Edited by

Mahmoud

On Thursday evening, May 23, 2013, the Interstate 5 Bridge over the Skagit
River in Washington state collapsed due to impact by an oversize truck, dumping
vehicles and people into the water. Fortunately, the bridge is located in a rural
area and nobody was killed in the accident, but three people were rescued after
their cars plunged into the frigid water of the Skagit River. According to
Washington state officials, the bridge was inspected last year and was not
structurally deficient, but collapsed because of apparent impact from an oversize
truck. Nevertheless, the collapse of the steel truss bridge renewed appeals for

Durability of Bridge Structures

greater investment in the nation’s aging infrastructure. These appeals are echoed
throughout the bridge engineering community worldwide, as the condition of
deteriorated bridges worsens with increasing traffic loads combined with lack
of proper maintenance.

Bridge engineers from different countries shared their experience toward

achieving durable bridge structures, during the 7th New York City Bridge
Conference, held on August 26-27, 2013. This book contains select papers that
were presented at the conference. These peer-reviewed papers are valuable
contributions and of archival quality in bridge engineering.

Durability of Bridge Structures

Edited by Khaled M. Mahmoud

an informa business
 DURABILITY OF BRIDGE STRUCTURES

FM.indd i 7/17/2013 5:08:23 PM

This page intentionally left blank
 Durability of Bridge Structures

Editor
Khaled M. Mahmoud
Bridge Technology Consulting (BTC)
New York City, USA

FM.indd iii 7/17/2013 5:08:23 PM

 Front Cover:
Jacques Chaban-Delmas Lift Bridge, Bordeaux, France
Photo courtesy of © REGIS DUVIGNAU/Reuters/Corbis
Back Cover:
Jacques Chaban-Delmas Lift Bridge, Bordeaux, France
Photo courtesy of Bernard Blanc
Cover Design:
Khaled M. Mahmoud
Bridge Technology Consulting (BTC)
New York City, USA

CRC Press/Balkema is an imprint of the Taylor & Francis Group, an informa business

© 2013 Taylor & Francis Group, London, UK

Typeset by V Publishing Solutions Pvt Ltd., Chennai, India

Printed and bound in the USA by Edwards Brothers, Inc., Lillington, NC

All rights reserved. No part of this publication or the information contained herein may be
reproduced, stored in a retrieval system, or transmitted in any form or by any means, elec-
tronic, mechanical, by photocopying, recording or otherwise, without written prior permis-
sion from the publisher.

Although all care is taken to ensure integrity and the quality of this publication and the
information herein, no responsibility is assumed by the publishers nor the author for any
damage to the property or persons as a result of operation or use of this publication and/or
the information contained herein.

Published by: CRC Press/Balkema

P.O. Box 11320, 2301 EH Leiden, The Netherlands
e-mail: [email protected]
www.crcpress.com – www.taylorandfrancis.com

ISBN: 978-1-138-00112-1 (Hbk)

ISBN: 978-1-315-85684-1 (eBook PDF)

FM.indd iv 7/17/2013 5:08:23 PM

 Table of contents

Preface vii

1 Cable-supported bridges
Main cable anchorages investigation at Forth Road Bridge 3
B.R. Colford
Fabrication and construction of self anchored San Francisco Oakland Bay
Suspension Bridge 17
M. Nader, G. Baker, J. Duxbury & B. Maroney
Structural cable designs and concepts 31
T.W. Klein
Latest developments in suspension bridge main cable dehumidification 39
M.L. Bloomstine
Design, analysis and construction of Puente Mercosur—third bridge across
the Orinoco River 55
K. Humpf, M. Müller & C. Silingardi

2 Bridge analysis & design

Development of economical short-span steel bridge standards 67
K. Barth, G. Michaelson, M. Barker & D. Snyder
Fatigue properties of steel-concrete composite slab using a cementitious adhesive 79
I. Yoshitake, A. Ogawa, Y. Fujimoto & Y.J. Kim
Development of a shallow press-brake formed tub girder for short-span steel bridges 87
K. Barth, J. Mash, G. Michaelson, M. Barker & D. Snyder
Load-carrying capacity of corroded end of I-section steel girder 99
E. Yamaguchi & T. Akagi
Uplift bearings—selection and design considerations 105
G. Moor, T. Spuler, N. Meng & C. O’Suilleabhain
Design of Idemli Bridges in landslide areas 115
C. Ozkaya, G. Cetin & F. Tulumtas

3 Bridge rehabilitation and retrofit

Precast concrete filled steel grid panels for deck replacement at the Newburgh-Beacon
Bridge—South Span 123
W. Moreau & S. Abdou

FM.indd v 7/17/2013 5:08:23 PM

 vi Table of contents

Blatnik and Bong Bridges wire rope replacement: Maintaining Minnesota’s

unique bridge inventory 135
L. Amundson, R. Haight & E. Lutgen
Rehabilitation of the Alexander Hamilton Bridge and Highbridge Interchange Ramps 147
M. Tughral Shaikh, H.-P. Ho & T. Bashir
Jointless ‘floating’ deck for the Walt Whitman Bridge 161
R. Prior, D. Griffith, Q. Ye & V. Pandya
Brooklyn Bridge orthotropic deck installation 177
R. Leo & P. Reynolds

4 Seismic analysis & design

Case studies on several isolated bridge configurations adopting a performance
based approach 185
D. Forcellini & M. Banfi
Assessment and design of seismically isolated bridges in proximity of major faults
and located on liquefiable soils 195
C. Ozkaya, G. Cetin & F. Tulumtas
Seismic retrofitting of curved highway viaducts 207
G. Moor, C. Mendez & T. Spuler
Retrofit procedures of bridge networks against seismic actions in Italy 215
A. Pipinato

5 Bridge monitoring
Investigation of feasibility of deterministic, theoretical predictive models
underpinning visual condition monitoring outcomes 231
A. Sagar, S. de Silva & S. Setunge
GPR testing of reinforced concrete viaduct beams 245
D. Beben, W. Anigacz & A. Mordak
Monitoring of soil-steel railway bridge under service loads 253
D. Beben & W. Anigacz

6 Bridge aesthetics & history

Artists’ portraits of the Sydney Harbour Bridge 265
S. Rothwell
Measures of structural art: A case study using Streicker Bridge 275
D.H. Sigurdardottir & B. Glisic
Lindenthal and the Manhattan Bridge eyebar chain controversy 285
K. Gandhi

Author index 301

FM.indd vi 7/17/2013 5:08:23 PM

 Preface

On Thursday evening, May 23, 2013, the Interstate 5 Bridge over the Skagit River in
Washington state collapsed due to impact by an oversize truck, dumping vehicles and people
into the water. Fortunately, the bridge is located in a rural area and nobody was killed in the
accident, but three people were rescued after their cars plunged into the frigid water of the
Skagit River. According to Washington state officials, the bridge was inspected last year and
was not structurally deficient but collapsed because of apparent impact from an oversize
truck. Nevertheless, the collapse of the steel truss bridge renewed appeals for greater invest-
ment in the nation’s aging infrastructure. These appeals are echoed throughout the bridge
engineering community worldwide, as the condition of deteriorated bridges worsens with
increasing traffic loads combined with lack of proper maintenance.
Bridge engineers from different countries shared their experience toward achieving durable
bridge structures, during the 7th New York City Bridge Conference, held on August 26–27,
2013. This book contains select papers that were presented at the conference. These peer-
reviewed papers are valuable contributions and of archival quality in bridge engineering.
The Proceedings lead off with a paper by Colford on the “Main cable anchorages investi-
gation at Forth Road Bridge”. The paper describes the investigation work carried out to try
to determine the remaining service life of the main cable anchorages at Forth Road Suspen-
sion Bridge. The anchorages on any suspension bridge are critical components and to be
able to inspect these components to evaluate deterioration and remaining strength is vital.
At Forth Road Bridge, the anchorages were constructed, on both the north and south sides,
by forming tunnels within the rock formations. Concrete has been cast within these tunnels
to form four large, individual, concrete plugs within the rock to which the cables are attached.
The concrete in the tunnels is reinforced using pre-tensioned, galvanized, high tensile steel
strands in grouted ducts. The strands are completely inaccessible and in a hostile environ-
ment. Concerns over the possible loss of galvanizing were expressed during construction and
further evidence came to light in 2005 that the condition of the strands may be worse than
previously reported. Given all the above, an investigation was carried out to try to determine
with some certainty the long term structural integrity of the anchorages and also to establish
that the risk of failure remains low and within acceptable levels. The investigation was consid-
ered to be a proactive measure to ensure that all parts of the structure that can be reasonably
accessed are inspected. The signature span of the new San Francisco Oakland Bay Bridge is
an asymmetric Self Anchored Suspension (SAS) Bridge, spanning 565 m. The superstruc-
ture hangs from an inclined cable system supported on a single four-shaft tower. Unlike a
traditional suspension bridge with parallel ground-anchored cables, the SAS has a single
3-dimensional cable that is anchored into the east end of the box girders, and loops over the
tower top and around saddles at the west pier. Box girders carry the cable tension; therefore
a temporary bridge supported them during erection, and throughout the cable construction.
The tower top, box girders, bearings and cable bands were advance positioned to fit the final
geometry. Geometric control during fabrication and construction monitored tower plumb-
ness, girder cambers, cable band twists, and suspender inclinations. In “Fabrication and
construction of self anchored San Francisco Oakland Bay Suspension Bridge”, Nader et al
discuss the challenges faced and the innovative solutions that were developed to construct
this unique structure. The use of zinc coated high carbon steel wires in the fabrication of
bridge cables has been practiced for decades. In “Structural cable designs and concepts”,

vii

FM.indd vii 7/17/2013 5:08:23 PM

 viii Preface

Klein discusses new and enhanced structural cable designs for use as tension members on
bridges and static applications. These developments include new wire coating processes for
increased corrosion resistance and the use of synthetic materials for amplified strength and
durability. Discussions regarding the use of synthetic materials in structural elements have
significantly increased recently leading to several hybrid cable concepts that provide greater
resistance to fatigue and strength without the vulnerability to corrosion. The paper also dis-
cusses existing technologies currently being used to enhance steel cable properties. The use of
shaped wires formed from high carbon steel provide high strength cables with locked outer
layers for increased protection from the elements. The paper examines the advancements in
these fields along with proven manufacturing practices shown to develop high performance
cables for use in new and existing structures. Degradation of suspension bridge main cables
is a very serious and widespread problem that compromises the safety of many bridges. The
deterioration process sets in due to the high stress in the wires in the presence of moisture
inside the cable. Dehumidification systems have been developed, mainly for parallel wire
cables, to provide a measure of protection against deterioration. It is important to articulate
that dehumidification should be regarded, at best, as a mitigating measure to slow down deg-
radation. The most efficient approach to validate efficacy of dehumidification techniques is
to perform periodic in-depth inspections of dehumidified cables. During inspection, samples
of wires are removed from the cable and tested to obtain their mechanical properties, which
are in turn used to estimate the remaining strength of the cable. This provides a quantitative
measure of the effectiveness of dehumidification techniques in protecting the cables against
deterioration. In “Latest developments in suspension bridge main cable dehumidification”,
Bloomstine presents a general description of corrosion protection by dehumidification, the
current worldwide status, examples of systems and the latest experience and developments
including full-scale testing, on-site testing and new optimal details. With the aim of develop-
ing the southwest of Venezuela and enabling its connections with the rest of the country, a
third bridge over the Orinoco River was designed between the cities of Cabruta, State of
Guárico in the north and Caicara del Orinoco in the south. The combined highway and rail-
road bridge Puente Mercosur will be the third crossing of the Orinoco with a total length of
11.125 km. The cable-stayed main bridge with 360 m main span and the adjacent approach
bridges with a length of 2 × 720 m are comprised of a steel composite truss girder with a
depth of 12.0 m. In “Design, analysis and construction of Puente Mercosur—third bridge
across the Orinoco River”, Humpf et al provide details of achieving this pioneering invest-
ment in face of difficult challenges to construct in such a fairly remote area of the country.
A considerable number of bridges in the United States are either structurally deficient or
functionally obsolete. Consequently, one of the main focuses of the bridge industry is on design
techniques that provide a high-quality and cost-effective product in a relatively short amount
of time, with less maintenance required and a longer life span. Standardized bridge plans sig-
nificantly reduce design time, increase and provide cost-effective solutions in the nation’s effort
to repair and upgrade the infrastructure. In “Development of economical short-span steel
bridge standards”, Barth et al provide details of a design aid for bridge engineers to assist in
optimizing the design process. In typical practice, mechanical shear connectors, such as headed
studs and L-shaped steels sections, are employed to provide composite action. Construction
of such composite bridge decks may be adversely affected by welding the shear connectors.
A new bridge deck system was recently developed, in which composite action is achieved using
a cementitious adhesive. Fatigue durability is a critical issue for bridge deck slabs subjected to
wheel load, so the property of the new slab system should be examined via experimental inves-
tigation. In “Fatigue properties of steel-concrete composite slab using a cementitious adhe-
sive”, Yoshitake et al examine the fatigue strength in a conventional bending test. In particular,
the experimental investigation aims to observe failure behavior of the composite slab subjected
to cyclic flexural load, and to examine the flexural fatigue strength. Prefabricated steel tub-
girder systems have been explored as a potential design solution for the short-span bridge
market for a number of years. Many previous research efforts have shown that these types of

FM.indd viii 7/17/2013 5:08:23 PM

 Preface ix

systems have the potential to be economical and competitive in the short-span range. In recent
years, the demands for accelerated bridge construction have been brought to the forefront
of design. Therefore, bridge systems such as the press-brake tub girder have once again sur-
faced as a viable alternative to conventional bridge fabrication. In “Development of a shallow
press-brake formed tub girder for short-span steel bridges”, Barth et al provide an overview
of experimental testing currently being conducted and further parametric analysis and design
studies focused on the refinement of cross-section dimensions. Corrosion is a crucial factor in
terminating service life of steel bridges. Since many of the corrosion problems are found in the
girder end, it is essential to know the load-carrying capacity of the girder end for efficient main-
tenance. In “Load-carrying capacity of corroded end of I-section steel girder”, Yamaguchi
and Akagi investigate the influence of corrosion at end of steel girder. Steel I-section girders
with various corrosion models are constructed, placing a vertical load at the upper flange
right above the support, and the girders are analyzed by nonlinear FEM. The vertical forces
exerted by bridge decks at supports are not always downwards; uplift can occur for a variety
of reasons. Upward forces are generally transient, lasting no longer, for example, than the
duration of a strong wind or of live loading of the relevant section of the structure. Although
the structure’s bearings must typically still be capable of carrying downward forces, facilitat-
ing rotations, and resisting horizontal forces and/or accommodating displacements, they must
be able to do all of this even under uplift conditions. And if the frequency of load reversal is
high, then the uplift load condition may become fatigue-relevant, adding yet another dimen-
sion to the design—especially if the load reversals would cause hammering of interfaces such
as the sliding surfaces of a sliding bearing. In “Uplift bearings—selection and design consid-
erations”, Moor et al describe key issues which must be considered in selecting and designing
bearings for uplift conditions. The number of bridges designed and built on landslide regions
is quite small since highway roads are normally designed to eliminate destructive effects of
landslides. Therefore, in literature, solutions and design guidelines for bridges on landslide
areas are scarce. In “Design of Idemli Bridges in landslide areas”, Ozkaya et al present details
of design of a highway viaduct located in a landslide area in northern Turkey.
The Newburgh-Beacon Bridge, officially the Hamilton Fish Newburgh-Beacon Bridge, spans
the Hudson River carrying Interstate 84 in New York State between the Town of Newburgh in
Orange County to the west and the City of Beacon in Dutchess County to the east. The New-
burgh-Beacon Bridge is one of six bridges over the Hudson River run by the New York State
Bridge Authority. The bridge consists of two parallel spans with similar appearance consisting
of deck and cantilever trusses. The original north span was awarded the 1965 AISC “Most
Beautiful Bridge” award for long span bridges. The bridge is the world’s 21st longest cantilever
bridge and with 25 million crossings per year, it is the most travelled of all spans maintained
by the New York State Bridge Authority. Over 40,000 m2 of deck on the south span will be
replaced with precast, high performance, concrete filled, galvanized steel grid panels for the
entire 2,374-meter length of the structure. The deck replacement will be constructed in stages.
Over 1,900 panels, topped with a one-inch polyester concrete overlay will replace the existing
concrete deck. An innovative aspect of the design involves the use of steel haunches to support
the deck on the existing girders/stringers. The top flanges of the existing stringers will be spliced,
providing continuity at 28 relief joints. The project also includes the modifications of existing
finger joints, major electrical work, new sign structures, new drainage systems and 4,779 meters
of new precast concrete barriers. In “Precast concrete filled steel grid panels for deck replace-
ment at the Newburgh-Beacon Bridge—South Span”, Moreau and Abdou describe the bridge
and its history, the development of alternatives and selection of deck type, innovative details and
construction stages. The collapse of I-35 W Bridge in Minnesota on August 1, 2007, triggered a
mandate within and outside of the state for the inspection of bridges focusing on the state’s frac-
ture critical inventory. Following that bridge collapse, emergency inspections performed imme-
diately on two suspended arches bridges; the John A. Blatnik and Richard I. Bong Memorial
Bridges located between Duluth, Minnesota, and Superior, Wisconsin, revealed deterioration to
suspender ropes. In “Blatnik and Bong Bridges wire rope replacement: maintaining Minnesota’s

