ratio of flanges, position of point loads, combination of loads and
the contact forces (i.e. the continuous beam support reactions) are not orthogonal support conditions. BS 5400 [92] provides effective width ratio coeflicients for
to the girder gravity axis. Moreover, as discussed in Section 2.8, the friction in uniformly distributed load both... for the interior spans of the continuous beam
launching bearings rotates the resultants and generates additional longitudinal and for the front cantilever. BS 5400 [92] also provides the effective web thickness
components. rm < tw to be used for bending stress analysis of beams without longitudinal
These systems of forces are easy to determine when the local launch plane stiffeners. '
gradients and the distribution of continuous beam support reactions are known. With the exception of circular sections and square thin-walled sections, warping
The longitudinalstresses generatedby these forcesmay behigh beforedispersionto , ,. of the cross-section occurswhen the sectiontwists [53]. If this warping is resisted,
 244 BRIDGE LAUNCHING COMPOSITE BRIDGES 245

restrained warping stresses are induced in the section. Simple equations are 4.4.4.3. Stress verification
proposed by some standards to predict peak warping stresses in box sections. When passing above launching bearings, the lower region of webs is subject to a
When the steel girder is curved in plan, the additional longitudinal stresses complex state of stress. Since the lateral guides of launching bearings keep support
produced by non-uniform torsion and restrained warping may be significant. reactions aligned with web axes, the transverse flexural stresses are relatively low,
AASHTO [89] prescribes reductions in the allowable normal stresses in the flange the state of stress in webs can be assumed as a plane one, and the ideal stress 51d
plates based on the geometric characteristics of the cross-section and the framing resulting from the Huber—Mises equation
system as well as the influence of non—uniform torsion.
As discussed in Section 3.5.2, eccentric loads deflect and rotate the box and aid : \/cr,7g — (7v + a; + 373332 (4.24)
distort the cross-sectional shape between internal diaphragms. Although distortion
is small in terms of displacement, it may give rise to significant transverse bending can be directly compared with the allowable stress. .
and longitudinal warping stresses. Several standards provide equations for the‘ , The longitudinal stress UK and the vertical compressive stress 02 generated by
analysis of these stresses as well. dispersion of support reaction have the same sign and decrease along the web
The overall load behaviour of the bottom flange plate of a U-girder can be depth. Even if the tangential stress s is lower at web edges than at mid—height,
predicted with an elastoplastic large—deflection analysis. The eflects of residual the highest values of 01d are almost always reached near the fillet welds of web to
stresses and the initial lack of flatness can be allowed for in the analysis, although bottom flange, which must therefore be checked accurately during fabrication of
it is difficult to measure the geometric imperfections. Residual stresses have a plate girders. .
predominant weakening effect on practical plates in compression. Analyses have Similar considerations apply for the bottom flange plate of U-girders:
shown that geometric imperfections can reduce plate strength significantly, but
cylindrical out-of—plane deformations have little weakening effect. Tests showed 01d 2 ‘/ 03% + 313%, (4.25)

that the transverse fillet welds between plates and diaphragms do not reduce
where the distribution of the longitudinal normal stress 0X depends on vertical and
plate strength and there is no significant weakening in overall strength, while
horizontal bending, restrained warping, and shear lag, and the torsional tangential
stepping at a butt Weld, particularly where a thinner plate is connected, reduces
stress e can be high when launch occurs along a curve. The combined result of
strength.
torsion, restrained warping, shear lag and normal flexure is difficult to determine,
The approach adopted in [5] states that for uniaxial compression the ultimate
and the worst case of either may be superimposed to obtain a conservative estimate.
stress is
EC3 [90] provides similar criteria for the ultimate limit state assessment of webs
01,1, =fy(l —
fy ) < kfy (4.23) under concentrated support reactions and design shear forces respectively below
816013,. and above 5.0%. of the plastic shear strengthofthe support section. In the first case,
the strength criterion is
where fy is the yield strength, 00, is the critical buckling stress, and k is the ratio of
the applied stress to cause first surface yield to the yield stress. The factor k is 2 2
Um and
+ __.._ _ 0m
_ 02a S 1
obtained from large-deflection elastic analysis and varies with the width-to— 4.26
(fyd) (fyd) fyd fyd ( )
thickness ratio, assumed initial imperfections and the degree of end restraint present.
For all these reasons the diagram of longitudinal normal stress in the support where am and 02a are the design values of local stresses (stress values multiplied by
sections of a continuous U-girder differs from the theoretical Navier diagram and the load factor) and
reaches much higher values at the lower edge. Some of the causes of stress non-
linearity depend on the tributary bridge length at launching bearings, and their ‘fg/d : y—hl) (4.27)
effects are therefore critical in the front support section when the nose tip reaches
the next pier. If the end span of the bridge is shorter than the internal spans, the where the material safety factor for steel is yMO : 1.1. Equations similar to (4.26)
length of the launching nose may be chosen so as to locate the stiflened region of apply in the case of design shear forces higher than 50% of the plastic cross-
the first internal support above the front pier when the nose tip reaches the next sectional shear strength, and when the flexural strength is determined by taking
pier. If the stresses due to this long cantilever are excessive, the nose may be into account the plastic stress redistribution within the cross-section.
lengthened and the stiffened region of the first internal support may be extended Finally, high transverse flexural stresses arise in bottom flanges and webs when
launchwards with additional cross-frames. As an alternative, a light front-stayed the support reactions are not centred along web axes. For this reason, the transverse
,-..system may be usedtocontrolrthe excess ofnegativemoment... ,, _.._.,_width.-..o.f.launching.bearings. shouldbe kept to, a minimum.
 246 BRIDGE LAUNCHING COMPOSITE BRIDGES 247

4.5. In situ casting of the concrete slab 4.5. I. Casting methods and cracking control
In most cases, the reinforced concrete slab of a composite bridge is cast in situ on The casting procedures for the concrete slab have technical and economic implica—
completion of launch of the steel girder. As the steel girder has a stable elastic tions and should therefore be studied and prescribed carefully. Concrete casting can
behaviour whereas the concrete slab is afiected by time-dependent phenomena, occur in one of two ways:
stress redistribution occurs within the structure over time.
0 on a temporary or left-in—place form as long as the entire bridge
Without prestressing, cracking of the concrete slab is practically unavoidable.
0 with a movable shuttering system (MSS) that follows a segmental casting
. For typical bridges, the most critical tensile stresses are induced in the concrete slab
sequence along the steel girder.
during construction. Even adopting specific sequences of segmental slab casting, the
portion of dead load resisted with composite action and the superimposed dead The costs involved in slab casting on a general falsework supported on the ground
load generate tensile stresses in the support regions of a continuous beam. The live are acceptable only in very particular cases, and in most cases the falsework is
loads increase negative bending tensile stresses, but longitudinal tensile stresses are replaced with left-in—place forms. Materials suitable for use as permanent formwork
often present even in mid-span because of drying and thermal shrinkage, creep and are: -
the relatively low compressive stresses resulting from composite action. Finally,
1 precast concrete
longitudinal cracks over the entire bridge length are frequent in the slab regions
precast concrete acting compositely with steel beams or lattices that are
above the steel girders in the case of long lateral cantilevers.
eventually embedded in the overlying in situ concrete
Slab Cracking affects the concrete cover of the top reinforcement. A cracked
3 galvanized profiled steel sheeting, alone or welded to steel beams that are
slab is subject to the adverse effects of water and de-icing salts, and this condition
eventually embedded in the in situ concrete
is particularly severe when the wearing surface constitutes an integral part of the
4 reinforced plastic or fibre cement sheeting or similar.
concrete slab. In most cases, an asphalt surface is used as a protection, but
watertightness cannot be assumed even with a protective membrane between Particular care must be exerted to prevent accidents from occurring when materials
the asphalt and the concrete slab. Several standards stipulate an increased con— of a fragile nature — e. g. some of those given in (3) and (4) ~— are used as permanent
crete cover to top reinforcement to provide for cracking and deck wear, but this is formwork. Solution (3) has been frequently adopted in Japan [97,98] to obtain
inconsistent with the practice of placing the reinforcement near the surface of the concrete-filled I—beam grid decks. In spite of weight saving and the possibility of
slab to control cracking. However, when the concrete slab is designed correctly, placing the semi-prefabricated forming-reinforcement system on to the steel girder
crack widths of about 0.3 mm are generally unable to cause corrosion of prior to launch, this solution is often limited to transverse spans of 3—4 m, and on
reinforcement steel. these spans it is more expensive than precast concrete planks.
Since slab cracking is caused mainly by flexural stresses and concrete shrinkage, Left—in-place forms are mostly made with precast concrete planks supported on
several standards provide instructions in both these directions. Adoption of mini- to the top girder flanges and reinfOrced With steel lattices or punched I—beams that
mum reinforcement ratios in the slab regions where the tensile stresses exceed the resist the load of fluid concrete, construction plant, and personnel Without excessive
effective tensile strength of concrete is often required. The minimum reinforcement deflection.
ratio is often set to 1% of the gross cross-sectional slab area, increased to 13—15% Precast planks may be considered as either structurally participating with the
in the support regions of the continuous beam. In these regions, the tensile stress in overlying in situ concrete under the action of loads imposed on the slab after
the reinforcement is limited to 60—70% of the allowable tensile stress to further casting, or structurally non—participating. In both cases, the composite slab
control crack width. Additional provisions rule quality and compactness of con- obtained by casting fresh concrete on to precast planks is highly sensitive to the
crete, the water-to-cement ratio, and the use of admixtures, as all these parameters time-dependent properties of concrete.
influence both tensile strength and permeability of concrete. " Drying and thermal shrinkage causes cracks at the interface between the planks
Crack width depends mainly on the distance to the nearest reinforcing bar placed and in situ concrete. Differential drying shrinkage may be studied with the conven-
perpendicular to the plane of crack, the distance to the neutral axis and the average tional age—adjusted effective modulus method. A method of evaluating the effects of
surface strain at the point considered. Several standards (EC2 [95], EC4 [85], BS thermal shrinkage is discussed in the following pages and may be easily adapted to
5400 [55,96]) provide instructions for estimating crack width and for adequately the interaction between precast planks and in situ concrete.
spacing the tension bars in moderate conditions of exposure. For instance, for Where composite action between the precast planks and the in situ slab is relied
nominal concrete cover not exceeding 30 mm, BS 5400 [55] determines maximum on, the design of the composite two-layered slab must satisfy several code require-
bar spacing based on the stress in the reinforcement and the amount of ments, which often turn out to be excessively restricting. On the other hand, wheel
,,r,e,dist,ribut_ion_,at the. ULS. as. W611’,&.S:QI,1.9aSti11g conditions of the ,..c,0ncrete slab. __-._1o_ads.___afiect__the..p.r_ecas.t_._planks. directly, and. durability of reinforcement is a prime
248 BRIDGE LAUNCHING COMPOSITE BRIDGES - 249

concern in both the transverse and longitudinal directions. This is especially true Precast planks are also expensive: their prime cost increases due to transport and
when the bridge has spaced girders Without intermediate longitudinal stringers that handling, and the plan framing of the steel structure is restrained by plank dimen—
create additional support lines for the planks. sions. Special attention must also be given to the provision of a suitable seal
Differential drying and thermal shrinkage affects structure durability, since between the steelwork and the planks to minimize the possibility of corrosion
humidity passing through full-depth cracks in the in situ concrete in the presence throughout the life of the bridge, and sealing causes additional costs. Several
of defects in the waterproof membrane gathers at this cracked interface. If the additional defects adversely affect the final result, as discussed in Section 4.6.2.
concrete cover of plank reinforcement complies with code requirements both Precast planks are often assembled proceeding from one abutment towards the
above and below the bar grid, the overall plank thickness soon becomes opposite one, as this facilitates access to the working area. Slab reinforcement may
excessive. Even neglecting shrinkage cracking at the interface between the two be assembled in the same direction to start earlier, and continuous slab casting in
slab layers, a minimum upper cover is necessary for plank robustness under the same direction is the most intuitive and easiest solution. With this casting
casting loads [99]. The overall thickness resulting from a minimum upper method, however, longitudinal tensile stresses (often in the range 1.5—3.0 MPa,
cover, a reinforcement grid designed for the flexural lower—edge stresses in the increasing with the span) arise in the continuous beam support zones when span
composite slab, and a bottom cover complying with code requirements, is often segments are cast.
excessive as well. / To limit tensile stresses and related full—depth cracks, segmental slab casting is
Thick planks also result in thin in situ slabs and fatigue problems at the trans- often prescribed aCCOrding to the discontinuous sequence of Fig. 4.14, in which the
verse joints between planks. Longitudinal continuity between precast planks may be segments in the positive moment regions are cast first, followed by pier segments.
achieved with lapping reinforcement projecting by units, post-tensioning, embedded This segmental casting sequence reduces or avoids dead-load tensile stresses in the
steel plates to be field—welded after plank assembly or high-strength bolts. All these concrete slab. After the concrete has cured rendering the composite action effective,
solutions may be affected by fatigue problems under repeated loading, are expensive only superimposed dead and live loads affect the composite structure. This casting
and involve a lot of field splices. For all of these reasons, the precast planks are procedure is more expensive than continuous casting, although it permits slight
often considered as structurally non-participating. Q weight savings in the steelwork because of the progressive onset of composite
Where the precast planks are structurally non-participating, account should be action.
taken of the effects of differential shrinkage or composite action that may adversely Complex step—by-step analyses of slab construction are necessary, since when a
affect the structure. Some standards (e. g. BS 5400 [55]) prescribe that requirements new segment is cast in a span, cracking tends to occur at the end of the hardened
for cover of reinforcement and crack control applicable to the in situ slab should be slab segment in the previous span [100]. Evaluation of precambers to be applied to
satisfied ignoring the presence of the planks. Moreover, the clear distance betweerr
precast planks and reinforcement embedded in the in situ slab should exceed the
maximum nominal size of aggregate used in the in situ concrete. This requirement is
often stringent, since it penalizes the lever arm of transverse reinforcement in the El><|><l><l><l><|><l><b§><l><l><|><l><l><l><l><l>§<l><l><l><t><l>4><lflf
middle of the span between the steel girders.
However, the weakest point of non—participating planks is self-weight. This
lIXNXIXMXl>§<I><l><l><l><l><l><l><lE

L47 LJ \_l
additional load does not produce any advantage in terms of flexural strength of
the composite section, and since haunched planks for varying-thickness slabs are
expensive, the cOnstant-thickness composite slabs further increase dead load of the @ @IXIXIXIXIXJXDi
bridge.
The cross-section of a continuous composite bridge with 60-m spans, 13-m—wide
deck, and two I-girders spaced about 7 m, might be composed of a25-cm in situ m am E

‘
slab, 5—cm non-participating (mid-span propped) planks, and steelwork for 200 kg/
m2. The same composite resisting section weighs 0.950 tonnes/m2 in the presence of
@

LJ
precast planks, and 0.825 tonnes/m2 without them. If superimpoSed dead and live
loads are about 0.750 tonnes/m2, the design load of the bridge increases from

r‘i
VLJ
1.575 tonnes/m2 to 1.700 tonnes/m2, and the weight of steelwork increases by
8%, i.e. 16 kg/mz.
Fig. 4.14. Segmental casting sequence for the concrete Slab
 250 BRIDGE LAUNCHING COMPOSITE BRIDGES 251

the steel frame is more difficult as well, because of differential concrete creep
between the slab segments. Stress evolution in the composite structure during
construction is analysed based on the modular ratio appropriate for the stage of
construction and the type of loading, and using an effective slab width compatible : E
with the expected shear-lag effects. 1
As an alternative, segmental slab casting can take place in a mobile formwork. In
medium-length bridges, a reasonable length of slab segments (say, about one-third
of the span) eases amortization of an MSS, which can therefore be of better quality.
The use of an MSS avoids the dead load increase and the other disadvantages of
precast planks, although discontinuous slab casting causes access problems through
the solutions in continuity in the slab (Fig. 4.15), and requires prudent casting
procedures that avoid torsional loads in the steel frame. This requirement is
particularly stringent when the form table between upper flanges is recovered by
frontwards extraction, Since this often requires removal of the horizontal bracing
between upper flanges before form setting. Finally, prevention of temporary tensile
stresses is complicated by the dead load of the casting equipment.
An MSS with hydraulic controls (Fig. 4.16) is rather expensive. Its own weight
may reach 250 kg/m2 of horizontal form surface, and the prime cost may be €25“
3.0/kg. Suspending the form tables to the carrying trusses with vertical through bars
reduces the total weight of the MSS but produces a great number of holes in the Fig. 4.16. Hydraulic MSSfor 22 m long, 14.3 m wide Slab segments (author)
concrete slab and complicates reinforcement assembly (Fig. 4.17). The assembly.
and dismantling costs, transport costs and crane demand, they all depend on the, -
weight of the MSS. Although the contractor’s amortization strategy governs the ‘1

JJFig- 4.]5.-Dis_cominuousslab .casting--(author), .. . I i - . ,.._.,,,,__.Ez'g._._4_.1.Z._..Im:ernal View of an MSS (author.)

-
252 BRIDGE LAUNCHING COMPOSITE BRIDGES 253

share of the equipment cost charged to the project, bridges shorter than 150—200 m The current trend to reduce labour costs at the expense of increasing the quantity of
are hardly compatible with such investments. steel, and with it the restraint action that the steel girders exert on the concrete slab,
As with precast planks, reinforcement assembly in an MSS requires transport of increases the practical relevance of the effects of thermal shrinkage.
materials to the working location. Access through the deck is difficult and causes Typical values of ,8 for recent two-girder bridges increase from 0.04 to 0.12 when
dead-times, and cage preassembly is infrequent due to delicate hoisting operations. the design span increases from 30 to 80 m. Laboratory tests and numerical analyses
In situ slab casting requires the use of concrete pumps and access to the areas below [103,104] show that the residual tensile stress a, in the concrete slab after hydration
the bridge, and the discontinuous segmental casting sequences require shifting varies in the range 0.5—0.8 MPa for sections with ,8 = 0.05, 1.2—1.5 MPa for ,8 =
of equipment back and forth along the deck, with additional costs. All these 0.08 and 1.6—2.0 MPa for ,8 = 0.12.
disadvantages may be avoided by incrementally launching the concrete slab on to These tensile stresses are significant when compared with the tensile strength of
the steel girder, as discussed in Section 4.6. young concrete (2.0—2.5 MPa), and can lead to full-depth cracking of the concrete
Without preventive interventions, the concrete slab of a composite bridge tends slab at an early age. They also reduce the available tensile strength of concrete ft,
to develop cracking over time. Early cracks (before hardening) are due to plastic and an effective tensile strength, fieff, should be used when considering the stiffness
settlement and plastic shrinkage. These two phenomena may induce important of the composite section, while minimum reinforcement for limiting concrete crack-
cracking, but preventive measures can be adopted to avoid them. Continuous ing should be determined based on code requirements (i.e. full tensile strength).
slab casting may cause cracking during construction, and discontinuous casting
fteff :ft _ O1r ‘ (4.29)
sequences may be prescribed. However, after hardening there exist at least five
additional causes of cracking under serviCe conditions: external loads, creep, drying A simplified relationship between the residual tensile stress in the concrete slab and
shrinkage, external temperature variations (daily and seasonal) and thermal shrink- ,8 can be expressed with the following equation [103]
age [101,102]. Most of these phenomena are covered by applicable standards. On
the contrary, thermal shrinkage of concrete is a specific problem of composite a. = armies —Ec2) (4.30)
bridges (although not limited to them), it is almost always disregarded by standard (fiEs + EcIBES + E02) -

specifications for both concrete and steel bridges, and may be a primecause of early where or is the coefficient of thermal expansion of concrete, E01 and E02 are the
cracking [103,104]. A mean values for elastic modulus of concrete during the heating phase and cooling
During cement hydration, temperature in the concrete slab increases of 15—300C period, respectively, ES is the elastic modulus of the steel girders, and AI is the
during the first 12—24 h. This heating phase is followed by a cooling period of 150— maximum difference between ambient and concrete temperature during cement
180 h. The development of mechanical strength of concrete is governed by cement hydration.
hydration [77], and the elastic modulus of concrete is therefore much higher in the A qualitative evaluation of the influence of ,8 on the effects of thermal shrinkage
cooling period than in the heating phase. of concrete leads to the following observations. Hydration effects have limited
If the concrete slab and the steel girders act compositely from the moment that influence on early cracking for 8 5 0.05. For 0.05 < 8 5 0.08, thermal shrinkage
concrete sets, then this composite action prevents the expansion of concrete during reduces the effective tensile strength of concrete (4.29) and produces a limited risk of
the heating phase and its contraction during the cooling period. The restraint that early cracking. For 0.08i< ,8 5 0.12, hydration effects reduce the effective tensile
the steel girders apply to the concrete slab can be modelled by assuming constant strength more markedly, early cracking is probable, and actions for reducing
but different values for the elastic modulus of concrete during the heating and residual tensile stresses should be considered. For 8 > 0.12, hydration effects
cooling periods. Restrained thermal expansion during heating phase generates significantly reduce the effective tensile strength, high risks of early cracking
compressive strains in the concrete that are recovered during cooling. The difference arise, and actions for reducing residual tensile stresses should be adopted.
in the elastic modulus of concrete in these two phases produces permanent tensile Possible corrective actions are all addressed to Controlling At. They involve
stresses that are high when compared to the tensile strengthvof young concrete. , reductions in the cement content (provided that cement quality is reasonably con-
The restraint action exerted by the steel girders can be represented by the ratio ,8 stant) and in the water-to-cement ratio [100], the use of low-heat cement, cooling of
of the cross-sectional areas of the steel girders, AS, and of the concrete slab, AC fresh concrete or during hydration by cooling pipes, and slight heating of the steel
[103,104]. girders. Laboratory tests showed that both the use of 10w-heat cement and concrete
cooling can limit residual tensile stress in the concrete slab to less than 1.0 MPa
,3 = _ (4.28) even in the case of high restraint level (8 = 0.12).
For technical and economic reasons, however, most of these interventions
are____impr_actical,_ and___th,e_ classic solution consists of a careful design of passive
 254 BRIDGE LAUNCHING COMPOSITE BRIDGES 255

reinforcement. In this case, full—depth cracks will open and it will almost always be 20—40 years are sometimes mentioned, but this should be confirmed by additional
impossible to keep their width below 0.1 mm. A suitable detailing of bar diameter research.
and distribution usually permits limiting the crack width to 0.15 or 0.2 mm, but The only way of improving the durability of the concrete slab is then to apply
crack width may increase with time because of progressive bond deterioration at the prestressing. Slab prestressing can reduce, or in some circumstances prevent, the
crack boundaries due to fatigue effects of live loading and temperature variations, cracking of concrete under service loading so increasing stiffness and improving the
with repercussions on structure durability. protection of steel from corrosion. Slab prestressing can be applied in the
Figure 4.18 summarizes, according to their chronological occurrence, the concrete- longitudinal direction, in the transverse direction, and contemporaneously in
slab tensile stresses in the negative bending zones of composite bridges [104]. Results both directions.
are given for slabs connected directly to the steel girders and cast with a continuous The prestressing concept follows two philosophies derived from PC bridges.
pouring sequence (from one abutment towards the opposite one). The effects of
0 One solution is to prestress the slab to control cracking under the combined
external temperature are not mentioned as they are essentially responsible only for
eflects of live load, dead and superimposed dead load, and related influences
surface cracks and not for full-depth cracks. The tensile stress due to traffic is
evaluated for a single 250-kN lorry, and values of concrete shrinkage are based on such as creep, shrinkage and differential temperature. The slab in this case is
measurements made on concrete slabs. This table shows the importance of tensile analysed as a homogeneous uncracked section, and the treatment constitutes
stresses due to hydration effects and the casting sequence in comparison with other total prestressing although small tensile stresses may be allowed.
0 An alternative is partial prestressing, whereby total prestressing is introduced to
actions. Considering the low tensile strength of concrete at an early age, the tensile
stresses due to these actions are likely to initiate cracking in the concrete slab of handle dead loads and long-term effects with uncracked slab, while some crack-
ing in concrete is accepted under the influence of live load. Limitations are-
30-m-span bridges, and to cause significant cracking in that of 80-m-span bridges.
placed on the crack opening by stipulating lower values for tensile stresses in
reinforcement steel. However, the risk of a progressive increase in crack width
4.5.2. Longitudinal prestressing of the concrete slab because of fatigue effects is still to be evaluated in detail.
In the presence of diflused full-depth cracks, the first protection of the concrete slab
Longitudinal prestressing of the concrete slab may be achieved in many ways:
is the use of a waterproof membrane of good quality, firmly anChored to a sound
concrete cover with low permeability, and perfectly sealed at the joints and at the 0 A moment may be applied to the steel structure in the same direction as it will
lateral kerbs. A waterproof membrane controls the two main causes of most act in the composite structure. The tension flange is then encased in concrete
chemical and physical processes affecting the durability of structures: transport and the moment relaxed when the concrete has adequate strength.
within pores and cracks, and water. Freezing and thawing cycles, eflects of 0 Jacking can be used to alter the relative levels of the supports of a few—span
de-icing salts, penetration of chemically aggressive agents, alkali—silica reactions, continuous beam after part or the wh6’1'E"’O'f the concrete slab has been cast and
concrete carbonation and reinforcement corrosion are all related to these two matured.
causes [101]. o The concrete slab or sections of the slab can be prestressed by tendons or
However, the stability of the protective action of the waterproof membrane over jacking while it is independent of the steel girder and subsequently connecting
time is a basic concern. Despite all precautions, de-icing salts will sooner or later them. '
cause chloride ions to penetrate into concrete. If the slab is cracked the rate of o The steel girder may be prestressed by tendons before concrete casting. The
penetration is very high and depassivation occurs quickly. In these circumstances, tendons may or may not be released after the concrete has matured.
structure durability is controlled by corrosion development. Service life values of o The composite sections may be prestressed by tendons or jacking.