FM.indd ix 7/17/2013 5:08:23 PM

 x Preface

unique bridge inventory”, Amundson et al discuss the configuration and existing condition of
the suspended deck system, suspender ropes removal and replacement procedure, the results of
the testing, and the maintenance recommended to extend the service life of the two bridges. The
Alexander Hamilton Bridge and Highbridge Interchange Ramps rehabilitation/replacement
project includes a total of nine (9) existing bridge structures on or spanning the Cross Bronx
Expressway from Amsterdam Avenue in New York County to Undercliff Avenue in Bronx
County, New York City. The estimated traffic on the bridge is in excess of 200,000 vehicles per
day. The bridge has not undergone major repairs since it opened to traffic on January 15, 1963.
Currently, the bridge is undergoing a major rehabilitation project, with a construction cost tag
of $407 Million, making it the largest single-contract construction project in the history of the
New York State Department of Transportation. In “Rehabilitation of the Alexander Hamilton
Bridge and Highbridge Interchange Ramps”, Tughral Shaikh et al present the complexities in
the designs of the permanent and temporary structures to successfully complete the proposed
rehabilitation work for the heavily travelled Alexander Hamilton Bridge and its associated High-
bridge Interchange Ramps. The Walt Whitman Suspension Bridge spans the Delaware River
from Philadelphia, Pennsylvania to Gloucester, New Jersey. The bridge, which is named after
the poet Walt Whitman, opened to traffic on May 16, 1957, and is owned and operated by the
Delaware River Port Authority. Water leakage at deck relief joints caused heavy corrosion to
the stringers and the top chord of the floor trusses. The poor condition of the framing, coupled
with the advanced age and the condition of the existing deck, resulted in deck replacement
studies. Several options, including a steel orthotropic deck and a concrete-filled steel grid deck,
were studied. The grid deck option was selected for its lower initial construction cost and the
feasibility of making it jointless. To make the grid deck jointless in each span, it had to be either
rigidly connected to the stiffening trusses through deck shear keys along the length of the span,
or allowed to move freely, i.e. “float”, relative to the stiffening trusses. The “floating” deck option
was determined to be more economical and have better performance, thus advanced through
final design and construction. In “Jointless ‘floating’ deck for the Walt Whitman Bridge”, Prior
et al provide details of the decision making and evaluation process used to determine the chosen
deck alternative for the suspended spans. Franklin Square Structure is located in Manhattan
on the Brooklyn Bridge Manhattan Approach, spanning Pearl Street. The site is very historic,
with the east abutment once being the location of the first presidential mansion in which George
Washington, his family, and his household staff lived from April 23, 1789 to February 23, 1790.
The Franklin Square Structure is a truss bridge that utilizes pin connections with eyebar mem-
bers, and was recently retrofitted with orthotropic deck panels. In “Brooklyn Bridge orthotropic
deck installation”, Leo and Reynolds present the advantages and disadvantages of orthotropic
deck panels. During installation, the orthotropic deck panels were aligned before being welded in
the flat and overhead positions. Multiple alignment tolerances needed to be met, including: verti-
cal alignment, root openings, diaphragm offsets and panel elevations. Welding processes, which
were used, include submerged arc (SAW) and shielded metal arc (SMAW). Also, the air carbon
arc cutting process was used to achieve a U-groove on top of the panels. The work was per-
formed during overnight closures without disruption of daily rush hour and peak traffic flow.
Performance-Based Earthquake Engineering aims to quantify the seismic performance and
risk of engineered structures using metrics that are of immediate use to both engineers and
stakeholders. In “Case studies on several isolated bridge configurations adopting a performance
based approach”, Forcellini and Banfi aim to assess the effects of isolation devices on a bench-
mark bridge in reducing the column displacements and repair costs adopting various configu-
rations of abutments and pier connections. The data was derived to relate response to damage
of individual components within the structure and tied to explicit repair procedures that could
be used for cost estimation and repair effort necessary to return the bridge to its original func-
tionality. The contribution of the isolation technique is assessed evaluating the resistance effects
applied to Peak Ground Acceleration levels in terms of cost and time repair quantities with
particular attention to the bridge-abutment configuration. Also on the topic of seismic isola-
tion, Ozkaya et al provide “Assessment and design of seismically isolated bridges in proximity

FM.indd x 7/17/2013 5:08:23 PM

 Preface xi

of major faults and located on liquefiable soils.” The authors study seismic isolation design of a
highway viaduct and a high-speed railway viaduct located within few hundred meters of North
Anatolian fault in Turkey. For the highway viaduct, a hybrid seismic isolation system is chosen
for the bridge consisting of four natural rubber and two lead-rubber bearings supporting the
six steel-girders at each substructure. Lead cores in the lead-rubber bearings are designed to
have an overall characteristic strength equal to 10% of the superstructure weight. This large
characteristic strength is intentionally chosen to limit the isolator displacements under near-
field effects. Further, the supplementary four natural rubber bearings over each substructure
provided additional stiffness. For the railway viaduct, the same design philosophy was followed
with considering the serviceability issues related to high-speed railway bridges. Curved surface
slider bearings having a dynamic friction coefficient of 12% with 2.54 mm yield displacement are
selected. In recent years, curved highway bridges, which are more prone to seismic damage than
straight ones, have become an important component in modern highway systems, as the most
viable option at complicated interchanges or river crossings where geometric restrictions apply.
Curved structures may sustain severe seismic damage owing to rotation of the superstructure or
displacement toward the outside of the curve line. A commonly adopted earthquake protection
strategy consists of replacing the vulnerable conventional bearings with seismic isolation devices.
In “Seismic retrofitting of curved highway viaducts”, Moor et al describe, with reference to a
current project, how seismic isolation, in the form of Lead Rubber Bearings, can be retrofitted
to curved highway viaducts. Italy has a rich heritage with magnificent historic structures and old
bridges. A large number of Italian bridges were built over one hundred years ago. In “Retrofit
procedures of bridge networks against seismic actions in Italy”, Pipinato presents an overview of
practical issues on the assessment of bridge networks in order to evaluate the structural integrity
of historical and deteriorated bridges, incorporating code provisions, analytical, mechanical and
structural topics.
Most condition assessment of bridge structures relies on visual inspection. However, visual
assessment alone is insufficient in providing a comprehensive evaluation of the condition of
the bridge, and crucial conclusions regarding the proper time of intervention and maintenance
measures. In “Investigation of feasibility of deterministic, theoretical predictive models under-
pinning visual condition monitoring outcomes”, Sagar et al discuss the viability of theoreti-
cal predictive models in identifying the most appropriate time for maintenance intervention
in critical structural elements of bridge assets. The key objective is in providing deterministic,
theoretical underpinning to visually inspected condition data. The authors propose to achieve
that objective by conducting a real-time theoretical simulation of flexural stresses, taking into
account the age dependent behavior of shrinkage, creep (under sustained loads) and imposed
loads. In doing so, crack widths can be theoretically estimated over the life cycle of the bridge
girders. The paper proposes a Theoretical Age of Intervention (TAoI), based on crack depth
reaching a critical value, which may not be otherwise conceivable to human eye. The paper
demonstrates that such a theoretically predicted time of intervention, when calibrated using the
condition monitoring data, would provide asset owners with a useful tool for maintenance plan-
ning. Many reinforced concrete bridges require effective repair due to deterioration. In some
cases, there is a lack of basic information on the cross-section of the bridge. Ground Penetrating
Radar (GPR) technique is a nondestructive testing technique to obtain this information. GPR is
a high-resolution electromagnetic (EM) technique, which originated in the 1970s. It is designed
for non-destructive investigations of the shallow subsurface of soil, structural elements, roads,
and bridges. In “GPR testing of reinforced concrete viaduct beams”, Beben et al present GPR
tests to obtain evaluation parameters for reinforced concrete viaduct beams. The subject of this
study is a three-span RC road viaduct located over railway lines. Due to lack of design docu-
mentation for the existing viaduct, it was necessary to obtain basic information on the concrete
beams. The main purpose of the GPR tests was to determine the geometric parameters of beams
and location of reinforcing bars. The tests also provided distances between reinforcing bars and
depth of their position. Precision microwave instrument are used for noncontact vibrations and
displacements monitoring of various engineering structures. In “Monitoring of soil-steel

FM.indd xi 7/17/2013 5:08:24 PM

 xii Preface

railway bridge under service loads”, Beben and Anigacz present the results of experimental
tests under dynamic loads that were conducted on a soil-steel railway bridge. On the basis of the
measured displacements, use of the Frequency Domain Decomposition method determined the
frequencies of the bridge. The microwave interferometric radar and inductive gauge were used
for monitoring of displacements of this railway bridge. The main aim of this paper is the evalu-
ation of possibility of the radar application for the measurements of soil-steel bridge under
dynamic loads. The measurements were made for all trains, which had been running over the
bridge during a 24-hour period. The dynamic amplification factors from tests were in the range
of 1.10 to 1.375. The frequencies of bridge did not exceed 5.5 Hz. The maximum displacements
of this bridge did not exceed 0.65 × 10−3 m. Conclusions drawn from the tests can be helpful in
the measurements and assessment of dynamic behavior of such bridges.
Throughout history, there have been fascinating developments in bridge forms and con-
struction techniques ranging from the primitive rope bridges to Roman aqueducts and the
timber trestle railway bridges of the American West, to today’s modern signature bridges.
As symbols of connection of communities and people, landmark bridges are also icons of art
and engineering elegance. The construction of the Sydney Harbour Bridge had a consider-
able impact on the development of a seminal modern art movement in Australia. When the
harbour bridge was being built, the art produced represented the hopes and aspirations of a
fledgling nation, and reflected the construction methodology adopted. In “Artists’ portraits
of the Sydney Harbour Bridge”, Rothwell reviews selected art works inspired by the bridge.
The paper also describes the development of the modern steel arch bridge to illustrate the intel-
lectual and aesthetic progress that the Sydney Harbour Bridge represented. Bridges serve the
function of providing a safe path over natural or manmade obstacles. Throughout the centu-
ries, bridges have assumed a landmark status and have become symbols of art. This symbolic
role of bridges has drawn non-structural engineers to the realm of conceiving the bridge form.
This, in many cases, has compromised the bridge function by the use of irrational structural
forms. In “Measures of structural art: a case study using Streicker Bridge”, Sigurdardottir
and Glisic examine the three measures of structural art, introduced by the Scottish engineer
Thomas Telford in the 19th century; namely, the scientific, social, and symbolic aspects of the
structure, referring to structural efficiency, economy, and elegance, respectively. A case study
using Streicker Bridge on Princeton University Campus is presented. The structural system is
shown to be efficient and then the structural behavior is verified using Structural Health Moni-
toring based on fiber optic strain sensors and a Finite Element Model. Manhattan Bridge
was the third East River bridge planned in 1899 for construction, alongside the Brooklyn and
Williamsburg Bridges. On January 1, 1902 Seth Low became Mayor of New York City, and
appointed Gustav Lindenthal as Commissioner of the Bridge Department. Design plans of
the Manhattan Bridge with wire cables as the main supporting elements were already at an
advanced stage, and the tower foundations were practically completed. Lindenthal rejected
the old plans and presented a new design, which instead employed eyebar chains. There was
an immediate strong opposition to this move. The Board of Aldermen refused to fund this
construction scheme. When Seth Low lost reelection in 1904, Lindenthal also lost his job. The
bridge was built as a wire cable suspension bridge and opened to traffic on December 31,
1909. In “Lindenthal and the Manhattan Bridge eyebar chain controversy”, Gandhi covers the
controversy between eyebar chain and wire cable, and presents his own view of Lindenthal’s
performance as Commissioner.
The archival quality of this volume is due mainly to the efforts of authors and reviewers of
papers, to all of whom the editor is grateful.

Khaled M. Mahmoud, PhD, PE

Chairman of Bridge Engineering Association (BEA)
President of Bridge Technology Consulting (BTC)
New York City, USA

New York City, August 2013

FM.indd xii 7/17/2013 5:08:24 PM

 1 Cable-supported bridges

NYBRIDGE_Book.indb 1 7/17/2013 8:41:18 AM

This page intentionally left blank
 Chapter 1

Main cable anchorages investigation at Forth Road Bridge

B.R. Colford
Forth Road Bridge, Scotland

ABSTRACT: This paper describes the investigation work carried out to try to determine
the remaining service life of the main cable anchorages at Forth Road Bridge. The anchor-
ages on any suspension bridge are critical components and to be able to inspect these com-
ponents to evaluate deterioration and remaining strength is vital. At Forth Road Bridge, the
anchorages were constructed, on both the north and south sides, by forming tunnels within
the rock formations. Concrete has been cast within these tunnels to form four large, individ-
ual, concrete plugs within the rock to which the cables are attached. The concrete in the tun-
nels is reinforced using pre-tensioned, galvanized, high tensile steel strands in grouted ducts.
The strands are completely inaccessible and in a hostile environment. Concerns over the
possible loss of galvanizing were expressed during construction and further evidence came
to light in 2005 that the condition of the strands may be worse than previously reported.
Given all the above, an investigation was carried out to try to determine with some certainty
the long term structural integrity of the anchorages and also to establish that the risk of
failure remains low and within acceptable levels. The investigation was considered to be a
pro-active measure to ensure that all parts of the structure that can be reasonably accessed
are inspected.

1 INTRODUCTION

1.1 Background and location

The Forth Road Bridge is a long span suspension bridge with a main span of 1006 metres and
spans each 408 metres long, Figure 1. The bridge crosses the Firth of Forth some 15 km west
of Edinburgh, the capital city of Scotland, and is a vital link in the country’s strategic road
network, Figure 2 (Institution of Civil Engineers 1965). The bridge was opened to traffic on
4 September 1964 and over 25 million vehicles now cross each year. The bridge deck supports
a dual two lane carriageway without hard shoulders or strips and there is a separate footway/
cycletrack on either side.
The historic importance of the structure to Scotland was recognised in 2001 when the
bridge was classed as a Category A listed structure.
Major strengthening works and remedial, replacement and improvement works have
been carried out over the past 20 years on the bridge (Andrew & Colford 2006), (Colford &
Cocksedge 2006), (Colford 2008), (Colford et al., 2009), (Colford & Clark 2010), (Colford
2011). In 2004, corrosion was found in the wires forming the main cables following an inter-
nal inspection and dehumidification and acoustic monitoring have been retrofitted to the
cables. The main cable anchorages are formed using concrete tunnels within rock. These
tunnels were pre-tensioned using galvanised steel strands within grouted ducts, and no access
to inspect, maintain or effectively monitor this system of pre tensioning within a hostile envi-
ronment was allowed for in the design.

NYBRIDGE_Book.indb 3 7/17/2013 8:41:18 AM

 4 Cable-supported bridges

Figure 1. Forth Road Bridge.

Figure 2. Location of the Forth Road Bridge.

2 GENERAL DESCRIPTION OF THE ANCHORAGES

The main cable anchorages secure the ends of the main cables of the bridge on the north and
south shores of the river, Figure 3. This is achieved by transferring the load from the cables to
rock strata through four tapered tunnels cut into the rock at an angle of 30° to the horizontal. On
the north shore competent rock is present near the surface of the ground and the rock tunnels
extend down for a length of approximately 57 metres. On the south shore rock is overlain with
soil deposits and the rock as a whole is less competent. Consequently, the tunnels of the south
anchorages were lengthened and are over 79 metres long. Each tunnel is infilled with concrete.
Most of the length of each tunnel is below the water table in a highly saline environment.
The ground conditions as indicated above are very different on the north and south
sides of the river. At the north anchorages the rocks in which the anchorages are formed

NYBRIDGE_Book.indb 4 7/17/2013 8:41:18 AM

 Main cable anchorages investigation at Forth Road Bridge 5

are part of the Permo-Carboniferous, intrusive volcanic Quartz Dolerite, which is intruded
into rocks of the Upper Oil Shale Group, including sandstones, siltstones, mudstones, shales
and limestones. However, the south anchorages are formed in rocks, which are part of the
Carboniferous, Upper Oil Shale Group, consisting mainly of mudstone, sandstone and
marine and freshwater limestone.
The connection of the main cable to the concrete anchor tunnels is made within the anchor-
age chamber. The main cable, comprising 11618 wires in 37 bundles, is split into 37 separate
strands at the splay saddle located just within the chamber. Each of the 37 strands then loops
round a strand shoe. The strand shoes are then connected in pairs, by 89 mm diameter tie
rods to mild steel crosshead slabs, two rods per shoe. The cross head slabs are clamped to the
tunnel infill concrete by 114 pre-tensioned tendons, six tendons per slab, Figure 4.
These 114 pre-tensioned tendons anchor the crosshead slabs and transfer the load from the
main cable to the concrete infill and as such their condition is critical to continued service-
ability of the anchorages. The six tendons per crosshead slab each consists of four 31.75 mm
wire strands each made from 19 wires, Figure 5. The total cross sectional area of the strand
is 625.61 mm2. Based on an ultimate tensile strength of 1545 N/mm2 this gives an ultimate
tensile capacity of 386.7 tonnes per tendon. This capacity assumes no reduction in the capac-
ity of the tendons due to the bends, which are present in the ducts. Based on this force the
cross head slab will have an ultimate capacity of approximately 2320 tonnes. Each tendon

Figure 3. Main cable anchorages (North Side). Figure 4. Cross head slabs.

Figure 5. Anchorage strands.

NYBRIDGE_Book.indb 5 7/17/2013 8:41:20 AM

 6 Cable-supported bridges

Figure 6. Original as built drawing southwest anchorage tunnel.

was tensioned to 152 tonnes, 912 tonnes in total on each crosshead slab, and this load is 25%
greater than the load applied from the main cable. Therefore, the load in the tendons will be
constant and will not be subject to stress changes due to changes in the main cable tension
and thus will be unlikely to suffer from fatigue failures. Loss of pre-tension may affect this
though given the low percentage that live load represents in the loading on the cable, the
stress ranges, if any, will be small and will have minimal influence on the strands.
The tendons were installed in 102 mm internal diameter mild steel ducts which were grouted
following the tensioning of the tendons. Hydraulic jacks applied at the lower end of the ducts
were used to tension the strands. Access to the bottom of the tunnel shafts was formed by
excavating a vertical shaft, which led to tensioning galleries located to the rear of the tapered
tunnels. These shafts and galleries were infilled with rubble and concrete on completion of
the works, Figure 6. In both north and south anchorages the last 3.65 m of ducting at the
lower end of the rock tunnel is 305 mm diameter, the duct being increased in size to allow
the strands forming the tendon to be splayed out for connection to the tensioning jacks and
for anchoring.

3 DESIGN & CONSTRUCTION INFORMATION

3.1 Available information

Concerns over the long term structural integrity of the pretensioning strands arose from an
examination of the design and construction information. These were:
• The tendons are formed from four independent strands, which are grouped and bound
together with spacers located at 3.05 metre centres. The combination of the multiple ropes
and the spacers may have resulted in voids in the grout placed within the ducts.
• The ducts are not straight, with bends occurring near the anchorage chamber and at the
end anchorage in the rock socket. The main cable is splayed out in the anchorage cham-
ber and the bends in the ducts are required to bring the tendons parallel. At the ten-
sioning location in the rock tunnel the strands forming the tendons require to be splayed
out in order that the tensioning jacks could be connected. Bends in the ducts could have
adversely affect the ability to fully grout the ducts and also introduce additional forces into
the strands, particularly at the lower end of the tendon where reinforcement was provided
in the concrete to resist the bursting forces, which would be induced.

NYBRIDGE_Book.indb 6 7/17/2013 8:41:21 AM

 Main cable anchorages investigation at Forth Road Bridge 7

• In the anchorage chambers, the tendons are socketed with all four strands being socketed
in a single steel casting. The individual wires are fixed using zinc and from experience this
raises a number of issues. The forming of the wire brush within the socket will be difficult.
With a single strand this can be achieved relatively easily with the wires of the rope being
removed from the lay and opened out to ensure the socketing metal fully bonds each wire.
A single strand can also be more easily held straight while the zinc is placed, thus minimis-
ing any potential bending stresses due to misalignment of the strand and the socket. With
multiple strands in a single socket this becomes significantly more difficult. It may have
been difficult to open out the wires of the strands effectively within the socket and avoid
misalignment.
• Another issue associated with the socketing is that when placing the hot metal the gal-
vanising protection to the wires will become molten and any corrosion inhibitors which
are provided just beyond the socket will be lost due to the heat associated with the mol-
ten white metal. This could lead to corrosion of the wires, which if associated with high
stresses can lead to the failure of the strands.
• The wires are anchored at the end of the rock tunnel by wedges, which were driven to
wedge and hold the strands once they were tensioned. Reports from the time of the con-
struction relating to water ingress suggest that these may be corroded and could have lost
their fixity or, alternatively, they may have tightened, damaging the wires.

4 CONSTRUCTION ISSUES

4.1 Excavation
During excavation of the south anchorage tunnels, the rock was noted to vary from soft shale to
sound sandstone. During excavation of the shale (carried out by controlled blasting and mechan-
ical removal), several gas issues from the rock mass were noted at below the Lower Explosive
Limit (LEL) for methane of 5%. However, while concreting of one of the rear working cham-
bers an explosion did occurr, with a second explosion occurring after 36 hours. To ensure that
all burning of methane was extinguished the shafts and galleries were allowed to fill with water.
This was subsequently pumped out and the shafts in-filled with rubble and concrete.
Shortly after completion it was noted that there was a leakage of gas from the cable sock-
ets which was subsequently found to have a high percentage of hydrogen. Investigations and
testing suggested that the cause was likely to be groundwater seeping through the interface
between the anchor block and the working chamber concrete and was thought to be attack-
ing the galvanising of the wires. It was considered likely that one of the products of this reac-
tion would choke any voids in the cable ducts such that the reaction would be self limiting.
Gas emission ceased within 12 months.

4.2 Hydrogen leakage

From the construction report on the southern anchorages it is apparent that there was a sig-
nificant problem with water ingress into the ducts even after they were grouted. The presence
of this water, which will be seawater, would have a detrimental effect on the galvanising on the
wires. This may have led to corrosion of the wires with associated pitting. Should this have
occurred there would be a potential for sudden failure of a strand and/or tendon.
With the tendons being enclosed in steel ducts they are not in direct contact along their
length with the concrete infill. This suggests that the engineering design anticipated that
the entire load would be applied to the embedded end of the tendons and, therefore, the
intention of the grouting of the ducts was solely to prevent water ingress and not to provide
anchorage for the tendons. This may not be the case in reality as, although the duct does form
a point of discontinuity, there will be a bond between the duct, grout and strands. The extent
to which load will be transferred by means of this bond is difficult to determine.

NYBRIDGE_Book.indb 7 7/17/2013 8:41:21 AM

 8 Cable-supported bridges

5 INVESTIGATION INTO THE CONDITION OF THE MAIN CABLE

ANCHORAGES

5.1 New information

Previously unseen records and papers from the office of the Resident Engineer were acquired
by the Authority in 2005. These related to the construction of the existing anchorages and
highlighted the various problems encountered during construction particularly in relation to
early depletion of the galvanising protecting the pre-tensioning strands.
These documents on the issues and conditions encountered during construction described
the anchorages, especially on the south side, as being in a potentially worse condition than
had been described in the published Institution of Civil Engineers Proceedings Paper. This
obviously caused further concerns and it was decided that there was a need to carry out a spe-
cial inspection or investigation to try to establish the existing condition of the pre-tensioning
strands.
Routine inspections were being carried out to monitor for movement within the anchorage
chambers at the tunnel/strand shoe interface and no signs of distress or movement had ever
been recorded. However, the new information obtained reinforced the view that an investiga-
tion was essential to try to determine with some certainty the long term structural integrity
of the anchorages and establish that the risk of failure remained low and within acceptable
levels. The investigation was also seen as a pro-active measure to ensure that all accessible
parts of the structure were inspected.

5.2 Appointment of Engineers and Engineering Peer Review Panel

As part of the feasibility study on the options for the replacement or augmentation of the
main cables on Forth Road Bridge, the Glasgow based Consulting Civil and Structural Engi-
neer, W A Fairhurst & Partners, were asked to carry out an investigation into the condition
and long term structural integrity of the main cable anchorages.
A Peer Review Panel was also engaged to audit and review the work being carried out
by Fairhurst. The panel members were David Mackenzie of Flint and Neill; Peter Sluszka
of the NY consultants Ammann & Whitney and Bill Valentine (now retired), Chief Bridge
Engineer, Trunk Roads Network Management, Transport Scotland.

5.3 UK guidance on inspection of post tensioned bridges

There is guidance from the UK Department of Transport for inspecting post tensioning in
bridges as it is acknowledged that there can be problems with this type of construction. The
guidelines refer mainly to the difficulties in establishing the condition of post tensioning
strand in bridge decks, which are of similar structural form and construction as the pre-
tensioned strands in the anchorage tunnels. However, these difficulties are exacerbated in a
tunnel and there is no guidance given on the inspection of anchorages.

5.4 Engineering recommendations

Following much discussion, research and consultation with various specialists, Fairhurst
concluded that three separate methods of investigation should be taken forward: These
were:
• excavation behind the anchorage chambers down to the top of the tunnel to expose,
inspect and test the pre-tensioning strands. The location of the excavation was fortunately
right within the tidal zone of the Forth. This would allow a visual inspection of the actual
strands within the likely worst environment for corrosion. Figure 7.
• the full scale load testing of a number of the sockets within the anchorage chamber.

NYBRIDGE_Book.indb 8 7/17/2013 8:41:22 AM

 Main cable anchorages investigation at Forth Road Bridge 9

Figure 7. Excavation—overburden, rock & concrete.

• a number of other methods of non destructive testing. Acoustic monitoring and the use of
radar were considered but these methods did not appear to show much potential in solv-
ing the problem alone. However, it was thought that they may provide some useful data
when combined with the first two methods. The use of radiography and magnetostriction
scanning were proposed if the ducts and strands could be successfully exposed following
excavation.

5.5 Peer Review Panel comments

The Peer Review Panel reviewed Fairhurst’s proposals and concluded that the investigatory
work should go ahead to allow some form of condition assessment to be carried out on the
anchorage strands. The main findings of the panel were as follows:
• The likelihood of significant corrosion in the anchorage strands was considered to be very
low, but not discountable. It is for that reason that the Panel concurred with the Author-
ity’s decision to carry out an intrusive inspection of the anchorages. At this stage, the
inspections were scheduled for only the south anchorage for reasons that are supported by
the contemporary records during and after construction. This decision is also supported
by the fact the south anchorage was constructed in much poorer ground conditions con-
sisting of raised beach deposits and shale that is subject to inter-tidal ground water move-
ment as compared to the hard igneous rock found at the north anchorage where there is no
evidence of water ingress. Furthermore, more hydrogen was reported as being issues from
the south anchorage. These inspections would give the Authority sufficient confidence at
this stage to either rule out the issue of corrosion altogether, or to consider suitable reme-
dial actions.
• If the intrusive inspection goes forward, it is recommended that a monitoring system
be installed to detect future corrosion of the inspected strands. One possible system is
the linear polarization resistor sensor, designed for remote reading of corrosion activity.
These sensors could be placed inside the reconstructed tendon ducts prior to backfilling
the excavation.
• Cutting of strands for sampling and testing should only be considered as a last resort.
• The full scale load testing of the anchorage sockets is not recommended to go ahead.
• The do nothing case was not considered to be a viable option. If a strand did fail due to
corrosion it would have to be accepted that there would be a strong possibility that further
strand failures would occur. There could be damage to the main cable within the anchor-
age due to the possibility of a socket becoming a projectile because of the release of strain
energy. Critically, there would be no means of predicting the timing of the next failure.
An investigation similar to the one being planned would have to be conducted, with the
additional pressure of further failures and the requirement that new anchorages may need

NYBRIDGE_Book.indb 9 7/17/2013 8:41:22 AM

 10 Cable-supported bridges

to be built. Without the means of predicting the safe capacity of the bridge there would be
no option but to close to traffic whilst the investigation was being carried out.
• Carrying out the investigation now, gives the Authority the ability to carry out a proactive
assessment of the safety of the bridge rather than having to react in an ad-hoc manner.
• It must be stressed that there is no current and direct evidence that the bridge is in any way
unsafe. Rather that this investigation is the only effective way to deal with concerns over
the longer term integrity of the anchorage strands.

5.6 The way forward

The comments and recommendations of the Peer Review Panel were accepted. The full scale
load testing was not taken forward but work on preparing tender documents for an inves-
tigation into the condition of the main cable anchorages on the south side of the bridge
was commenced. The main thrust of the work would consist of excavating down behind the
southwest and southeast anchorage chambers to expose the top of both the west and east
tunnels. The concrete forming the tunnels would be carefully removed using hydrodemolition
in order to expose the nine ducts lying at the crown of each tunnel, Figure 8.
All 18 ducts would then be carefully opened using a milling or precision cutting process,
circumferentially, in two places, about 10 metres apart. Then a longitudinal cut would be
made between the two circumferential cuts and the duct carefully split open by hand. The
grout would be washed out and then, depending on the condition of the wires, a decision
would be taken on the need to cut wires for sampling and testing.

5.7 Potential outcomes from the investigation

From the investigation, an assessment would be made to provide a determination of the
likely current strength of the anchorages and the condition survey would assist in the difficult
task of estimating future strength.
Depending on the findings of the investigation, it was recognised that challenging engi-
neering decisions may have to be taken. However, as all strand sockets were firmly in bearing
with the anchorage plates, this was taken to indicate that there has been no complete loss of
pre-tensioning in any one tendon but, as confirmed by the audit, significant corrosion could
not be discounted.

Figure 8. Exposure of tendons at crown of tunnel.

NYBRIDGE_Book.indb 10 7/17/2013 8:41:22 AM

 Main cable anchorages investigation at Forth Road Bridge 11

It was also recognised that if the investigation did reveal significant deterioration in the
steel strands within the anchorage tunnels then, depending on the level of that deterioration,
measures may have had to be considered to limit loading on the bridge.
Contingency planning based on a number of possible scenarios was developed prior to the
inspection to ensure that all foreseeable events are taken into account. This work involved
discussion and consultation with other agencies and stakeholders.
To assist in this planning, Fairhurst recommended that Statistical Condition and Strength
Models of the anchorages were developed with the assistance of the University of Strathclyde
in Glasgow, prior to the inspection work taking place on site. These models would be used to
assist in determining the likelihood of the condition and strength of the anchorages being at
a certain level based on the information obtained from the inspection.
Any further work, including extending the investigation to the north anchorages, would
depend on the results of this examination.

5.8 Expected difficulties and safety

The excavation was expected to be difficult as the ground conditions vary and there is meth-
ane present within the shale. The work was further complicated by the close proximity of the
foundations to the viaduct piers, Figure 6. The safety of the workforce, the bridge and users
would govern all aspects of this complex project.

5.9 Traffic and community issues

Work on the anchorages was not anticipated to involve major disruption to bridge traffic.
However, given the nature of the works there would be significant environmental issues to
be dealt with especially with regard to noise, dust and discharge from the excavation and
hydrodemolition. To assist the community it was a requirement of the contract that that the
contractor appointed employed a named member of staff as community liaison officer.

6 INVESTIGATION WORK ON SITE

6.1 Tender award

Following a tendering exercise, a bid submitted by John Graham (Dromore) Ltd for the sum
of £3,497,849.00 to carry out the anchorage investigation works at the southern anchorages
only was accepted by the Authority in April 2011.

6.2 Extent of contract

The contract included the excavation down through overburden and rock to the top of both
tunnels to expose the concrete forming the tunnels over a length of around ten metres. The
maximum depth of excavation was also around ten metres. This work was to be carried
out in a carefully controlled manner utilising a combination of mechanical excavation and
hydrodemolition techniques. Side slopes were strengthened by subsurface soil nailing and
rock anchors. Further careful hydrodemolition was to be used to remove the tunnel concrete
to expose the 18 steel ducts housing the pre-tensioning strand. The opening of the ducts and
removal of the grout washed out to expose the strands to allow inspection, testing and moni-
toring was also included within the contract.