One method frequently used involves the concept of vertical adjustment of the
30-m span 80-111 span 1 intermediate supports. These are raised before slab casting and lowered after the
e = 0.04 0 = 0.12
concrete has cured and a stable composite action has been achieved. Bending
Thermal shrinkage 0.8 2.0 moments are thus introduced in the composite structure and the concrete is com»-
Casting sequence (continuous method) 1.8 2.7
Road surfacing 0.8 1.3 pressed in the negative moment regions. The effectiveness of countering shrinkage is
Traffic 0.3 0.1 only partial, as the benefits are balanced by creep. Time-dependent stress redistri-
Drying shrinkage 0.8 1.4
bution Within the composite structure depends on the relative stiffness: the more
. Fig-4.]8....Tensilestress .(MPa)- in the concrete slab__at..inlermediare supports rigid the steel girder, the higher the slab compression lost over time. Therefore, this
256 BRIDGE LAUNCHING COMPOSITE BRIDGES 257

system of slab prestressing is rarely used for bridges with large steel sections continuity must be provided for a concrete slab composed of precast segments.
(,3 3 0.12) as prestress loss is extremely high, especially for two-span bridges. However, launch of the concrete slab on to the steel beam solves most of these
Other methods involve temporary prestressing of the steel beam. Stability of the limitations and makes longitudinal prestressing much more affordable.
steel beam during prestressing is a prime concern, and the stresses in the steelwork
and in concrete at transfer should not exceed the allowable stresses prescribed by 4.5.3. Transverse prestressing ofthe concrete slab
standards. When the slab is compressed by releasing a temporary prestress in the The high torsional stiffness of a composite boX girder, its ability to resist high
steel beam, BS 5400 [55] requires that the compressive stress in the concrete at negative moments, and the weight saving and handling requirements, generally
transfer (before losses) does not exceed 0.5fci, where fOi is the cube strength at result in narrow U-girders [105]. Consequently, the lateral cantilevers of the
transfer. An increase up to 0.6fci is allowed when the strain in the prestressing concrete slab are often wider than with two-girder steel frames, and transverse
steel before transfer does not exceed 0.25%. If the composite superstructure is slab prestressing may be advantageous.
prestressed by an external system or by tendons not directly bonded to the concrete, Transverse prestressing by post-tensioning cables is introduced to control long-
the calculation of prestressing forces should also account for the deformation of the itudinal cracking and creep deformations of the concrete slab under dead and
whole structure. superimposed dead loads. Either complete or partial prestressing may be used.
Longitudinal prestressing with internal tendons in the concrete slab is seldom Transverse post—tensioning is expensive and has been used occasionally for wide
employed because it adds considerably to the bridge cost. Although transverse slab decks. It is effective when the deck width exceeds 15—18 m as it permits increasing
prestressing may result in some weight saving, longitudinal prestressing produces the transverse Spacing of the main girders at the concrete slab level up to 10 m. In
irrelevant structural advantages and is introduced mainly for durability purposes. narrower bridges it is adversely affected by prestress losses and the cost of
Prestressing tendons are not required for the composite cross-sectional flexural anchorages.
strength at the ULS. because the steel flanges and the slab reinforcement are Compared with transverse prestressing of PC bridges, tendons are less powerful
sufficient to guarantee the resistance to negative moments. (often T15 or T158 monostrand tendons with prestressing forces of 150—160 kN)
Further problems derive, from cracking at the anchorages of post—tensioning and less spaced (20—60 cm in relation to the distribution of reinforcement and the
cables, and from the difficulty of controlling prestress losses due to the composite intensity of transverse flexural stresses). The use of paired polyethylene—coated
action. A part of prestressing is lost in the steel beam at tensioning, prestress monostrands with 80~cm spacing may approximately reduce the passive reinforce-
migration increases over time because of concrete creep, and buckling prevention ment ratio from 250 kg/m3 down to 180—190 kg/m3 [106]. As the concrete slab is in
requires additional stiffeners in the girder webs. Since the time-dependent redis~ this case thin, the prestressing tendons are often designed with a straight layout.
tribution of prestressing forces depends on the relative stiffness, and longitudinal Tensioning anchorages are alternatively placed on both sides of the concrete slab as
prestressing is mainly addressed to relatively long spans with high values of ,6 (4.28), in Fig. 3.53. Transverse prestressing of wide deck slab-s can be obtained with
limitations are evident. 4-strand tendons placed in transverseribs, withadditional plate—action advantages.
Although composite bridges have become standard practice in several countries Launching the concrete slab on to the steel beam permits avoidance of any forming
and standards and specifications are available, few include provisions for bridges problems related to transverse ribs.
with longitudinal PC slabs. In most cases, engineers must therefore resort to codes
for PC and structural steel. Experience indicates that, in assessing the structural
safety of a composite section, the ultimate moment capacity should form the basis 4.6. Incremental launching of the concrete slab onto the steel
rather than allowable stresses, because prestressing and long-term effects such as beam
creep and relaxation do not affect the ultimate flexural strength. Where sufficient In many instances, the cost of left-in-place forms and MSS warrants consideration
plastic rotation cannot be predicted because of local instabilities (rotation capacity of alternatives based on the concept of prefabricated elements. Precast slab
decreases abruptly when the web slenderness increases from Class 1 to Class, ,2 segments of short length and, when feasible, as wide as the entire deck, may offer
according to the EC3 [90] classification), safety against this factor should be competitive solutions.
assessed. In addition, the fatigue strength must be evaluated, and because the stress Segmental slab precasting offers the same advantages of segmental precasting of
range is a controlling factor, the design should address the live load effects. This is PC bridges, such as dimensional stability due to long curing, accurate dimensions,
more important in high-strength-steel bridges where no increase in fatigue strength and high quality resulting from industrialized casting processes. The main disad—
should be expected for the higher-grade steel at welded details and connections. vantage of slab precasting is the small segment width deriving from transport
For all these reasons, the present trend is to avoid longitudinal slab prestressing requirements. This increases the number of construction joints in the deck slab,
unlessdepthrestriQEiOHS Controlawhcre slab gracking mustbc, prevented,,,9r where ,, withrelated.costsand.durabili.ty_problems.. ._
 258 BRIDGE LAUNCHING COMPOSITE BRIDGES 259

Powerful ground-based cranes or special gantries running along the steel beam
further increase the final cost. For these reasons, segmental slab precasting is mainly WINXWWXIXIXIXIXIXIXI v
used in cold countries such as Russia, to minimize the amount of field operations
involving fresh concrete [107].
An alternative construction method for precast slab segments involves in-yard

l
construction followed by jacking onto the steel beam. Construction of the deck slab
by incremental launching involves the following main steps [5]:
><r><lmxrxm><I><I>§g><l><i><1><r><l><l><lgfl

H
0 Cast a given length of concrete slab (20—25 m) in the bridge axis behind the WIXIXJWIXIXIXIXIXQE
abutment or at mid-span, depending on constructional convenience.

I
0 Jack this slab section onto the girder flanges, thus liberating the casting bed.

\__ll \q
0 Prepare the casting bed for concreting a new slab segment.

I:
.A
The finished deck is produced in a series of segments in incremental casting, and
when in place the continuous slab plate thus obtained is connected to the steel “

girders by shear attachments. The basic advantages are:

Fig. 4.19. Segmental Slab Casting behind the abutment and incremental launching
c The same formwork is used repeatedly.
0 Concrete pouring is done at a fixed location.
0 The entire slab is made structurally continuous by overlapping the reinforcing This solution, adapted directly from the Russian experiences with precast seg—
bars at the ends of each segment. mental construction of the deck slab [107], is affected by the same limitations.
Concrete filling of small openings distributed along the deck surface is expensive
During jacking, the freshly hardened slab sections slide on special strips, usually and requires protected curing. Connection forces are concentrated at the slab open—
built on standard I-shapes. The form table may be directly supported on the ground ings and this tends to cause slab cracking over time. The cost of field welding of
when the casting is done outside the bridge, or it may be suspended from the main
girders for interior casting.
The movement of slab segments normally requires the use of hydraulic/jacks. For
bridges longer than about 100 m, sliding shoes ease movement, but better sliding
details can be developed from a consideration of the sliding forces. A sliding surface
having a central groove for lubrication can reduce friction considerably. Deck
steering is achieved by the combined action of the main and auxiliary jacks in
conjunction with visual checking.

4.6. l. Technological features

In most cases, the slab is cast in a fixed formwork placed behind an abutment
(Fig. 4.19). The slab is composed of segments progressively match-cast against
each other with through reinforcement. Segment casting is followed by the launch
of the slab section, and this sequence (construction of a new segment and launch-of
the entire slab) is repeated until slab construction is complete. With a last launch
the slab is placed into its final position on the steel beam, and is finally connected to
the girder flanges.
In the very first applications of this construction method, the concrete slab was
launched directly on to the upper girder flanges. When the slab reached its final
positiOn, shear connectors were welded to the flange plates within predetermined
,,,,Qpenings, (Fig_._._4_._2_0)_,_ which .Were,,,eventually. filled with noeshrink concrete,-.-,-,, , , 3 ._,____.Fig-_.4.-20.._.D.irect..slab .,.launch--onto.. the flange plates and discontinuous connection
 260 BRIDGE LAUNCHING COMPOSITE BRIDGES 261

shear connectors into slab. discontinuities is high, and the absence of concrete fillets been used for construction of the access spans of the mobile bridge over the
into which the top flange is embedded is a structural deterrent as the lateral support Tancarville Canal in France, with the slab arrangement shown in Fig. 4.21.
of compression flange is concentrated at a few points. Continuous openings avoid concentration of connection stresses and improve
In turn, slab launching causes new problems: lateral stability of compression flange, but do not solve the remaining weak points
of the original scheme (field welding of shear connectors, need for a flat flange
0 The high sliding friction along an inaccessible launch surface generates
surface), and create new problems. Transverse reinforcement bars that join the
anomalous stresses in the steel girders and increases the cost of thrust devices.
slab strips must prevent instability of the lateral cantilevers. Since the vertical
0 Since the precambered vertical profile of the steel beam is attained by joining
lever arm is small and the transverse forces are therefore high, the compression
straight units to shape, the support lines of the concrete slab are irregular. The
bars may buckle and this might trigger progressive overturning of the entire canti-
flexural stiffness of the concrete slab is significant at a local level, and support
lever. Compression bars are therefore replaced with bar lattices or stiffened by
irregularities generate stress concentrations that may cause cracking. As these
increasing their diameter, in both cases with additional costs.
phenomena are hardly quantifiable, the remedy is a generalized increase in
In the horizontal plane, relative movements between the slab strips (caused for
slab reinforcement. Moreover, the irregular loading on the steel girders causes
instance by irregular frictional forces or stick-Slip effects along the launch lines) may
transverse bending in the upper flanges, which may require adoption of thicker
also cause buckling of compression bars. It is therefore necessary to brace the slab
flange plates and vertical web stiffeners that protect the fillet welds of webs to
strips in the horizontal plane, as shown in Fig. 4.21. Finally, the launch shoes are
flanges nearby girder splices. ,
located at the edges of the flange plate, and this causes high transverse flexural
0 Adoption of upper flanges with splice plates, bolted joints, and varying-
stresses and local torque because of transverse load diflerences.
thickness plates with discontinuous upper surface, is impossible, since the
An alternative solution is based on the use of vertical steel plates embedded into
irregularities in the upper flange surface compromise longitudinal continuity
the concrete slab and emerging from its bottom surface for the entire bridge length.
of the slab support lines.
On completion of slab launching, these plates (one near each girder flange) are
During launching, the concrete slab is supported on launch shoes that slide along welded to the flange edges, thus achieving structural continuity. In their embedded
the flanges, and the distribution of launch stresses/depends on the transverse width
of the sliding shoes. Any eccentricity should preferably be less than half this width
to limit transverse bending of the top flange. With larger eccentricities, flange
bending may become critical [5].
The following stresses arise in the steel girder near a launching shoe:
0 a longitudinal stress in the top flange and web due to the longitudinal flexural
effect of the dead load of the concrete slab
a vertical compressive stress generated by dispersion of the support reaction
0 a transverse bending stress produced by transverse eccentricity of the support
reaction
0 tangential stresses caused by shear and torque.
An eCcentric shoe also causes a slight twist of the flange locally, and this can lead to
yielding over a limited area.
Slab eccentricity during launching must be detected before it becomes detrimen—
tal. An eccentric position of the launching-shoes can be avoided if the shoes are
provided with lateral guides that take contrast against the flange edges, even if this
increases their cost. Since the shoes transfer concentrated loads at any point along
the beam, the use of web stifleners is not practical, and buckling theories must be
considered and interpreted with caution. In addition, the introduction of additional
web stiffeners might not be justified given the temporary nature of this loading.
After some trials, continuous launch openings filled with no-shrink concrete on
.. completion of,_,launch_have beenintroduced in the, concrete slab. ,.This-.solution-hasW _._...._fig.__5{.21_.._...D.irect.slab....Zaunchonto-.the.fiange. plates and continuous connection
 262 BRIDGE LAUNCHING COMPOSITE BRIDGES 263

portion, the steel plates can support headed stud connectors, or are punched with (Fig. 4.23). These continuous launch rails are welded to top flanges with a contin-
round holes to attain a concrete dowel action [108]. uous contact or with interposition of steel plates [110] (Fig. 4.24). Short I—beams
This solution avoids launch openings in the concrete slab, thus improving struc- join the slab strips through the launch openings and are supported on launch shoes
tural durability. It also distributes connection forces uniformly, and facilitates that control sliding friction. During launching, the dead load of the concrete slab is
future replacement of the concrete slab by pulling it backwards after weld grinding. transferred to the main girders through the transverse I-beams, the launch shoes,
However, this solution requires stringent dimensional tolerances both in the con— the longitudinal rails, and their bearing plates.
crete slab and in the steel girders, and extensive overhead welding from mobile The flexural and shear stiffness of transverse I-beams eases transfer of slab
platforms, with additional field activities and controls. Moreover, lateral fillet- support reactions. Their axial and flexural stiffness, in cooperation with transverSe
weld connection prevents transfer of transverse bending between the concrete top reinforcement, avoids instability of cantilever slabs and limits their rotations.
slab and the steel girders, and this might trigger fatigue phenomena [109]. Flexural stiffness in the horizontal plane avoids the need for additional bracing
Widening the launch openings as shown in Fig. 4.22 permits workshop welding between the slab strips.
of headed stud connectors, but concentrates the slab launch forces at the extreme The bearing plates of the launch rails compensate for flange surface irregularities
flange edges. Stability of the cantilever slabs also worsens due to longer com— (splice plates, bolts, varying-thickness flanges, angular precambering discontinu-
pression bars in the launch openings and a shorter vertical lever arm; however, ities) and concentrate the launch forces transferred to the main girders in both
transverse bar instability is often less critical than instability in the horizontal plane, the longitudinal and transverse direction, thus simplifying design of steelwork
which is always controlled by cross-braces. This solution still presents some of the and any necessary local web stiffening.
weak points of the original scheme, and increases the flexural stresses in the top The launch shoes reduce friction when launching uphill and control friction when
flange and in the fillet welds of web to flange. Finally, it can hardly be applied to launching downhill. It is thus possible to launch the slab with low forces in the first
curved bridges, as the polygonal alignment of top flanges reduces the transverse case, and to avoid braking in the second case. The launch rails may be covered with
width of the slab support lines excessively. materials that cooperate with the launch shoes in controlling friction, and as the
The disadvantages of the solutions examined so far may be avoided by launching launch surface is perfectly accessible, any unforeseen event can be easily solved.
the concrete slab on to longitudinal I-beams supported on the top flanges

, . Fig- 4.22. ,Directslab__launching with headedsrud connectors welded in..w0rksh0p. , . . Fig.4_._23.1ndireet sla-bvvvldunching-w - -
-
 264 BRIDGE LAUNCHING COMPOSITE BRIDGES 265

In very long bridges, the launched slab may be divided into sections separated by
transverse joints. Joint jacking produces relative movements between the slab sec-
tions that are used for slab alignment on to the girders in case of high thermal or
time-dependent deformations. These temporary joints are also filled with concrete

n—I
~—-.
on completion of slab adjustment. The case of varying plan curvature may be

.
solved by launching the slab from the opposite abutments with central closure,
or by suspending the forming system from the main girders in subsequent positions
fit to launch slab sections with constant curvature.
In the presence of many I-girders (Fig. 4.25) a concrete slab as wide as the deck
may be launched by taking support on all girders, but this requires a wide form
table that is hard to amortize in small bridges. Parallel slab strips may be built by
shifting a small form table in the transverse direction, and since slab panels must
bear on two adjacent beams for stability, bridges should preferably have a framing
system of 2, 4 or 6 I-beams.
Fig. 4.24. Detail of launch opening
4.6.2. Advantages with respect to in situ casting
On completion of slab launching, the longitudinal launch openings are completed Incremental launching of the concrete slab on to girders has several advantages
with reinforcement cages, and pouring of no—shrink concrete joins the slab to the over casting in situ. Safety of workers, construction costs, structure quality,
main girders without secondary stresses. Filling concrete must have limited cement durability and aesthetics are all improved by slab launch.
content, limited water-to-cement ratio and high tensile strength, and must be cured The first advantage is avoidance of the cost of a mobile formwork. A hydraulic
and protected as for a concrete slab cast in situ. MSS is expensive, and short and medium-length bridges often do not permit
Details of the reinforcement cage for the continuity kerbs depend on the tem- adequate investment amortization. In contrast, the investments for slab launching
porary or permanent nature of the launch rails. When the launch rails are left in are limited to a form table supported on the ground, a system of launch rails and a
place, the transverse I-beams are raised to recover the launch shoes and to insert pair of hydraulic jacks that apply the thrust force, and can be easily amortized in
shim plates that definitely adjust the vertical alignment of the concrete slab. The just one project.
bearing plates of the launch rails create a vertical clearance that may accommodate The use of an MSS may also be avoided with left-in-place fOrms, but precast
transverse reinforcement from the very beginning of the slab launch [110]. planks are expensive and increase the weightwof steelwork. In contrast, the launched
Longitudinal reinforcement may be placed on to these stirrups before launching, slabs are entirely reacting, the composite cross-sectional efiflciency increases, and the
or may be tied to the, transverse I—beams to be distributed into the launch opening weight of steelwork decreases. Even neglecting these cost savings, avoidance of the
without difficulties. By inserting coupling bars between the bottom bars emerging cost of precast planks and their handling may cover the equipment cost for slab
from the slab and the raised U—stirrups it is possible to join the compression launching even in short bridges.
reinforcement with lower costs than with threatened couplers or welded splices, Slab launching ofiers additional advantages. Structure durability improves sig—
although transverse I-beams resist most flexural stresses. In the presence of nificantly, as a monolithic slab is more stable than a two-layered composite slab, the
transverse prestressing, polyethylene-coated strands or ducts cross the openings bottom concrete cover may be thicker, and adequate inspection and maintenance of
during launching (Fig. 4.23). the bottom slab surface can be carried out after segment extraction from the form-
The concrete slab may also be connected to cross—beams by pouring concrete into work, and then over time. Imperfections and defects can be immediately detected
transverse openings that arrive on to cross-beams on completion of launch. The top
flanges of cross-beams, equipped with headed stud connectors, are used as support

e' T
surface for concrete casting. W\\\\\\\\W;R\\\\\\\\ \\\\\\\\)=-1——.;R\\\\\\\V==M

l i J";
Figures 4.23 and 4.24 show upper flanges with headed stud connectors welded in
% ar—J l—v—I
the workshop. This scheme is not restrictive, and different types of connectors
(L-beams, T-beams, hooks, hooks with shear-resisting plate) may be used. The
launch rail itself may act as a connection device by carrying horizontal studs or
m i. .1:
.,....with punched ...h.ole.s_.ai.med at producing a .concrete...dowels.shear..-transfer-[.108]-W - Fig. 4.25. Slablcrnnching. anmnny [girders
 266 BRIDGE LAUNCHING COMPOSITE BRIDGES 267

and corrected with minimum costs, and protective treatments can be applied in any cast in situ, while its introduction is easiest during segmental construction of a
moment with the same ease. launched slab. Finally, prestress migration towards the girders is governed by the
Since the slab is cast in a fixed formwork, the top surface can be finished with a residual creep and shrinkage after onset of the composite action and is therefore
roll-compacting unit (Fig. 3.35). Roll compacting expels the bleeding water from much lower: this improves the structural efficiency and reduces costs.
the top layer of concrete, thus locally reducing the water-to—cement ratio. This High-performance concrete (HPC) can be used without excessive worries, with
increases compactness and tensile strength of the concrete cover to top reinforce- additional economic advantages and without risks of thermal-shrinkage cracking
ment and reduces permeability, with enhanced protection. Roll compacting also in spite of the great development of hydration heat. The use of HPC permits
reduces the number of workers for concrete finishing, and this cost saving soon significant reductions in the slab thickness, and 12.2-m-Wide precast slab panels
covers the investment for equipment. made with 80—MPa concrete and with thickness ranging from 14 cm in mid-span-
Since the vertical profile of top girder flanges is a precambered polygonal line, in to 22 cm above the girders have already been used in the Roize Bridge in France.
situ slab casting on precast planks results in a varying-thickness slab, while the Additional aspects of HPC are discussed in Chapter 5.1.1.
launch openings in the launched slab absorb alignment irregularities in girder Slab launching also oflers technological advantages. Slab casting in situ requires
flanges and ease control on the final vertical profile of the deck. handling of loose materials and exposes personnel and the areas below the bridge to
Aesthetic advantages are not less significant. Detachment of planks from bar risks of fall. In contrast, slab launching concentrates construction activities on the
lattices, unpredictable plank deformability, grout spilling from many joints, ground, the presence of workers, materials and equipment on the deck is mini—
different colours of planks: all of these aspects compromise bridge aesthetic. mized, and there is no need to create accesses and working areas on the ground.
Moreover, placing planks is diflicult in curved bridges, and bolted field splices This last aspect is particularly advantageous when the steel girder is in turn
in the main girders cause additional irregularities. In contrast, launched slabs have launched, and the same site facilities can be used both for the steel girder and for
a regular and ordered aspect, and launching permits adoption of pleasant the concrete slab. Finally, slab casting in situ is slowed by concrete curing require-
haunched shapes and rounded corners with minimum costs. Surface finishing ments, whereas neither concrete curing (temporary cantilever supports permit
(sandblasting, high-pressure washing) is easiest at form stripping, and deeper earlier form stripping) nor weather conditions (formwork can be sheltered and
surface finishing (board-marking, engraved textures) is inexpensive because of heated) affect the time—schedule of a launched slab.
the small form dimensions, in both cases with perfect control on the final aesthetic Incremental slab launching can easily be adapted to the technological require-
aspect of the structure. ments of segmental precasting. Precast slab segments With emerging reinforcement
The steel cages of launched slabs can be entirely preassembled (in the workshop can be joined with no-shrink concrete, and match-cast segments equipped with
or on site) or handled in prefabricated grids to improve labour rotation. Concrete alignment keys are glued and prestressed longitudinally prior to launching.
pouring with buckets or conveyor belts is less expensive than pumping, the areas A last consideration about construction risks: the dead load of the concrete slab
below the bridge are unaffected, and personnel productivity is enhanced by is about four times greater than that of the Steel girder. During slab launching, load
better work conditions (no risk of falling, possible use of mobile coverings, roll eccentricity and concentration of support reactions may be risky, and most of these
compacting instead of hand finishing) and repetitive operations. risks are hardly quantifiable. Consequently, slab launching should always be carried
Launched slabs undergo most of the dimensional variations due to drying and out in close collaboration with the designer of the structure, and unexpected events
thermal shrinkage before being connected to the girders. This results in low residual should be recorded, investigated, and shared between all parties involved.
stresses, and the jacking action imparts a definitive compression to concrete. The
use of no-shrink concrete for filling launch openings also reduces the residual 4.6.3. Stress analysis
stresses in these slab areas, where the expansion action of concrete, contrasted by Slab launching simplifies structural analysis and makes its results more reliable,
the slab strips, generates a diffused compression that prevents shrinkage contraction since the slab is joined to the steel girders without longitudinal stresses, and its
from creating cracks. Creep stresses areloweras well, owing to the absence of thickness is constant in the absence of deflections in the forming system. When
longitudinal long-term stresses in the concrete slab and the dimensional stability longitudinal prestressing is adopted, the contribution of the concrete slab to the
resulting from long curing at joining. composite action is also constant along the bridge.
As already discussed, longitudinal prestressing of the concrete slab offers many Factors to be considered in the design are related to the physical process of slab
advantages. It improves structure durability by avoiding full-depth cracks and launching and to the effects of the launching forces on the concrete slab and the
makes the slab contribution to the carrying capacity of the composite section steel structure. In the final position, the loads and composite action are as in
constant along the bridge, which simplifies structural analysis and makes it more conventional construction, and because the shear connectors are activated when
, ,,I,el,iab1¢,- Longitudinql prestressing involves, diflicult and risky operations in ,a,,slab .vvwnumgm .. M _ ___- _. ._._,.____the____slabis.._in.._._place.,. the concrete. weight. is resisted by the steel frame alone.
 268 BRIDGE LAUNCHING COMPOSITE BRIDGES 269

The jacking forces are applied parallel to the plane of the slab and must overcome
the frictional resistance. For a body moving at low speeds, this reaction is discon—
tinuous and is manifested in a series of ‘sticks’ and ‘slips’. Experience with staged
launching shows that the stick—slip effects depend on the mass of the slab, the length
of the launch strokes, the mass of the steel frame, the elasticity of the links and the
jacking velocity. Steering the slab along a prescribed path may also cause the slab to
veer off its course, and in this case the path must be corrected with the help of
transverse forces.
The simple beam shown in Fig. 4.26 is assumed to have a uniformly distributed
friction load of intensity r. The beam has a cross-section with the centre of gravity
at mid-height. If the beam is held longitudinally by a tie fixed at point C, the
eccentricity is small and the only force acting at distance x is an axial force N = rx.
In the continuous beam shown in the top of Fig. 4.27, distributed friction forces
of intensity r uniformly load n consecutive spans along their entire length. The
beam is held at the level of the upper flange at point F. The internal forces may Fig. 4.27. Continuous beam under uniform friction loading
be determined as follows. 1
First, the continuous beam is loaded as in the original beam but is held at mid-
height point C as shown in the middle of Fig. 4.27. In this case, the only internal where cosh y z 2. Using appropriate trigonometric functions, an expression for MJ
force in the beam is the axial tension N = rLtot, where Ltot is the total length of the may be written in terms of continued fractions. Based on this analysis, the rotation
concrete slab supported On to the steel frame. of support sections can be derived and used to handle more complex loading cases.
Then, the continuous beam is not subject to friction but to a terminal couple If the jacking force is symmetrical with respect to the longitudinal bridge axis, the
M = NH/2 applied at the end as shown in the lower part of Fig. 4.27. For equal friction forces are essentially distributed between the two main girders.
spans and constant flexural stiffness, the bending moment generated in each section Asymmetrical driving of the concrete slab causes flexural stresses about the vertical
by the terminal couple M may be obtained from! a second-order linear difierence axis and warping of the steel sections. Therefore, launch jacks should preferably be
equation. For n equal continuous spans, the bending moment MJ over the support J fed in parallel by a common hydraulic circuit.
is [5]

MJ 2 —M 91)a (4.31)
Sinh ny

. Fig.4.26. Simple span underuniform. friction loading

 5. Prestressed composite bridges

._
‘
.0:
1;
.1
_
....
A
5.|. Introduction

W-—
Whatever the construction technique of a bridge may be, the limitation of dead load
is a prime requirement. Dead load is the most important loading on the structure,

*W’
and its reduction creates a reserve available for live loads. Dead lead also governs

——Wr——'—'—"~—
the design of construction equipment, the amortization of which is a prime
component of the construction cost of a bridge.
The dead load of a PC box girder is the sum of the weight of the top slab, the
bottom slab and the webs, and these three components can be treated separately to
lighten the cross-section. The thickness of the top slab depends on the live load and
on the need to reach adequate punching strength, and therefore cannot be too much
reduced. The thickness of the bottom slab often depends on the need to contain
internal tendons, but its design is generally less restrained.
Using internal prestressing, the webwthieknessis often governed by the need +n LU

contain and deviate the tendons, and in narrow box girders the web area can reach
30% of the cross-sectional area [44,111]. The webs reduce the cross—sectional
flexural efficiency because of their position close to the gravity axis, and increase
the cost of materials and labour since they are the most difficult element to cast of
the whole cross-section. However, webs are necessary for shear transfer, and their
shear efficiency mainly depends on the mechanical properties of concrete. '
The efficiency of a material can be evaluated with the ratio of the strength to the
specific weight (Equation 4.1). A 45-MPa concrete has a compressive efficiency of
6M 2 1800m and a low tensile efficiency; consequently, shear efficiency is poor as
well. The tensile efficiency of a Fe 510 EN-10025 steel plate is much higher, 6M %
4600m at yielding, although instability penalizes the compressive efficiency.
Prestressing steel is definitely more efficient: a T158 EN-10138 strand reaches 6M
§ 21 500 m at 0.1% load, and cable is generally the most effective way to use steel in
tension. By relating the efficiency of structural materials in their optimum work
conditions to that of prestressing steel, 45-MPa concrete reaches 8.4%, and high-
strength steel plates reach 21.4%.
272 BRIDGE LAUNCHING PRESTRESSED COMPOSITE BRIDGES 273