6.3 Commencement of work on site and unforeseen difficulties

The investigation of the anchorages project commenced on site in August 2011 and has not
involved any disruption to bridge traffic. It was recognised that by the very investigatory

NYBRIDGE_Book.indb 11 7/17/2013 8:41:23 AM

 12 Cable-supported bridges

nature of this project that the estimate of cost was always going to be difficult to determine.
Due to the unforeseen nature of the ground conditions encountered, which included a large
amount of unrecorded mass and reinforced concrete, and the location of the ducts being
deeper and lying at a steeper angle than shown on the record drawings, the contractor has
had to supply additional resources to carry out the works.
The original proposal was that the anchorage tunnel would be excavated to a level in order
to maintain a minimum of 500 mm cover over the ducts. As the condition of the strands
was unknown, and it was considered that there could be significant deterioration within the
strands, the construction of permanent inspection chambers was included in the contract to
allow for future monitoring. Additionally, the chambers would also have provided a more con-
trolled environment for the current investigation work. Figure 9 shows the anchorage site.
When reviewing the work to remove the existing tunnel concrete, the contractor proposed
that, both on safety and economic grounds, all concrete was removed from around the ducts
prior to the construction of the chambers. Temporary protection was proposed to cover the
ducts prior to the chambers being constructed.
This early exposure of the ducts, Figure 10, offered the opportunity to view the strands
a lot earlier than had been programmed, albeit over a more limited length, and in October
2012, all the ducts were exposed at the crown of both tunnels.

6.4 Results from visual inspection

The external surfaces of all the exposed ducts were found to be in very good condition, Fig-
ure 11. The ducts are not galvanised and are reliant on the alkaline nature of the concrete
to protect the steel. There were welded joints within the ducts and these were fabricated to a
good standard. In addition, bitumen strips were found around the welds. These findings were
taken to be indicative that the ducts were well protected by the tunnel concrete (which was of
very good quality) and that the anchorages were not allowing the ingress of water. Once the
ducts were exposed, radiography was carried out along their full length and circumference in
order to locate the strands, to allow cutting to be carried out without damaging wires, and to
locate any voids within the grout. Although expensive and time consuming, the radiography
proved to be very worthwhile.
In November 2012, the first ducts were cut open and the grout was found to be in remark-
ably good condition and very difficult to remove using hand tools. After carefully removing
the grout using hydrodemolition, the strands, where exposed, were also found to be in remark-
ably good condition with no visible signs of corrosion of the wires, Figures 12 & 13. Where
access permitted, the diameter of the strands was checked and no evidence of a change in
diameter was found suggesting no internal corrosion had occurred. Careful wedging using

Figure 9. Anchorage site. Figure 10. First ducts exposed.

NYBRIDGE_Book.indb 12 7/17/2013 8:41:23 AM

 Main cable anchorages investigation at Forth Road Bridge 13

Figure 11. All ducts exposed in southwest tunnel. Figure 12. Stands exposed.

Figure 13. Strands wedged. Figure 14. Tendon with misaligned strands.

hand tools and hardwood wedges was used to separate the strands to enable a visual inspec-
tion to be made around the circumference of the strands. The dull grey colour of the galva-
nising on the surface of the strands indicated oxidation of the zinc coating had occurred and
localised darker areas on the strands were noted suggesting that these areas had oxidised to a
greater degree. However, overall the strands that were inspected were in good condition with
no evidence to suggest that they had deteriorated from the time they had been grouted.
On one of the ducts, the strands were found to be misaligned and on one strand the lay of
the outer wires was out of alignment with one wire bulging up clear of the rest of the strand,
Figure 14. The strands all appear to be loaded and within approximately one metre of the
top circumferential cut were realigned. Inspection revealed that the displaced wire and the
strands were fully grouted with grout visible between wires and strands. This indicates that
the damage occurred during installation of the strand and prior to the grouting operation.
A temporary acoustic monitoring system had been installed within the anchorage cham-
bers on some of the strand sockets and anchor plates to monitor for activity during the
investigation. Only one event was recorded during the 18 months of the investigation work
and that was outwith the ducts being opened. Whether or not that event represented a wire
break could not be confirmed.
Nine ducts out of a total of 114 were opened in each tunnel of the southern anchorage. That
is around 8% of the pre-tensioning ducts were opened during this investigation. In comparison
only 4% of the main cable panels were opened during the 2012 main cable internal inspection
and from the visual inspection, the condition of the wires in the anchorage strands was signifi-
cantly better than the condition of the wires inspected in the worst panels in the main cables.

NYBRIDGE_Book.indb 13 7/17/2013 8:41:24 AM

 14 Cable-supported bridges

7 REVIEW OF INVESTIGATION WORK ON SITE

7.1 Hold point established

By the end of January 2013, all the strands within the 18 ducts were exposed and all were
found to be in a similarly good condition. Given the good condition of all the strands exposed,
a hold point was established to determine whether or not further investigation work was going
to provide significantly different information, balanced against the further removal of ducts
and grout which appears to be providing very good protection to the strands. The further work
that was considered by Fairhurst, the Peer Review Team and the Authority was as follows:
• Sampling of strands/wires
• Further radiography
• Magnetostriction scanning
• Load testing
There was agreement between the Fairhurst, Flint and Neill and the Authority that, given
the unexpectedly good condition of the strands, sampling of strands/ wires; further radiog-
raphy and magnetostriction scanning were no longer required.
However, Fairhurst considered that it would be worthwhile undertaking load test to try to
establish measure of the load in the strand although these tests would be limited in extent.
The Peer Review Team comments on the proposed load testing were as follows:
• The proposed test would be very difficult to carry out and probably inaccurate without
calibration of the equipment in a laboratory with a strand of the same age and construc-
tion and would risk breaking the bond with the grout over an unknown length.
• The tests would provide little benefit to the assessment of the overall strength of the
anchorage and would simply affirm the initial pre-tensioning force in the strands.
• As there is no evidence of movement in any of the anchorage plates within the anchorages,
there is no reason to suspect that any plate was not properly stressed.
• Load testing would only show that the force in a strand may have reduced; any loss of
section would not result in a reduction in load, but would give rise to an increase in stress
commensurate with the loss of area.

7.2 Limitations of the investigation

The investigation has been limited in nature, and it was not possible to access the wedging area
at the bottom of the anchorage tunnels. In addition, the possible source of hydrogen emission
is still being considered. However, in the absence of any significant loss of galvanising on the
strands, it has been postulated by the Peer Review Team that the hydrogen gas and corrosion
products observed at the time of construction arose from the corrosion of the zinc metal at the
boundary of the socket and zinc socketing material. This interface would not necessarily have
been blocked completely after casting of the zinc within the socket as the zinc shrinks slightly
when cooling. During construction the anchorage tubes were open to water prior to grouting
and with any small gaps in the socket head a path for corrosion was available. However, as has
been witnessed from the anchorage excavations, the grout was effective in blocking the strands
and there is no evidence of continuing corrosion at the anchors. Any gaps between the zinc
and the socket would be plugged by the corrosion products. This is not considered to create
vulnerability in the structure as the mass of zinc in the socket is very substantial.
It was recognised that this was just a postulate, but one that is supported by the data and
the recent inspections, something that could not have been done without the excavation. The
only further additional investigations that could be carried out are the wholesale destructive
testing of the anchorage around the anchorage plates. The cost of this would be prohibitive
and given the lack of evidence currently available as to which anchorages were observed to
be bleeding off hydrogen gas, is as likely to imperil a sound anchorage as to discover evidence
of corrosion in another.

NYBRIDGE_Book.indb 14 7/17/2013 8:41:27 AM

 Main cable anchorages investigation at Forth Road Bridge 15

7.3 Decision to conclude the investigation

As a consequence of all of the above, after careful consideration, it was considered that
enough evidence of the condition of the anchorage strands had been gathered during the
investigation to enable a conclusion to be reached that the anchorages on the southern
bridgehead were in a satisfactory condition.
Based on the information available from the investigation, it was also considered that the
risk of the structural integrity of the southern anchorages being compromised had been
reduced significantly as a result of these findings.
In addition, given the history of the construction of the anchorages both north and south,
and the better conditions on the north side, it was decided that there was no need to carry out
any investigation work at the north anchorages.
It was concluded from the information obtained during the investigation that the risk of hav-
ing to replace the anchorages during the remaining service life of the bridge was relatively low.

7.4 Reinstatement of site

Decisions have had to be made about the reinstatement of the anchorages and two options
were selected for consideration. These were:
• Reinstate Ducts and Grout followed by Sequenced Mass Concrete Infill:
Reinstatement of the ducts and the grout under pressure, and re-concreting of the anchor-
age tunnel, replicates the existing protection, which has been shown to maintain the strands
in a good condition for 50 years. There would be no future access to the strands.
• Construct Limited Chambers over exposed Length of Ducts:
One of the key considerations when the chambers were first proposed at the design stage
was that the strands were likely to be in a relatively poor condition and would require future
inspection and monitoring. good condition of all the strands exposed during the investiga-
tion raised questions over the need to build these chambers. The chambers add significant
cost not only during construction but also to the future operation and maintenance.
Fairhurst consider that limited chambers with dehumidification would provide the best
protection to the exposed strands.
The Peer Review Team reported that they considered that the inspection chambers were
no longer required as their purpose was to provide the means to monitor a damaged strand,
one that might need intervention at a later date. The Peer Review Team considered that
the exposed strands had demonstrated the value of the corrosion protection system in the
inspected region. They reported that it had been extant for almost 50 years and there was
no reason to suppose that it would corrode for a similar length of time, assuming that the
corrosion protection system could be reinstated. They also considered that the proposal to
reinstate the grout under pressure should provide the necessary protection to the strand, fol-
lowed by controlled re-concreting of the anchorage shaft locally.
After some consideration, the view was taken that given the good condition of all the
strands exposed, there was now no real need to provide future access to the strands and that
as the existing system of grout and ducts appears to have provided excellent protection for
over 50 years then that system should be replicated for reinstatement.
Therefore, in April 2013, the Contractor was instructed to proceed with reinstating the
ducts and grout followed by sequenced mass infill concrete and works are expected to be
completed in late summer 2013.

7.5 Future work and monitoring

Once the concrete has been reinstated, there will be a programme of monitoring the anchorages
for any signs of corrosion emanating from the sockets or anchorage plates along with checks for
movement of the anchorage plates. The purpose of these is to determine if there is any evidence

NYBRIDGE_Book.indb 15 7/17/2013 8:41:27 AM

 16 Cable-supported bridges

of corrosion or if there has been a failure of one of the strands restraining the anchorage plates.
It was considered prudent to continue monitoring of the anchorages to inspect for products of
corrosion and to inspect for any evidence of movement of the anchorage plates. Both of these
could provide evidence of any on going corrosion, the latter being indicative of a potential loss
of capacity of the anchorages that would require a reduction in the load carrying capacity of the
bridge. Whereas it is considered that the risk of damage to the bridge is low, this monitoring will
provide a low cost method of gaining assurance that the anchorages are not deteriorating.
FETA’s own staff will continue to inspect and monitor the anchorage sockets and plates
within the existing anchorage chambers, on the north and south sides, as part of the on going
inspection of the bridge. The frequency of those inspections will reflect the criticality and
vulnerability of the anchorages and will take account of the investigation results.
Some further statistical analysis of the results of the investigation will be carried out by
Fairhurst assisted by staff from the University of Strathclyde have set up of a statistical
model for the investigation.

8 CONCLUSIONS

As the owners and operators of suspension bridges are aware, the main cable anchorages are
critical parts of the suspension system. The inspection and determination of possible loss of
strength of these anchorages is particularly difficult to determine if the anchorages are buried.
When the anchorages at Forth Road Bridge were designed, they were described as innova-
tive. The use of pre tensioned strand within a concrete tunnel anchorage in a corrosive envi-
ronment was certainly unusual, however, there seems to have been little thought towards the
future inspection of these vital structural elements.
Given the information made available in 2005 about the difficulties in constructing the
anchorages at Forth, it was considered that some means must be found to determine their cur-
rent condition. Various methods of NDT from acoustic monitoring to radar were examined
and found not to be suitable. In the end the decision was taken to expose the pre-tensioning
strands behind the anchorage chambers by excavating down from ground level.
The excavation and exposure of the ducts was a challenging but relatively straightforward,
though expensive, civil engineering project. The duct cutting, exposure of the strands and the
decisions taken on the extent of the investigation were perhaps more complex as there were
many issues to be considered.
These investigations often only allow a small sample of a bridge component to be inspected
and significant engineering decisions have to be taken based on that sample.
In this case, the consistent good condition of all the strands examined gave sufficient con-
fidence that the anchorages at Forth Road Bridge have not deteriorated and that the risk of
having to replace the anchorages within the service life of the bridge is low.

REFERENCES

Andrew A.A.S. & Colford B.R. 2006. Forth Road Bridge—Maintenance challenges, Fifth International
Cable Support Bridge Operators Conference New York City.
Colford B.R. 2008. The Forth Road Bridge—Maintenance and Remedial Works, Institution of Civil
Engineers, Bridge Engineering.
Colford B.R. 2011. Forth Road Bridge Main Cables and Anchorages, IABSE-IASS Symposium, London.
Colford B.R. & Clark C.A. 2010. Forth Road Bridge Main Cables. Replacement/augmentation study,
Institution of Civil Engineers, Bridge Engineering.
Colford B.R. & Cocksedge C.P.E. 2006. Forth Road Bridge—First internal inspection, strength evalu-
ation, acoustic monitoring and dehumidification of the main cables, Fifth International Cable Sup-
ported Bridge Operators Conference New York City.
Colford B.R., Jones S., Timby D. & Brown K. 2009. The maintenance of the main expansion joints on
the Forth Road Bridge, Fifth New York City Bridge Conference.
The Forth Road Bridge, 1965. The Proceedings of the Institution of Civil Engineers.

NYBRIDGE_Book.indb 16 7/17/2013 8:41:27 AM

 Chapter 2

Fabrication and construction of self anchored San Francisco

Oakland Bay Suspension Bridge

M. Nader, G. Baker & J. Duxbury

T.Y. Lin International, San Francisco, CA, USA

B. Maroney
California Department of Transportation, Sacramento, CA, USA

ABSTRACT: The signature span of the new San Francisco Oakland Bay Bridge is an
asymmetric Self Anchored Suspension (SAS) Bridge, spanning 565 m. The superstructure
hangs from an inclined cable system supported on a single four-shaft tower. Unlike a tradi-
tional suspension bridge with parallel ground-anchored cables, the SAS has a single 3-dimen-
sional cable that is anchored into the east end of the box girders, and loops over the tower
top and around saddles at the west pier. Box girders carry the cable tension; therefore a
temporary bridge supported them during erection, and throughout the cable construction.
The tower top, box girders, bearings and cable bands were advance positioned to fit the final
geometry. Geometric control during fabrication and construction monitored tower plumb-
ness, girder cambers, cable band twists, and suspender inclinations. This paper discusses the
challenges faced and the innovative solutions that were developed to construct this unique
structure.

1 INTRODUCTION

The seismically vulnerable East Span of the San Francisco-Oakland Bay Bridge is being
replaced with a dual east bound and west bound 3.6 km long parallel roadway structure. The
signature span consists of a self-anchored suspension bridge (SAS), situated between the
Skyway and the Yerba Buena Island (YBI) Transition. Some of the resulting unusual features
of this structure presented challenges to the Design Joint Venture (T.Y. LIN, International/
Moffatt & Nichol Engineers), the Contractor (American Bridge Company/Fluor Enterprises
Inc., A Joint Venture) and the California Department of Transportation. This paper dis-
cusses some of those challenges and the innovative methods that were developed during the
construction of this unique structure. (Nader et al., 2002).
Four distinct structures will make up the bridge crossing: a low-rise post-tensioned concrete
box girder near the Oakland shore; a 2.4 km long segmental concrete box girder (Skyway);
a self-anchored suspension signature span; and a post-tensioned concrete box girder that
connects to the east portal of the Yerba Buena Island tunnel (Tang et al., 2000), Figure 1.

Figure 1. The new San Francisco Oakland Bay Bridge.

17

NYBRIDGE_Book.indb 17 7/17/2013 8:41:27 AM

 18 Cable-supported bridges

2 SELF-ANCHORED SUSPENSION BRIDGE

The Self-Anchored Suspension (SAS) portion of the new East Span of the San Francisco-
Oakland Bay Bridge consists of dual box girders suspended from cables, which are supported
on the 160 m tower located off of the eastern shore of the Yerba Buena Island. The SAS
spans 565 m between the piers E2 and W2, with a 385 m main span, over the navigational
channel, and a 180 m back span (Figure 2).

3 STRUCTURAL FEATURES OF THE SELF-ANCHORED SUSPENSION

BRIDGE TOWER

The 160 m main tower is composed of four shafts interconnected with shear links along
its height. Seismic performance is enhanced by the redundancy of four shafts to support
the bridge’s full vertical weight, and by the cross-braces and shear links that tie the shafts
together and provide redundant horizontal framing (Nader et al., 2000), Figure 3.

4 SUPERSTRUCTURE

The superstructure consists of dual hollow orthotropic steel box girders (OBG), Figure 4.
These girders are in longitudinal compression (reacting against the cable tension force) and
are a part of the gravity load system. The box girders are connected together by crossbeams
that carry the transverse loads between the hangers (span of 72 m), (Baker et al., 2000).

5 SUSPENSION SYSTEM

The main cable consists of 137 prefabricated parallel wire strands (PWS). Each strand consists
of 127 galvanized wires of 5.4 mm in diameter, with a tensile strength of 1760–1960 MPa and
a ductility of 4% in 250 mm. Wires were specified to be straight to 230 mm in a 3 m length
and the wire chemistry limits phosphorus and sulphur to 0.025% to improve ductility. The
configuration of the main suspension cable arose from key aesthetic and structural considera-
tions. Strands of the single main cable are anchored inside the two box girders at the east pier.
The cable then passes over the tower top, through one side of the twin-trough tower saddle.

Figure 2. Elevation of the Self Anchored Suspen- Figure 3. Single tower with four shafts inter-
sion Bridge. connected with shear links.

NYBRIDGE_Book.indb 18 7/17/2013 8:41:28 AM

 Fabrication and construction of self anchored San Francisco Oakland Bay 19

Figure 4. Typical bridge cross section box girders with cross beam.

Figure 5. Three-dimensional cable.

The cable loops below the deck at the western pier, and returns over the tower to the east
anchorage, Figure 5. The Hangers consist of two wire ropes looped over a cable band, and
anchored into a bracket on the box girder by threaded rods into the rope sockets. The hangers
are spaced at 10 m and lie in two sloping planes. Within each plane all hanger ropes are paral-
lel and exert no longitudinal force with respect to the deck girders or cable, as later shown in
section 17.

6 EXPANSION JOINT PIPE BEAMS

The hinges in the transition spans between the SAS and Skyway as well as the SAS and
YBI structures are designed to allow the structures to move relative to each other in the
longitudinal direction and to rotate about the longitudinal axis of the bridge. The hinges are
comprised of compact steel beam pipe sections capable of transferring live loads and seismic
loads, see section 17.
The Skyway was built under a separate contract. The construction engineering for the
alignment and the final connection of the two structures is described below.

7 ERECTION OF THE BRIDGE

In a classic suspension bridge design, the entire design dead load is assumed to be carried by
the suspension system, while the stiffening trusses or girders only serve to distribute live loads

NYBRIDGE_Book.indb 19 7/17/2013 8:41:29 AM

 20 Cable-supported bridges

and limit local deflections. The stiffening system typically has very small bending moments
under the design dead load. The hangers are typically vertical, and the longitudinal compo-
nent of cable tension is a constant. The suspension cable and hangers are erected first. The
deck segments are then hung from the hangers, and little falsework is needed. Erection is
facilitated by the following factors:
1. The suspension system supports the deck during erection;
2. The hangers are vertical;
3. The deck segments have no dead load moments built into the construction.
In the SAS these conditions do not apply. Since the box girder maintains the tension in the
cable, it must be erected on falsework prior to the cable erection, Figure 6. Given the sloping
hangers, it is a highly indeterminate problem to find the profile of the suspension system
hangers, even when the hanger supports don’t move. Due to the support conditions at the
end piers there are moments in the box girders throughout the length of the bridge, and these
are determined by design. The consideration of the dead load moments that are built into the
deck segments is described below.
Traditionally, cable supported bridges are analyzed “backwards”, starting from their
intended final configuration, to find the necessary initial conditions from which to base
construction. Particularly with respect to the cable, this technique was employed for the
SAS. The main analysis of the SAS has been a “forward” analysis, however, starting from
known, or computed, initial conditions. This has facilitated determining critical steps in the
construction, evaluating alternative methods of erection, and tracking the progress of the
construction.
The analysis and the erection control considered the staged construction of the bridge in
great detail, including the following major steps:
1. Box girder erection,
2. Tower erection,
3. Cable erection,
4. Hanger installation,
5. Connection to the Skyway,
6. Addition of superimposed dead load.

8 TOLERANCES AND GEOMETRY CONTROL DURING FABRICATION

AND ERECTION

The unusual layout and supporting system of the SAS required fabrication to close toler-
ances within the controlling interfaces of the following locations: the foundation piers, the

Figure 6. Temporary truss falsework. Figure 7. Box girder fabrication.

NYBRIDGE_Book.indb 20 7/17/2013 8:41:29 AM

 Fabrication and construction of self anchored San Francisco Oakland Bay 21

connection to the adjacent Skyway, and the entry of the main cable into the orthotropic box
girders (OBG). It is important to realize that all points on the superstructure, except for the
foundation connections, were initially erected at other than their final positions. This fact
required extensive erection analysis both before the construction, as well as in response to
unanticipated field conditions.

9 FABRICATION OF BOX GIRDERS

The geometric control started during fabrication in China and was continued on site in San
Francisco. The cambered geometry of the three-dimensional girder model was projected onto
the shop floor in the total station set-up of the fabrication jigs, which set the lower faces of
each girder segment. The geometry of the floorbeams and upper faces were controlled from
the lower faces, Figure 7.

10 FABRICATION OF TOWER

The tower shaft segments, lying prone, were similarly assembled from surveyed jigs. For the
base of each shaft it was necessary to ensure that all of the 574 anchor rods and dowels in
the tower foundation in San Francisco would match the holes fabricated into the tower base
plates. Figure 8 shows the tower foundation. The chosen solution was to survey the as-built
foundation and to fabricate a steel template in match-marked quarters. From the survey the
holes were cut in the template, which was fitted onto the actual foundation prior to shipping
to China. The template was matched to the four tower shaft bases to control the coring of the
holes for the foundation connections.
When the tower segments were fabricated they were trial assembled over the tower base
template, aligning all the holes, and then the shear link connections between each segment
were fit up and match drilled. Once the four segments of each tier were assembled, adjacent
tiers were placed one on the other for trial fit-up of the segment splices between tiers, see
Figure 9. The splice plates for the tower splices were match drilled in position. This geometric
control facilitated alignment and assembly on site.

Figure 8. Tower foundation. Figure 9. Tower tier trial fit-up.

NYBRIDGE_Book.indb 21 7/17/2013 8:41:30 AM

 22 Cable-supported bridges

11 ALIGNMENT OF EAST SADDLE GRILLAGE

The East Saddle Grillages transfer and distribute cable forces from the east saddles to the box
girders, through a framework of connected plates with thicknesses varying between 75 mm
and 100 mm. The grillage comprises five plates parallel to the saddle base plate, three plates
in the plane of the transverse floorbeams, and two plates tying the grillage to a longitudinal
supporting member, Figure 10.
The plates are connected to each other and to the box-girder through full penetration
welds. The detailing of the saddle grillage plates was based on the cable geometry, so as to
transmit the various components of the saddle bearing forces into the girder. Given the mag-
nitude of these forces, tolerances for the alignment were specified, and a full-size mock-up of
the grillage was fabricated to verify welding and inspection access, and to establish a welding
sequence that would minimize distortion and meet the tolerances.

12 DEAD LOAD CAMBER

In the design of the SAS dead load moments were built into the girder, in order to reduce
compression in the top fiber of the box girders, Figure 11. These dead load moments were
developed in the girders during SAS fabrication and erection by use of a downward “dead
load camber”, which results in forced upward flexure of the boxes during erection. The dead
load vertical camber is the deflected shape of the girders that results from the moment dia-
gram, in the absence of all other loading. Each girder was fabricated with the camber shown
in Figure 12. The actual fabricated shape of the girders was determined by the combination
of the final grade line and the dead load camber. The cambered shape was obtained by detail-
ing small angle breaks between segments of the box girder.

13 ORIENTATION OF BOX GIRDER SPLICES

During the fabrication the segments were constructed as straight segments of 20–25 m in
length. As segments were constructed they were aligned and surveyed in the fabrication yard
to implement the camber profile. The ends of each segment were trimmed to the alignment
angle, and the bolted stiffener splices were fit up and match drilled. During the erection
on site the segments were erected on falsework and oriented by jacking to align the bolted
splices. Starting at the west end, the girder was aligned to the final profile grade and con-
nected monolithically to the concrete pier.

Figure 10. East saddle grillage. Figure 11. Girder dead load moment.

NYBRIDGE_Book.indb 22 7/17/2013 8:41:31 AM

 Fabrication and construction of self anchored San Francisco Oakland Bay 23

Figure 12. Dead load camber.

The camber shown in Figure 12 is downward sag, and if the west end of the camber profile
(left in the figure) is aligned to level, it can be seen that the girder in its cambered shape would
rise significantly. For this reason, as the segment splice bolting was sufficiently complete each
segment was lowered to its final profile, thereby introducing the Dead Load moments into
the girders.
Erection of the subsequent sections and partial bolting advanced as welding proceeded on the
skin plates of the previous girders. Partial bolted splices were completed with temporary bolts,
and about half of the box perimeter stiffeners were left unbolted to facilitate the welding. After
welding was complete on each girder face, the permanent bolted splices were installed, allowing
for the removal of adjacent temporary bolts, until the perimeter splicing was complete.

14 GIRDER LENGTH CONTROL SURVEYS

In addition to the vertical camber, a dead load axial camber was also provided. Axial camber
consisted of detailing the box girders longer than the final erected length. The axial compres-
sion strain on the box girders, resulting from the cable tension, was computed to be about
300 mm over their length during cable erection and load transfer, and these calculated strains
were included in the fabricated girder geometry.
During the erection on site, the leading edges of the constructed girders were surveyed.
Due to fabrication and erection tolerances, and shrinkage in the field welds, the cumula-
tive length error in the first half of the total girder span was determined. The last three lifts
leading up to the eastern pier were corrected for length as their fabrication was completed
in China. The error that was measured and projected to the end of construction was divided
among these three lifts. Excess length (green) that was detailed into each segment allowed
this to be done.
When the bearings at the east end of the main span (pier E2) were installed after load trans-
fer, the alignment error was only 5 mm over the 565 m distance between piers W2 and E2.

15 TOWER ERECTION

The tower shafts were erected one segment at a time. Starting at the base each shaft segment
was placed on the tower foundation. The use of the tower template ensured that each of the
anchor rods and dowels entered its hole in the tower base. The four shafts in the first tier were
aligned and plumbed on the temporary blocking prior to grouting. With the base aligned the
assembly of the remaining segments followed the pre-drilled splice formed in China.

NYBRIDGE_Book.indb 23 7/17/2013 8:41:32 AM

 24 Cable-supported bridges

16 TOWER TIE-BACK

The main cable is supported on the top of the tower in a cable saddle. As in other suspension
bridges, the saddle tends to move during the erection of the suspended structure.
Therefore, the tower saddle was diverted 0.5 m towards the side span from its final, verti-
cal position by tying back the tower. Figure 13 shows the stay cables used to tie-back the
SAS tower. These stays were anchored in the ground on Yerba Buena Island, and gradually
released during the erection of the box girder to minimize the tendency of cable strands to
slide through the tower saddle. The forces in the stays were reduced in each step of the analy-
sis in accordance with the erection schedule.

17 MAIN CABLE STRAND ERECTION TOLERANCES AND CONTROL

17.1 Cable and hanger length

The geometry of the main cable is determined by the final configuration of the bridge under
dead load. This configuration includes the final grade of the box girder, the total weight of
the box girder, and the horizontal component of the force in the cable, which is set by the
design. The geometry of the cable and the forces in the cable and the hangers were computed
to satisfy equilibrium of the structure. The final geometry of the bridge and the forces locked
into the bridge are determined by the fabricated lengths of the cable and the hangers.
Great care was taken in the weight take-off of the fabricated structure. All components
that were to be included in the final structure were tracked, including the rolling tolerance
on the thickness of the plates used to fabricate the structure. Through a survey made in the
fabrication shop by use of ultrasonic thickness gauges, it was determined that plates were on
average 2% thicker than specified. This factor was considered in the weight take-off. For the
SAS accurate weights are required, since the cable is anchored into the box girders, which
compress under additional load, and this affects the cable final geometry.

17.2 Cable displacement

The suspension cable was erected as a free-hanging cable (FHC), in order to have good com-
paction. The parallel wire strands (PWS) of the cable were towed by means of a tramway
system around the path shown in Figure 5. It was determined that the cable would then move
over 9 meters from its initial to its final position, when the inclined hangers are attached and
fully loaded. These movements are illustrated in Figure 14, which shows both the free-hanging

Figure 13. Tower tie-back.

NYBRIDGE_Book.indb 24 7/17/2013 8:41:33 AM

 Fabrication and construction of self anchored San Francisco Oakland Bay 25

Figure 14. Deflection of the main cable from the free-hanging to the fully-loaded position.

and the fully-loaded positions of the cable. The importance of nonlinear geometric effects
may be appreciated from this figure and the values of cable displacement.

17.3 Load transfer suspender tension and length monitoring

The SAS cable movement from its initial to its final position, when the inclined hangers are
attached and fully loaded, complicated the attachment of the hangers. Most of the main
span hangers were too short to reach between the cable and the hanger brackets on the
girder. During erection these hangers were gripped by clamps and rods and tensioned using
strongbacks on at the base of the hanger brackets. In addition, hanger inclinations vary con-
tinuously during the process, and as a result the cable is twisted about its axis. The twisting
during load transfer required the installation of all the cable bands at pre-determined angles
difference from their final orientations. This is described below.

17.4 Load transfer monitoring

The key issue for the erection of the SAS was the load transfer of the box girders from the
temporary trusses to the cable system. This took place after the box girder had been assem-
bled on the trusses, the main cable was erected and compacted, the cable bands bolted in
place, and the hangers hung from the cable. The load transfer was accomplished by progres-
sive jacking of the hangers until they could be socketed to the box girders, Figure 15.
The load transfer was accomplished in three main phases, as illustrated in Figure 16. Only
about one-half of the total hangers were installed in Phase 1 (top of Figure 16), to simplify
the construction staging. However at this stage of erection approximately 90% of the weight
of the box girder (less paving, utilities and bikepath) has been lifted by the main cable and the
Phase 1 hangers. It is clear that the erection forces in these hangers are on average 90% higher
than the forces occurring during the service life of the bridge (25% of the breaking load),
which is acceptable temporarily. Great care was taken with the exact sequencing of the Phase
1 hangers, in order to avoid overstressing them by an uneven distribution of load between
individual hangers. Phase 1 was divided into about 50 steps, and throughout the load transfer
the jack pressures and the remaining length to be jacked was monitored, in order to control
interim hanger tensions. Note that in Phase 1 three hangers at the tower remained unin-
stalled. These were placed in Phase 2 (center of Figure 16).
Phase 2 included just the three main span hanger positions adjacent to the tower, and this
phase was introduced to control cable bending, as described in the next section. The timing of
the Phase 2 hanger erection had an effect on the overall lift-off of the bridge from the falsework.
The deflection of the box girder at the end of Phase 1 of hanger erection is shown in Figure 17.

NYBRIDGE_Book.indb 25 7/17/2013 8:41:33 AM

 26 Cable-supported bridges

Figure 15. Hangers shown dur- Figure 16. Three phases of load transfer via hanger installation.
ing load transfer.

Figure 17. Deflection of box girder after phase 1 of load transfer.

It may be seen that the box girder has lifted off of the temporary truss over most of its length,
but remains in contact with one cradle near the tower and with one cradle near the East Pier.
The two last supporting cradles needed to be checked for interim wind and seismic loadings for
the period of time between Phase 1 and Phase 2. This effort was justified by the need to mitigate
cable bending, as described below.
Phase 3 of load transfer (bottom of Figure 16) completed the installation of the hangers
brought the suspension system to near its final position.

17.5 Cable bending

One normally thinks of the main cables of suspension bridges as carrying tensile forces; and
this is their primary behavior. But cables may also carry bending moments. The issue of cable
bending is complex. Bending moments may arise in a cable through the interaction of three
factors:
• Rotation of the cable due to large displacements in space.
• The tension in the cable.