A continuous-beam bridge is mainly designed for flexural stresses, and prestres— Generally, CBC materials aim to reach the highest mechanical strength permitted
sing steel is therefore used to create an eccentric compression that controls edge by the absence or the minimum presence of defects or pores in the matrix micro—
stresses. This requires the presence of two wide flanges to compress without instabil— structure, and can be grouped into two classes:
ity, and the use of reinfOrced concrete offers a reasonable compressive efliciency at
low cost. Once the flexural demand has been met, tendon deviation can reduce the 0 Macro defect free (MDF) materials, characterized by a gummy rheology of the
shear stresses in the webs, with beneficial eflects on web thickness and, conse— mixture that permits its processing by extrusion or calendering between rolls.
quently, on the flexural efliciency of the cross-section. , However, the geometric a Densified with small particles (DSP) materials, characterized by a more
requirements of internal tendons often control web thickness. conventional rheology and based on the principle of filling the voids among
External prestressing produces a first improvement of the cross-sectional effi— the cement granules (dimensions 1—50 mm) with sub—micron—sized particles
ciency and reduces the cost of materials and labour. In turn, the compressive (0.01—0.l um). Their processing involves both humid powder pressing and
eflflciency of reinforced concrete can be improved by increasing the strength with pouring mixtures of plastic consistency into moulds.
the same density (i.e. by adopting high—performance concrete), or by reducing DSP materials may be further divided into two subtypes: the current high perfor-
density with the same strength (i.e. by adopting structural lightweight concrete). mance concrete (HPC), and reactive powder concrete (RPC), which represents its
Once even these steps have been taken, the PC box girder concept is abandoned peak of technological development.
for the non-prestressed composite beam made of a steel girder completed with a HPC is currently produced in many countries. It is characterized by a compres-
reinforced concrete slab. A steel plate just a few millimetres thick resists the same sive strength of 60—120 MPa obtained by using the following materials:
shear force as a conventional concrete web, although it weighs only the 1/ 15 or 1/20
as much. However, instability penalizes the use of steel in compression, and the use 0 High-strength cements with fairly low alkali content mixed with mineral
of a concrete bottom slab for double composite action in the negative bending additives with high pozzolanic activity such as silica fumes, rice husk ash and
regions is generally limited to the longest spans. In most cases, therefore, the microfine slag.
torsional stresses are resisted by cross-frames and horizontal bracing, thick bottom o Super—water-reducing admixtures (superplasticizers and retarders) able to pro-
flanges carry the negative-bending compressive stresses, and compression flange duce a lowest water—to-cement ratio (generally prescribed in the range 0.26—0.30,
instability is a main concern. considering as cement both the cement itself and the mineral additives).
Similar efliciency may be accomplished with prestressed composite sections that o High-quality rough surface crushed coarse aggregates (basalt, granite and dia-
combine the use of external prestressing, reinforced concrete slabs and light steel base) along with washed silica sand that assures high mechanical strength and
webs. This results in eflicient cross-sections (composed of masses far from high adhesion to the cement matrix. In fact, one can obtain such a dense and
the gravity axis and prestressed in an effective way) that make the most out of compact matrix as to exceed the mechanical strength of conventional aggregates.
prestressing steel and are light and easy to build. Along with the best mechanical perfOrmance,"hOWever, HPC has some limitations
when strained to the ULS because of the typical brittle rupture of the ceramic
materials, characterized by low fracture energy. The low ductility of the material
5. I . I. High performance concrete should be taken into account in the structural design and compensated by means of
In the last 20 years, international research in high performance concrete has led to a an adequate distribution of reinforcement.
great variety of new materials, with new abbreviations and acronyms. Among these, The introduction of RFC is aimed at eliminating these defects. Although it is only
probably the most widely used is chemically bonded ceramic (CBC), a name that in the research stage, the results are very promising. The soft-casting production of
emphasizes the ceramic character of the material (inorganic but not metallic) and a concrete with a compressive strength of 200 MPa and ductility, bending strength
the prevalently chemical nature of the bond [112]. The association of CBC materials and fracture energy much higher than the present HPC will open new horizons in
with the ceramics family is also justified by their production processes, much more the use of concrete and will represent a valid alternative to structural metals.
similar to those for conventional neoceramics (powder pressing, high-shear mixing, At present, the design standards of many countries allow the use of HPC, with
lamination, extrusion, calendering) than to those for ordinary concrete. Only a few maximum strength of 80 MPa in Japan and Sweden, 100 MPa in Finland, 105 MPa
of these materials are suitable for the construction of large and complex works in Norway and the Netherlands, 115 MPa in Germany and without any limits in
using conventional techniques (mixing and transport in concrete mixers, casting the USA and Canada.
in forms). Others are reserved for more sophisticated uses such as the production The advantages of using HPC in PC launched bridges are numerous. The com-
of small regular-shaped elements. bination of external prestressing and a higher concrete strength results in thinner
slabs and webs. with increased. .flcxural efliciency. Preliminary studies [5,66] show
274 BRIDGE LAUNCHING PRESTRESSED COMPOSITE BRIDGES 275

that the concrete quantity may be reduced by 25—30% through the use of HPC in for conventional 35-MPa concrete. Based on an increase as the cube root of
the superstructure, leading to a 20% volume reduction in the substructure. compressive strength, the short-term elastic modulus of 60-MPa concrete is 1.19
Economic comparisons appear to indicate that, considering the higher unit price times that of 35—MPa concrete, and the short-term modular ratio is therefore about
but the reduced concrete volume, HPC also results in lower initial costs. 5 instead of 6. The long—term modular ratio is 12.5 for a 60-MPa HPC without silica
Lightening the deck reduces the influence of the launch stresses, with cost savings fumes and 9.0 with silica fumes instead of 18.0. Reduction in the long-term modular
in the launch equipment and longer spans that can be overcome without temporary ratio is therefore 31% without silica fumes and 50% with silica fumes. This reduc-
piers. Lightening the deck also increases the relative quantity of parabolic tendons, tion results directly in either lower stresses in the steel frame or lower weight of
and with it the overall efliciency of final prestressing. A thinner cross-section results steelwork, which adds to the weight savings resulting from a thinner slab.
in more slender and pleasing structures, and the total quantity of reinforcement also Design strengths of 60—120 MPa do not subject the concrete handling to parti—
decreases, although not in linear proportion to the self-weight reduction. cular restraints. However, the production of hydration heat must be controlled
Structure deformability and prestress losses due to elastic shortening, shrinkage since it can produce microcracks. The following interventions may be adopted [66]:
and creep decrease as well. Some standards allow significant increases in the elastic
modulus compared with conventional concrete, and the Finnish standards also allow 0 Limit the maximum temperature inside the concrete. In the 80—MPa HPC for
a 30% reduction in the creep coefficient. Reduced drying-shrinkage cracking and the Elorn Bridge in France (97 MPa average strength was actually reached),
increased frost—thawing resistance, abrasion resistance and imperviousness to water internal temperatures up to 80°C were measured inside the pylons during hydra-
and gas improve structure resistance to aggressive agents and increase durability. tion. However, the effect of the high temperature was lower than that expected
The advantages of using HPC in non—prestressed composite bridges are less in conventional concrete, with a reduction of the 28-day strength of about 10%.
evident but not less significant. The use of HPC is often related to a higher Some vertical cracks appeared in the pylons due to thermal effects, but their
compressive strength, which appears pointless when compared with the low long- widths were limited to less than 0.3 mm.
itudinal stresses in the concrete slab in the positive moment regions. In addition, 0 Limit thermal gradients. In the Great Belt Link in Denmark, maximum differ-
both short-term drying shrinkage and thermal shrinkage of HPC are higher than ences of 200C were allowed between the external surface and any internal point,
that of conventional concrete. V and of 120C between the mean temperature of the setting segment and the
The interest in the use of HPC is mainly related to segmental precasting of the previous one. These conditions were achieved and maintained by combining
deck slab or to slab casting in situ with incremental launching onto the steel girder. heating or cooling of the fresh concrete, heating of the adjacent segment and
These construction methods are not frequently adopted, so the advantages of the cooling of the setting concrete. When the cross-section is cast in two phases (a
use of HPC have not been completely exploited. bottom U-segment and the top slab later), heating may also be necessary for the
Total shrinkage of HPC is similar to that of conventional concrete, but its first phase to control thermal shrinkage cracks at the horizontal joints.
evolution with time is different. With 70% relative humidity and an average 0 Postpone form stripping to reduce the superficial shrinkage cracks. In the Elorn
cross-sectional radius of, say, 0.15 m, 28—day shrinkage of conventional concrete ‘ Bridge, formworks were maintained against the young concrete for 72 h.
is about 17% of total shrinkage, while a 60-MPa concrete with silica fumes reaches 0 Immediately protect the exposed surface of the top slab with plastic sheets.
31% of total shrinkage at the same age. If composite action is attained at 28*day a Operate with the lowest possible water-to-cement ratio, checking the number of
curing, the residual drying shrinkage is 20% higher in conventional concrete, and microcracks measured in drilled cores.
thermal shrinkage is not a problem in either case. Shrinkage stresses depend on the 0 Adhere to quality assurance procedures with training programs for the owner’s
residual shrinkage strains after onset of composite action and on the elastic staff and the contractor.
modulus of concrete. Since EC increases with the cubic root of compressive strength, 0 Carefully programme the whole production process, including methods and
long-term tensile stresses are slightly lower in HPC deck slabs, and the tensile limit times for transport, pouring and vibration of concrete, as well as the
strength is definitely higher. . , . possible mixing of admixtures in the yard.
The long-term creep of HPC is smaller than that of conventional concrete. The
creep coefficient may decrease to 1.5 without silica fumes and to 0.8 with silica
fumes, compared with 2.0 for conventional concrete [86]. The ratio of short-term to
5. I .2. Lightweight concrete
long-term modulus is therefore 2.5 and 1.8 for HPC (respectively without and with Concrete with lightweight aggregates is used when dead load constitutes a high
percentage of the design load of the bridge [113]. Reduction in dead load increases
silica fumes) compared with 3.0 for conventional concrete. The long—term modular
the structural efficiency, reduces the cost of reinforcement, prestressing and launch
ratio (Equation 4.2), is equal to the short-term ratio multiplied by the ratio of short-
equipment, and lightens piers and abutments.
-,,term to long:te_rm__...e.1a,stic modulus ofconcrete) (Equation.4-3),.- It is Often takan as 18
 276 BRIDGE LAUNCHING PRESTRESSED COMPOSITE BRIDGES 277

The savings so obtained soon exceed the higher cost of concrete. In bridge No. 33 Since concrete is concentrated at the extreme edges of the cross—section, the
on highway A55 in France [36], built by full—span incremental launching, a value— radius of gyration increases, the cross—sectional flexural efficiency increases with
engineering proposal with lightweight concrete was 3% less expensive than the quadratic ratio (Equation 2.13), and prestressing decreases significantly
conventional-weight concrete design because of the lower cost of prestressing. (Equation 3.3). The fiexural efficiency is even higher than that of conventional
Lightweight concrete is more expensive than conventional-weight concrete non-prestressed composite sections.
because of the higher production cost of the aggregates and the few sources of The contribution of materials is specialized: the concrete slabs resist the bending
supply, which increases the cost of transport. This concrete easily reaches a com- moment thanks to prestressing, whose deviation reduces the shear forces to values
pressive strength of 60 MPa with a density less than 18.5 kN/m3 [114]. In the main that can be Carried with light steel plate webs. Each material works in uniform rather
l63—m span of the cable-stayed Salhus Bridge in Norway, lightweight high-strength than triangular stress pattern (the concrete slabs are uniformly compressed and the
concrete was prescribed with an average strength of 73.5 MPa and density lower web plates resist uniform shear stresses), with enhanced individual efficiency.
than 19.2 kN/m3. It is a very adaptable material and its use can often be convenient, Compared with non—prestressed composite bridges, the webs are thinner and the
such as for bridges in seismic areas or very high above the ground. flange plates are much smaller. The weight of steelwork is 15—20% of that of a
The physical properties of light aggregates are rather specific and many details of conventional composite bridge, and the unit cost of steelwork is similar since the
the project must be adapted to the material. The lower ductility requires dense cost of web plate folding is balanced by avoidance of welded stiffeners.
reinforcement to increase the deformation capacity in the plastic domain and to Field activities are very simple. Compared with PC bridges, casting of the
dissipate energy before collapsing. The low thermal conductivity of the light aggre— concrete webs is avoided. Compared with conventional composite bridges, field
gates reduces the loss of hydration heat and accelerates the growth of mechanical assembly activities of steelwork are definitely simpler, and tolerances less stringent.
strength, so that the ratio of 7—day to 28-day strength is generally 0.8—0.9 compared Construction duration decreases and the casting cell, the launching nose and the
to an average value of conVentional concrete of about 0.7. In some instances this thrust systems are less expensive than those for PC bridges, although launch
may be an advantage, in other cases (box girders cast in two phases, deck slabs of bearings are costly because of the higher structural flexibility.
composite bridges cast in contact with the steel girders) delayed concrete cooling Box girder bridges with prestressed composite section may be divided into two
may require specific casting procedures to reduce the effects of thermal shrinkage. types, the main difference between them consisting in the transmission of shear
Shrinkage is about 50% higher because of the lower rigidity of aggregates and stresses:
creep is about 20% higher, and this further increases prestress losses. The tensile
0 Trussed-web and space—frame bridges (Fig. 5.1) are aimed at eliminating
strength (which affects adhesion between concrete and reinforcement and therefore
material not working in the Morsch lattice scheme.
the number and the width of cracks) is 15—30% lower than conventional concrete.
0 Platenweb bridges benefit from the higher shear efficiency of steel plates
The elastic modulus is 25 to 50% lower, and this increases prestress losses but
compared with reinforced concrete webs,
reduces the influence of secondary stresses. BS 5400 [96] prescribes a reduction in
the elastic modulus of conventional—weight concrete under short-term loading equal The advantages in terms of structural efficiency of trusseduweb and space-frame
to (V/2300)2 for lightweight aggregate concrete having a dry density y between 1400 bridges over conventional PC box girders Can be significant, and the interesting
and 2300 kg/m3. Some codes (BS 5400 [96], DIN 4227 [115]) also provide design aspects of these sections are not limited to the savings in materials, since many
instructions for prestressed lightweight concrete structural components. qualitative aspects are also involved [116,117].

5. I .3. Introduction to prestressed composite bridges

The combined use of external prestressing, reinforced concrete slabs and steel webs
results in efficient cross—sections that make the most out of prestressing steel and are
light and easy to build by incremental launching.
Compared with a conventional PC box girder, the dead load decreases without
penalizing the carrying capacity, and the structure efliciency (Equation 2.10)
increases immediately. On a 40-m Span, a PC box girder with internal prestressing
requires about 0.55 In3 of concrete per square metre of deck surface, which
decreases to 0.45 m3/m2 with the use of external prestressing. A prestressed compo-
, ,site section-_(PC_S).....b9x. girder requires,,,on,1y,Q-35 Inf/m2 andlightening is 25—35%.
278 BRIDGE LAUNCHING PRESTRESSED COMPOSITE BRIDGES 279

Most PC box girders have only two webs because of the forming costs and the
high dead load of multicellular sections, and the longitudinal fiexural stresses in the
slabs differ from the average values of elastic calculation because of shear lag. In a
trussed—web bridge, the web number is generally increased until a space-frame
behaviour is attained (Fig. 5.2), the elastic analysis of which is much more reliable
[118——120]. The uniform transverse deflection of a space—frame PCS also makes the
assumption of an unidimensional beam correct, whereas distortion is not negligible
in PC box girders.
In a conventional box girder (CBG), differential shrinkage due to the difierent
thickness of webs and slabs causes stress gradients at the cross-sectional nodes. In a
space-frame bridge with concrete diagonals, this phenomenon is reduced by the
similar thickness of the structural elements (Fig. 5.3). Finally, load dispersion
occurs through several paths, and in a case of collapse of a member, the stresses
are redistributed along alternative paths thanks to the higher grade of redundancy.
However, limitations derive from the construction method. Migration of support
reactions discourages the launch of trussed—web and space-frame bridges owing to
the local stresses that arise in the bottom chords [121,122]. When the longitudinal
spacing of the bottom slab nodes is smaller than the pier cap width, movable launch
bearings integral with the Superstructure and placed under the bottom nodes may
be used, although this requires the discontinuous launch sequence shown in
Fig. 1.5. A value-engineering design involving incremental launching on 53.2-m Fig. 5.3. Sylcms Bridge (courtesy Bouygues)
spans of a PCS bridge composed of two paired Warren steel trusses, two compo-
site—tube bottom chords and a common deck slab has been proposed in France. To
date, however, the most significant applications of these structural systems have Bridge [123,124], Glaciéres and Sylans Bridges [125—127]) or, in the case of small
been built by segmental precasting with launching-truss field assembly (Bubiyan bridges, by in situ casting on falsework (Arbois Bridge, Fig. 5.4). To the author’s
knowledge, the incremental launching construction method has been applied only
to PCS box girders with steel plate webs.

5. I .4. Cross-sectional efficiency

The efiiciency of PCS bridges has been compared to that of conventional PC box
girders with a statistical analysis of 76 continuous—beam highway bridges with
constant—depth single-cell cross—section and internal prestressing [128]. The
structural efficiency has been evaluated using Equation (2.10), and the cross—
sectional fiexural efficiency has been evaluated using Equation (2.13).
The flexural efficiency of CBG is represented by white triangles along with their
regression line in Fig. 5.5. The flexural efificiency of CBG sections was recalculated
by neglecting the contribution of webs both in the moment of inertia and in the
area. Then, the slab area so obtained was increased by 5% to account for the weight
of the steel webs of PCS, and the results are represented by white dots in Fig. 5.5.
The few actual PCS bridges (black dots) distribute conservatively among the white
dots and have the same static behaviour. This appears to validate the use of a
common regression line, which shows an average improvement in the flexural
Fig. 5.2.. BubiyanBridge.(reproduced.with. permission,_BouygueS) efficiency of about 22%.

l
s
5
i
5
 280 BRIDGE LAUNCHING PRESTRESSED COMPOSITE BRIDGES 281

30

unit dead load (kN/sq.m)

20

x>
ot>o>
O
o

o
10 O ,—0 43'

20 4o 60 ' so
span (m)

Fig. 5.6. Dead load of CBG arid PCS bridges

Finally, the structural efliciency is compared in Fig. 5.8. The improvement ranges
Fig. 5.4. Arbois Bridge (courtesy Virlogewe—ACI) from about 15% for 30—m spans to 27% for 70—m spans. This better performance
opens new perspectives in bridge construction. The prime cost is similar since
In Fig. 5.6, the white triangles represent the dead load of CBG (including support savings in concrete, reinforcement and prestressing balance the higher cost of the
diaphragms and for the database average deck width of 12.6 m) and the white dots steel webs, construction does not involve particular costs and savings in equipment
represent the dead load of the derived PCS. Again, the black dots of actual PCS balance the risks of inexperience.
bridges validate the use of a common regression line, which shows significant
savings in dead load that are practically constant with the span.
The distributed service loads (superimposed dead and live load) are represented
5.2. PCS bridges with stiffened-plate webs
The most intuitive way to substitute the concrete webs of a PC box girder with steel
along with their mean value and the dead load of CBG and PCS bridges in Fig. 5.7.
webs is to use stiflened-plate webs (Fig. 5.9). The superstructure has a conventional
"i'id‘m
aspect since tendens are hidden insidethe‘boxceii, the oenaviour o L115 1 ‘ lUlVlULl'cU.
I 1

-
0.90
20
deadload
CBG
PCS
0.70
flexural efficiency

unit loads (kN/sq.m)

deadload
PCS
CBG
10 BS 5400
0.50
DIN 1072
AASHTO

0.30 .
. .. 2o 40 60 80
' ; , span (m) 0
all”, in; -:.?""‘"&‘ ‘-‘ 0', , fie... 20 40 60 80
.é’
‘ i span (m) "
Flg55£Flexural eficiency pf of conventional box girders (CBG) and prestressed
composztesectionS(PCS) . . . . . . ,. . . . .. L 1‘0. 5.7..Deadandservice loads.
‘. ' ; .:-_ t
 282 ’ BRIDGE LAUNCHING PRESTRESSED COMPOSITE BRIDGES 283

0.60
reaches 20—25% of the total prestressing force. This force is wasted and requires the
use of additional stiffeners to prevent web buckling, since the use of thicker plates,
often advantageous in terms of workshop costs, makes things worse because of the
higher restraint level exerted by AS.

structural efficiency
Stress analysis is complex as creep decreases progressively due to load shedding
towards the steel girders. The variation of the modular ratio with time also modifies
the mechanical properties of the cross—section, the position of the centroid and the
I \

R— CBS

eccentricity of the prestressing force, with accentuating effects on the migration of

prestressing towards the steel girders.
A cross—section composed of two plate webs and two concrete slabs would be
excessively deformable, and vertical stiffeners must be introduced to stabilize web
0.00
20 40
‘
60 80
panels against buckling and to attain a transverse flexural stiflness of the webs
span (m) similar to that of the slabs. The use of braces (steel diaphragms or cross-frames
are light but deformable, and concrete diaphragms are excessively heavy in such
Fig. 5.8. Structural efi‘icierzcy ,0S 0f CBG and PCS bridges refined structures) is necessary to complete the action of the tendon deviation
diaphragms in controlling distortiOn. Finally, longitudinal stifleners are often
materials is well known and the combined use of reinforced concrete slabs and steel- necessary to control web panel buckling in the most compressed areas, with
plate girders has been amply tested in conventional composite bridges. However, additional costs.
this simplicity is only apparent and several aspects discourage the adoption of this The connection between the steel girders and the bottom slab must be studied
structural scheme. carefully, as it is the weakest point of the structure. Longitudinal corrosion lines
First of all, the steel girders absorb a significant portion of the prestressing force have been frequently observed at the deck paving level in U-framed bridges with
applied to the composite section. Under an axial force F, common longitudinal lateral plate girders joined by a concrete slab supported on bottom flanges. In a
strains at the flange-slab nodes govern the distribution of the compressive stresses PCS box girder the situation is similar, although only condensation water inside the
between the concrete slabs, of total area AC, and the steel girders, of area AS. The cell can flow along the webs and gather at their lower nodes. ,
portion of prestressing resisted by the steel girders at tensioning is Corrosion prevention discourages placing triple contacts (air, concrete and steel)
near main structural elements. Therefore, although the upper cross-sectional nodes
can be detailed by conventional schemes (migration of prestressing forces super—
S ‘ A. + (EC/Esme
F — AS
(5’1)
F
imposes on longitudinal shear foi‘ces lO'wer”'t'han in conventional composite bridges)
the lower nodes present specific problems.
where EC is the elastic modulus of concrete at stressing. The increase in axial strains
due to creep of concrete may be evaluated with an age—adjusted modulus that
decreases with time; consequently, Fs increases with time and in practical cases it
5.2. I. Web—slab lower node
The node between the steel girder and the bottom slab can be developed by using
the schemes shown in Fig. 5.10.
The simplest solution, scheme A, consists of casting the slab on to the bottom
flange, and is penalized by the presence of a triple contact on a main structural
element. This scheme may be improved by welding a longitudinal stiffener at the top
surface level of the slab (scheme B) to move the triple contact away from the web.
A second solution is to place the slab under the flange plate (scheme C). This
moves the triple contact away from the web, reduces the total weight of steelwork
and offers a better aesthetic result. Concrete pouring and vibration under the
bottom flange may be simplified by inclining the girder (scheme D), with better
adhesion between concrete and the flange plate and better work conditions of shear
connectors. Additional cost savings derive from the smaller width of the bottom
Fig. 5.9,. Stiflened plate..PCS. slab __ _. _ . _ . _. , '
 284 BRIDGE LAUNCHING PRESTRESSED COMPOSITE BRIDGES 285

A B The concrete slabs cannot effectively control web panel buckling, since they only
point 01“ attack longitudinal increase the grade of restraint of the flanges and reduce the web panel depth (when
or“ corrosion stli-‘Fener‘
the schemes C and D of Fig. 5.10 are adopted). In addition, the compressive
stresses produced by prestress migration worsen panel stability in the upper web
region in mid-span, and in the lower web region in the negative moment areas.
These additional stresses may require the use of closely spaced Vertical stiffeners to
prevent buckling, and this requirement is particularly stringent during the launch
since axial prestressing does not decrease the shear forces and every cross—section is
subject to both positive and negative moments; an alternative is the use of long-
C itudinal stiffeners located at the thirds of the web panel depth. However, web panel
buckling is not a potentially dangerous situation like that due to instability of
difficult concrete
casting slender compression members or lateral—torsional buckling, since the concrete
slabs and the vertical web stifleners restrain the web panel edges, and membrane
stresses resist the growth of out-of-plane elastic deformations after buckling.
/////////////

During the launch, an I—girder can also develop the local modes of instability
Fig. 5.10. Node between the steel girder and the bottom slab discussed in Section 4.4.4. The local modes depend on the overall support reac-
tions, and since the cross-section is launched finished, support reactions are much
greater than those of non—prestressed composite bridges. On the other hand,
The stresses transferred between the steel girders and the concrete slabs are however, PCS bridges are not precambered, and one of the causes of increased
similar to those of non-prestressed composite bridges. They consist of distributed support reactions is avoided. When the schemes C and D of Fig. 5.10 are used,
longitudinal forces (shear and prestress migration) and of orthogonal forces in the the deck surface that comes in contact with launching bearings is match—cast
regions of the continuous beam affected by discontinuities in the shear force dia— against rigid extraction rails as for PC bridges, and tolerances at flange splices
gram. COnventional shear connectors may be used when the concrete slab is cast on are avoided as well. Finally, when the thrust force is applied to the bottom slab,
to the flange plate, and the connection devices for suspended slabs must also resist the longitudinal components of the support reactions due to friction and the
, the dead load of the slab and must therefore be embedded more deeply into con- longitudinal gradient remain into the bottom slab and do not affect significantly
crete. The different efficiency of connectors in relation to the direction of concrete the lower region of webs.
pouring should always be taken into account in Overhead connections [129—131]. The local modes of instability at the application points of the launch support
reactions are the web piasticization (iocai yieiding) and the creation of a local
mechanism in the bottom flange (local buckling) (Fig. 4.13). Both modes are pos-
5.2.2. Web instability sible in spite of the presence of the bottom slab, since the support reactions are high
Instability is a basic issue in steel construction, and steel plate girders are often and the web plates are thin because of the shear force reduction obtained with the
designed to satisfy stability more than strength criteria. tendon deviation on completion of launch. Local-yielding and local—buckling
An I-girder subject to mainly flexural stresses can develop the modes of in- strength can be assessed as discussed in Section 4.4.4, by adjusting the length of
stability discussed in Section 4.4.3. However, construction of a PCS bridge the rigid patch loading and the width and thickness of the bottom flange according
by incremental launching occurs by casting the slabs in the casting yard and by to the expected contribution of the concrete slab.
launching the finished section, and stability of I-girders improves significantly when Local yielding is controlled by the dispersion of the support reaction through
they are fixed to concrete slabs at both flanges. - - the bottom slab, whose thickness is significant at a local level (Fig. 4.12). The
Lateral stability of the compression flange is guaranteed by the concrete slabs, bottom slab also controls local buckling by stabilizing the bottom flange and by
which act as a diflused lateral bracing, and by the cross-frames or diaphragms, increasing its effective flexural strength. Both relieving eflects are distributed along
which avoid relative movements between the slabs and stabilize the vertical web the entire bridge length and do not depend on the longitudinal spacing of vertical
stiffeners. The concrete slabs also control torsional instability of the compression web stiffeners. Finally, the thickness of web plates is almost constant along the
flange through the diffused stabilizing effect of connectors, and general warping is bridge, and the design of launching bearings can therefore be optimized for most
prevented by the high cross—sectional moment of inertia about the vertical axis and support situations.
by the torsiQnal._Stifiness (lithebox,,section._. . _
 286 BRIDGE LAUNCHING PRESTRESSED COMPOSITE BRIDGES 287