NYBRIDGE_Book.indb 26 7/17/2013 8:41:33 AM

 Fabrication and construction of self anchored San Francisco Oakland Bay 27

• The lengths of cable bands, which impart flexural and shear rigidity to a portion of the
cable.
• The spacing of the cable bands.
The cable bands rotate with the cable as it swings out and down from the FHC position.
Rotation of the cable bands generates force couples due to the tension in the cable acting
through the relative displacements of the ends of the cable bands. These couples are equili-
brated by bending moments in the cable.
For the SAS, this issue was studied using the ADINA model shown in Figure 14. The cable
modeling included all the individual PWS strands (the cable was pre-fabricated in 137 strands)
to form the cross-section of the cable. The model also includes the cable bands to their actual
lengths, matching the actual cable bands in sufficient detail to compute the bending moments
and rotations in the main cable arising from the phenomenon described above.
This analysis indicated that load transfer could potentially generate unacceptably high
bending stresses in the cable in the main span near the tower. This section of the cable is
restrained from rotating by the tower saddle (which is fabricated to the final geometry) and
so attracts bending stress resulting from the cable movements. These stresses were avoided by
delaying the installation of three hangers at the tower, and the tightening of their cable band
bolts. Without these bands installed, the cable was provided a 35 m length that was very flex-
ible. As a result, the large displacements of phase 1 were accommodated while generating less
stress. Phase 2 consisted of installing these 3 hangers.

17.6 Cable twist

Related to the phenomenon of cable bending is cable twist. As the cable was displaced from
its free-hanging to its fully loaded position, it twisted about its own axis, as was observed
earlier in the Yeongjong Bridge, (Cho et al., 2001). The problem was solved by accurate
predictions of cable twist, and the specification of compensating camber in the cable band
installation. The ADINA model shown in Figure 14 was used to study the twist of the cable
and to compute a twist camber to be applied to the cable bands before load transfer.

17.7 Cable band installation angles

The computed rotations of the main span cable bands are shown in Figure 18—this is labeled
“structural response” in the figure. The curve labeled “total rotation” is the intended final

Figure 19. Cable band rotation dur-

Figure 18. Cable twist and camber. ing load transfer.

NYBRIDGE_Book.indb 27 7/17/2013 8:41:35 AM

 28 Cable-supported bridges

inclination of the cable bands to match the hanger angles between the cable and the edge of
the box girder (see Figure 15 and Figure 14). The difference between these is called the “cable
band camber”. The cable bands were placed on the cable in a rotated position in accordance
with this camber (see Figure 19). The camber varies from about 20 degrees near the tower
and anchorage (where the cable rotates little) to about 5 degrees at midspan (where the cable
rotates a lot). It may be observed in Figure 18 that the rotation of the cable even reverses
direction near the tower.
The as-built measurements of the completed cable after load transfer demonstrated the
success of the construction analysis and the use of cable band camber. Only two cable bands
deviated from the analyzed value by more than 1.5 degrees, and these cases may possibly be
attributed to installation error.

17.8 Connection of SAS to skyway

As the SAS was being constructed, the as-built Skyway was itself calibrated by jacking opera-
tions to determine the stiffness of the spans that were to be connected to the SAS. The Sky-
way was carefully surveyed and then lifted using load-calibrated jacks. The verified stiffness
and as-built elevations were then used to calibrate the global model of the combined SAS-
Skyway bridges, including the elastic modulus and creep coefficients consistent with the sur-
vey history and final measured properties. The refined model served to guide the operations
that connected the two structures together.
The tips of the box girders—where they meet the Skyway on the right of Figure 17—curl
up at this stage. This is due to the complex interaction between the weight of the box girders,
their initial camber, the absence at this stage of any reaction from the Skyway, and the pull of
the cable which is anchored inside the girders at this location.
Subsequent to the load transfer, the SAS and Skyway were connected, using the strong-
back system shown in Figure 21. The strong-back beam was used to align the SAS and
Skyway tips by jacking the SAS down to the Skyway. After alignment, the pipe beams illus-
trated in Figure 20 were pulled through the joint and connected with bearings to both the
SAS and Skyway. (The pipe beams had been erected previously and stored inside the Skyway
while the SAS was constructed.).
These steps were carefully included in the analysis. The elements representing the pipe
beams and the bearings were added to the model as the strong-back jacking was staged. The
daily thermal movements at Hinge A were measured, and by analysis the reactions on the
as-built pipe beams were determined. Temporary shimming and the timing of grouting for
the pipe beam bearings were developed, such that the grout could develop sufficient strength
prior to the peak thermal reactions.

Figure 21. Strongback used to align the SAS

Figure 20. SAS to skyway hinge. and skyway.

NYBRIDGE_Book.indb 28 7/17/2013 8:41:36 AM

 Fabrication and construction of self anchored San Francisco Oakland Bay 29

18 CONCLUSIONS

The fabrication and construction of the self-anchored suspension bridge has been a
challenging undertaking. It has required specialized analysis, and extensive monitoring of
tolerances and geometry throughout the work. Furthermore, it has required the investigation
of issues not often encountered in bridge erection, like cable bending and cable twist. The use
of detailed modeling served to:
• Verify the on-going construction quality.
• Address deviations in expected behavior during construction.
• Confirm the serviceability of the final structure.

ACKNOWLEDGEMENTS

The authors would like to acknowledge the invaluable contributions of the Caltrans
Construction Team and the American Bridge/Fluor Team.

REFERENCES

Baker, G., et al. 2000. Design of the Superstructure of the New San Francisco-Oakland Bay Bridge Self-
Anchored Suspension Bridge. 17th International Bridge Conference, Pittsburgh, Pennsylvania, 2000.
Cho, C-Y., Lee, S-W., Park, S-Y., Lee, M. 2001. Yeongjong Self-anchored Suspension Bridge. Structural
Engineering International, February 2001, n. 1 v. 11: 21–23.
Design Criteria of the San Francisco-Oakland Bay Bridge East Span Seismic Safety Project (1999), TY
Lin International—Moffatt & Nichol Joint Venture.
Nader, M., et al. 2000. Seismic Design Strategy for the New San Francisco-Oakland Bay Bridge Suspen-
sion Span. 12th World Conference on Earthquake Engineering, New Zealand, 2000.
Nader, M., et al. 2002. San Francisco-Oakland Bay Bridge Self-Anchored Suspension Bridge Steel
Design Challenges and Solutions. National Steel Bridge Alliance Conference, Chicago, Illinois, 2002.
Tang, M.C., et al. 2000. Design Concepts of the New San Francisco-Oakland Bay Bridge. 17th Interna-
tional Bridge Conference, Pittsburgh, Pennsylvania, 2000.

NYBRIDGE_Book.indb 29 7/17/2013 8:41:36 AM

This page intentionally left blank
 Chapter 3

Structural cable designs and concepts

T.W. Klein
WireCo WorldGroup, Kansas City, MO, USA

ABSTRACT: The paper discusses new and enhanced structural cable designs for use as
tension members on bridges and static applications. These developments include new wire
coating processes for increased corrosion resistance and the use of synthetic materials for
amplified strength and durability. Discussions regarding the use of synthetic materials in
structural elements have significantly increased recently leading to several hybrid cable con-
cepts that provide greater resistance to fatigue and strength without the vulnerability to cor-
rosion. The reader will also be informed about existing technologies currently being used to
enhance steel cable properties. The use of shaped wires formed from high carbon steel pro-
vide high strength cables with locked outer layers for increased protection from the elements.
The paper will discuss the advancements in these fields along with proven manufacturing
practices shown to develop high performance cables for use in new and existing structures.

1 INTRODUCTION

Steel wire is the strongest most cost effective and versatile product used in the cable industry
today. The main draw back of the product is the susceptibility to corrode as time passes if not
adequately protected. The most effective method to overcome this problem is to galvanize the
wire forming a zinc layer between the environment and the steel. The use of zinc coated high
carbon steel wire in the fabrication of structural cables for the bridge industry has been prac-
ticed for many years. The goal of this paper is to educate the bridge industry about advance-
ments in cable technology and help them to take full advantage these items.

1.1 Wire production

The building block of structural cables is the galvanized wire. The current specifications for
these items mandate the physical requirements for the wires. The requirements include the
mechanical properties, stress under load, elongation, tensile strength, ductility, zinc coat-
ing weight, coat adherence, and wire surface finish. Typically to improve on one property
another must be sacrificed.
The process of producing galvanized wires has seen several improvements and refinements
in the past ten years leading to increased strengths and improvements in quality. Some of the
processes once thought to be necessary have been removed or replaced to reduce cost and
stream line the manufacturing. The manufacturing begins with the rod production. Carbon is
the major steel element in the rod used for wire production. Higher carbon elements directly
correlate to higher tensile strengths as well as the resistance to abrasion. The negative side of
this relates to the elongation and ductility of the wire. The rod production begins with con-
tinuous cast blooms rolled to diameter and control cooled to provide a clean homogenous
pearlitic grain structure suitable for direct drawing. The controlled cooling eliminates the need
to patent the rod prior to drawing the wire. The drawing of the wire is the process to reduce
the diameter of the rod by passing through a number of successive dies. This process includes
simple round wires as well as complex shaped wires. The drawing process aligns the grain

31

NYBRIDGE_Book.indb 31 7/17/2013 8:41:37 AM

 32 Cable-supported bridges

structures into a fibrous geometry substantially increasing the tensile strength of the wire. To
ensure ductility does not become a factor clean steel without inclusions is required. Since the
drawing process is a cold working process the amount of reduction that can be performed is
limited by the steel. Tensile strength increases dramatically with the amount of reduction but
the ductility of the wire also becomes very poor. With the quality of the steel available some
rod maybe reduced up to 95% of its original area. However in some cases the wire must be
patented between cycles. During patenting the steel grain structure becomes more uniform
and ductile making the wire suited for further drawing. Several advancements in the patenting
of wire have been introduced and incorporated in to the process. The use of lead to quench
the steel is no longer a viable solution in the United States due to stringent EPA regulations.
Out these regulations have come new methods using salt, sand, and steam as viable quenching
methods. These methods provide superior refinement of the wire grain structures without the
harmful environmental effects that are associated with the use of lead.

1.2 Corrosion protection of wire

Corrosion is an electromechanical process that takes place when there is a different electri-
cal potential between two points connected by an electrolyte. The most effective method to
overcome this destructive process is to galvanize the steel wires forming a zinc layer that is
alloyed with the steel surface. The galvanizing process protects the steel by removing it from
the external agents while the zinc becomes the sacrificial agent.
The majority of the wire supplied for use in structural cables is galvanized by the hot dip
method. This process requires the wire to be submerged in molten zinc providing a very uni-
form coating with an aesthetically pleasing appearance. This method deposits large amounts
of zinc onto the surface of the wire. The main drawback to this type of coating method is
strength loss due to the annealing of the wire while submerged in the molten zinc. Approxi-
mately 10% of the wire strength will be lost during the hot-dipped coating process. The other
means of zinc coating the wires is electroplating the zinc to the surface of the steel. This
method has the ability to provide a much heavier and clean surface that is acceptable for use
in static and dynamic applications.
The most rapidly growing hot dip coating process for structural steel wire is made with
the Galfan® process (Goodwin, F., Wright, R., 1983. The Process Metallurgy of Zinc-Coated
Steel Wire and Galfan® Bath Management). This process requires a Zn-5% Al coating be
applied to the surface of the wire by the hot-dip method. Basic Galfan® coated carbon steel
wire is addressed in ASTM A856 and EN 10244. The coating is a eutectic Zn-Al mischmetal
resulting in a more uniform crystalline microstructure with improved mechanical character-
istics. Ageing tests have shown this process to provide improved corrosion resistance in com-
parison to conventional galvanized coating. Combining the passive corrosion inhibition of
aluminum oxidation with the active and passive effects of zinc results in approximately two
times the amount of corrosion protection compared to standard zinc coated wire. The coat-
ing also provides an anodic feature that heals over the exposed steel when the wire is abraded
or scratched exposing the base steel. This process was developed in the mid 1980’s with the
majority of the production and use taking place in Europe.

2 STRUCTURAL CABLE CONSTRUCTIONS

2.1 Structural strand construction

Zinc Coated Structural Strand produced per ASTM A586 (ASTM A586-04a. 2004. Zinc
Coated Parallel and Helical Steel Wire Structural Strand) is used in many different struc-
tural applications. This specification also covers Zinc Wires for Spun-In-Place Strand. The
process of manufacturing structural strand consists of laying consecutive layers of round
wires around a center, changing the direction of the wire lay of each layer. The opposing lay

NYBRIDGE_Book.indb 32 7/17/2013 8:41:37 AM

 Structural cable designs and concepts 33

direction increases tensile strength and Modulus of Elasticity while providing a balanced
strand. Figure 1 shows a cross section of a typical design. The wires used in each layer may
or may not be preformed. The preforming of wires consists of mechanically forming the wire
ensuring a tight layer around the core or base layer. Some strand constructions will have
multiple wires laid to form a unitized center, sometimes referred to as the Parallel Contact
Core, to increase the static fatigue of the cable. Tension-tension fatigue testing has shown
this construction reduces internal nicking and bending stresses which occur in conventional
cross lay designs. This testing has shown cracks in the wire initiated at the cross-wire contact
point of inner wires, eventually leading to tear in shear. Increasing the number of wires in the
strand layer requires smaller diameter wires however this is been proven to increase the axial
fatigue life of structural strand. The increased fatigue life is related to less contact area and
increased ductility of the strand. However it has also been shown that wire grade and mean
loading have little or no effect on the fatigue life of the strand. Structural strand is designed
to be used in straight tension only. Deflections in structural strand could reduce the strength
and severely reduce the fatigue strength of the strand if not tensioned properly.

2.2 Structural wire rope construction

Zinc Coated Steel Structural Wire Rope produced per ASTM A603 (ASTM A603-98. 2003.
Zinc Coated Steel Structural Wire Rope) is regularly used where there are requirements for
the structural element to be bent or deflected around supporting members. Traditional sus-
pension bridges are the best example for the use of this product. The construction of the
product allows deflection of the product without a significant loss in the tensile strength
properties. Due to the mechanical requirements of the specification structural wire rope is
required to be constructed differently from standard six or eight stranded running ropes. The
individual outer strand lay and rope lays are lengthened to provide increased strength and
Modulus of Elasticity. Since the wire rope is utilized in static tension there are construction
options available to enhance the strength or flexibility. A single strand core or independent
wire rope core can be provided without a reduction in the mechanical properties of the rope.
The strand core provides a simpler rope construction which on average will have an increased
Modulus of Elasticity due to fewer elements. The use of structural rope has been prominent
in the bridge industry for more than 100 years. The historical data available for fatigue and
life prediction is very well documented in the domestic market.

Figure 1. Traditional spiral strand.

NYBRIDGE_Book.indb 33 7/17/2013 8:41:37 AM

 34 Cable-supported bridges

2.3 Locked coil cable construction

Locked Coil Cables are becoming more prevalent in structures throughout the world (EN
12385 Part 10. 2008. Spiral Ropes for Structural Applications). This is product is commonly
referred to as Locked Cable due to the locking action of the outer wire layers. Very similar to
structural strand the process of manufacturing consists of laying consecutive layers of wires
around a center, changing the direction of the lay of each layer. The differentiating factor being
the interlocking z-shaped wires stranded on the outer layers. The interlocking wires provide a
barrier for corrosive media penetration and a dense metallic area. The increased metallic area
results in higher breaking forces compared to Spiral Strand or Wire Rope as well as a substan-
tial increase in axial stiffness. The axial stiffness is approximately 30% more than traditional spi-
ral strands with a small added benefit of improved lateral stiffness. Good clamping abilities are
available due to the smooth outer surface. Figure 2 shows how the interlocking wires connect
forming a solid layer. The connection of the wires in the layer is also critical for wire retention
in the event of wire fracture. Figure 3 shows the formation of the wires into a layer.

Figure 2. Interlocking wires.

Figure 3. Outer layer of interlocking wire.

NYBRIDGE_Book.indb 34 7/17/2013 8:41:37 AM

 Structural cable designs and concepts 35

3 HYBRID CONSTRUCTIONS

3.1 Hybrid design concept

Hybrid constructions incorporate high strength synthetic fibers in the inner core of structural
cable with multiple layers of shaped or round wires on the outer layers for increased strength
and fatigue life. One of the main goals in the design of the Hybrid construction is to take
full advantage of the performance characteristics of the fiber element. The introduction of
a fiber member in the core area adds superior fatigue characteristics of the material as well
as the high strength. Balancing the load sharing between the fiber and the steel elements
requires the lay of the outer steel strands to be matched perfectly with the fiber braids. These
constructions also reduce the weight per foot of the cables significantly. Although this may
not be an important characteristic in all structures the significant weight reductions for long
spans can become a critical factor.
Figure 4 and 5 show the conceptual idea of structural Hybrid cables. Matching synthetic
materials with steel outer wires combines both the positive and negative attributes of both
materials.

Figure 4. Hybrid strand with round wires on the outer layers.

Figure 5. Hybrid strand with round and shaped wires on the outer layer.

NYBRIDGE_Book.indb 35 7/17/2013 8:41:38 AM

 36 Cable-supported bridges

Table 1. Weight per length reductions and strengths for wire rope
construction.

Min break
Description Diameter Lbs./Ft force (KN)

6 × 25FW XIP Steel 38 mm 4.16 1014.20

Flex-X Compacted 38 mm 5.01 1112.06
6 × 31 DGXIP PC 38 mm 3.64 956.37
8 Strand Hybrid 38 mm 3.66 1262.41

3.2 Tensile strength and elongation properties of hybrid wire rope

Table 1 below lists the typical stranded wire rope constructions and Minimum Breaking
Force (MBF) requirements as specified in the wire rope user’s manual (WRTB. 2005. Wire
Rope Users Manual 3rd Edition). The traditional wire rope construction is listed in Table 1 to
provide a bench mark for the Hybrid constructions. The Hybrid construction provides a 24%
increase in strength from the base 6 × 25FW construction with a 12% reduction in the weight
per foot. The core strength member provides a tremendous increase in strength due to the
load sharing capabilities of the core member. This ultimately leads to a substantially smaller
diameter wire rope with increased tensile strengths and comparable weight per lengths.
The measured Modulus of Elasticity of the Hybrid wire rope construction is approxi-
mately 103,421 MPa. This is comparable to the traditional Polypropylene and Polyester fiber
cores. The Hybrid cable construction provides a smooth load curve with continued elonga-
tion during the loading.

3.3 Fatige performance

The added benefit of the Hybrid construction is the fatigue performance of synthetic material
over that of traditional steel cable. This is applicable to both running ropes and static appli-
cations. The selection of the synthetic material and the construction of the braid affect the
mechanical properties. For structural applications high strength with high modulus material
are required to be combined with dimensional stability and resistance to chemical elements.

4 INSPECTION AND MAINTENANCE

4.1 Inspection
The inspection of structural cables in service is critical in evaluating the serviceability and life
of the wire rope. Visual inspections are the easiest to perform but depending on the condition
of the cable this may not be feasible. The cleaning of the ropes to provide a visible surface
may require expensive and time consuming processes. The disposal of this material may also
require the handling of hazardous waste and in the case of most bridges this operation must
be performed over a waterway. Therefore it is recommended to have the cables inspected with
non-destructive means to evaluate the condition prior to the removal of any coatings.
Many types of non-destructive testing methods for the inspection of wire rope have been
experimented with over the years to include acoustic emissions and radiography. The use of
electromagnetic inspection is the only proven practical way to efficiently inspect wire rope
however this requires either the tool or the cable to be moving. The mining industry has used
electromagnetic inspection on wire rope since the early 1950’s. In many countries, including
the United States and Canada, the use of electromagnetic inspection is mandated for the
inspection for running wire ropes. In addition, the visual method of inspection must be used
as an important aid to electromagnetic testing.

NYBRIDGE_Book.indb 36 7/17/2013 8:41:39 AM

 Structural cable designs and concepts 37

All inspections and maintenance cycles are intended to keep the cables and their components
functioning as intended. The results of any inspection should be written and filed with the
structure maintenance manual.

5 CONCLUSIONS

This paper is by no means a complete analysis of new structural cable constructions or

designs. This paper was written to educate the bridge engineering community on the develop-
ments in the field of structural cables. The current ASTM specification covering these types
of cables allows the manufacturer the latitude to produce cables without restrictive manufac-
turing parameters. Innovative materials are now available to increase the corrosion resistance
of the cable. Some of the materials are applied during the manufacturing of the strand to
individually coat wires and some are applied to the cable once it is installed and tensioned
onsite. The fatigue tests presented in this paper show that standard preventative maintenance
on bridge cables can significantly extend the life of cable. Variations in the construction of
the cable have also been shown to increase the life cycle of the cable.

REFERENCES

ASTM A586-04a. 2004. Zinc Coated Parallel and Helical Steel Wire Structural Strand.
ASTM A603-98. 2003. Zinc Coated Steel Structural Wire Rope.
EN 12385 Part 10. 2008. Spiral Ropes for Structural Applications.
Goodwin, F., Wright, R. 1983. The Process Metallurgy of Zinc-Coated Steel Wire and Galfan® Bath
Management.
WRTB. 2005. Wire Rope Users Manual 3rd Edition.

NYBRIDGE_Book.indb 37 7/17/2013 8:41:39 AM

This page intentionally left blank
 Chapter 4

Latest developments in suspension bridge main cable

dehumidification

M.L. Bloomstine
COWI A/S, Denmark

ABSTRACT: Hidden corrosion of suspension bridge main cables is a very serious and
widespread problem that compromises the safety of many bridges. In order to protect against
corrosion, dehumidification systems have been developed, tested and proven over the last
18 years. This paper presents a general description of corrosion protection by dehumidifica-
tion, the current worldwide status, examples of systems and the latest experience and devel-
opments including full-scale testing, on-site testing and new optimal details.

1 INTRODUCTION

Corrosion protection of suspension bridge main cables has generally been carried out
according to a method developed by Washington Roebling in the 1880s for the first large
scale suspension bridge—the Brooklyn Bridge. The system is comprised of galvanization
of the cable wires, a layer of red lead paste on the bundle of wires, galvanized wrapping
wire and paint on the outer surface. This is called a traditional system and was (with some
modifications) applied to nearly all suspension bridges until 1998. Many examples from
countries around the world have shown that the traditional system at best only slows
down corrosion, but does not prevent it. Worldwide experience shows that many suspen-
sion bridge cables have wide spread corrosion and even broken wires. While this condi-
tion may be expected on very old bridges, it has also been found on numerous bridges
under 30 years old. This is a very serious problem, as it compromises the safety of the
bridge, especially on the younger bridges, which are designed with a lower safety factor.
Originally the design safety factors for main cables were in the range of 4–5, whereas
nowadays the design safety factor may be as low as 2. This serious condition also neces-
sitates regular in depth inspections by wedging and strength evaluation, which are very
expensive.
Based on extensive successful experience with corrosion protection of steel bridge struc-
tures by dehumidification since the 1960s, systems for main cables were developed in the
1990s and applied for the first time on the entire length of the main cables in 1998. Since then
many more systems for main cables have been designed and applied and operation experi-
ence has shown them to be effective and durable. Main cable corrosion protection by dehu-
midification is already a standard procedure in many countries and is well on the way to
becoming a worldwide standard. This is the only system that provides such a high level of
corrosion protection. The effectiveness of the system is documented by a monitoring system,
which documents a noncorrosive atmosphere inside the main cables. It should be noted, that
whereas a properly designed and installed dehumidification system can provide a high level
of protection against further corrosion, it can not “repair” the damage that has already been
done. Therefore, it is still necessary to inspect and evaluate the main cables before installing
the system.

39

NYBRIDGE_Book.indb 39 7/17/2013 8:41:39 AM

 40 Cable-supported bridges

2 BACKGROUND OF CORROSION PROTECTION BY DEHUMIDIFICATION

The concept of dehumidification was known for many years and has been successfully applied
to a wide range of applications before it was adapted to bridge structures. Dehumidification
has with over 40 years experience from steel bridge structures been proven to be the optimal
solution for corrosion protection where it is applicable. Before the development of dehu-
midification systems for steel structures, painting was the only means available for corrosion
protection. Dehumidification has been proven to be superior to painting in all respects, i.e.,
technically, economically and environmentally. To start with the most widespread application
for bridges was the inner surfaces of the closed box bridge girder. This application has virtu-
ally become a worldwide standard with systems in service since 1970. The main advantages
of dehumidification are:
• A properly designed and applied dehumidification system is virtually 100% effective, pro-
viding a much higher level of protection than painting. Painting is difficult to perform
correctly, always includes many inherent defects and breaks down with time.
• The initial cost of dehumidification is only a fraction of the cost of painting.
• The maintenance costs of dehumidification are also only a fraction of those for painting.
• Dehumidification is environmentally friendly and beneficial for health and safety, as it
does not have the environmental problems which are caused by blasting and painting.
The first application on a bridge was the Little Belt Suspension Bridge in Denmark.
A system of four dehumidification plants was developed for the 1,080 m long bridge box
girder. The system has been continuously in service for 43 years and has only required a
minimal amount of maintenance. As the method was not proven by similar experience at that
time, it was decided to paint the internal surfaces as an extra precautionary means. In order
to control the effectiveness of the dehumidification system shiny sand blasted steel plates
were hung up at various locations with low air circulation. After 43 years these test plates
are still shiny and without sign of corrosion, which proves the excellent protection that is
provided. Plants were also installed in each of the four anchor chambers to protect the main
cable strands, anchorages and the splay saddles. These plants have also performed excellently
and the components in the anchor chambers look as new.
Dehumidification of anchor chambers has also grown to become an accepted worldwide
standard and many bridges throughout the world have been fitted with dehumidification
systems for anchor chambers, including many retrofits. Further applications for steel bridge
structures have been developed and successfully applied. These include cable saddles on sus-
pension bridges, anchorage boxes on cable-stayed bridges, and hollow truss bridge girders.
Dehumidification of main cables is a natural extension of this extensive experience and
development of systems for main cables commenced in the 1990s. Further experience with
main cables is the main topic of this paper and is presented in the following sections.

3 PRINCIPLES OF CORROSION PROTECTION BY DEHUMIDIFICATION

The basic principle upon which corrosion protection by dehumidification is based, was devel-
oped by Professor H.H. Uhlig at the MIT Corrosion Laboratory.
During this research it was established that atmospheric corrosion of iron is negligible
when the relative humidity is below 40%, that corrosion starts above RH 40% and the corro-
sion rate increases slightly until RH 60%, where after the corrosion rate increases dramati-
cally, as illustrated in Figure 1. A dehumidification system for a steel bridge structural element
is composed of relatively few elements. A dehumidification plant provides sufficiently dry air
and circulates it inside the structural element, ensuring that the inner surfaces are protected
from corrosion. The main components of the dehumidification plant are an electrical/control
panel, a dehumidification unit and a fan unit as illustrated in Figure 2.

NYBRIDGE_Book.indb 40 7/17/2013 8:41:39 AM

 Latest developments in suspension bridge main cable dehumidification 41

Figure 1. Relationship between RH and rate of corrosion (Prof. H.H. Uhlig, MIT Corrosion
Laboratory).

Figure 2. Typical dehumidification plant and diagram of active sorption dehumidification unit.

The dehumidification unit is generally based on active sorption, as it is efficient for virtu-
ally all air conditions, i.e., there are practically no temperature and relative humidity limits.
This method works by binding the moisture in the process air to a hygroscopic material
(a sorbent). A dehumidification unit based on active sorption contains a rotor which is built
up of many small pipes, coated with a sorbent, most commonly lithium chloride. The process
air is forced through the rotor and its moisture is absorbed under this process, resulting in
dry air. The rotor turns very slowly, allowing time for the process. On the opposite side of the
rotor, intake air is heated and blown through, which dries out the sorbent coating. This air
becomes moisture laden and is subsequently discharged. A dehumidification unit with active
sorption is shown in the picture and diagram in Figure 2.

4 GENERAL DESCRIPTION OF SYSTEMS FOR MAIN CABLES

A system for corrosion protection of main cables by dehumidification consists of the follow-
ing three main sub-systems:
• A dry air system capable of producing and blowing dry air through the main cables.
• A sealing system for the main cables, including cable bands, saddles and other connected
components.
• A control and monitoring system.

NYBRIDGE_Book.indb 41 7/17/2013 8:41:39 AM

 42 Cable-supported bridges

4.1 Dry air system

The dry air system produces dry air and blows it through sections of the main cables. The
system assures overpressure inside the sealed cable system and minimizes intrusion of mois-
ture through possible small leaks. The dry air system is made up of the following main
components:
• Dehumidification plant(s).
• Injection and exhaust points.
• A layout, such as shown later in the paper, is developed for the dehumidification system.
The layout defines the locations of the dehumidification plant(s), buffer chamber(s), injec-
tion and exhaust points as well as the flow sections.
The main components of a dehumidification plant for main cables are a dehumidification
unit, a fan, an electrical board, filters and ducting, such as illustrated in Figure 3. Injection
points are established by either modifying existing bridge components, such as the saddles or
by designing purpose suited injection and exhaust collars, see Figure 3.

4.2 Sealing system

We have carried out extensive research, development, workshop testing and on-site testing to
determine the best systems for sealing the main cables, cable bands, saddles and other con-

Figure 3. Dehumidification plant for main cables in buffer chamber and exhaust sleeve.

Figure 4. Applying elastomeric wrap with wrapping machine and bonding the wrap with heat
blankets.

NYBRIDGE_Book.indb 42 7/17/2013 8:41:40 AM

 Latest developments in suspension bridge main cable dehumidification 43

nected components. This has been supplemented by up to ten years of operation experience
with sealing systems installed on bridges with dehumidification of main cables. We have con-
cluded that the best system to seal the cable panels is the Cableguard™ Wrap System from the
D.S. Brown Company. This is an elastomeric wrap with a thickness of 1.1 mm and a width of
200 mm. It is applied with slightly more than 50% overlap, so the total thickness is 2.2 mm. It
is applied under tension with a special wrapping machine. After wrapping a section the wrap
is heat bonded with a special heat blanket, which melts the two layers together and shrinks
the material slightly, giving an even tighter fit. The wrapping and bonding work is illustrated
below in Figure 4. Special details have been developed to ensure sealing at the transition to
the cable bands and to give a uniform appearance.

4.3 Control and monitoring system

The control and monitoring system allows remote control/adjustment of the system and data
from the system documents that the system is performing properly and that the cables are
protected from corrosion. Instrumentation is arranged at the dehumidification plant(s), in
the buffer chamber(s), at monitoring sleeves and at injection and exhaust points. Key data
to be monitored includes system functionality, relative humidity, temperature, flow and pres-
sure. To further ensure monitoring of moisture condition inside the main cables, the absolute
water content in the air is calculated and monitored; see section 7.2 for further information.
Generation of a number of standard and custom graphs should be integrated in the moni-
toring system allowing an even more effective overview of the systems functionality. Results
from the monitoring system are essential to document the initial drying out process, the
corrosion protection of the main cables, possible development of leakage and the general
functionality of the dehumidification system.

5 CURRENT WORLDWIDE STATUS FOR MAIN CABLE SYSTEMS

5.1 Systems in service

There are currently 25 suspension bridges in a total of 7 different countries where dehumidi-
fication has been installed on the main cables. As far as we know these systems are generally
performing well, though some problems have been reported and suitable modifications have
been carried out. The development of dehumidification systems for main cables began in the
1990s with studies being carried out in both Japan and Denmark. The systems for main cables
were based on tried and proven technology as a natural extension of earlier bridge projects.
The first bridge with a full system for dehumidification of the main cables was the
Akashi-Kaikyo Bridge in Japan, the world’s longest suspension bridge with a main span of
1,991 m opened in 1998. The dehumidification system was integrated in the bridge’s design
and installed during construction. Based on the experience from the Akashi-Kaikyo Bridge
the operator Honshu-Shikoku Bridge Expressway Company Limited (HSBE) decided to
install dehumidification on the main cables of all their suspension bridges. To date they have
installed systems on 6 new suspension bridges and 9 existing suspension bridges in Japan. In
2008, after 10 years of service, the main cables of the Akashi-Kaikyo Bridge were opened and
wedged for inspection. The purpose of the inspection was to verify the effectiveness of the
dehumidification system. According to HSBE Newsletter No. 38, July 2009, the cables were
completely dry and the wires were in the same condition as at the time of construction.
The first suspension bridge outside of Japan to be retrofitted with dehumidification of main
cables was the Little Belt Suspension Bridge in Denmark with a main span of 600 m. The bridge
opened in 1970 and the dehumidification system was installed on the main cables in 2003.
Research and testing of sealing systems and on-site testing of injection and flow parameters
were carried out over several years before the final project was developed and tendered. This

NYBRIDGE_Book.indb 43 7/17/2013 8:41:42 AM

 44 Cable-supported bridges

system has been performing well for 10 years and no leakage has developed, as documented by
the monitoring system. The drying out process was also documented by the monitoring system
and took only a few days. The dehumidification plants have been inspected once a year and so
far the only maintenance has been the replacement of filters and one sensor.
Also during 2003 a dehumidification system was installed on the main cables (made up of
helical strands) of the 40 year old Aquitaine Bridge in France during a retrofit project where
the entire cable system was replaced. This system was based on the experience obtained from
the Little Belt Bridge project.
In Sweden there are two suspension bridges retrofitted with dehumidification. The
Högakusten Bridge has a main span of 1,210 m and the retrofit was carried out in 2005. Fur-
ther information is included in section 6.1. The Älvsborg Bridge has a main span of 414 m
and cables made up of helical strands. A dehumidification system with just one injection
point per main cable was installed in 2011. Further information is included in section 6.2.
In Great Britain there are three major suspension bridges and dehumidification systems
have been installed on the main cables of all three of these during 2007 to 2010. These are the
Forth Road Bridge opened in 1964 with a main span of 1,006 m, the Severn River Crossing
opened in 1966 with a main span of 988 m and the Humber Bridge opened in 1981 with a
main span of 1,410 m. The Humber Bridge was also earlier retrofitted with dehumidification
systems for the box girder and the tower saddles in the 1990s.
Dehumidification of main cables has also been installed on suspension bridges in China
and Korea. In China dehumidification of main cables was incorporated in the design of the
Rung Yang Bridge. In Korea dehumidification of main cables was incorporated in the design
of the Gwan Yang Bridge and installed on the existing Yong Jyong Bridge.
It should be noted that a dehumidification system does not provide full protection before
the initial drying out process has been completed. During this process the relative humidity will
be progressively diminishing and the rate of corrosion will be correspondingly decreasing.