5.2.3. Launching bottom region of web panels is high owing to the raised location of the gravity axis
Incremental launching of the finished section is probably the simplest and most for short-term composite action. Consequently, the total compressive stresses in the
rational construction method for medium-span (50—70 m) PCS bridges. Compared web region immediately above the bottom flange are rather high.
with the launch of the steel girder of a conventional composite bridge, dead load is The webs also resist the entire dead load shear forces until their reduction with
much greater due to the presence of two concrete slabs, and the web plates are tendon deviation on completion of launch, and the shear stresses are high even in
thinner since the final shear stresses are reduced by deviation of prestressing the absence of the torsional and distortional components due to placement
tendons. Therefore, dispersion of the launch support reactions within the web plates tolerances of launching bearings.
is a prime concern. ' Alignment of the support reactions along the web axes is another prime issue. If
When the longitudinal spacing of vertical web stiffeners is smaller than the pier the concrete slab is cast on to the bottom flange, centring the flange plate and the
cap width, the launch can occur with movable bearings integral with the super- web along a common axis as in Fig. 5.11 (scheme A) is not enough to align the
structure and placed every time under the next stiffener. This solution requires the support reaction, since the contact stresses transferred by the launch bearings
discontinuous launch process of Fig. 1.5, which is more expensive and much distribute according to stiffness, and the transverse stiffness of a concrete slab is
slower than continuous launching. On the other hand, continuous launching much greater than that of a thin flange plate. On the other hand, widening the
involves migration of the support reactions along the contact surfaces of the flange under the concrete slab so as to apply the support reaction uniformly to the
superstructure. concrete (scheme B) creates a transverse couple. Stiffening the external side of the
Support reactions applied far from the vertical web stifleners produce high flange as in scheme A of Fig. 5.12 improves the transverse equilibrium of the
vertical compressive stresses in the web plates. These stresses may govern the design bottom node and the dispersion of support reactions in the web plates. However,
of the web plates, and their dispersion is therefore a prime concern. Even from this these cost increases are not justified by permanent structural requirements.
point of View some types of lower node are more efficient. When the girder flange is Even from this point of view, the solution with a lower slab (scheme B in
placed above the concrete slab (schemes C and D of Fig. 5.10), the slab thickness Fig. 5.12) is more efficient. During launching, the neo—flon pads can be inserted
facilitates dispersion of support reactions, and the buckling strength of web plates under the intersection of the mid-planes of webs and bottom slab, the concrete node
increases due to the smaller panel depth. Generally, however, this is not suflficient improves dispersion of support reactions, and slab casting in situ compensates for
and dispersion of support reactions is attained by lengthening the launch bearings. the geometrical irregularities in the girder units and in their field splices.
The flexural stiffness of a PCS box girder is high and the flexural rotations of the When passing over launching bearings, the lower web region is subject to a
support sections of the continuous beam are therefore small. In spite of this, the complex triaxial state of stress. Lateral guides and a suitable design of the bottom
significant length of launching bearings requires rocking about the transverse axis cross-sectional nodes keep the support reactions aligned with the webs, and the
to distribute the contact stress uniformly. Pinned bearings are commonly used in transverse flexural stresses are relatively low in normal launch conditions (correctly
the launch of the steel girders of conventional composite bridges. In this case, levelled launch bearings, reasonable geometric accuracy, low thermal gradients). In
however, pin articulations are inadequate, since much higher support reactions this case, the state of stress in the web plates can be assumed to be plane, and the
result in expensive devices, and the long cantilevers in the rocking arms adversely
aflect the distribution of support reactions with their deflections. Steel frames
lodging multilayered elastomeric bearings may be placed onto jacks fed by a closed
hydraulic ring so as to create an articulation. However, the stiffness of launching
bearings is limited by economic reasons, and dispersion of the support reactions is
far from being uniform. In the most delicate cases, therefore, the vertical compres—
sive stresses in the web plates should be evaluated with 3D finite-element models of
the bottom cross-sectional node above the articulated launch bearing, and verified
by strain gauges attached to the web plates during launching. Such analyses show a
dispersion of the support reaction that differs from the 450 theoretical value, with
local concentration above the stiffer regions of the launch bearings.
When launching affects the finished section, axial prestressing is necessary to
control the edge stresses in the slabs, and both the concrete slabs and the steel
girders are subject to uniform compression. In the support zones of the continuous
, beam: the Comprsssivsstress producedbythc dendloadbending. momentinthe- Flg 511.. Transverse distribution of support reaction
288 BRIDGE LAUNCHING PRESTRESSED COMPOSITE BRIDGES 289

"Tu-"HTTP"

J
——-q-—
l
.i
I
\

\
\
\

II l/I/
II [III
\

l/////////////
//l I /

Fig. 5.13. Salbris Bridge (courtesy Virlogeux)

l/I/I/l/Il/l/l

W
Ill/ll/ll/l/ll

i
This concept may be extended to axial forces by folding the steel plates so as to
Fig. 5 .12. Transverse distribution of support reaction
reduce their in-plane axial stiffness without affecting the shear behaviour.
Corrugated steel or aluminium plates are commonly found in naval and aero—
allowable stress can be directly compared with the ideal stress resulting from the nautical engineering, and several fuselages and wing structures have been built with
Huber—Mises equation (4.24). Principal stresses in the concrete slabs are assessed by undulated or corrugated plates. With the same aim, the stiffened-plate webs of a
conventional criteria. PCS box girder can be replaced by steel plates mechanically folded so as to create
After some tests carried out in France [116], a small PCS bridge (Fig. 5.13) with the axial deformability necessary to prevent migration of prestressing forces and
stiffened-plate webs has been launched on a simply supported 37.2-m span with the internal moments, and the transverse stiifness necessary to prevent shear buckling
help of a temporary pier. Web plates were 12 mm thick, and external prestressing of web panels without additional stifieners (Fig. 5.14).
tendons were deviated at span thirds. However, the incremental launching construc- Fmite-element analyses, tests and the first actual structures confirmed many
tion method is generally applied to PCS bridges with corrugated-plate webs. advantages of the corrugated-web PCS over the stiffened-plate ones [4,132]:

o The lowest axial stiffness of webs avoids migration of prestressing forces and
5.3. PCS bridges with corrugated webs internal moments and the onset of long-term stresses due to the shortening of
The worst defect of a prestressed composite box girder with stiffened-plate webs is the concrete slabs over time. Consequently, the state of stress in the steel webs is
the capacity of the steel girders to resist axial stresses. Web stiffness causes migra— almost pure shear stress.
tion of prestressing forces and bending moments towards the steel girders. Without 0 The reduction in plate thickness permits a web design according to strength
these parasitic stresses, the webs would-resist shear forces. with almost constant criteria, and the shape of corrugations can be chosen so as to avoid buckling.
tangential stresses along the panel depth, and the concrete slabs would resist bend- Achieving this goal is practicable and economic when using 8-mm—thick web
ing moments through uniform axial stresses acting with a long lever arm. The axial plates, but the advantages are relevant even when a stiffened-plate web would be
stiffness of the steel girders should therefore be reduced without affecting the shear only lO-mm thick. On the other hand, however, corrugated plate webs are
stiffness, and this requires orthotropic behaviour of the steel girders. slightly more flexible in shear than plane webs.
Orthotropic plates with different flexural properties in the orthogonal directions 0 Pure shear stresses govern buckling stability instead of the combined shear—
are commonly used in steel bridges. Figure 4.4 shows a U-girder with webs and compressive stresses of stiffened—plate webs. However, local shear—compressive
bottom flange. heavily -stiffenedalong ,,,th.e direction of principal fleXural stresses. stress interaction arises above the launch bearings.
-
290 BRIDGE LAUNCHING PRESTRESSED COMPOSITE BRIDGES 291

5.3. I . State ofstress in the corrugated webs

The application of vertical forces to a PCS box girder generates bending moments
and shear forces. As with conventional PC box girders, the webs resist most of shear
forces. The shear force V applied in the plane of each web can be assumed as
uniformly distributed over the web, so that the applied tangential stress is
V

_ zwhw
(5.2)
where [W is the web plate thickness and hw is the web depth measured clear between
flanges.
Fig. 5.14. Corrugated—plate web PCS Letting hf be the transverse depth of the folds (Fig. 5.15), the axial deformation
modulus for the corrugated web is
t 2
EX 2 ES (—0 (5.3)
o The transverse flexural stiflness of the corrugated plates avoids the need for hr
additional stiffeners (except at supports), reduces the number of cross-frames
In practice, EX is about 1 / 1000 of ES. This reduction in the axial deformation
or diaphragms and permits savings in materials and labour. These sayings,
modulus governs the behaviour of webs in terms both of elastic work and of
along with the reduced plate thickness, are generally competitive compared
instability. Because of the axial flexibility of the corrugated plates, the webs do
with the cost of mechanical corrugation.
not resist flexural stresses and two equal and opposite longitudinal forces F3 in
0 Fewer tendons are needed to prestress the concrete slabs and, in such a
the concrete slabs balance. the external moment [133].
situation, fewer shear connectors.
o The sensitivity to premature buckling due to geometrical defects is decidedly F, =
h
(5.4)
lower. The defects of planarity must be compared with the amplitude of the
folds instead of the plate thickness, and their effects are 10—20 times lower. For the same reason, prestressing forces remain within the concrete slabs. Finite-
Local deformations due to welding are concentrated at the panel edges, which element analyses and tests carried out in France [116] showed that the corrugated
are fixed by the concrete slabs. The local stresses due to plate corrugation are webs resisted less than 1% of the prestressing force, and that the axial forces
uniform along the panel depth and do not intervene in the elastic equilibrium. In produced by prestressing in the concrete slabs can be calculated based on the
addition, the steel webs act in their theoretical design dominion. gravity axis of a cross-section composedofoniy the two concrete slabs.
o The marked flexibility of girder units facilitates the construction and improves The longitudinal compressive stress ax is negligible in the central region of
field assembly conditions, avoiding any specific tolerance (the web panels are the web panels. Since the vertical stress oz is also negligible far from the support
placed in contact with a minimum overlap and fillet-welded on both sides). sections, the state of stress in the corrugated webs is one of pure shear.
Welds between vertical edges of web plates are slightly more frequent than In a web element with 0X : 0 and oz 2 0, the two principal stresses 01 = ‘L' and
for plane webs owing to handling requirements during plate corrugation; 02 = —‘L' are inclined at 450 and confirm the Morsch lattice action of this type of
however, most of these are shop welds.
o The lower torsional stiffness of the corrugated web girders is not a problem
when they are used in box girder bridges because of the hollow-section torsional
behaviour of the composite section.

On the other hand, however, if the use of corrugated-plate webs in a PCS box girder
i...\ 113
bridge is to be economic, the job must include suflicient repetition to cover any one-
ofl costs for equipment and any know—how incurred by the fabricator. Therefore,
a bridge length of some hundreds of metres is necessary to obtain a good cost
balance. ‘ Lb F—"

Eig.___.5_..15.. _.Corrugated.._web geometry .__

 292 BRIDGE LAUNCHING PRESTRESSED COMPOSITE BRIDGES 293

plate. Along a vertical fold, the two inclined forces F1 2 01 tw and F2 2 02$“, do not the webs that was far from uniform, and to a more significant reduction in GS. Since
produce any longitudinal resultant, and the vertical resultant the concrete slabs of PCS box girders prevent flange deflections, the eflective shear
modulus of Equation (5.7) may be directly used in the analysis of the continuous
F, = (71t/5 . (5.5)
beam.
resists the shear force along the vertical fold [134]. In the transverse direction, the The application of axial compressive forces and bending moments to a box girder
action of the compression fibre, which tends to push the fold outwards, is balanced with corrugated-plate webs activates the specialized response of the individual
by the action of the tension fibre, which tends to pull it inwards. materials. Because of the axial deformability of the webs, the concrete slabs resist
This behaviour characterizes the central region of the webs, far from the flanges. the whole axial force and bending moment. The contribution of the webs to flexural
Shrinkage of concrete, the elastic slab shortening at prestressing and over time, and behaviour was found in finite—element analyses to be small, and in tests to be
the additional deformations due to live loads and thermal gradients, generate long- negligible. The immediate consequences are the high flexural efliciency of the
itudinal stresses in the web regions near flanges due to compatibility of longitudinal cross-section in the absence of bending-resistant material close to the centroid,
strains. The more constant is the corrugation of the web plates, the smaller are the and the behaviour of each material in uniform rather than triangular stress
fluctuations of these local stresses. They do not seem to cause fatigue, as indicated patterns. Moreover, the force of a prestressing tendon placed in a slab remains
by the rarity of problems in conventional composite bridges, which are subject to concentrated in that slab without affecting the opposite slab and the webs, nor
the same stresses. does elastic shortening at tendon tensioning induce compressive stresses in the
The flange plates do not contribute much to the flexural stiflness of the composite steel plates.
section, so their cross-sectional area should be as small as possible. As corrugation Finite-element analyses and tests confirmed that the longitudinal stresses resisted
of web plates increases the transverse width of flanges, their thickness must be by the corrugated-plate webs are negligible. On the other hand, they also showed
reduced to a minimum. Therefore, shear connectors are generally closely spaced that the shear forces resisted by the slabs, negligible in most of the superstructure,
to avoid buckling of compression flanges. increase significantly near the discontinuities of the shear force diagram [4,134].
For a geometrically perfect corrugated plate in pure shear, the mean shear strain Corresponding to a discontinuity AV in the shear force applied to the web, the
of an element several wavelengths long is shear stress discontinuity Ar tends to deform the web to a lozenge, i.e. to produce a
sudden angular discontinuity Ay in the longitudinal web axis.
V=— (56)
_A7:
A (5.10)

G6 = — (5.7) The concrete slabs are subject to axial loads that do not produce vertical deflec—
tions, and their flexural stiflness opposes sudden” angular discontinuities in the webs.
The Shear modulus GS depends on the elastic modulus of the steel and on the Consequently, local systems of orthogonal forces arise at the slab—flange interface
Poisson’s ratio 1) to transfer the shear forces necessary to bend the slabs until a common curvature
(Fig. 5.16) is reached.
Es The local forces required for the shear curvature of the slabs may be significant
and may require additional connectors in the tension zones of the interface. Slab
The ratio AW of the initial plate length to its final length after corrugation is reinforcement must be strengthened as well, since the flexural stiffness of the slabs is
_ 2Z7f (5.9)
locally considerable and produces high edge stresses. Finite-element analysis per-
mits evaluation of these local stresses in the web plates, the flange-slab interfaces
W—af‘l—bf
and the slabs, as well as the parametric assessment of approximate methods for
where af‘ is the longitudinal projection of the sloped web panel and bf is the distance their hand calculation.
between two folds, which is generally kept equal in the longitudinal and sloped This phenomenon of shear curvature is due to A V and therefore occurs near the
direction to achieve equal panel slenderness (Fig. 5.15). application points of vertical concentrated loads such as support reactions and
Tests and finite-element analyses Carried out in the UK [132] showed that the tendon deviation forces. Since support reactions migrate during the launch, the
reduction in GS obtained by replacing a flat web by a corrugated one of the same connectors welded to top flanges and the longitudinal reinforcement in both slabs
thickness exactly corresponds to the value of AW. In the same tests [135], the should be detailed so as to resist these local stresses over the entire length of the
-,,flexibility of unsupported flange,,,plates led to a distribution of. shear.-,stre,sses,,,in bridge . .
294 . BRIDGE LAUNCHING PRESTRESSED COMPOSITE BRIDGES 295

flexural stiffness of its plates, which together oppose distortion. In practice,

torsional forces are low in rectilinear box girders, and distortion fades rapidly.

NI'TI
The behaviour of a prestressed composite box girder with corrugated webs is

2 ____._._ conceptually similar, but the effects are different because of the different stiffness of
the plates. The transverse flexural stiffness of a corrugated web is smaller than that
of a concrete web, and the cross~section is therefore more deformable. The trans-
verse flexural stiffness of webs is also much smaller than that of the concrete slabs,
the cross-section behaves as hinged at the nodes, and the contribution of the cross—
sectional frame action (term k in Equation 3.18) is minimum.
Corrugated webs are unable to resist in—plane flexural stresses. The warping
cross—sectional stiffness (term Cd in Equation 3.18) decreases since it is generated
only by the concrete slabs, and the shear deformability of webs makes things worse.
slob-web connection forces Consequently, distortion generated by an eccentric load is mainly resisted through
1 opposite in—plane deflections of the concrete slabs.
To study distortion, an eccentric load F applied to the deck slab can be
decomposed into two symmetrical loads F/2 applied to the upper nodes of the
theoretical distribution of shear stress cross—section (nodes 1 and 2 in Fig. 3.1) and two hemisymmetrical loads Fh,
equal in value and opposite in sign, applied to the same nodes.
Fig. 5.16. Slabmweb interaction near discontinuities 0f the shear force Symmetrical loads cause longitudinal bending and shear. The shear force pro-
duced by F1, in each web, assumed to be hinged at the edges and with the symbols of
Fig. 3.1, is
This phenomenon is not a peculiarity of PCS bridges. Similar effects occur in the c
support regions of non-prestressed composite bridges, where the web plates are V = Fh (5.11)
12
thicker but the shear forces are higher since they are not reduced by deviation of
prestressing tendons. In a non-prestressed composite bridge, however, two facts This force is higher than the shear force produced by a torsional moment
reduce the local stresses produced by shear curvature: T = Fha (5.12)
0 At the flange—slab interface, the effects of Ar are related only to superimposed and it is equal to that due to a moment
dead and live loads as dead load is rarely resisted with composite action.
0 These local systems of forces appear to depend inversely on the square root of T = Fh(a + b) (5.13)
the web area, and are therefore higher in thinner webs. Therefore, the shear force in the upper slab can be obtained by superimposing a
The assumption that plane sections remain plane is not satisfied in these types of force equal to
structures. It may be adopted in the preliminary design phases and to determine a
Vt = Fh (5.14)
longitudinal bending moments, shear forces and prestressing forces in the con- n
tinuous beam. However, analysis of the state of stress in the slabs and the webs on the shear force resulting from the torsional moment (5.13) — the negative sign is
of the support regions and near the deviation points of prestressing tendons to remind us that this force is opposed in direction to the torsional shear force in the
requires more refined approaches. " - - - '- slab [136]. The shear force in the bottom slab can be obtained by superimposing a
force equal to
5.3.2. Cross-sectional deformability
In a PC box girder, the effects of eccentric loads deviate from Bredt’s theory of pure b _
Vb 2—"t * (5.15)
torsion, and the cross-Sectional deformation results from the superimposition of a
rigid rotation and a distortion. These effects depend respectively on the torsional on the torsional shear force. Taking into account the rigid connection between webs
stiffness of the box girder, which limits its capacity to rotate about its axis, and on and slabs slightly reduces these shear forces, and the torsional moment (5.13)
-theinteraction between- the framewstiffness of the crossssectionrandmthe-insplaneww . .—.»—v-slightly—. decreases-aswell-,although— it remains much higher than the theoretical one.
296 BRIDGE LAUNCHING PRESTRESSED COMPOSITE BRIDGES 297

Under these horizontal forces, the concrete slabs deflect according to their reactions migrate during the incremental launching of the superstructure, these
in-plane flexural stiffness and the constraint action exerted by cross-frames or local forces affect the entire length of the bridge.
diaphragms, and the edge stresses may reach high values. The hinged—rod action 0 Connectors must transfer transverse bending moments generated by in-plane
of webs and the low torsional stiffness of the slab plates superimpose large rotations bending of the slabs when the cross-section distorts.
on these transverse deflections [4,116]. o Connectors welded to the bottom flanges must resist local flexural stresses
The shear forces in the two slabs are similar (a is not very different from b), but caused by transverse eccentricity of the shear forces in the webs from the
the in-plane flexural stiffness of the slab plates is very different, and the webs are application axis of the support reactions (Fig. 5.17).
therefore subject to large transverse deflections. Although their transverse flexural Connectors are designed according to conventional criteria for non-prestressed
stiffness is low, the lever arm hf of the folds is short, and the superficial stresses may composite sections. The longitudinal shear forces between the steel girders and
reach the same values as the tangential stresses. These transverse bending moments the concrete slabs depend on the shear force diagram in the continuous beam,
propagate to the slabs, and both the connection devices and local slab reinforce— Which is corrected by the deviation forces of prestressing tendons. A high level of
ment require a specific design. Transverse reinforcement in the upper slab is often polygonal prestressing is necessary to reduce the shear stresses in the webs and to
adequate, since this slab resists the service loads as simply supported on the webs. force the state of stress in the unloaded structure to approach an axial compression,
Transverse reinforcement in the bottom slab may require some integration. thus ensuring the longest durability. PCS box girder bridges are flexible structures,
To summarize, application of eccentric - loads to a PCS box girder with and a high level of polygonal prestressing is also necessary to control mid-span
corrugated—plate webs produces specific stresses: deflections, since incremental launching of the finished superstructure prevents the
application of precambers.
0 The torsional shear forces in webs and slabs are higher than those resulting from
Prestressing techniques for PCS box girders are the same as those discusSed in
the classical theory of torsion.
Section 3.7 with regard to external and mixed prestressing. Design criteria for-
0 Longitudinal warping stresses arise in both slabs, especially in the bottom slab,
launch and service prestressing are the same as well. The most effective approach
corresponding to bending moments about the vertical axis. ‘
is to balance the shear forces due to permanent loads and 50% of live loads with
0 Bending moments about the longitudinal axis arise at the connections between 4
tendon deviation, so that the shear stresses in the webs are due only to the presence
slabs and webs, with related local effects in both. These effects are due to cross-
or absence of live loads. Additional prestress demand aimed at controlling the edge
sectional distortion and are unfavourable compared with conventional PC box
stresses may be met with internal tendons, placed in the bottom slab in mid-span
girders.
and in the top slab above the piers. Internal tendons are more effective than external
Cross-sectional distortion must be limited by frequent cross—frames or diaphragms tendons because of the longer lever arm; they also require smaller anchor blisters
that integrate the action of support and tendon deviation diaphragms. Because of and do not cause congestion of the box cell. Adoption of varying longitudinal
the fundamental role of prestressing in reducing shear forces, the spacing of the
main diaphragms should be defined for optimum tendon deviation, and distortion
can be further reduced by additional lighter devices.

5.3.3. Connection between webs and slabs

Connectors welded to the flange plates have to meet several requirements.

0 The area of top and bottom flanges is reduced to a minimum in order not to
accumulate axial forces and to better control thermal shrinkage of concrete with
a low )3 factor (Equation 4.28). Wide flanges are necessary to accommodate the
web corrugation and to contain concrete, so the plate thickness is reduced to a
minimum. Consequently, connectors must also control buckling of compression
flanges.
o Connectors transfer longitudinal shear forces and systems of orthogonal forces.
The dead load of the suspended bottom slab and the local stresses due to shear
curvature. tend.......to. separate the slabs from the girders, and «Since. ...Supp0rt.-... Fig. 5J7. Local eccentricity of support reactions
-
 298 BRIDGE LAUNCHING PRESTRESSED COMPOSITE BRIDGES 299

prestressing does not cause problems or indeterminate eflects, as the prestressing (5.16)
P11 2 Pu -— %
force applied to a slab affects only that slab and does not migrate towards the webs
and the opposite slab. Finally, thin upper slabs may require a light transverse
where P,l is the nominal static ultimate shear strength. BS 5400 [55] provides the
prestressing in the case of very inclined webs [137]. ,
nominal static strength of different types of shear connectors (headed studs, bars
Permanent polygonal tendons may be subject to stress variations and related
with hoops, channels, and friction grip bolts) for different concrete strengths. BS
fatigue effects because of the enhanced flexibility of the superstructure. Therefore,
5400 [55] also requires that the overall height of a connector should be not less than
the use of devices aimed at avoiding tendon slippage into the grouted ducts at
100 mm, and that the surface of connectors that resists separation forces should not
deviators may be advisable. Even when a low level of polygonal prestreSsing is
extend less than 40 mm beyond the transverse reinforcement of concrete slabs.
adopted, reduction in the shear forces is significant and flange-slab interfaces are
unaffected by prestress migration. Consequently, shear connection between slabs
and flanges is not a major problem.
5.3.4. Web instability
_
The general modes of instability of prestressed composite box girders with
Several design standards rule the maximum spacing of flange connectors. BS
corrugated webs are similar to those already discussed for PCS bridges with
5400 [55] requires that longitudinal spacing be not greater than 600 mm or three
stiffened—plate webs. With regard to panel buckling, however, the webs are to be
times the slab thickness or four times the height of connectors. This last require-
assessed with specific criteria.
ment is likely to prevail; however, a smaller spacing should be adopted in the case of
The design of stiffened-web PCS and conventional composite bridges is often
thin flanges (10 or 12-min plates) to control flange buckling in the regions far from
based on stability rather than strength criteria. In contrast, the normal design
connectors. j
approach for a corrugated-web PCS is to define the plate thickness based on the
Headed stud connectors are not much suitable for thin flanges. BS 5400 [55]
yield stress in shear ry of the steel, and then to determine the shape and amplitude of
requires that the diameter of connectors welded to a tension flange should not
the undulations so that the ultimate tangential stress for instability exceeds ry. This
exceed one and half times the plate thickness. Where the flange plate is not
allows reaching the desired margins of safety with marginal differences in corruga—
subject to tensile stresses the diameter of stud connectors should not exceed
tion costs. Achieving this goal is often practicable and economic; when impossible,
twice the plate thickness. In the presence of partial launch prestressing, any region
the ratio td/ty, where rd is the design ultimate shear stress, should lie between 0.8
of both flanges is subject to tensile stresses during launching, so the first require-
and 1.0 for efficient use of steel.
ment prevails. This results in a lot of small connectors, which is impractical and
The Von Mises yield criterion can be used, so the yield stress in shear is
increases fabrication costs. L—shaped connectors with two through bars at the
ends to control transverse bending or channel connectors are more practical,
and stabilize the flange plate better. L—shaped or channel connectors are welded
z = f—y : 0.577 f, (5.17)
y «5
to the flange in correspondence to the web folds to stiffen the flange near folds
and to prevent compressive buckling at prestressing from locally detaching the where fy is the tensile yield strength of steel. If there is no buckling, the design
flange from the slab. Therefore, their longitudinal spacing (af and bf alternately, ultimate stress rd is equal to ‘L'y.
Fig. 5.15) is not constant and varies between 250 and 400 mm, with smaller Corrugated plates show three difierent buckling modes related to the ratio of the
values near permanent support sections. depth of the folds to their width. One mode (local buckling) corresponds to the pure
BS 5400 [55] also requires that the leg length of the welds joining the connectors shear instability of a flat web panel supported between two vertical folds. It is
to the flange should not exceed half the thickness of the flange plate, and that the reversible and preponderant in webs with deep folds, and does not cause significant
distance between the edges of shear connectors and the edges of the flange plate risks of general instability. The longitudinal panels and the sloped panels of the
should be not less than 25 mm. When the flange plates are wide to ease forming of corrugated webs have the same width bf to achieve constant slenderness, and setting
the upper slab, their small thickness suggests not reaching this code limit, and the Poisson’s ratio 1) = 0.30, the critical shear stress for local buckling of a web panel of
transverse width of L-shaped or channel connectors should not greatly exceed the indefinite depth (in practice, hw >> bf) is determined by the Timoshenko equation:
transverse fold depth hf. 2
I
Because of these combined requirements, in the presence of significant orthogo- Trim 2 4.826ES(—W) (5.18)
nal forces and thin flange plates, integrative headed stud connectors may be added bf
without modifying the longitudinal spacing of main connectors. In this case, BS Since the web panel is hinged to elastic constraints and small out-of—plane move-
5400 [55] requires that the nominal static ultimate shear strength P1, of stud ments are possible, the critical shear stress for local buckling rod is slightly lower.
,,,,,Connect0rs.. in ._th_¢____pre_s.en¢epf trnsiqnlfu should bentaken. as, ., , French researchers proposed the following expression [132]:
300 BRIDGE LAUNCHING PRESTRESSED COMPOSITE BRIDGES 301

2
Z negligible over a length of the same order as the depth of the web. In a PCS bridge,
{CLLI = 0.88'L'Tim 2 4.247ES (ID—w) (5.19)
f
both flanges are fixed to concrete slabs and can therefore provide significant
restraint either to warping or to rotation about a longitudinal axis. Therefore, it
Studies of the effects of geometrical imperfections and residual stresses have led to a
appears that the value K = 60.4 (used for the design of the Pont de la Corniche)
second expression for the critical shear stress for local buckling. According to EC3
may be appropriate.
[90,138] it is
A third mode (interactive buckling) has been identified in tests on corrugated
Tcr.L2 = 09V TcrLl Ty (520)
plates carried out in Sweden. It is a sudden phenomenon characterized by a sharp
snap and steel plasticization along the folds. It is critical in webs with folds of
that is medium depth. The following formula, based on a series of tests, has been proposed
l‘ for the critical shear stress fer-I for interactive buckling (however, some warping and
:0a = 1.40917: E93, (5.21) imperfections were presumably present in the tests).
The critical shear stress for local buckling $a is taken as the smaller of 70,141 and 1 l l
ram (the latter generally prevails). rcrl Tcr.Ll I7cr.G
A second mode (global buckling) corresponds to the pure shear instability of the
web panel and is critical in webs with shallow folds. It is characterized by a regular Allowance should be made for imperfections and residual stresses. French research-
wide wavelength deformation, progressiVe in the onset and development if in the ers [116,138] proposed taking a design value for interactive buckling stress equal to
absence of local buckling, and similar to that of an orthotropic plate. If local T1 2 0.6712311 ‘ (5.26)
buckling occurs in the postcritical dominion, the corrugated plate may lose the
stability necessary to reach shear yielding. Easley investigated global buckling and a design value for global buckling stress equal to
theoretically and with tests on 0.4—mm aluminium plates. Tests on steel plates
carried out in France [133] and UK [132] eventually confirmed Easley’s results. :6 = 050,7c ' (5.27)
The transverse fiexural stiffness of a unit web length is determined first (hf is the These margins of safety have been adopted for the design of the Pont de la Corniche
fold depth). [138]. A higher margin from global buckling (IG 2 0.33ICLG) has been adopted for
the Charolles Bridge. Further research might lead to relaxation or refinement of
Dx 2 ES twhf
(522) these rules. The design value for local buckling stress might be based on the EC3
3(ar + bf)
[90] provisions (yMl = 1.1)
The longitudinal flexural stiffness, reduced by AW (Equation 5.9) is
:1, = 0.91%], (5.28)
DZ=+M3_W_:0.
Ez‘3 :3
12(l—v2))tw 09 1E_W
6 SAW (5 _ 23) Different margins of safety have been proposed for the three buckling modes to
-
reflect the different postcritical dominions. When global buckling occurs, the shear
Where DX 2 50D, (as usual in practice), the critical shear stress for global buckling strength decreases and the plate must assume the large deformations required for
is defined by the Easley equation stretching the folds before being able to transfer diagonal tension forces through
«4/ DZDi membrane action. Tests carried out in France confirmed the existence of a
Tcr.G : K (524) postcritical dominion for corrugated plates. However, the diagonal deformations
twhév necessary to achieve postcritical stability are often excessive for bridge structures.
Where kw is the net depth of the web. panel clear between flanges. K varies from 36.0 The margins of safety from global and interactive buckling should therefore be
for a hinged panel to 68.4 for a perfectly fixed one according to Easley [139], and higher than that for local buckling, for which the postbuckling behaviour is
from 32.4 for a panel fixed to elastic slabs to 60.4 for stiff unsettling slabs according good. As EC3 [90] allows for the postcritical behaviour of conventionally stiffened
to Bergfeld [140]. The hinged-panel assumption (K = 36.0) neglects the restraint to plates, the use of yM1 = 1.1 appears adequate only for local buckling.
warping of horizontal cross-sections of the web panel that is provided by the con- The use of higher safety factors does not have economic consequences, as the
crete slabs. Finite—element analyses have shown that this restraint action greatly critical shear stresses for global and interactive buckling are often higher than the
increases the global buckling stress. The use of the no-warping model (K = 60.4) critical shear stress of an equivalent conventionally stiffened plate. In addition,
implies thatthc ___fl_anges are so stiff normal to- their plane. that, the. deformation, is, , adoption of deeper folds may increase the critical shear stresses for global and
 302 BRIDGE LAUNCHING PRESTRESSED COMPOSITE BRIDGES 303

4,00 l
interactive buckling with minimum repercussions in terms of corrugation costs and
-<>- Global
of shorter final plate length due to higher values for AW. + Interactive

critical shear stress/yield stress

The design ultimate shear stress rd is taken as the smallest of the yield stress in 3,00 — + Local
shear (5.17), the local stress (5.28), the global stress (5.27) and the interactive stress
(5.26). The shear stress computed in the web plates at the ULS should be smaller
than rd to allow for load differences between the two webs due to torsion and 2,00

1
distortion.

0
The ratio of the design shear stresses for the individual buckling modes to the

>0
I>
yield stress in shear for the fold profile of the Pont de la Corniche (8-mm—thick 1,00

webs, hf = 220 mm, K 2 60.4, and the safety factors discussed above) is plotted in
Fig. 5.18 versus kw. In this case, interactive buckling prevails. With the same
0,00
corrugation parameters and 10-min webs, Fig. 5.19 shows that yielding prevails. 1 000 1500 2000 2500 3000 3500 4000
The situation of the Charolles Bridge [141—-—144] is different because of the small web plate depth (mm)
flexural stifiness of the bottom slab (a 610—mm composite tube) and the shallow
folding of the corrugated webs (hf = 150 mm). With K = 32.4 and the higher safety Fig. 5 .19. Design shear stresses for the P011: de [a Corniche, IO-mm web plates
factor from global buckling adopted in the design, global buckling prevails on
yielding for hw beyond 2900 mm (Fig. 5.20). With the fold profile adopted for
the Ginzan-Miyuki Bridge [145] (discussed below) and K = 60.4, yield stress in A procedure has been proposed in [135]:
shear prevails in the 9-mm plates (Fig. 5.21). The shear behaviour of the Cognac 0 Assume a value of bf/rw.
Bridge [146,147] is intermediate between these last two cases. 0 Determine the critical local buckling stress from Equations (5.19) and (5.21).
A corrugated-plate web is normally designed for a known depth 11W and a known The related design shear stress for local buckling (Equation 5.28) usually
Vertical shear. The corrugation angle is normally between 250 and 320. In a plate eXceeds the yield stress in shear and does not govern design, because interactive
girder, the ratio of ca to bf (Fig. 5.15) may be influenced by the need to limit the buckling is more critical (Figs 5.18—5.21).
maximum flange outstand [132], but this need does not arise in a PCS box girder. 0 Determine the critical interactive buckling stress rod from Equation (5.26) by
Web panels have the same width bf in the longitudinal and sloped direction to setting T1 equal to the required value of rd (for instance rd 2 0.9ty).
achieve the same slenderness. For efiicient use of steel, the ratio rd/ry should not c When both tori and 1‘a are found, determine "ECLG from Equation (5.25).
be less than 0.8. It remains to find 1,, and bf.

4,00 i _ A
4,00 , , I
-o-(mobal V\\\\ —0— Global
-0— Interactive + Interactive
critical shear stress/yield stress

critical shear stress/yield stress

3’00 __ “'11— Local
3,00 \ -A- Local ,_

2,00 2,00 \ \\0\

1,00 .. 1,00
II I’-
I>

0h

0,00 . 0,00 ,
1 000 1500' 2000 2500 3000 3500 4000 1 000 1500 2000 2500 3000 3500 4000
web plate depth (mm) web plate depth (mm)

,,,,,Fig. 5.18- Designs/war Stressesfor the. Pom de la CorniCh.e.,,,.8-mm .web,,pla,t,es-, Fig. 5.20. Design shearszresses for. the Charolle Bridge, 8-mm web plates
304 . BRIDGE LAUNCHING PRESTRESSED COMPOSITE BRIDGES 305

4,00 J
where the local shear stress I is determined by taking into account the reduc-
—0— Global
tion in the shear force resisted by webs because of the ‘shear—curvature’ effect in

es \
+ interactive

critical shear stress/yield stress

3’00 \ + Local _ the concrete slabs. Finally, local buckling of flange plates is centrolled by
reducing the longitudinal spacing of the main shear connectors or by adding
intermediate headed studs. Adjusting flange plate thickness to the width is
2,00 inadvisable.

.
3%. ._.
A J A A \E‘i A A -
.1 H 1.: H z... ‘—I 5.- mi A-I VF

1,00 5.3.5. Fabrication of corrugated plate girders

Problems of large-scale fabrication using webs at least 8 mm thick should be
addressed to provide sufficient robustness and durability. Higher cost of fabrication
0,00
1 000 1 500 2000 2500 3000 3500 4000 is generally perceived to be the main disadvantage of corrugated-plate webs [148].
web plate depth (mm) However, for a fabricator with profile—following welding equipment and with
experience, the difference should be small, since vertical and longitudinal stiffeners
Fig. 5.21. Design shear stresses for the Ginzari-Miyuki Bridge, 9-mm web plates are eliminated.
Corrugated web girders are fabricated by welding flange plates to the top and
bottom edges of the web plates after their pressing to shape. Shear connectors may
be welded to the flange plates prior to welding the latter to web plates. Web plate
0 From Equations (5.23) and (5.24),rcr_G can be expressed as a_ function of bf/tw dimensions depend on commercial standards, and it is reasonable to assume that
and kw / 1W. Therefore, design curves with bf/tw in the abscissa and hw/ 1,, in the steel plates of width hw up to about 3 m, thickness tw up to 12 mm and length up to
ordinate can be drawn for the desired ratio rd/ry and the chosen shape of web. 15 m, are available for corrugation.
During the launch, the corrugated webs are subject to vertical compressive stresses The least expensive method for forming the corrugation is probably cold
caused by the dispersion of support reactions. Preliminary values for az may be pressing. The web plates for the Cognac Bridge were corrugated by overturning
obtained by assuming a 450 dispersion through the bottom slab and the bottom the plate after each pressing. Eventually, one complete 4-fold corrugation was
flange. In the most delicate cases (thin webs, thin bottom slab, flexible launch cold-pressed at a time, by placing the plate on rollers to avoid stretching
bearings), more realistic values for 02 may be determined by finite-element analyses (Fig. 5.22).
of the interaction between the launch bearings and the lower deck region. The
systems of forces applied to these local models should take into account differences
in elevation of launching bearings and irregularities in the contact surfaces of the
bottom slab. The boundary conditions should be such as to generate the direct . , / .
/ /
eflccts of transverse unbalance of the support reactions (different vertical forces / // /
/
/
/
applied to the contact surfaces) and the related indirect effects (cross-sectional ’ / //
/

distortion with transverse flexural stresses in the web regions adjacent to the bottom
flanges). 1 .lilllj n
llllllll
IIIIIIIIII

Values for the local yielding strength and the local buckling strength may be
determined using Equations (4.16) and (4.18) by taking into account the load
dispersion through the bottom slab and flange. Then, the lower value of the local
vertical strength may be used to determine the design ultimate compressive stress ad
according to the expected load dispersion.
Taking into account the shear—compressive stress interaction at the ULS, the
following limitation has been proposed:

(i) +(_) 51.8

0 2 'L' 2
(5.29)
, ,,: Ed Id .. ., . . . . , , . . ,, Fig- 522- Preqiiflgeqalpmerli and plastic biases in the web plate
306 ' BRIDGE LAUNCHING PRESTRESSED COMPOSITE BRIDGES 307

Presses for steel plates up to 4 m wide are available in Europe. The pressing force 5.3.6. Field activities
F can be estimated by assuming four plastic hinges in the plate. Allowing for strain The launching yard organization for a prestressed composite bridge is similar to
hardening, by replacing the yield stress by the ultimate tensile stress fu, the peak that for a conventional PC box girder, as concrete casting and prestressing require-
force is [148] ments g0vern in both» cases.
Double fillet welding of the field splices of corrugated webs is so inexpensive that
eh... (5.30) the use of short girder units may be advisable to decrease the hoisting capacity of
Fmax :fu
61f the field cranes. Whatever the length of girder units may be, the length of the slab
segments should be kept as constant as possible to avoid the need for frequent
and the plastic work for a complete corrugation is (or in radians) [148] shifting of the form shutter. Field splices of the steel girders and construction joints
of the slabs can of course be staggered.
Wp = furivhwoz (5.31) The use of precast slab segments does not appear particularly advantageous, as
discontinuous continuity openings filled with concrete cast in situ concentrate the
For f, = 510 N/mmz, 1W = 8 mm, hw = 1000 mm, 01 = 300 and b, :- 300 mm, the continuity forces excessively for such slender flanges. However, specific continuOus
force and work requirements are modest: Fmax = 126 kN and Wp : 17.1 kJ. The
connection systems might be envisaged, and the longitudinal shear forces are
rate of pressing is mainly governed by the time taken to move the plate one wave-
reduced by deviation of prestressing tendons. The need for total prestressing in
length between each stroke. Even if the cycle time is as short as 10 s, the mean
the absence of continuous longitudinal reinforcement in the concrete slabs causes
power required is only 17.1/10 = 1.7 kW. At this rate, corrugation of a plate
additional costs.
14.5 m long would take about 2 min and provide a 13.5-m length of web (12
Two-phase casting in two adjacent formworks is an effective yard organization for
complete corrugations with 1120-mm wavelength, AW 2 1.072). Heavier equipment
PCS bridges. In the rear casting cell, the girder units are placed on adjustable
would be needed for a 12mm web plate, 2.0 m wide. As bf is likely to be propor-
supports. Plan and vertical alignment of the girder units is facilitated by their
tional to tw, the force Fmax is proportional to twhw, and so would be 378 kN in this
flexibility. Along with in situ casting of the concrete slabs, this avoids the need for
case.
stringent geometrical tolerances in the steelwork and simplifies field assembly activ—
The angle or is generally between 250 and 320. This enables a generous bend
ities. After alignment, the corrugated plates are double fillet welded to the rear end
radius, such as 101W, to be used for the corners, which also reduces the stress
of the steel girders, and flanges may be either butt joint welded or left discontinuous.
concentration in the flange—to—web welds at these points. Flange plates are
Preassembly of the bottom slab reinforcement is diflicult in the presence of
welded orthogonal to the corrugated plates and the box girders with inclined
L-shaped connectors with through bars. Channel connectors are preferable for
webs are obtained by shaping the surface of both concrete slabs with local
this purpose. In-form assembly of the reinforcement cage is expensive due to the
haunches. Research on the stress concentration in flange-to-web welds is in
restrained WOrk environment and the needtowork below the bottom flanges. After
progress in France. Submerged-arc profile-following equipment has been used
completion of reinforcement assembly, concrete is cast by conventional methods.
in France, and profiled gas-shielded fillet welding in Austria for 3-mm webs
The bottom form is composed of two longitudinal extraction rails and a central
used in buildings. Manual fillet welding appears to be the only alternative.
collapsible form table. Precambering is not used in PCS launched bridges because of
Conventional radiographic or magnetoscopic controls are easily applicable to
the significant reductiOn in the permanent deflections obtained by deviating the
these welds. .
polygonal tendons. The absence of precambers facilitates the launch, and the
Transport requirements and the wavelength of web folds govern the optimum
extraction rails are directly set along the local vertical profile of the launch plane.
length of the girder units. Girder segments 12 m long have been used for the
Extraction rails may be cambered upwards to create a diffused positive moment
Charente Bridge in France, and 10—14-m segments have been used for the
at segment extraction. This solution may effectively compress the top slab, and the
Charolles Bridge along with 12-m slab segments. For the Ginzan-Miyuki Bridge
corresponding tensile stresses in the bottom slab are advantageous in the negative
in Japan 5.5-m girder units have been uSed along with ll-m slab segments. ' i ' '
bending regions, and without major consequences in mid—span. Similar results may
Web units can be joined by overlapping and double fillet welding or by friction-
also be obtained by shifting the resultant force of launch prestressing.
grip bolting, with or without cover plates. The cost of field splices is much lower
In the front casting cell, the top slab is cast directly on to the steel girders.
than in conventional composite bridges. Specific procedures for precambering
Integral cage preassembly or the use of preassembled bar grids for bottom and
corrugated webs have been proposed in [148]; however, precambers are not used
top reinforcement are both possible. Compacting-roll finishing can be used
in PCS bridges built by incremental launching.
extensively for the top slab surface. The forming system is composed of two
external forms supported on modular towers for the cantilevers, and a central
 308 BRIDGE LAUNCHING PRESTRESSED COMPOSITE BRIDGES 309

form table suspended to the steel girders and extracted backwards after launching The elastic deflection of the front cantilever may be recovered by a hydraulic
to be repositioned for the new slab segment. Cantilever forms can shift along the pantograph applied to the front end of the superstructure. Cantilever deflection is
extraction rails to cast several short slab segments, and in this case the bottom and much larger than for a PC bridge, and this discourages the use of hydraulic jacks
top slab segments can be cast together because of the small concrete quantities supporting launching saddles.
involved. In the case of longer slab segments, casting can be staggered over time. All the thrust systems examined for PC bridges may be used, back-thrust systems
The launch prestressing tendons are anchored near the construction joints. The included. When friction thrusters are used, great care should be exerted when
use of internal tendons is advisable as they can be joined with couplers, do not setting the electronic controls for the low—speed final approach of the hoisting
require anchor beams and may be tensioned after a short concrete curing. If the saddles, not to apply excessive vertical forces to the superstructure. The winch-
length of the curing zone is suflicient, the launch tendons may cover several slab based systems for steel girders are excessively slow when significantly de—multiplied.
segments. This solution reduces the cost of couplers and is requested by those codes The use of launch bearings integral with the superstructure (Fig. 1.5) is expensive
that limit the number of tendons coupled in any cross-section to not more than and time-consuming. When fixed launch bearings are used, however, the support
50% of the total number. reactions migrate along the contact surface of the bottom slab. A bottom node
As with conventional PC bridges, the PCS superstructure is launched directly at its scheme with the flange plate placed above the concrete slab is practically unavoid-
final level, since the high self-weight discourages vertical shifting. Launching bearings able (Fig. 5.17), since the minimum thickness of web plates requires accurate
are removed on completion of launch to cast the bearing seats and insert the dispersion of the concentrated actions, and this can be easily attained by locally
permanent bearings under the superstructure without major hoisting. When rocking increasing the slab thickness with a transverse haunch.
launch bearings are used, the vertical clearance requested by their depth is attained by The corrugated profile of the web plates makes the support reactions eccentric
temporarily engraving the pier caps. As an alternative, higher permanent bearing with respect to the web panels, and the bottom slab must therefore resist local
seats may be used, although they are often not aesthetically pleasing. transverse couples. Local thickening of the bottom slab (Fig. 5.17) increases the
transverse flexural stiffness and improves bottom flange stability far from connec-
tors. It also improves the longitudinal dispersion of support reactions, and the
5.3.7. Launching . allowable stresses in the web plates increase due to the smaller panel depth. This
Incremental launching is likely to be the most competitive construction method for scheme of node also improves corrosion protection. The use of wide neo—flon pads
medium—span (50 to 70 m) PCS bridges some hundreds of metres long. In these improves the transverse distribution of support reactions, provided that their length
bridges, launching the complete section is simpler and more rational than in situ is suflficient for adequate longitudinal dispersion of vertical compressive stresses in
casting of the concrete slabs on to the steel girders, although the dead load of the the web plates. Lateral launch guides should always be adopted at every pier to
finished superstructure is significant. . avoid additional eccentricities.
The dead load is about five times greater than the steel girder of a non-prestressed A PCS bridge is launched with finished slabs, the absence of precambers aVOids
composite bridge, and the launch support reactions increase in the same propor— angular discontinuities in the launch surface and few modes of instability can arise-in
tion. Conversely, the webs are thinner, as the corrugated plates are designed for the web panels. Elastic stability of a corrugated plate is better than that of a stiffened
strength rather than stability, and the permanent shear stresses are reduced by plate since the distance between the folds is much shorter than the usual spacing of
tendon deviation. vertical stifleners. However, the corrugated-plate webs are thinner, and longitudinal
The launching nose must be quite long to reduce the anomalies in the launch dispersion of support reactions often requires the use of long launch bearings.
flexural stresses. This requirement is not particularly expensive, as the steel girders The flexural stiffness of the finished section reduces the rotations of the support
themselves can be used as a launching nose. In this case, the bottom slab is cast in the sections of the continuous beam. In spite of this, however, the significant length of
launching yard for its entire length, while the top slab of the front span is cast in situ on the launch bearings requires rocking about the transverse axis to distribute the
completion of launch. The U-section thus obtained in the front zone is much lighter vertical stresses uniformly in the web plates. Pinned hinges are inadequate since
and particularly effective under negative moments, and the additional costs are only the long cantilevers of the rocking arms disturb the distribution of support reac-
related to stifler top flanges and a temporary horizontal bracing between them. tions, and mechanical articulations are expensive. Rigid frames lodging steel skids
When the bottom slab is wide and the front U—section is therefore excessively covered with tensioned stainless steel sheets or multilayered elastomeric bearings
heavy, the use of a temporary cable—stayed system may be less expensive than a (Fig. 4.11) may be placed on to jacks fed by a common hydraulic circuit to create
specific launching nose. This solution, however, requires the presence of a distribu- an articulation. Control valves between the two hydraulic support circuits at each
tion diaphragm under the tower of the stayed system, and the concentrated load pier make it possible to optimize the tranSverse distribution of the support reac-
, disturbs, the, 1a,un_9h___m9ment.ems/10116., , tiOIlS, thus avoiding overloading the. steel webs. Sliding bearings based on neo-flon
 310 BRIDGE LAUNCHING PRESTRESSED COMPOSITE BRIDGES 311

pads are preferable to roller bearings because of the overall support reactions and The principal compressive stresses and the local stresses in the concrete slabs can be
the higher flexibility; rollers may be useful for unattended segment extraction from assessed by conventional criteria.
the casting cell.
Economic factors limit the stiffness of the launch bearings, and the longitudinal 5.3.8. Case studies
dispersion of support reactions is far from being uniform. Local stress concentra— Prestressed composite sections with corrugated-plate webs are particularly suitable
tions also arise during the hoisting operations for maintenance of launch bearings for balanced cantilever construction on medium spans. In varying—depth super-
and for insertion of permanent bearings on completion of launch. The slenderness structures, the extreme weight reduction in the webs results in significant increases
of web plates and the flexibility of launching bearings sometimes suggest analysis of in flexural efficiency. A 496—111 bridge (Pont de la Corniche) has been built in France
the support reaction dispersion through the bottom slab. Finite-element models of with five 80-m continuous spans and two 48—m end spans [138]. The top slab is
the interaction between the support regions of the superstructure and the launch 14.5 m wide and the bottom slab width increases from 5.15 m at supports to 7.2 m
bearings also provide useful information about the vertical stresses at the flange— at mid—span to keep a constant web inclinatiOn of 17". The overall depth of the
slab interface generated by shear curvature (Fig. 5.16). As a result of these internal superstructure is 2.5 m at mid-span and 5.5 m at supports. The clear depth between
forces, the concrete slabs resist a significant portion of the shear force applied to the flanges of the corrugated webs ranges from 1081 mm in mid-span to 4011 mm at
cross—section, and the tangential stresses in the webs are therefore lower than the supports. Plate thickness is 8 mm for typical web units, 10 mm near the piers and
theoretical values. Tests and finite-element analyses also showed that transverse 12 mm near the deviation diaphragms.
bending due to the cross-sectional distortion produces higher axial stresses in the The web plates are welded to 350—mm—wide l4-mm flange plates equipped with
web panels parallel to the longitudinal axis, and higher tangential stresses in the 150 x 15 L-shaped connectors spaced 430 and 370 mm (bf and af, respectively) and
sloped panels. 200 mm wide compared with a transverse fold depth (11,) of 220 mm. Fold wave-
In the SUpport regions of the continuous beam, the bottom slab compressive length is 1600 mm and the girder units are 3280 mm long for an 8'0-mm overlap at
stresses due to the dead load bending moment are significant because of the raised field splices.
position of the gravity axis for short-term composite aCtion, and axial launch pre- Long-term flexural efficiency is 0.625 at mid-span and 0.788 above the intermedi—
stressing further increases the state of stress. Although the central region of the web ate supports. Longitudinal prestressing consists of three families of tendons;
plates, far from the concrete slabs, is unaffected by longitudinal normal stresses, the 0 Straight 12T15 cantilever tendons anchored at the top slab gussets during
flange plates and the adjacent web regions undergo the same longitudinal strains as balanced cantilever construction.
those of the concrete slabs, and local longitudinal stresses arise. These stresses may 0 Straight 12T15 mid-span continuity tendons in the bottom slab.
be easily computed on the basis of the cross—sectional longitudinal strain pattern, o Polygonal 19T15 external tendons for service loads, tensioned on completion of
and their dispersion may be analysed by the finite—element models of the support the superstructure.
regions. These stresses do not contribute to the elastic equilibrium and they can be
reduced by cutting the flange plates at a regular distance. At the field splices, the The following quantities of materials have been reported per square metre of deck
flange plates may be discontinuous, with additional advantages in terms of flexibility surface: 0.57 m of 35-MPa concrete, 26 kg of T15 strand, 124 kg of reinforcing steel
of the steel girders and of welding costs. This solution has been adopted in the field and 39 kg of steelwork (shear connectors included).
splices of the Pont de la Corniche, with 80-m spans. Construction on falsework is a viable solution only for short bridges that do not
The PCS superstructure resists full dead load shear forces until their reduction by allow cost amortization of the incremental launching technology. A three-span PCS
deviated tendons on completion of launch. Consequently, the shear stresses in the bridge has been built on falsework at Cognac, France (Fig. 5.23) on a main span of
corrugated webs are high during launching, and the additional tangential stresses 43 m and two bank spans of 31 m. The top slab is 11.7 m wide, the bottom slab is
generated by torsion and distortion may be high as well. Tests have shown that the 4.17 m wide, and the overall depth of the cross—section is 2.285 m. Thickness of the
cross-sectional distortion may generate shear forces in the webs similar to the shear top slab varies from 0.23 to 0.33 m.
forces resulting from the thin-walled hollow—section torsion theory. The web plates are 8 mm thick, their folds have longitudinal and sloped distance
When passing over the launch bearings, the lower region of the corrugated webs (bf) of 353 mm and are 150 mm deep (hf), and the net depth of the web panels is
is subject (to a complex triaxial state of stress. The lateral guides of launching 1770 mm. Transverse inclination of the steel girders is 400. Flange plates are 10 mm
bearings and a careful design of the bottom cross-sectional node facilitate the thick, and their width is 250 mm fOr the top flange and 330 mm for the bottom
control of transverse flexural stresses, and the local state of stress in the webs flange. Flange plates are equipped with 250-mm-wide 100 x 9 L-shaped connectors,
may often be assumed to be plane. In this case, the ideal stress resulting from the longitudinally spaced 353 mm in mid—span and 159 mm near supports. The
Huber—:Mises- equatiOn. (4.24) may. be- directly compared 7 with the allowable stress. short-term flexural efficiency of the composite section is 0.659. Totally external
-
 312 BRIDGE LAUNCHING PRESTRESSED COMPOSITE BRIDGES 313

.w
"W-h— ... .
..
._,...._...~-«~--—..
.
- ' Educ:
hu ,
Fig. 5.23. Cognac Bridge (reproduced with permission, Campenon Bernard SGE
Fig. 5.24. Internal View of the Cognac Bridge (reproduced with permission, Campenon
Photote‘que)
Bernard SGE Phototéque)

prestressing (eight 19T15 tendons) required the use of concrete deviation diaphragms
(Fig. 5.24), and the resulting tendon is pseudo-funicular of a distributed load higher The top slab is 10.75 m wide with 2.80—m lateral cantilevers; thickness ranges
than the permanent load. The following quantities of materials have been reported from 0.20 m to 0.30 m above flanges. Because of the triangular cross—section,
per square metre of deck surface: 0.41 m of 40-MPa concrete, 15.3 kg of T15 strand, vertical loads generate transverse tensile stresses in the slab portion between flanges.
67.3 kg of reinforcing steel and 31.5 kg of steelwork (connectors included). These stresses have been resisted with transverse prestressing. Rectilinear poly-
So far, few PCS bridges with corrugated—plate webs have been built by ethylene-coated T15 monostrands have been tied below the top reinforcement
incremental launching. Construction of the Charolles Bridge gave researchers the grid, and tensioned alternatively on both slab edges.
opportunity to compare different types of PCS bridges. The tender design proposed The superstructure has been built by field welding 10—13-m girder units, and by
two different solutions (a stiffened-plate PCS built by incremental launching and a casting the corresponding segments of the concrete slab and concrete filling of the
trussed-web PCS built by balanced cantilever construction), but the final choice was bottom tube. Pouring concrete into the steel tube required the use of a vibrating
left to the bidders. feeding pipe. On filling completion, the hole let by the Vibrating pipe was filled with
Among numerous solutions, the PCS proposal with corrugated-plate webs of
Fig. 5.25 was chosen for construction of a seven-span continuous beam with a
53.6-m main span and a total length of 327 m by full—span incremental launching.
With a top slab width of 10.75 m, the overall cross-sectional depth is 3.05 m, and

_..0_

--1I~—
the span-to—depth ratio is 17.6.
A steel tube 610 mm in diameter and 20.6 mm thick replaced the bottom slab. l/f/l/I/I/Il/IIIIIIIIIII III/llI/l/ll

The concrete filling of the tube stiffened against local instability andimproved

0.20 :
lllll Ill
IIIII Ill/Ill I’ll/l
ll Ill/I
I

0.20
dispersion of support reactions during launching. The steel girder is composed of
two inclined webs shop-welded to the bottom tube, so that complete V-segments
were delivered for field assembly. The steel webs have a transverse inclination of
Folded plate I ,x'
450. The webs are 8-mm-thick corrugated plates with longitudinal and sloped panel
width bf of 284 mm, transverse fold depth (hf) of 150 mm, and wavelength of :25 0.61 ' +
1050 mm. Flanges are 300 x lO—mm plates with 100 X 10 L-shaped connectors
,,,,W¢1ded in Corre_8p9n.d6nce,,with th¢,,W¢b,f01dS-, Fig 5.25. Cross section of the Charolles Bridge
314 BRIDGE LAUNCHING PRESTRESSED COMPOSITE BRIDGES 315

high—pressure grouting to avoid tensile shrinkage strains and to create a residual

radial compression in the concrete filling.
The triangular cross-section is very stiff transversely (Fig. 5.26) and does not
suffer distortion during launching, as no vertical misalignment between the support
points is possible. However, the moment of inertia is small, more similar to that of a
noneprestressed composite section than to that of a PC box girder. Large rotations
of the support sections of the continuous beam and the use of just one central
launch bearing required the design of a particularly refined hydraulic—bearing
device. On the other hand, the concrete-filled steel tube provided a prime
contribution to the dispersion of the support reactions.
The launch prestressing was obtained by permanent horizontal external tendons
(12—14 6T13 in relation with the span), placed near the bottom surface of the tOp
slab because of the raised location of the short—term cross-sectional centroid.
Launch tendons were anchored to transverse beams emerging from the bottom
surface of the top slab above the support diaphragms. Spacing of transverse
prestressing strands was reduced near the anchor beams to enhance dispersion of
the temporary stresses transferred to the thin deck slab. Before tensioning of the
balancing tendons of the new segment, unbalanced anchor stresses were so high
that steel props supported on the bottom pipe were necessary to reduce the tor— Fig. 5.27. Lightenedfront portion of the superstructure during launching (reproduced
sional effects in the rear anchor beam. The concrete slab of the front span and the with permission, Campenon Bernard SGE Phototéque)
front—end support diaphragm were cast in situ on completion of launch, so that this
lighter portion of the superstructure could be used as a launching nose (Fig. 5.27).

Fig. 5.26. Pier cap detail of the Charolles Bridge (reproduced with permission, Fig. 5.28. At the end of launch, the end diaphragm is reaching the abutment (repro-
7 Campenon Bernard“ SGE Photo téque) duced with permission, Campenon Bernard SGE Phototeque
316 BRIDGE LAUNCHING PRESTRESSED COMPOSITE BRIDGES 317

On completion of launch, prestressing was completed with four 19Tl5 external that the influence of shear strains on the vertical modes is significant and should be
tendons anchored to the end diaphragms (Fig. 5.28) and deviated by steel dia— taken into account in FE modelling. In contrast, the influence on the torsional
phragms over the piers, and by steel saddles welded to the bottom tube at the vibration modes of a torsional stiflness determined based on the corrugated web
span thirds. Polygonal prestressing balanced more than 80% of the permanent profile rather than on the average web alignment (equivalent flat plate) appeared to
loads, thus enhancing structure durability and recovering most of dead load be negligible. Vibration frequencies of the external tendons are much higher than
deflections. Prestress losses in the central region of the bridge were high owing to the main structure frequencies, and this prevents reasonance eflects. The equivalent-
the 330-m tendon length, and jacking the deviation saddle of the centralsupport damping ratio (1 to 2%) fell between the vibration characteristics of steel bridges
diaphragm upwards recovered most of prestress losses. After almost 40 years, one of and PC bridges.
the most refined launched bridges ever built had recourse to the shifting techniques -
of tensioned tendons adopted in the very first launched bridge on the Rio Caroni.
The use of a composite tube instead of a bottom concrete slab guaranteed a good
distribution of the support reactions and avoided secondary transverse stresses. It
also posed new problems due to the long launching bearings, the transverse stability
of the superstructure during launching, and the need to guarantee a perfect
connection between the steel tube and the concrete filling to control local buckling.
In spite of the high distortional stiffness, the triangular section is expensive and the
construction method appears to reach the limit span for this type of section. The
following quantities of materials have been reported per square meter of deck
surface: 0.25 m of concrete, 1.5 kg of transverse prestressing strand (poly-
ethylene—coated), 5.7 kg of launch prestressing strand, 8.5 kg of polygonal prestres-
sing strand, 58.6 kg of reinforcing steel, and 95.2 kg of steelwork. Steelwork is
composed of 47.8 kg of corrugated—plate webs, 28.3 kg of steel tube, 2.3 kg of
connectors, and 16.8 kg of support diaphragms.
A small PCS bridge has recently been built in Japan. This five—span continuous
beam has a total length of 210 m and includes three 45.5—m internal spans. A
detailed cost comparison confirmed the competitiveness of a PCS bridge with
corrugated-plate webs built by full-span incremental launching over a conventional
non-prestressed composite bridge.
The cross section is composed of a 9.3-m-wide, 0.30—m thick top slab and of
a 3.9um—wide bottom slab whose thickness ranges from 0.25 m in mid-span to
0.50 m in the support regions. The superstructure has a constant depth of 3.0 m
with a span—to-depth ratio of 15.2. Web plate thickness varies from 9 to 12 mm, the
plate folds have longitudinal and sloped breadth bf of 300 mm, and the transverse
fold depth (hf) is 150 mm. External prestressing tendons are as long as two spans to
reduce friction losses at the deviation points.
The bridge is composed of 19 segments with standard length of 11.0 m. Again,
the use of a specific launching nose was avoided by casting the 33-m-long . front
portion of the top slab (3 standard ll-m segments) as well as the front support
diaphragm after completion of the launch. The presence of the bottom slab, how-
ever, increased the negative moment in the front cantilever and required the use of a
temporary cable-stayed front system, which was based on a steel tower placed 2.2 m
behind the front support when the front end was reaching the next pier.
Comparisons between the dynamic bridge behaviour resulting from FE analysis
and the vibration modes measured with dynamictests on the field [145] confirmed”
 References and bibliography

References
[11 BAUR, W., LEONHARDT, F. (1963). Die Agerbrucke. Die Baatechnik, September.
[21 BAUR, W., LEONHARDT, F. and TRAH, W. (1966). Briicke uber den Rio Caroni, Venezuela.
Beton and Staltlbetonbau, February.
[31 FERNANDEZ, A. (1965). Construction d’un pont sur 1e Caroni (Venezuela). Travaax, July—
August.
ROSIGNOLI, M. (1998). Launched Bridges. ASCE Press, Reston, VA.
XANTHAKOS, PP. (1994) Theory and Design of Bridges. JOhn Wiley & Sons, New York.
MULLER, J.M. and PODOLNY, W. JR (1982) Construction and Design of Prestressed Concrete
Segmental Bridges. John Wiley & Sons, New York.
ROSIGNOLI, M. (1993). Ponti stradali in c.a.p. per getto cadenzato con varo di punta. Le
Strade 1296, September.
ROSIGNOLI, M. (1993). Incremental launching method in PC. bridge construction. Italian
Building and Construction 54/55, July—December.
ROSIGNOLI, M. (1997). Incremental bridge launching. Concrete International, February.
ROSIGNOLI, M. (1995). Evoluzione di una tecnica. Le Strade, October.

Nationale des Ponts et Chaussées, Paris.

ROSIGNOLI, M. (2001). Deck segmentation and yard organization for launched bridges.
Concrete International, February.
SEIFRIED, G. and WITTFOHT, H. (1979). Die Brucke fiber den Shatt—al Arab in Basrah
(Iraq). Beton and Stahlbetonbaa, April. '
FRIZZI, D., GIOVANNINI, B. (1977). Ponte sul P0 21 Casale Monferrato per 1’ Autostrada dei
Trafori. L’Indastria Italiana del Cemento, April.
AUI‘HOR UNKNOWN (1977). Extruded span cut bridge costs in. half — Mainbriicke
Mainfiingen. Engineering News Record, 17 November.
PLACIDI, M., VIRLOGEUX, M. (1991). I1 ponte di Trellins: un arco in ca. sul fiume Isere.
L’Indastria Italiana del Cemento, October.
[17] FLUGEL, K. (1987). Der bau der Maintalbrucke Veitshochheim, ein neuer Weltrekord im
taktschieben. Tiefbaa Berufsgenossenschaft, August.
[18] LEONHARDT, F., ANDRA & PARTNER (1991). Ponte ferroviario sulla valle del Meno a
Veitshochheim. L’Industria Italiana del Cemento, February.
'[19] THEINER, J. (1987). Weltrekord im Taktschiebeverfahren. Maintalbriicke Veitshoechheim
fertiggestellt. Beton, January.
320 BRIDGE LAUNCHING REFERENCES AND BIBLIOGRAPHY 321

[20] AUTHOR UNKNOWN (1986). Noch nie zuvoe so weit im Takt geschoben. Strabag bau AG [44] DEZI, L., MENDITTO, G. and ROSIGNOLI, M. (1982). In tema di predimensionamento dei
techn. federfiihrung und WTB erzielen Weltrekord beim bau der Maintalbrficke ponti a cassone in calcestruzzo armato ed in calcestruzzo armato precompresso.
Veitshéchheim. Bauwirtschaft 49/50. L’Industria Italiana del Cemento, June.
[211* ZILCH, K. (1987). Maintalbrficke Veitshéchheim — Maschinen und Rusttechnik fur eine ROSIGNOLI, M. (1997). Influences of the incremental launching construction method on the
ungewohnliche Bauaufgabe. BMT Baumaschine und Bautechnik, January. sizing of prestressed concrete bridge decks. Proceedings of the Institution of Civil Engineers
[22] MATILDI, P. and MATILDI, G. (1990) Ponti Metallici, Esperienze Vissute. Cimolai SpA, — Structures and Buildings, August.
Pordenone. ROSIGNOLI, M. (1999). Presizing of prestressed concrete launched bridges. American
[23] GREISCH, F. (1993). Il ponte strallato di Wandre sulla Mosa in Belgio. L’Industria Italiana Concrete Institute — Structural Journal, SeptemberMOctober.
del Cemento, May. ROSIGNOLI, M. (1999). Nose optimisation in launched bridges. Proceedings of the
[24] MARTINEZ, Y., CABRERA, F. and ROSIGNOLI, M. (2001). Il nodo di Via Palizzi a Milano. Institution of Civil Engineers — Structures and Buildings, November.
L’Industria Italiana del Cemento, April. ROSIGNOLI, M. (1995). L’interazione avambecco — impalcato nei ponti in c.a.p. realizzati
[25] COLOMBO, A., GENTILE, R., MARTINEZ y CABRERA, F ., ROSIGNOLI, M. and Rossr, S. (1997). per varo frontale. L’Industria Italiana del Cemento, March.
Il ponte strallato di Via Palizzi a Milano. Giornate AITEC 1997. ROSIGNOLI, M. (1998). Nose-deck interaction in launched prestressed concrete bridges. _
[26] ROSIGNOLI, M. (1998). I tre ponti in ca. precompresso del nodo di Via Palizzi. Le Strade, American Society of Civil Engineers — Journal of Bridge Engineering, February.
June. , KONCZ, T. (1968). Vorschub-Freivorbau bei einer hohen Talbriicke, dem Semorile-
[27] ROSIGNOLI, M. (1998). Site restrictions challenge bridge design. Concrete International, Viadukt. Beton und Stahlbetonbau, March.
August. ARCANGELI, A. (1990). Viadotto sulla vallata dell’ Our 3. Steinebruck, Belgio. L’Industria
[28] CALVI, G.M., PRIESTLEY, M.J.N. and SEIBLE, F. (1996). Seismic Design and Retrofit of Italiana del Cemento, May.
Bridges. John Wiley & Sons, New York. AUTHOR UNKNOWN (2001). Wide load. Bridge Design and Engineering, 25.
[29] AASHTO (2000). Guide Specifications for Seismic Isolation Design. American Association of BELLUZZI, O. (1988) Scienza delle Costruzioni. Zanichelli Editore, Bologna.
State Highway and Transportation Officials. PETRANGELI, MP. (1993) Progettazione e Costruzione di Ponti. Masson, Editoriale ESA,
[30] KELLY, J .M. and NAEIM, F. (1999). Design of Seismic Isolated Structures. John Wiley & Milan. ‘
Sons, New York. BRITISH STANDARDS INSTITUTION (1979). Steel, Concrete and Composite Bridges — Part 5.
[31] ABEYSINGHE, R., GAVAISE, E., ROSIGNOLI, M. and TZAVEAS, T. (2002). Non-linear pushover Code of Practice for Design of Composite Bridges. BSI, Milton Keynes, BS 5400: part 5.
analysis of Greveniotikos Bridge. American Society of Civil Engineers — Journal of Bridge LACROIX, R. (1967). La methode des matrices—transfert. Annales ITBTP, March—April.
Engineering, March. ROSIGNOLI, M. (1997). Solution of the continuous beam in launched bridges. Proceedings
[32] ATC—40 (1996). Seismic Evaluation and Retrofit of Concrete Buildings. Applied Technology of the Institution of Civil Engineers — Structures and Buildings, November.
Council, RedWood City, CA. 7 ROSIGNOLI, M. (1999). La risoluzione della trave continua nei pOnti realizzati per varo
[33] GIOVANNINI, B. (1972). Un impalcato estruso: il Viadotto in Val Restel. L’Industria delle frontale. L’Industria Italiana del Cemento, March.
Costruzioni, September—October. ROSIGNOLI, M. (1999). Reduced transfer matrix method for analysis Of launched bridges.
[34] BADOUX, M., BURDET, 0., FAVRE, R. and LAURENCET, P. (1999). Design of a curved American Concrete Institute — Structural Journal, July—August.
incrementally launched bridge. IABSE Structural Engineering International, May. CESTELLI—GUIDI, C. (1987) Cemento Armato Precompresso. Hoepli editore, Milan.
[35] ROSIGNOLI, M. (2000). Thrust and guide devices for launched bridges. American Society of DIN (1985). Road and Foot Bridges —D’eSign’ Leads. DIN 1072.
Civil Engineers — Journal of Bridge Engineering, February. BRITISH STANDARDS INSTITUTION (1978). STEEL, Concrete and Composite Bridges w Part 2.
[36] GILLET, G. and JACQUET, P. (1988). L’ouvrage 33 sur l’autoroute du Littoral a Marseille. Specifications for Loads. BSI, Milton Keynes, BS 5400: part 2.
Annales ITBTP, October. - SALES, J .J . de, MARTINS GONCALVES, R., MALITE, M. and TAKEYA, T. (2000). Monitoring of
[37]. ROSIGNOLI, M. (1996). I1 ponte spinto contro la piena. Le Strade, June the Parana River Bridge during launching. IABSE Structural Engineering International,
[38] ROSIGNOLI, M. (1998). Serio River Bridge, creep and incremental launching. Proceedings of March.
the Institution of Civil Engineers _ Structures and Buildings, February. CEB-FIP (1993). Model Code 1990. Thomas Telford, London.
[39] CHEMERINSKI, 0.1., SELIVERSTOV, V.A. and ZI—IURAvov, L.N. (1996). Launching steel NAM SHIU, K. and RUSSELL, H.G. (1983). Knowledge gained from instrumentation of the
bridges in Russia. IABSE Structural Engineering International, August. Kishwaukee river bridge. PCI Journal, September—October.
[40] EUROCODE 8 (1994) Design Provisions for Earthquake Resistance of Structures. Part 2: CEB-FIP (1994) Application of High Performance Concrete. Bulletin d’Information "222.
Bridges. ENV 1998-2. ' Federation Internationale du Beton, Lausanne.
[41] BRITISH STANDARDS INSTITUTION (1983). Steel, Concrete and Composite Bridges — Part 9. CEZARD, C. and SERvANT, CL (1988). ChariX Viaduct. Contribution of the French'Group.
Bridge Bearings. Section 9.1: Code of Practice for Design of Bridge Bearings. BSI, Milton AIPC—IABSE Congress, Helsinki. AFPC Publications, Zurich.
Keynes, BS 5400: part 9. BATTERMANN, W. (1987). Elastomer-Verformungsgleitlager mit lastabhaengiger
[42] AASHTO (1998). Guide Specifications for Design and Construction of Segmental Concrete Volumenver- draengungS—Schmierung. Die Bautechnik, January.
Bridges. American Association of State Highway and Transportation Officials. ROSIGNOLI, M. (1981). Ponti a cassone: la distorsione della sezione trasversale secondo la
[43] ROSIGNOLI, M. (1996). Sul dimensionamento degli impalcati da ponte in cap. realizzati analogia della trave su mezzo elastico. University of Ancona.
per varo frontale progressivo. L’Industria Italiana del Cemento, December. CAMPBELL, D. and WEDGWOOD, J .L. (1971). Need for diaphragms in concrete box girders.
American Society of Civil Engineers — Journal of the Structural Division, March (discus—
si (ms in December. 19.7.1...
322 BRIDGE LAUNCHING REFERENCES AND BIBLIOGRAPHY 323

[71] ROSIGNOLI, M. (1999). Prestressing schemes for incrementally launched bridges. American 196] BRITISH STANDARDS INSTITUTION ( 1990). Steel, Concrete and Composite Bridges — Part 4.
Society of Civil Engineers -~ Journal of Bridge Engineering, May. Code of Practice for Design of Concrete Bridges. BSI, Milton Keynes, BS 5400: part 4.
[72] ROSIGNOLI, M. (1998). Misplacement Of launching bearings in PC. launched bridges. [97] KURITA, A., MATSUI, S., NAKAI, H. and YODA, T. (1998). Trends in steel-concrete compo-
American Society of Civil Engineers — Journal of Bridge Engineering, November. Site bridges in Japan. IABSE Structural Engineering International, February.
[73] ROSIGNOLI, M. (1997). Tolleranze di pOSizionamentO degli appoggi di varo nei ponti spinti [98] MATSUI, S., MORI, K., OHTA, K. and TAKAGI, M. (1997). Design method and fatigue
in c.a.p. L’Industria Italiana del Cemento , September. strength Of large Span concrete-filled I—beam grid deck. IABSE PIP-CEB International
[74] ABDEL SAMAD, S.R., ROBINSON, AR. and WRIGHT, RN. (1968). BEF analogy for analysis Conference, Innsbruck, Austria.
of box girders. American Society of Civil Engineers — Journal of the Structural Division, [99] ROSIGNOLI, M. (1999). Estrusione delle SOlette nei ponti in acciaiO e c.a. Le Strade, May.
July. [100] BERTHELLEMY, J ., LACOMBE, J .M. and POINEAU, D. (1997). Cracking control in the concrete
[751 HETENYI, H. (1971). Beams on Elastic Foundation, Theory with Application in the Fields of slab Of the Nevers composite bridge. IABSE FIP—CEB International Conference,
Civil and Mechanical Engineering. University Of Michigan, Ann Arbor. Innsbruck, Austria.
[76] GODART, B., LENOIR, B., LEVEILLE, A., NERON, P. and TONNOIR, B. (1989). Viaduc d’ [101] CREMER, J .M., DOTR'EPPE, J.C., ERNENS, M., LOTHAIRE, A. (1997). Cracking and durability
Amiens. Conception et realisation de l ouvrage. Suivi experimental sur site, pendant la of concrete slabs Of composite bridges. IABSE FIP—CEB International Conference,
construction. Annales ITBTP, March—April. Innsbruck, Austria.
[771 COLLEPARDI, M. (1980) Scienza e Tecnologia del Calcestruzzo. Hoepli Editore, Milan. [102] NAVARRO, M.G. and LEBET, J .P. (2001). Concrete cracking in composite bridges: tests,
[78] ROSIGNOLI, M. (2000). Creep effects during launch Of the Serio River Bridge. Concrete models and design proposals. IABSE Structural Engineering International, August.
International, March. [103] DUCRET, J.M. and LEBET, J .P. (1997). Effects Of concrete hydration on composite bridges.
[79] LEONHARDT, EC. (1979) CA. e C.A.P. Calcolo di Progetto e Tecniche Costruttive — Vol. 4 — IABSE FIP—CEB International Conference, Innsbruck, Austria.
I Ponti, Dimensionamento, Tipologia, Costruzione. Edizioni di Scienza .e Tecnica, Milan. [104] DUCRET, J .M., LEBET, J .P. (1999). Behaviour of composite bridges during construction.
[80] BREEN, J. and NAAMAN, A. (1990) External Prestressing in Bridges. Special Publication IABSE Structural Engineering International, August.
120, American Concrete Insitutute, Detroit, Mi. ‘ [105] BERTHELLEMY, J ., LACOMBE, J .M. and POINEAU, D. (1997). The Nevers Bridge: design Of the
[81] CAPITANIO, S. (1985). The King Taskin Bridge in Bangkok. L’Industria Italiana del steel concrete composite box girder. IABSE FIP—CEB International Conference, Innsbruck,
Cemento, September. Austria.
[82] COMBAULT, J ., LEVEILLE, A., NERON, P. and THIBONNET, J .L. (1987). Incrementally [106] BERTHELLEMY, J ., PLACIDI, M. and VIRLOGEUX, M. (1997). Prefabrication and prestressing
launched bridges with total external prestressing. Contribution Of the French Group. Of concrete Slabs in composite bridges. IABSE FIP—CEB International Conference,
AIPC—IABSE Congress, Paris— Versailles. Innsbruck, Austria.
COMBAULT, J., LEVEILLE, A., NERON, P. and THIBONNET, J .L. (1987). Fonts poussés a [107] MONOV, B. and SELIVERSTOV, Y. (1997). Erection of composite bridges with precast deck
précontrainte totalernent extérieure. AIPC—IABSE Congress, Paris— Versailles. Slabs. IABSE FIP—CEB International Conference, Innsbruck, Austria.
SAUL, R. (1997). Erection methods for long-Span steel composite bridges. IABSE FIP- [108] KRAUS, D. and WURZER, O. (1997). Bearing capacity Of concrete dowels. IABSE PIP-CEB
CEB International Conference, Innsbruck, Austria. » 7 International Conference, Innsbruck, Austria.
EUROCODE 4 (1997). Design of Composite Steel and Concrete Structures, CEN, Bruxelles. [109] AKESSON, B., EDLUND, B. and SHEN, D. (1997). Fatigue cracking in a steel railway bridge.
C81-63 (1981). Relative au Reglement de Calcul de Pont Mixtes Acier-Be'ton. Circular 81- IABSE Structural Engineering International, May. .
63. [110] ROSIGNOLI, M. (1998). 11 mm detie" SOiette ' nei ponti compositi. L’Industria delle
RPX-95 (1996). Recomendaciones Para el Proyecto de Puentes Mixtas Para Carreteras. Costruzioni, November.
RPX-95, Madrid. [111] ROSIGNOLI, M. (1996). Una specie ibrida. Le Strade, November—December.
MELO, L.G., REIS, A.J. (1997). Composite plate girder bridges: safety and serviceability. [112] COLLEPARDI, M., COPPOLA, L. and TROLI, R. (1995). Materiali cementizi innovativi: dagli
IABSE FIP—CEB International Conference, Innsbruck, Austria. HPC verso gli RPC. L’Industria Italiana del Cemento, March.
AASHTO (1993). Guide Specifications for Horizontally Curved Highway Bridges. American [113] AC1 (1992). Structural LightWeight Aggregates Concrete Performance. Special Publication
Association Of State Highway and Transportation Oflicials. 136, American Concrete Institute, Detroit, Mi.
EUROCODE 3 (1993). Design of Steel Structures — Part I .1 .° General Rules and Rules for [114] COLLEPARDI, M., COPPOLA, L. and MONOSI, S. (1995). Il calcestruzzo leggero. Il Nuovo
Buildings. ENV 1.1. Cantiere, October.
SAUL, R. (1996). Bridges with double composite action. IABSE Structural Engineering [115] DIN (1986). Prestressed Concrete ~— Prestressed Lightweight Concrete Structural
International, February. Components. DIN 4227 Part 4.
' BRITISH STANDARDS INSTITUTION (1982). Steel, Concrete and Composite Bridges — Part 3. [116] VARIOUS AUTHORS (1987). Innovation dans le domaine des structures mixtes: ponts mixtes
Code of Practice for Design of Steel Bridges. BSI, Milton Keynes, BS 5400: part 3. metal—beton precontraint. Annales ITBTP, October—November.
HOECKMAN, W. (2001). Bridge over the River Loire in Orleans, France. IABSE Structural [117] RICHARD, P. (1984). Structures triangulées en béton. Annales ITBTP, January.
Engineering International, May. [118] BRUNEAU, J ., CAUSSE, G., RADIGUET, B., RASPAUD, B. (1984). Test loading of a concrete
VARIOUS AUTHORS. (1984). Innovation dans le domaine des ouvrages d’art en béton pré— . truss. Contribution Of the French group. AIPC—IABSE Congress, Vancouver.
contraint. Annales ITBTP, January. [119] BRUNEAU, J., CAUSSE, G., RADIGUET, B. and RASPAUD, B. (1984). Experimentation d’ une
EUROCODE 2 (1992). Eurocode 2 — Design of Concrete Structures — Part I .1 .' General Rules structure spatiale en trellins. AIPC—IABSE Congress, Vancouver.
and Rules for Buildings. ENV 1.1. [120] MIRELES, A.A., PODONLY, W. (1983). Kuwait’s Bubiyan bridge. A 3D precast segmental
spaceframq PCI Journal,_Jan11.ary::Fcbruary., _
324 BRIDGE LAUNCHING REFERENCES AND BIBLIOGRAPHY 325

[121] ROSIGNOLI, M. (2000). Ponti a cassone in c.a.p. ad anime reticolari. L’Industria Italiana [147] DUVIARD, M. ( 1987). Le premier pont a ames plissees a Cognac. AIPChIABSE Congress,
del Cemento, May. , Paris— Versailles.
[122] RoeNOLI, M. (2001). Trusses instead of solid webs? Concrete International, May. [148] CAFOLLA, J. and JOHNSON, RP. (1997). Fabrication of steel bridge girders with. corrugated
[123] COCOMELLO, S. (1984). A spatial lattice in prestressed concrete for the deck Of the webs. The Structural Engineer, April.
Bubiyan bridge in Kuwait. L’Industria Italiana del Cemento, August.
[124] MARTIN, R.A., RASPAUD, B. and RICHARD, P. (1982). Le pont de Bubiyan au Koweit.
Travaux, November. Bibliography
[125] AUTHOR UNKNOWN (1990). I viadotti di Sylans e Glacieres con struttura a traliccio spaziale AC1 (1974) ACI Fall Convention: Symposium on Segmental Bridge Design and Construction,
in c.a.p. realizzati per 1’ Autostrada A40 nel massiccio del Jura. L’Industria Italiana del Atlanta, GA.
Cemento, June. ACI (1977) Concrete Box Girder Bridges. Monograph 10, American Concrete Institute, Detroit,
[126] BOUDOT, J., RADIGUET, B. and THAO, P.X. (1987). The Sylans and Glacieres Viaducts. Mi.
Contribution of the French Group. AIPC—IABSE Congress, Paris- Versailles. AMAT, M., ANDRE, M. and CARSIGNOL, M. (1983). Le pont de la RN. 504 a Rossillon. Travaux
[127] BUODOT, J., RADIGUET, B. and THAO, P.X. (1987). Les Viaducs de Sylans et des Glacieres. 504, July—August.
AIPC—IABSE Congress, Paris— Versailles. ARCANGELI, A. (1990). Il viadotto sopraelevato Hamburg ad Amburgo. L’Industria Italiana del
[128] ROSIGNOLI, M. (1997). Prestressed composite box girders for highway bridges. IABSE Cemento, January.
Structural Engineering International, November. ASCHRAFI, M., LUSTGERTEN, P., RINNE, KB. and SAUL, R. (1992). Verbundbriicke mit rekord-
[129] ARIBERT, J .M. (1988). Connexion au cisaillement des poutres fatigue — determination spannweite uber den Rio Caroni/Venezuela. Der Stahlbau, January.
experimentale de la resistance. Annales ITBTP, September. ASWAD, ‘A. and ASWAD, G.G. (1991). Rational prediction Of bridge girder reinforcing and
[130] JOHNSON, RP. (1997). Shear connection for composite bridges and Eurocode 4: Part 2. strength. PCI Journal, May—June.
IABSE FIP-CEB International Conference, Innsbruck, Austria. ATTAL, A., CALGARO, J ., CAUSSE, G. et al. (1984). New composites structures. Concrete and steel.
[131] KRISTEK, V., VITEK, IL. (1997). Performance and stud failure in steel-concrete composite AIPC—JABSE Congress, 1984, Vancouver.
beams. IABSE FIP-CEB International Conference, Innsbruck, Austria. AUTHOR UNKNOWN (1977). Box girder Will be launched in the US. Engineering News Record, 27
[132] CAFOLLA, J. and JOHNSON, RP. (1997). Corrugated webs in plate girders for bridges. January.
Proceedings of the Institution of Civil Engineers — Structures and Buildings, May. AUTHOR UNKNOWN (1978). Arch slipformer shuns ground support to cross valley. Engineering
[133] MOREAU, P. and THIVANS, P. (1983). Composite structures steel—prestressed concrete. News Record, 1 June.
IABSE Proceedings, February. ' AUTHOR UNKNOWN (1978). Segmental box girder bridges make the big time in US. Engineering
[134] ROSIGNOLI, M. (1999). Prestressed concrete box girders with folded steel plate webs. News Record, 2 March. ' r
Proceedings of the Institution of Civil Engineers Structures and Buildings, February. AUTHOR UNKNOWN (1986). Prestressed concrete in Switzerland, 1982—1986. FIP Congress, New
[135]
-
CAFOLLA, J. and JOHNSON, RP. (1997). Local flange buckling in plate girders with corru- Delhi. ’
gated webs. Proceedings of the Institution of Civil Engineers — Structures and Buildings, AUTHOR UNKNOWN (1986). Remarkable structures 1983—1986 — Spannbetonbau in der
May. Bundesrepublik Deutschland 1983—1986. FIP Congress, New Delhi.
[136] COMBAULT, J. (1988). The Maupre Viaduct near Charolles. AISC NEC/COP National AUTHOR UNKNOWN (1986). Sovrappasso ferroviario in cemento armato lungo la linea Wuppertal —
steel construction conference, 1988, Miami Beach, Florida. Colonia nella Germania Federale. L’Industria'ltaliana del Cemento, May.
[137] CAUSSE, G. and DUVIARD, M. (1987). Ouvrage a ames plissees mis en place par poussage, AUTHOR UNKNOWN (1986). Three-dimensionally curved prestressed concrete bridge built by the
Charolles. AIPC—IABSE Congress, Paris— Versailles. incremental launching method for railway trafl-ic .. Spannbetonbau in der Bundesrepublik
[138] REINHARD, J.M. (1994). Pont de la Corniche a Dole. Ouvrages d’Art 19, November. Deutschland 1983—1986. FIP Congress, New Delhi.
[139] MOREAU, P. and THIVANS, P. (1984). Structures composites acier-béton precontraint. AUTHORS UNKNOWN (1986). Monographies d’ouvrages. Travaux, January.
Annales ITBTP, January. AUTHOR UNKNOWN (1988). Bau einer raumlich gekrummten spannbetonbrficke fur den
[140] BERGFELT, A., LEIVA, L. (1984). Shear Buckling of Trapezoidally Corrugated Girder Webs. Eisenbahnverkehr im taktschiebeverfahren. Tiefbau — Berufsgenossenschaft, May.
Chalmers University of Technology, Sweden. AUTHOR UNKNOWN (1988). Evolution recente des ponts en beton precontraint. IABSE
[141] BONNET, M., BOUVY, B., CAUSSE, G. et al. (1988). Viaduc du vallon de Maupre, a Proceedings, November.
Charolles (Saone-et-Loire). Travaux, October. AUTHOR UNKNOWN (1988). Le passage superieur n. 8 sur 1’ autoroute A 71. Travaux, October.
[142] CAUSSE, G. and DUVIARD, M. (1987). Maupre Viaduct near Charolles. A bridge deck with AUTHOR UNKNOWN (1989). Autobahnbriicke, Frankreich. Der Bauingenieur, November.
corrugated steel plate webs constructed by the incremental launching technique. AIPC— AUTHOR UNKNOWN (1989). Sviluppi recenti nel campo dei ponti in c.a.p. L’Industria Italiana del
IABSE Congress, Paris— Versailles. Cemento, June
[143] COMBAULT, J. (1988). Viaduct du vallon de Maupre, a Charolles. Travaux, January. AUTHOR UNKNOWN (1989). Un sovrappasso autostradale di concezione innovativa. L’Industria
[144] NADAL, P. (1987). Le Viaduct de Maupre a Charolles. Chantiers de France, November. Italiana del Cemento, October.
[145] ISIGURO, W., MURATA, Y., SUGO, T. and UEHIRA, K. (1997). Stability of prestressed AUTHOR UNKNOWN (1990). Eine briicke uber die Donau. Der Baukasten, September.
concrete bridge with corrugated steel web. AIPC—IABSE Congress, Innsbruck. AUTHOR UNKNOWN (1990). Entwicklungen im Spannbetonbau. Ruckschau und ausblik. Betonwerk
[146] CASSIAU, A., COMBAULT, J., DUVIARD, M. et al. (1988). Pont sur la Charente a Cognac. und Fertigteil-technik, April.
Travaux, October. AUTHOR UNKNOWN (1990). Il ponte sull’ estuario di Mandurah. Primo ponte australiano realizzato .
___.--.c.on.il.._p.1:o.cedimento.-a...spi.nta...LfIndustria Italiana. del Cemento, February.
 326 BRIDGE LAUNCHING REFERENCES AND BIBLIOGRAPHY 327

AUTHOR UNKNOWN (1990). Il viadotto Lirone per il tratto SS. 229 — Gattico dell’ Autostrada dei COLAs, M., FRESNAIS, P., HIRSCH, D. et al. (1983). Le nouveau pont sur le Loir a La Fleche.
Trafori a Borgomanero, Novara. L’Industria Italiana del Cemento, April. Travaux, July—August.
AUTHOR UNKNOWN (1990). Ponti a conci prefabbricati in ca. Ultime realizzazioni e prospettive in COMBAULT, J. (1986). Evolution et developpement des ponts modernes a precontrainte totalement
Florida. L’Industria Italiana del. Cemento, May. exterieure au beton. Travaux, January.
AUTHOR UNKNOWN (1992). Baustelle eines Weltrekords. Die pylone der Pont de Normandie. COUCHMAN, G. and LEBET, J .P. (1996). A new design method for continuous composite beams.
Bauzeitung, November. IABSE Structural Engineering International, February.
AZPARREN, C., HINOJOSA, IA. and RUI—WAMBA, J. (1997). The Spanish recommendations for the CREMER, J.M. (1997). Variety of composite bridge construction. IABSE FIP-CEB International
design of composite road bridges. IABSE FIP-CEB International Conference, Innsbruck, Conference, Innsbruck, Austria.
Austria. CURBACH, M. and SEITZ, G. (1983) Unusual applications of the incremental launching method.
BARFOOT, J. (1986). Incremental launching at Storeham bridge. Concrete, November. IABSE Proceedings, Zurich.
BARTELS, H.J., BERz, P., JASCHKE, W., SCHWARZ, J. and MEYER, H. (1987). Die neue Weserbrficke DEINHARD, J ., SCHAFER, H. and SCHLOSSER, H. (1987/ 1988) Hochstrasse Ratsweg in Frankfurt
Bodenwerder. Der Bauingenieur, February. a.M. im zuge der BAB 661. Beton und Stahlbetonbau, November 1987; January and February
BARTOLI, C. and ROSIGNOLI, M. (2002). Il ponte sul fiurne Serio a Bergamo. L’Industria Italiana 1988.
del Cemento (in press). DIMITROV, D.G. (1986). Spannungsverteilung querbelasteter Stegbleche mit abgestuften Dicken
BAUM, P., CHASSAGNETTE, P., LAUNAY, P. and STOUF, H. (1970). Une nouvelle méthode de con- unter einseitiger begrenzter gleichmassiger last. Der Stahlbau, August.
struction des ouvrages de franchissement en béton précontraint: le poussage. Annales ITBTP, DIMITROV, D.G. (1986). Zum taktschiebeverfahren bei Verbundbriicken. Der Stahlbau, October.
May. DORNER, M. and FRANZI, J. (1986). Der bau des Taluberganges Enzenreith. Osterreichische
BAUMANN, H. and SCHWICHOW, V. (1982). Vom Bau der Steinaggertalbriicke. FIP Congress, Ingenieur und Architekten Zeitscltrift, March.
Stockholm. DORTON, RA. (1986). Safety considerations for the Burlington Skyway project. AIPCWIABSE
BAUR, W. (1977). Bridge erection by launching is fast, safe and eflicient. American Society of Civil Congress, Tokyo.
Engineers — Civil Engineering, March. . DUREIL, P. and PLACIDI, M. (1982). Le pont sur le Drac a Grenoble. Annales des Ponts et
BAUR, W. and LEONHARDT, F. (1962). Neue Verfahren zur Herstellung weitgesspannter, mehrfel— Chaussees, second quarter.
driger Balkenbriicken aus Spannbeton. Beton und Stahlbetonbau, July. DUTOIT, D. and MULLER, J. (1984). Projects Marta CS 360 and CN 480. Atlanta mass transit:
BAUR, W. and LEONHARDT, F. (1971). Erfahrungen mit dem TaktschiebeVerfahren im briicken Marta CS and CN 480. Railways and mass transit systems bridges. French experience. AIPC—
und hochbau. Beton und Stahlbetonbau, July. IABSE Congress, Vancouver.
BAUR, W. and LEONHARDT, F. (1978). Taktsclziebeverfahren. ATC, Stuttgart. ‘ FALEMPIN, M., GLASSON, J .P., LACROIX, A. et al. (1988). Le pont sur la voie 3 de Clergy Pontoise
BEGUSCH, W. (1987). Grossbrucken der Wechselautobahn. Osterreichische Ingenieur und (Val d’ Oise). Travaux, October.
Architekten Zeitsclzrift, May. , FERRETTI TORRICELLI, L. and FURLANETTO, G. (1997). Il viadotto Lijak presso Nova Gorica nella
BEGUSCH, W. (1995). Ponte sulla vallata di Edlitz per 1’ Autostrada del Sud A2 in Austria. Repubblica Slovena. L’Industria Italiana del Cemento, June
L’Industria Italiana del Cemento, June FREYSTEIN, W. and WALLAs, H. (1987). Planung der Eisenbahnbrficke uber die Werra mit uber-
BELLERO, C., COLAZINGARI, C. and Gozzr, V. (1995). Ponte Posadas-Encarnacion sul fiurne bau in der verbundbauart. Beton und Stahlbetonbau, April.
Parana tra Argentina e Paraguay. L’Industria Italiana del Cemento, February. FRITSCH, P. (1988). Schragkabel und taktschiebebriicken in Korea und Taiwan. Osterreichische
BERNARD, C. (1986). Due viadotti per 1’ autostrada Esterel — Costa Azzurra. L’Industria Italiana Ingenieur und Architekten Zeitschrzft, Septemberw'éOc'IOber.
del Cemento, April. . FRITSCH, R. and SCHEIDLER, J. (1980). Donaubrficke Woerth — Taktschiebeverfahren mit
BERNARD, J. (1997). The Rocher de l’Escalade Viaduct, France. IABSE Structural Engineering Querverschub. Der Bauingenieur, May.
International, February. FROEHLICH, H. (1984). Neubau der Donnersberger Briicke in Munchen unter Anwendung weitern
BEST, K.H., KINGSTON, RH. and WHATLEY, M.J. (1978). Incremental launching at Shepherds entwickelter Lager bei der Herstellung in Taktschiebeverfahren. Beton und Stahlbetonbau,
House Bridge. Proceedings of the Institution of Civil Engineers, February. January.
BIBERSCHICK, P., HUFNAGEL, W., WENZEL, H. (1997). Composite bridges in Austria. IABSE FIP— GISI, R. (1986). Die Birsbrucke Rutihard bei Miinchenstein. Prestressed concrete in Switzerland,
CEB International Conference, Innsbruck, Austria. ' 1982—1986. FIP Congress, New Delhi.
BRITISH STANDARD (1978). Steel, Concrete and Composite Bridges — Part 7. Specification for GIULIANI, G. (1970). Il nuovo ponte per 1’ oleodotto dell’ Europa Centrale, realizzato sul fiume Po
Materials and Workmanship, Concrete, Reinforcement and PreStressing Tendons. BSI, Milton con il sistema dell’ autovaro. L’Industria Italiana del Cemento, March.
Keynes, BS 5400: part 7 - GOBEL, W., MAUCH, A., MULLER, B. and SCHULDT, O. (1989). Brucken Hetzdorf * Die bautech-
BRITISH STANDARD"'(1980).' Steel, Concrete and Composite. Bridges — Part 6. Specification for nologie fur das Taktschiebeverfahren und ihr einfluss auf die bautechnische losung. Strasse,
Materials and Workmansbip, Steel. BSI, Milton Keynes, BS 5400: part 6. February.
CAPRA, R. and MARTINEZ y CABRERA, F. (1988). Tecnica della prefabbricazione per la realizza- GRANT, A. (1975). Incremental launching of concrete structures. ACI Journal, August.
zione di sezioni strutturali a cassone. L’edilizia e L’industrializzazione, January—February. GREISCH, F. (1991). Il ponte di Ben-Ahin sulla Mosa in Belgio. L’Industria Italiana del Cemento,
CECCHERELLI, D. (1996). Il ponte sul fiume Brembo della SS 342 Briantea. Le Strade, January— April.
February. GRENNERAT, Y. (1988). Le viaduct de Poncin sur l’autoroute A40 (Ain). Travaux, October.
CHEYREZY, M. and COMBAULT, J. ( 1996). Composite bridges With corrugated steel webs — achieve— HANEL, D. and KUSCHE, R. (1993). Sonderlosung zur Aufnahme grosser Horizontalkrafte bei
ments and prospects. IABSE Congress, Brussels. Eisenbahnbriicken — Beispiel Hetzdorfer Viadukt. Die Bautechnik, January.
328 BRIDGE LAUNCHING REFERENCES AND BIBLIOGRAPHY 329

HANSVOLD, C. and NILSSON, H. (1999). Uddevalla Bridge, Sweden. IABSE Structural Engineering MARTINEZ Y CABRERA, F. and MENNI, C. (1974). I ponti a cassone monocellulare a profile defor-
International, May. mabile. Technical Note 33, ISTC, May, Politecnico di Milano. Tamburini Editore.
HARRIES, H., KINKEL, H., KLEINMANNS, R. and PETRI, H. (1986). Zwei trogbriiken uber die Sieg MATHIVAT, J. (1987). L’ evolution recente des ponts en beton precontraint. Ingenieurs et
bei Hennef. Beton und Stahlbetonbau, September. Architectes Suisses, March.
HARRIES, H., KINKEL, H. and PETRI, H. (1986). Sliding formwork launching gantries. Transferring MATHIVAT, J. and VIRLOGEUX, M. (1990). Le pont en arc sur la Rance — la technique francaise du
the principle of incremental launching to formwork launching gantries in bridge construction. beton precontraint. FIP Congress, Hamburg.
Spannbetonbau in der Bundesrepublik Deutschland 1983—1986. FIP Congress, New Delhi. MOREAU, P. and THIVANS, P. (1984). Structures composites acier-beton precontraint. Annales
HEYWOOD, P. (1987). Corrugated box—girder web lowers bridge weight, cost. Engineering News ITBTP, January.
Record, 17 December. MOSCHNER, H. (1986). Der bau der Landlbriicke im Zuge der B 25 Erlauftal. Bundesstrasse
HURD, M.K. (1986). Segmental box-girder bridge construction. Concrete International, Zement und Beton, February.
September. MUNZ, HT. (1987). Zwei briicken ans einem guss vermessungsarbeiten beim bau von gekoppelten
HURD, M.K. and WATSON, SC. (1990) Esthetics in Concrete Bridge Design. Monograph taktschiebebriicken fur die B 224 n bei Wulfrath. Der Vermessungsingenieur, January.
Publication 1, American Concrete Institute, Detroit, Mi. NADAL, P. (1988). L’ Echangeur Al/Parc Astérix. Chantiers de France, September.
KLEIN, W. (1986). Montage der Sauertalbriicke im taktschiebeverfahren. Der Stahlbau, O’CONNOR, C. (1971) Design of Bridge Superstructures. John Wiley & Sons, New York.
December. PEREVOZNIKOF, B. and SELIVERSTOV, VA. (1999). Reconstruction of the Mosow Ring Road,
KORN, H. (1975). Bridge construction by extrusion sliding. Concrete, May. Russia. IABSE Structural Engineering International, May.
KRAUTWALD, W., SCHUTT, K. and THORMAHLEN, U. (1990). Berbke Valley Bridge, incremental PETERS, H. (1993). Vom Bau der Pont de Normandie bei Le Havre. Beton und Stahlbetonbau,
launching and external prestressing. Spannbetonbau in der Bundesrepublik Deutschland January.
(1987). FIP Congress, Hamburg. PLACIDI, M. (1982). La construction d’ouvrages d’art avec mise en place par rotation. FIP
KREBS, F. and SCHAMBECK, H. (1986). Grossdoordamm flyover in Hamburg — Incremental Congress, Stockholm.
launching under difficult geometrical conditions — Spannbetonbau in der Bundesrepublik PLACIDI, M. (1983). La construction au sol et la mise en place par deplacement. Annales ITBTP,
Deutschland 1983—1986. FIP Congress, New Delhi. December.
KROPPEN, H. and SCHAMBECK, H. (1982). Die Donaubriicke Metten. Beton und Stahlbetonbau, PLACIDI, M. (1988). Evolution et tendances des méthodes de construction des tabliers de ponts
May. poussés. Annales ITBTP, October. : =
KUSCHE, R. and VERCH, W. (1986). Briickenbau in der VDR Jemen. Bauplanung, Bautechnik, PLACIDI, M. (1990). La mise en place par rotation. L’Ingenieur Constructeur. June—July.
January. POIER, P. (1986). Eisenbahnbrtickenbau im Iran. Zement und Beton, March.
LAVIGNE, C., PLACIDI, M. and VIRLOGEUX, M. (1985). The new bridge over the Loir river at La POPOF, O.A. and SELIVERSTOV, VA. (1998). Steel bridges on Ankara’s peripheral motorway.
Fleche. L’Industria Italiana del Cemento, April. , IABSE Structural Engineering International, August.
LEHMANN, W. (1992). Ein Schieber mit Vielen Trumpfen. Viadukt Lowenberg n. 1. Ban, April. POPPY, W. (1989). Neue technische verfahren: strassen, briicken und tunnelbau. Osterreichische
LEONHARDT, F. (1964). Prestressed Concrete. Wilhelm Ernst & Sohn, Berlin. Bau Zeitung, 20.
LEONHARDT, F. (1969). Abminderung der Tragfahigkeit des betons involge strabformiger, PORTLAND CEMENT ASSOCIATION (1978). Wabash River Bridge, Covington, Indiana. Bridge Report
rechtwinklig zur Druckrichtung angeordneter Einlagen. Stahlbetonbau, June. SR 201.01E. Portland Cement Association , Skokie, IL.
LEONHARDT, F. (1973). Procédé de construction par cycles de bétonnage en coffrage fixe et cycles POST TENSIONING INSTITUTE (1976). First Incrementallwaunched Post Tensioned Box Girder
de' poussage. Annales ITBTP, January. Bridge to be Built in the United States. Bridge Report, December. Post Tensioning Institute,
LEONHARDT, F., ANDRA & PARTNERS (1987). Il ponte ferroviario sul fiume Meno a Gemunden. Phoenix, Arizona. - '
L’Industria Italiana del Cemento, December. - POZZATI, P. (1980) Teoria e Tecnica Delle Strutture. UTET, Turin.
LEONHARDT, F., ANDRA & PARTNERS (1994). Il viadotto ferroviario sul fiume Enz presso RADEMACHER, CH. (1986). Die montage der neuen Hammer Eisenbahnbriicke.
Enzweihingen (Germania). L’Industria Italiana del Cemento, October. Eisenbahningenieur, November.
LIBBY, J .R. (1990). Modern Prestressed Concrete, Design Principles and Construction Methods. RAMAKRISHNA, A. and SANKARALINGAM, C. (1997). Panval Nadhi Viaduct, India. IABSE
Van Nostrand Reinhold, New York. Structural Engineering International, August.
LLOMBART, J.A. and REVOLTOS, J. (1996). Cable-stayed pedestrian bridge, Spain. IABSE RAOUL, J. (1991). Some remarks concerning the use of corrugated webs for bridges. European
Structural Engineering International, November. Convention for Constructional Steelwork, October.
LLOMBART, J .A. and REVOLTOS, J. (2000). Petra Tou Romiou Viaduct, Cyprus. IABSE Structural RASPAUD, B. (1990). Les couvertures de 1’ aerogare n. 2 de 1’ aeroport Charles de Gaulle. AIPC—
Engineering International, November; ‘ ' ' ‘ IABSE Congress, Wien.
LOSINGER LTD (1977). The Incremental Launching Method in Prestressed Bridge Construction. REICHELT, A. ( 1989). Briickenbau in der VDR Jemen — Bauausfiihrung. Die Strasse, January.
April. Losinger Ltd, Berlin. REIFEGERSTE, L. and ZBYNEK, P. (1990). Externe Vorspannung im Briickenbau. Die Strasse,
MABOUNGA, D., SAKALA, L., SERVANT, C., SIBIDE, Y., TRAN V0 NHIEM, and ZUMAGLINI, M. (1986). December.
Pont sur le Kouilou. Travaux, January. REINBECK, U., VOSS, W. (1986). Entwurf und Bau der Strassenbriicke uber den Bahnhof und die
MADDISON, M. (1978). Crossing the cutting with segments at Sonning. Concrete, February. Ruhr in Meschede — Anwendung des Taktschiebeverfahrens in einer Wendelinie —
MARCHETTI, M. (1982). Specific design problems related to bridges built using the incremental Spannbetonbau in der Bundesrepublik Deutschland 1983—1986. FIP Congress, New Delhi.
launching method. 2nd Conference on Structural Engineering, Bangkok. RICHTER, K., SCHMACKPFEFFER, H. and HERSTELLUNG von, LP. (1988). Blechen und deren
__ _ ..____Ve_r_wendung..im .Briickenbau- DerStahlbau, February.
-
330 BRIDGE LAUNCHING REFERENCES AND BIBLIOGRAPHY
331

RIEDL, R. (1970). Das Taktschiebeverfahren im Bruckenbau. Zement und Beton, May. VRIEND H DE and DUI), H.J.C. (1987). Unieke brug ~— tunnelverbinding in
ringweg om
RIEDMANN, E. (1991). Mainbrucke Oberhaid Bundesautobahn A 770 Schweinfurt + Bamberg. Amsterdam. Cement, August.
Tiefbau — Berufsgenossenschaft, July. WALTER, S. (1997). 10.000 mg Fertigteile u’ber den Bahnhof in Basel. Bauwirtschaft,
12 May.
RIEDMANN, E. and SCHEWIOR, E. (1986). Talbrficke Marktbreit — Taktschiebebrucke von zwei WITTFOHT, H. (1984). Building Bridges. Beton Verlag, Dusseldorf.
Widerlagern aus eingeschoben. — Spannbetonbau in der Bundesrepublik Deutschland 1983— WITTFOHT, H. (1986). Wohin entwickelt sich der spannbeton-brfickenbau? Analysen und
tenden-
1986. FIP Congress, New Delhi. zen. Beton and Stahlbetonbau, February.
RIMBOEUF, M., SALZMANN, C. (1990). Ouvrage de franchissement de 1’ Azergues. AIPC—IABSE WOLFF, H.J. (1981). Spannbetonbrficke im Taktschiebeverfahren. Beton, May.
Congress, Wien.
ROSIGNOLI, M. (1998). Ponti ad arco a struttura tubolare composita. L’Industria delle Costruzioni,
January.
ROTH, C. (1984). Neue Wege im Eisenbahnbrfickenbau. Beton, June
ROWLEY, EN. (1993). Incremental launch bridges: UK practice and some foreign comparison.
The Structural Engineer (GB), April.
SCHAUCH, G. (1992). Von Lyon nach Valence mit rasender Geschwindigkeit. Stassen and Tiefbau,
October.
SCHEIDLER, J. (1990). Bauverfahren und ihre kritischen Montagezustande bei Grossbriicken.
Tiefhau — Berufsgenossenschaft, May.
SCHELLER, A. (1992). Die Donaubriicke Fischerdorf. Tiefbau — Berufsgenossenschaft, January.
SCHRADER, R. (1986). Erneuerung der Schwarzbachtalbriicke bei Dusseldorf unter
Aufrechterhaltung des Verkehrs auf der BAB A 3 — Spannbetonbau in der Bundesrepublik
Deutschland 1983—1986. FIP Congress, New Delhi.
SERVANT, C. (1982). Le pont sur 1’ Armancon ed le canal de Bourgogne — Ligne T.G.V. FIP
Congress, Stockholm.
SPIELMANN, A. (1988). Il viadotto di Roquebilliere a Cahors. L’Industria Italiana del Cemento,
October.
STICH, W. (1986). Rekonstruktion der Stadtbahnbogen in Berlin, Hauptstadt der DDR. Die
Strasse, July.
SVENSSON, H. and ZELLNER, W. (1983). Incremental launching of structures. American Society of
Civl Engineers — Journal of Structural Engineering, February.
TAKASHI, S., TEI, K. and TOMITA, M. (1999). Shin-Karato Bridge in Kobe, Japan. IABSE
Structural Engineering International, May.
TREFFOT, G. (1995). Le viaduc de la Corniche a Dole. Ouvrages d’Art 19, September.
TWIEHAUS, W. (1990). Eisenbahnbrficke uber den Rhein zwischen Dusseldorf und Neuss.
Baktuell, January.
TWIEHOFF, R. (1991). Stahlverbundbriicke uber den Guterbahnhof Dortmund. Der
Vermessungsingenieur, June
VERCH, W. (1990). Taktschiebeverfahren im Spannbetonbrfickenbau der DDR. Bauplanung,
Bautechnik, April.
VIRLOGEUX, M. (1981). Développement de 1’ innovation dans le domaine des grands ouvrages.
Travaux, December.
VIRLOGEUX, M. (1982). Evolution de la conception des ponts de portée moyenne. FIP Congress,
Stockholm.
VIRLOGEUX, M. (1983). La précontrainte extérieure. Annales ITBTP, December.
VIRLOGEUX, M. (1985). Bilan de la politique (1’ innovation dans le domaine des ouvrages (1’ art.
Travaux, March.
VIRLOGEUX, M. (1986). La precontrainte exterieure au beton. FIP Congress, New Delhi.
VIRLOGEUX, M. (1988). Die externe vorspannung. Beton and Stahlbetonbau, May.
VIRLOGEUX, M. (1990). Les ponts de portee moyenne. L’Ingenieur Constructeur, June—July.
VIRLOGEUX, M. (1991). La conception et la construction des ponts a précontrainte extérieure au
béton. Annales ITBTP, November—December.
VOLKE, E. (1992). Funktions und ausfuhrungsgerechte planung von stahlbriicken.
Eisenbahningeoieur, August. ,. , ,.
 Index

Entries in italics refer to a page where a figure appears apart from

the relevant text.

AASHTO requirements back thrust systems 32, 135, 234, 309

braking/locking forces 56 balanced cantilever construction 12—13
concrete covering 167 Basrah Bridge 21
curing time 145 beams
flange plate stresses 244 continuous 19, 21
PC bridges 53 on elastic foundations (BEF) 161—163
post-tensioning tendons 181 parallel 23
tensile and compressive stresses 190 simply supported 15
top flange bracing 220 bearings
unsupported flange length 213, 215 damaged 29—30
abutment diflerential settlement 119
as anchor element 32 fixed 19—20
seismic demands 28' low—damping rubber laminated 28
use in launching 32 mobile 6, 22
aesthetic considerations 126, 266 rolling 234, 310
Ager Bridge 3 sliding 19—20, 234, 235
algorithms 7, 110 bending moment envelope 62, 77, 118—119,
alignment corrections 76 178—179, 190
see also lateral guides bending stresses, during launch 16
Amiens Viaduct 165', 176, 1:87, 188 Boivre Bridge 9]
anchor blocks, precast 1'32, 137, 179, I80, bolt splices 223, 228, 229
183 bottom plate/slab
anchor boxes, prestressingxbars 8] composite box girder 218
anchor devices, launching‘nose 82, 83, 85—86 compressive stresses 310
antagonist tendons 185418983, 188, 202—203 concrete 15, 16, 272
Arbois Bridge 279, 280 connection to steel girders 283
arch local thickening 309
deck launching on to 937.4103 thickness 271
stabilization 98—99 boundary conditions, finite—element analysis
Aronde Bridge 47 304
assembly bearings, adjustable: 227 box girder, cross-sectional efliciency 279—281
assembly yard, organizationv225~229 box girders
axial launcher 44
composite 17, 61, 217—221
334 BRIDGE LAUNCHING INDEX
335

box girders (Coma) concrete slab 16, 246—257 concrete corrosion 207, 217, 246, 249, 254, 283
conventional (CBG) 278—281 desk slab 61 bottom slab 272 corrugated plate girders, fabrication 305—306
flexural efficiency 286 disadvantages 265—267 cracking 132, 179—181, 247—254, 267 corrugated web see web
monolithic casting 132—141 in falsework 279 creep 170, 172—173 costs
parallel 22, 23 and incremental launching 13 curing 149—151 casting yard 24
prestressed composite section 277—279 support diaphragms 29 design strength 132 corrugated plate girders 305—306
reinforcement 206—207 temporary piers 95 elastic deformation 49, 120, 121—123, 173, and dead load/service load 61
box stifieners 77, 78, 79, 79 casting yard 276 labour costs 8, 50, 253
braking, during launch 32, 41, 43, 56—57 costs 24 hardening time 131—132 launching devices 40—41
breakaway forces 56 fixed 98 high-performance (HPC) 132, 267, maintenance costs 19, 197
Bredt’s torsion theory 160, 163 foundation 32 272—275 reduced by segmentation 133
BS 5400 industrialization 100, 128, 130 high-strength 179 couplers, staggered 181, 182
compressive stress 256 long 88 lightweight 275—276 coupling, tendons 180—182
flange plate width 236 organization 128—131, 145—146 local heating 179 cracks
geometric tolerances 228 placing 32, 41 mechanical strength 252 avoiding 179, 247—254
horizontal bracing and cross-beams 238 Champigny-sur-Yonne Bridge 165 monolithic strength 197 composite bridge 252, 254
lightweight concrete 276 Charente Bridge 306 no-shrink 258, 260, 264, 266 horizontal 132
maxiumum bar spacing 246 Charix Viaduct 91, 145, 204 pouring 266, 283, 284 tensile 180—181
shear connectors 298—299 Charolles Bridge 301, 302, 303, 306, 312—316 pumping 266 creep
thermal gradients 114 chemically bonded ceramic (CBC) 272—273 reactive powder (RFC) 273 concrete 170, 172—173
vertical compressive stress 240 climates, severe 132, 258 shear efiiciency 271—272 effects on launch 174—177
web panel stiffening 216 clothoid section 22 shrinkage 171—172, 252—253 HPC 274—275
Bubiyan Bridge 278—279 Cognac Bridge 302, 305, 311—312, 313 slab see slab critical stress, web panels 238—239
buckling composite bridges 209—269 thermal expansion coefficient 123 cross-beams 264
bottom plate 218 advantages 15, 209—210 time—dependent behaviour 101, 120, headed stud connectors 264
global 300 aesthetic considerations 266 170—177 spacing 216—217
interactive 301 box girders 217—221 connectors cross-braces, spacing 213
local 241, 300 curved 220 BS 5400 298—299 cross-sectional deformability 294—296
safety margins 301 dead load 61 corrugated plates 306 cross-sectional efficiency 279—281
web panels 238, 242, 285 deck presizing 210—221 requirements 296—299 curing rail 134—136
bush-hammering, construction joints 130, design aspects 210 constraint eccentricity 29 curing yard 149—151
132 fatigue 256, 262 construction equipment see equipment curvature
butt weld joints 223 incremental launching 15—1 6 construction
'
jomts
' ' +
ruz
10'),
constant 21
see also welding launch stresses 60—61, 225 composite bridges 257 radius 31
launching nose 64, 87 horizontal and vertical 132 variable 20
cable bearings 233 multigirder 211—212 improved techniques 209 curved alignment 30—31, 42, 53, 220,
cable-stayed front system 88—94 recent developments 209 launching truss 102—103 230—231, 238
Cadore Bridge 26 safety factors 256 position 129, 142, I44
cantilever devices 6, 63—64 service stresses 225 precast segements 15 1—152 dampers 28, 29
cap tendons 195 slab casting stresses 225 weak points 100 dead load 11
carrying capacity 61 two—girder 212—217 see also bolt splices; field splices; welding composite box girder 61
casting compression flanges, instability 230, 272, construction risks, and dead load 267 concrete slab 16
and crack control 247—254 284 construction stresses, and dead load 11 and construction risks 267
monolithic 132—141 compressive stress construction time 140 and construction stresses 11
phases 131—146 AASHTO requirements 190 contact plates 51, 52 and costs 61
segmental 10 launching bearings 166—169 contact stresses, launching nose 85 deck slab 210
two—phase 141—146, 307 longitudinal 168—169, 242—244 continuous beam high 30
casting cell neo-flon pads 167 analysis by RTM method 110—1 18 and launch stress 210
adjustable 30 steel bridges 10—11 launched on to arches 23 limiting 271
aerial 14], 147 vertical 166—188, 240—242 static system 115 prestressed composite bridges 276,
fixed 134, 135 computer methods, advances 1 conventional box girder (CBG) see box 280—281, 308—311
casting in situ computer software 110
girder..._._ , - reducing 275—276
336 BRIDGE LAUNCHING INDEX 337

dead load (Cont) extensometers 121, 123 safety factors 49 incremental launch
superstructure 61 extraction rails 146—149, 307 working cycle 45 advantages 102—103, 265—267
deck friction launching, heavy superstructures and casting in situ 13
cast in situ 61, 210 fabricated elements see materials 43—53 composite bridges 15—16
concrete slab 10, 16 falsework 13, 100, 247, 279 friction load-transfer devices 35, 36 concrete slab 257—269
dead load 210 fatigue, composite bridges 256, 262 friction thrusters 107, 309 cost savings 7~8
incremental launching 99 field activities, prestressed composite bridges friction-lock devices 37—38 deck 99
launching on to arches 97—103 277, 307—308 frictional resistance, discontinuous 268—269 Kufstein Bridge 5
lightening 274 field splices over an arch 98 ‘
presizing, composite bridge 210—221 corrugated plates 306 gantry crane 130, 225, 226 . prestressed composite bridges 16—17,
segmental casting 102—«103, 128—131, 211 design characteristics 223 geometric constraints 26, 29—32, 38—40 286—288, 308—311
soflit, damage 150 horizontal 77 geometric tolerances, allowable 228 prestressed concrete bridges 13—15
densified with small particles (DSP) nose—deck joint 87 Ginzan-Miyuki Bridge 302, 304, 306 techniques 10—13
materials 273 welding 228—229, 307 Glacieres Bridge 279 Val Restel Bridge 30
depth finite-element analysis gradients see uphill/downhill launch Wandre Bridge 27
of bridge 127 boundary conditions 304 Great Belt Link (Denmark) 275 see also launch
constant 30 continuous beams 165 Greveniotikos Bridge 29 Isere River Bridge 24
design assumptions, monitoring 120—124 corrugated webs 289, 291, 292, 293 guide devices
design-level earthquake 28, 53, 54 launch bearings 310 and launch bearings 53—55 jack systems 33—39
diaphragms, spacing 213 launch support reactions 286 see also lateral guides jack-bearings 94—95, 100, 107
DIN 1072 114 launching nose 79 gusset plates 80, 84 jacking force, slab launching 2684269
DIN 4227 276 longitudinal compressive stress 169 jacks, hydraulic 154, 258, 269, 286, 310
distortion stresses 220 spreadsheet analysis 159 headed stud connectors 263, 264, 298, 306 joints see construction joints
double composite action 219, 220 support region of deck 121 highway bridges 15, 127, 128
downhill launch see uphill/downhill launch truss nose 236 hinge bearings 28 kerbs 32
draw systems 32, 38—40 vertical compressive stress 168 hoisting jacks 155—156, 168 Kufstein Bridge 5, 9
drawbars 34 flange connectors, spacing 298 Hoover Dam Bridge Bypass 98
drying, shrinkage 247—248 flange plates HPC see concrete, high-performance L-girder 227, 229
dynamic analysis, non-linear 29 in corrugated webs 292 hydration heat 275 labour rotation see workforce, organization
dynamic response, high velocities 28 dimensions 213—214, 220 hydraulic launch devices 32, 40, 4] lateral guides 30, 53, 54, 231, 260
direct slab launch 258 see also jacks, hydraulic launch
EC2 246 in prestressed composite bridges 277 back thrust systems 32, I35, 234, 309
EC3 238, 239, 241, 242, 245, 256, 301 flange stiffeners 78, 79, 79 I-girders cantilever 63—64
EC4 211,216, 239, 246 flexural efficiency 62, 277, 286 assembly bearings 227 curved alignment 230—231, 238
edge stresses 176, 178, 190, 191—194", 272 formwork concrete-filled 247 discontinuous 6
elastic behaviour double 142—146 cross-beams 216—217 downhill see uphill/downhill launch
concrete 49, 120, 121—123, 173, 276 external 141—142 instability 230, 238, 284, 285 effects of concrete creep 174—177
steel girders 232 fixed 126 local yielding strength 241 equipment 15, 32, 152
elastomeric bearings 230, 233, 234, 286, for incremental launch 13 modelling 237 fiction forces 56—58
309—310 internal 11 parallel 212—214 forces 32, 55—58
Elorn Bridge 275 left—in—place 133, 247 in slab launching 265 guides 26, 30, 155
environmental considerations 100, 101, 102, mObile 134, 250, 265 stress—strain relationship 243 incremental see incremental launch
103 recovery and reuse 129, 130, 132, 141—142, torsional rigidity 231 openings 259—262, 264, 266
equipment ~ 258, - - ‘ transverse 263, 264 parallel I44
amortization 14, 15, 16, 32, 34 self—extracting 133, 134 web plates 219 precast segmental bridges 15, 152
custom-made 152 shifting 136 ICE, high-speed railway bridges 21, 24, 103 prestressing 60, 157, 178, 179wl89,
heavy 133 short 140 He Falcon Bridge 31 190—194
inexpensive 32 friction forces, at launch 56—58 in situ casting see casting in situ problems 30
reuse 10, 13, 130 friction launchers 32, 43—44, 45, 122 in-yard construction, precast slab segements rails 262—263
specialized 14 advantages 49—53 258 reactors 107, 108
standard 13, 130 design 169 inaccessibility l3 speed 42, 50
expansion joints 28 . ___ _ . , , ”positioning .194 inclinometers 106 support, adjustable 230
338 BRIDGE LAUNCHING INDEx 339

launch (C0m‘.) Rio Caroni Bridge 4 nose~decl< interaction 64—74, 87, 193 dead load 276, 280—281
support reactions 286 segmentation 76 numeric control 48, 50 efiiency 10
techniques 9, 15, 32—53 sizing 73—82 field activities 277, 307—308
on to arches 23, 974103 sliding over superstructure 22 Oli Bridge 20—21 flexural efliciency 277
uphill see uphill/downhill launch steel girder 231, 235—236 Olifant’s River Bridge 20, 51, 53 incremental launching 16—17, 286—288,
see also incremental launch stiffness 68—69 308—311
launch bearings trussed 64, 65, 73, 98, 102—103 P-delta effects 109 launch stresses 60
adjustable 31—32 types 64, 73 Palizzi Bridge 27, 49, 56, 97, 142, 160 launching bearings 286
compressive stress 166—169 see also nose—deck interaction pantograph, hydraulic 309 launching nose 64
damaged 159 launching saddles 137—138, 154 parabolic prestressing 194—195 stiflened-plate webs 281—288
design 285 launching shoes 260, 263 PC box girder, over an arch 97—103 structural efficiency 281, 282
finite-element analysis 310 launching tendons 34—35, 178—179 PC bridges technological advances 10
and guide devices 53—55, 231 live load 11, 61, 157, 194 aesthetics 126 thrust systems 309
insertion 147 local stresses, above launching bearings launch bearings 153—170 yard organization 307—308
integral 309 18 165—470 launch stresses 58—60, 125 prestressed composite sections, shear forces
location 62—63 lock forces 55—58 launching nose 64 17
low-friction 55 lock-up devices (LUD) 28, 29 prestressing 82—85, 178~205 prestressed concrete, competitiveness 1
overload 159, 160 longitudinal compressive stress 168—169, shear force 85 prestressed concrete bridges
PC bridges 153—170 242—244 standards 53 historical aspects 1—10
prestressed composite bridges 277, 286 superstructure presizing 125~128 incremental launching 13—1 5
and PTFE sliders 26 macro defect free (MDF) materials 273 PC section, flexural capacity 125 launch techniques 9
rocking 232, 307 Mainflingen Bridge 23 peak ground acceleration (PGA) 27, 54 prestressing
steel girder 231—234 materials Piave River Bridge 26 axial 286—287
support reactions 157, 166 fragile 247 pier—cap arrangement 155—156 centroidal 19, 204
transverse shift 154 maximum length 306 piers cost 152
transverse width 245 standard dimensions 222 launch stresses 103—110 external 10, 272
vertical misalignment 156—165 transport 218, 222, 252 temporary 21, 26, 63, 92—97, 125, 126, see also tendons, external
launch stresses Megalorema Bridge 29, 167 140, 14], 155-—156 extradosal 22
composite bridge 60—61, 225 modelling, two— or three-dimensional 236, pins integrative 101
correction 58—103 237 preloaded 73, 78, 79 mixed 201—205
and dead load 210 moment of inertia 61—62 through 37, 153 modelling 111—1 12
local 165—170, 236—245 monitoring, to verify design assumptions plate girders 223, 238 PC bridge 178—205
PC bridges 58—60 120—124 plate—web bridges 277 recent, developments 201—205
piers 103—110 movable shuttering system (MSS) 98, Pont de la Corniche 301, 302, 303, 310, steel 128, 271, 272
prestressed composite bridge 60 100—102, 247, 250—252, 265 311 temporary joints 189
steel frame 210—211 multigirder composite bridge 211—212 post-tensioning, transverse 257 transverse 199
launching nose preassembly area 130 vertical 200
anchor devices 82, 85—86 Neckar River Bridge 24, 25 precambers 232, 237, 285, 306, 307 prestressing bars, anchor boxes 8]
bracing system 75—76 neo-flon pads precast planks 247—249, 265 prestressing jacks, drawing devices 32
composite bridge 87 clearance 168 precast segments prestressing techniques
contact stresses 85 compressive stress 167 advantages 257—M258 advances 1
coupling to superstructure 82—87 construction 153 Ager Bridge 3 PCS box girders 297——298
cross-frames and bracing 80—82 excessive wear 29 assembly and launch 32, 151—152 prestressing tendons, Rio Caroni bridge 7
design stresses 77 incorrect insertion 40, 49—50, 106, 155, 168 in-yard' construction 258 progressive placement 12
devices 83 insertion 147, 154—155, 287 precast towers, modular 95 PTFE
disassembly 82 new technique 6 pressing equipment, steel plates 306 importance in launching techniques 103
finite-element analysis 79 removal 50 prestressed composite box girders sliders 26, 28, 233
flexural stiffness 73 reuse 149 prestressing techniques 297—298
optimum length 70—71 wide 309 web instability 299—305 quality control 13, 14
presizing 72 nose see mainly launching nose prestressed composite bridges 271—316
prestressed composite bridges 308 nose tip deflection, recovery 87—88, 236, corrugated webs 288-316 radial guides, uphill/downhill launching 54
reuse 64, 73, 76 309 cross-sectional efficiency 279—281 __ radius of gyration 62, 126
340 BRIDGE LAUNCHING INDEX 341

railway bridges 128, 132‘ service prestreSsing 194—205 span—by-span construction 12 backwards sliding 38
rainfall drainage 227 service stresses 225 spans, irregular 119 constant-depth 31
reactive powder concrete (RFC) 273 settlement, differential 19 splices see field splices construction 14
rear launch system 39, 40, 4] shear connectors 284, 306 spreadsheet analysis 159 coupling to launching nose 82—87
reduced transfer matrix (RTM) method shear curvature 293—294, 305 static analysis 1 dead load 61
1 10—1 18 shear efficiency 271 static system division 20
reinforced concrete, compressive strength shear forces continuous beam 115 front zone 193—194
272 corrugated webs 291—294 modification 21 geometry 29—32, 49
reinforcement envelopes 118—1 19 * and seismic design 26—30 heavy 43—53
box girder 206—207 PC bridges 85 superstructure 19—20 hoisting 155—156
concrete cover 246, 248 prestressed composite sections 17 variations 11 light 33—43
highway/railway bridges 128 steel 272 stays, temporary 92 limit support conditions -59
watertightness 246, 248 shear keys steel long segments 13
reinforcement ratio, minimum 246 concentrated 80 relaxation 173—174 presizing 125—128
ribbed slabs, monolithic casting I33 distributed 86, 132 shear efficiency 271—272 rear zone 191—193
Rio Caroni Bridge 3—5, 7, 184, 316 sacrificial 28—29 technological advances 9, 15, 209 shear force reduction 198
road bridges see highway bridges sliding 29 towers 95—96 static system 19—20
Roize Bridge 267 transverse 55 steel cage, preassembly 129—130, 133, 143, structural system 1926
roll compacting 172, 266 shear lag 218, 236, 278 266, 307 suspension 20
roller bearings 233, 234, 310 shear stress, design 301—302 steel girders transverse shifting 23, 24
RPX-95 211, 216 shrinkage elastic deflections 232 support, differential 58
concrete 123—124, 252—253 flexural strength 213 support diaphragms, cast in situ 29
safety factors cracking 267 launch 52, 76, 229—245 support reactions
buckling 301 drying 247—248 launching nose 23 5—236 alignment 287
composite bridges 256 HPC 274 maximum length 306 eccentricity 233
friction launchers 49 lightweight concrete 276 segmentation 221—225 finite-element analysis 286 .
incremental launching 101, 102, 103 thermal 252—254 steel plates incremental launch of PCS bridges 286
launch prestressing 178—179 silica fumes 275 axial stiffness 288—289 launch bearings 157, 166 ».
launch tendons 34—35 single-line analysis, 2D 236, 237 corrugated 289 local eccentricity 297
Salbris Bridge 289 Sintal Bridge 21 orthotropic 288—289 longitudinal compressive stress 168—169
Salhus Bridge 276 ' Skye Bridge 14 pressing 306 steel girder 231—232
Sart Canal Bridge 103 slab standard dimensions 222 vertical compressive stress 166—168 .
Sathorn Viaduct I86, 187 cast in situ 16, 246—257 transport 22.2 surface finishing 266
Schnaittach Bridge 21 casting stresses 225 steel structures Sylans Bridge 279
Schroetal Bridge 2, 21, 34 conneCtion to webs 296—299 launching 2
segmental construction 13—1 5 cracking 246, 252 stress resistance 10—11 Tanaro River Bridge 23, 56
casting 249 dead load 16 weight 217 Tancarville Canal bridge 261
cycle 131 durability 254—255 steel—Teflon contact 32, 56—57, 103, 148, technological advances 1, 9, 10, 15
precasting 257—258 incremental launching 16, 257—269 153, 155 Teflon see PTFE; steel—Teflon contact
reduces construction costs 133 launch openings 259—262, 264 Steine Bridge, temporary piers 96 temperature elfects 56»
techniques 11—13 launching stress analysis 267—269 strain gauges 120, 121, 123—124 temporary joints, prestressing 189
segments prestressing 254—257 strand pulling 49 tendons
casting in situ 32 segmental casting 249 stress anchoring 197
extraction 146—149 segmental precasting 257—258 analysis, in slab launching 267—269 antagonist 185—188, 188, 202—203
length 13, 173 segments, length 306 in corrugated webs 291—294 deviated 310
seismic factors 19, 26—30, 53, 55 support 221 transverse 245 external 4, 100, 184—185, 196—199,
self-clamping devices 32, 40—43 tensile stress 253 verification 169—170, 245 201—205
self—resistance 10, 11, 12 thickness 127, 211 structural efl‘iciency, PCS bridges 281, 282 internal 100, 195—196, 272, 307
Serio River Bridge 48, 49, 50, 130, 159, waterproofing 254 structures, presizing 61, 63 overlapping 182—183
177 . sliding bearings 234, 235, 310 superstructure parabolic 194—195, 197
service life 254—255 sliding materials 57, 103 alignment corrections 76 polygonal 298, 307
anchoring to piers 29 post-tensioned 178
servi,c.e,,load 61, 1.9.4. ' ...space-framebridges277a278
342 BRIDGE LAUNCHING

tendons (Cont) uphill/downhill launch

pretensioned 34—35 average gradient 56
recovery/reuse 183, 187, 201 braking/control 33, 34, 38, 42
straight, coupled 180—182 curved alignment 54
tension 4—5, 11, 173—174 friction forces 57—58
tensioning cycle 35 friction launchers 53, 108
tensile stress 10, 253, 254 mechanical contrast 40
tensioning=34—35, 84 pier cap forces 105—106
TGV Atlantique bridges 20 radial guides 54
thermal conductivity, lightweight concrete superstructure locking 34, 36
276
thermal correction 123 Val Restel Bridge, incremental launching 30
thermal expansion coeflicient 120, 123 Veitshoechheim Bridge 24, 25, 26, 103
thermal gradients 114, 119, 120, 121, 123 vertical compressive stress 166—188, 240—242
thermal shrinkage see shrinkage
thermal stresses 58
Wandre Bridge 26
thermocouples 120
warping 238 ,
thrust devices 6, 10, 32, 39—40, 309
stress 220, 243—244
see also hydraulic launch devices-
watertightness 246, 254
Tiziano Bridge 39, 137—138
web
top surface, finishing 266
buckling 238, 242, 285, 299
tow systems 32, 33—39, 234
connection to slabs 283—284, 296~299
tower crane 13, 129, 132
corrugated 2884316
towers, modular 141—142
critical stress 238—239
transfer matrix method 110—1 11
cross-sectional deformability 294~296
transport
instability 284—285, 299—305
critical dimensions 222
shear 271, 272
fabricated elements 218
stiffened 128, 213, 216, 218—219, 281—288
steel plates 222, 306
thickness 127—128, 196, 271, 277, 285
to working location 252
welding
transverse shifting 137—138, 154
automatic 229
transverse stresses 245
field splices 223, 228—229, 307
truss launching nose, steel girders 235—236
West Coast Main Line Railway bridges 210
trussed-web bridges 277, 278
winches 32
two-girder composite bridge 212—217
wind, effects 53, 54
Wire mesh 206
U-girder
workforce, organization 142—143, 144, 145
assembly bearings 227
working areas, specialized 226
bottom flange 218
workshop, organization 222
in composite box girder 217
Worli-Bandra Sea-link project 27
compression flange stress 243—245
cross-frames 220
field splice 229 yard
incremental launching 239 operations 13, 14
instability 230 organization 307—308 ' ' '
local yielding strength 241 see also assembly yard; casting yard;
parallel 221 curing yard
torsional rigidity 231, 237 yielding, local 285
web plates 219
U-segments, two—phase casting 145—146 ZTV-K88 standards, PC bridges 54
 received his degree with hOHOUrS in Structural Engineering from the University
of Ancona, ltaly in 1981.
A site director, project manager, and for ten years head of the bridge department of a prime
italian contractor, he managed the design and construction of a large number of prestressed
concrete, composite, and steel bridges. He has been a freelance consultant since 1997; his
activities include the design and independent design check of bridges and specialized bridge
construction equipment, bidding assistance, and constructabiiity review.
Twenty years of independent research on design and construction methods for bridges have
been summarized in two books on launched bridges and numerous technical papers.

ISBN D~?E??-3LHb-? 4 9‘T1?”*5Q7Ui: ¢“**,~r‘k5a

' :1 781072? 731%}!