5.2 Current projects and standards

At the current time there are at least 11 projects for dehumidification of main cables under
various states of progress as described below. Furthermore, authorities in two countries,
Japan and Norway, currently require dehumidification of main cables as a standard. As men-
tioned above HSBE in Japan requires dehumidification of main cables on all their suspension
bridges. In Norway the state authorities have required dehumidification of main cables in
their bridge design standard (Bruprosjektering, Normaler, Håndbok 185) since 2009. To the
best of the author’s knowledge the current status for main cable dehumidification projects
in progress is:
• USA: 4 retrofits planned
• Norway: 2 new bridges
• Denmark: 1 retrofit
• Qatar: 2 new bridges
• Italy: 1 new bridge
• Turkey: 2 new bridges
When these projects have been completed the worldwide total will be 37 bridges in 12 dif-
ferent countries.

6 EXAMPLES OF SYSTEMS

Dehumidification systems for main cables can be divided in two main categories:
• Systems for only the main cables
• Integrated systems that provide corrosion protection for one or more other bridge ele-
ments besides the main cables

NYBRIDGE_Book.indb 44 7/17/2013 8:41:42 AM

 Latest developments in suspension bridge main cable dehumidification 45

Dehumidification systems for main cables provide the lowest possible life cycle cost for the
main cables. Integrated systems provide even greater savings in the life cycle cost, as other
bridge elements are protected by the same system.

6.1 Systems for main cables

The Högakusten Bridge has a main span of 1,210 m and opened in 1998. The cables are
made up of parallel wires and have an outer diameter of 640 mm. Five years after the bridge
opening there were already serious problems with water intrusion and corrosion of cable
wires. A dehumidification system was installed on the main cables during 2005 and a massive
amount of water in the cables was dried out and the relative humidity brought down below
the corrosion threshold of 40% RH. The system has been performing well, though there has
been a problem with the transitional shrouds outside of the anchor chambers which has been
remedied, see section 7.6 for further information. An overview of the bridge and the layout
for the system are shown below in Figures 5 and 6.
The system includes 3 dehumidification plants, one in the upper cross beam of each pylons
and one in the box girder at the middle of the main span. These 3 locations act as buffer cham-
bers (see section 7.7) for the production of dry air. The buffer chamber in the box girder is suf-
ficiently large to perform optimally and provides savings by reducing the running time of the
dehumidifier by about 70%, giving savings in electrical consumption as well as reduced wear.

6.2 Integrated systems

The following two examples illustrate some of the possibilities for integrated systems. The
first example has a system that protects the main cables and the strands in the splay cham-
bers. The second system protects all the enclosed steel surfaces; main cables, strands in splay
chambers, box girders and tower.

Figure 5. Overview of Högakusten Bridge.

Figure 6. Layout of dehumidification system.

NYBRIDGE_Book.indb 45 7/17/2013 8:41:42 AM

 46 Cable-supported bridges

A retrofit of the 46 year old Älvsborg Bridge in Gothenburg, Sweden was completed in
2011, see Figure 7. The main cables are made up of helical strands and the strands are only
coated with paint, i.e., no galvanisation. The corrosion protection of the main cables needed
rehabilitation and a design study was carried out to determine the optimal method of reha-
bilitation. Dehumidification was chosen, as the study concluded that it was superior in all
aspects.
The integrated system encompasses the main cables and the strands in the anchorage
chambers. Part of the southern anchor house is enclosed as a buffer chamber and a dehu-
midification plant that serves the entire system is located here. Ducting connects the buffer
chamber with injection points at the middle of the main span. Dry air flows through the main
cables about 400 m in both directions and finally flows through the anchorage chambers. At
the southern end the dry air returns to the buffer chamber and is re-circulated, giving a highly
effective system.
The twin Lusail suspension bridges in Qatar are currently being constructed and have fully
integrated dehumidification systems, see Figure 8. The inner surfaces of all steel elements are
protected from corrosion by one integrated system with just one dehumidification plant. The
bridges each have two steel box girders, a circular steel tower and main cables made of strands.
A dehumidification plant is located in one box girder and it produces dry air and blows it
through one box girder, then through a duct at the end of the bridge connecting the two box
girders, through the other box girder and finally back to the plant through a duct at the other
end of the bridge. The entire collective volume of the box girders serves as a buffer chamber.
Part of the dry air in the box girders is injected into the tower from one side and circulates
through the tower to the opposite box girder. The air in the tower is slightly over pressured,
which causes a controlled amount of air to flow through the main cables, where it finally flows

Figure 7. Älvsborg Bridge, overview and system layout.

Figure 8. Fully integrated system on Lusail Bridges.

NYBRIDGE_Book.indb 46 7/17/2013 8:41:43 AM

 Latest developments in suspension bridge main cable dehumidification 47

out through the anchor chambers. In this manner the insides of the box girders and the towers,
the main cables and the cable strands in the anchorages are all protected from corrosion by just
one integrated system. This is a fully optimised and extremely economical system.

7 LATEST EXPERIENCE AND DEVELOPMENTS

7.1 Injection and flow testing

Injection and flow testing should generally be carried out when developing the layout for
a main cable dehumidification system. The purpose is to ensure a functional design of the
system with the lowest possible life cycle cost. The length of the flow sections should prefer-
ably be as long as possible in order to minimize the amount of equipment (dehumidifica-
tion plants, injection/exhaust sleeves, sensors, communication wiring and air piping). This
minimizes the cost of construction as well as maintenance costs, giving the lowest possible
life cycle cost. At the same time the acceptable maximum pressure of approximately 3,000 Pa
should also be held to avoid overloading the sealing system and minimize electrical consump-
tion. Hence injection and flow testing should be carried out with pressure in the range of app.
2,000 to 3,000 for main cables made up of parallel wires and from app. 1,000 to 3,000 Pa for
main cables made up of helical strands. The lower bottom pressure level for cables of helical
strands is due to the lower resistance (pressure loss) in this type of cable, which can allow suf-
ficiently long flow sections at a lower injection pressure then in cables of parallel wires.
The testing is relatively inexpensive to perform. The main equipment that is required
includes an interim injection sleeve, a fan, connecting hose and sensors to measure flow and
pressure, as illustrated above in Figure 9. When a constant pressure has been established at
each level, inspectors walk along the main cables in both directions and inspect for leaking
air. Soap water is sprayed on areas that are expected to leak and bubbles indicate air leaks and
that air is flowing to the actual position. The inspection is carried out progressively further
away from the injection point until no more leaks/bubbles are found. The last position with
signs of leakage is the maximum flow length for the test. This will be a conservative length,
as it must be assumed that the main cable will be much better sealed in connection with the
dehumidification works, hence the air will be able to flow somewhat longer.
The Great Belt Bridge in Denmark has main cables made up of parallel wires with an outer
diameter of 834 mm and has the 3rd longest main span in the world, 1,624 m. Testing was
carried out in April 2012 before developing the tender design. A test rig was installed close
to the center of the main span and testing was done at three pressure levels; 2,000, 2,500 and
3,000 Pa. Inspection of the main cable indicated flow lengths of respectively app. 320, 345
and 390 m. This was substantially longer than was expected and can be attributed to the main

Figure 9. Setup for injection and flow test and bubbles from escaping air at drain hole at cable band.

NYBRIDGE_Book.indb 47 7/17/2013 8:41:43 AM

 48 Cable-supported bridges

cable’s good air-tightness. Before the testing a layout with a maximum flow length of app.
210 m was proposed in accordance with the current best practice. The testing documented
that a substantially longer flow length was viable and a new layout was developed with a
maximum flow length of app. 280 m, which is still well below the test flow length of app.
320 m at 2,000 Pa. This update of the layout gives the following advantages:
• The design flow length is verified, ensuring that the system will function well.
• The low design pressure will minimize electrical consumption and leakage.
• The number of injection/exhaust sleeves with corresponding sensors, wiring and air pipes
has been reduced from 24 to 14.
• The number of dehumidification plants is reduced from three to one and this one plant
will actually be a modification of an existing plant in the bridge box girder.
All these changes give lower construction, operation and maintenance costs, i.e., a sub-
stantially lower life cycle cost. The difference between the outline design layout and the ten-
der design layout is illustrated below in Figure 10.
The Älvsborg Bridge in Sweden has main cables made up of helical strands. Injection
and flow testing was carried out in 2009 in connection with the outline design. In this case
it was planned that there should only be one injection point at the middle of the main span
for each main cable. From here the air should flow approximately 400 m in both directions
and exhaust in the respective anchor chambers, see figure 7. The testing was carried out at
2,000 Pa and trace gas was injected through the interim sleeve. Measurement of the trace gas
in the anchor chambers demonstrated that air could flow over this distance in both directions
and the flow time was about 20 minutes. Further, the main cable was reasonably air tight at
this pressure as only a small amount of minor leaks were discovered.

7.2 Improved monitoring with absolute water content

Monitoring of dehumidification systems has generally included relative humidity (RH) and
temperature. Relative humidity is defined as the ratio between the absolute water content and
the maximum possible water content at the actual temperature. Relative humidity is depend-
ent on the temperature, i.e., for a certain water content in a certain volume of air, the relative
humidity will vary according to the temperature. Warm air can contain more water than cold

Figure 10. Original layout at top, tender layout at bottom.

NYBRIDGE_Book.indb 48 7/17/2013 8:41:44 AM

 Latest developments in suspension bridge main cable dehumidification 49

air, hence when the air temperature increases the RH will fall and when the air temperature
decreases the RH will increase. The following examples illustrate this.
• RH 40% at 10°C—Temperature increases to 30°C, RH falls to 12%
• RH 40% at 10°C—Temperature decreases to app. -3°C, RH increases to 100% (condensa-
tion point)
In order to give improved understanding of moisture transportation inside the main cable
we have developed the use of monitoring absolute water content (AWC). AWC can not be
measured directly by sensors, but is calculated on the basis of a standard formula and the
measured values of RH and temperature. The advantage of AWC is that it eliminates the
uncertainties caused by fluctuations of RH due to temperature variation. When the tem-
perature at the injection and exhausts points are different (which often occurs) it is somewhat
incorrect to compare RH. Utilizing AWC eliminates this uncertainty and always presents truly
comparable values at these points. Further, the use of AWC allows the best possible and most
accurate monitoring of the drying out process. The monitoring system is programmed to
automatically generate graphs illustrating the ratio between exhaust and injection AWC for
each flow section. When the value of the ratio falls to one, the cable has been dried out. In this
way the drying out process will be clearly illustrated and it will be possible to roughly foresee
when each section of the cable will be dried out. The ratio may again increase to over one,
which indicates that water is again being removed from the cable. This can be due moisture
intrusion during periods with extreme weather or a quick fall in the temperature. When the
temperature falls quickly the dehumidification plant will produce air with a lower AWC, which
will replace air in the cable that has a higher AWC and the ratio will be over one for a period.
Monitoring of AWC on the Högakusten Bridge in Sweden has shown that the initial dry-
ing out period after commissioning was nearly 2 years, as illustrated below in Figure 11. The
long drying out period was accounted to two factors: 1. The cables were extremely wet to
start with, and 2. The water in the cables was frozen to ice during the long cold winters, slow-
ing down the drying out process. In more moderate conditions, main cables of parallel wires
usually dry out during about one half year.
The data from the monitoring also allowed a calculation of the amount of water removed
from the cables during the initial drying out. For a flow section of 300 m roughly 500 liters of
water were removed, corresponding to about 3% of the void volume.

Figure 11. Ratio between absolute water content in exhaust air and injection air.

NYBRIDGE_Book.indb 49 7/17/2013 8:41:45 AM

 50 Cable-supported bridges

7.3 Monitoring sleeves near anchor chambers

Thorough monitoring of the dehumidification system is essential in order to document its
effectiveness. This is quite straightforward for all points except at the exhaust flow in the
anchor chambers. At this point it is possible to measure the relative humidity and the tem-
perature, but it is not generally possible to measure the flow, as the exhausting air continues
to flow between the wires or strands and out into the anchor chamber. It is in some cases
possible to catch some of the exhausting air by building an enclosure around the strands
and this has been done e.g., on the Älvsborg Bridge and Högakusten Bridges in Sweden, see
Figure 12.
In order to improve the quality of the exhaust monitoring we have developed the con-
cept of a monitoring sleeve, which is placed on the main cable a short distance from the
anchor chamber. The monitoring sleeve includes sensors for relative humidity, temperature
and pressure. During the commissioning of the dehumidification system the relation between
flow and pressure can be accurately established and the monitoring system will then be able
to indirectly measure the flow close to the anchor chambers. In order to still monitor the
short stretch between the monitoring sleeve and the anchor chamber the sensors for relative
humidity and temperature are still applied inside the anchor chamber. This solution provides
better monitoring of the flow and saves the difficult and somewhat ineffective enclosures.
It has been applied on several projects which are either under construction or will soon be
constructed, such as the Great Belt Bridge and the Macdonald Bridge.

7.4 Pressure curve adjustment

Monitoring results of the exhaust air have shown that the humidity conditions inside the
cable vary somewhat during operation. Under perfectly ideal conditions, with a 100% airtight
cable and completely constant absolute water content (AWC) in the injection air, the exhaust
air should have the same AWC as the injection air. In the real world this is not the case, as the
cable can never be 100% air-tight and the AWC varies somewhat as the systems are controlled
by relative humidity, which gives variations according to variations in the temperature.
The overpressure inside the main cables along a given flow stretch varies more or less
linearly from a maximum at the injection point to zero at the exhaust point. The low over-
pressure in the area near the exhaust points makes these areas susceptible to intrusion of
moisture during extreme weather conditions such as high atmospheric pressure, hard rain
and high speed turbulent wind. Due to the height, exposure and locations of many suspen-
sion bridges, these types of weather conditions occur relatively often. In order to counteract
this effect we have developed a pressure curve adjustment solution. By applying a damper
in the exhaust pipe on the exhaust sleeve a small overpressure can be maintained instead of

Figure 12. Exhaust enclosures in anchor chambers, Älvsborg Bridge at left and Högakusten at right.

NYBRIDGE_Book.indb 50 7/17/2013 8:41:45 AM

 Latest developments in suspension bridge main cable dehumidification 51

falling to zero, and thereby better protect the sensitive areas from moisture intrusion. This
solution has been incorporated in our later projects and operation experience will soon be
forthcoming.

7.5 Sensitivity to leaks

Experience from field observations during operation of main cable systems as well as dur-
ing injection and flow testing indicates that main cables made up of parallel wires are not
especially sensitive to leaks, whereas main cables made up of helical strands are somewhat
more sensitive. The most likely explanation for this is that a main cable of parallel wires has
thousands of small voids over which the air flow is distributed. If there is a leak in the seal-
ing of the cable only the small amount of air flowing through the nearby voids will leak out.
This was clearly demonstrated during the flow testing on the Great Belt Bridge, as described
in section 7.1. Besides minor leaks at all the cable bands, there was a rather large opening at
the bottom of each band, an app. 20 × 100 mm drain hole as seen in figure 9. These relatively
large leaks, which are much larger than any leaks that will be found after sealing the main
cables, did not stop the air from flowing over long distances. This and similar experience from
other bridges has shown that the dehumidification systems on main cables of parallel wire
are very robust with regards to leakage development and that several relatively large leaks will
have negligible effects.
There is less experience from main cables of helical strands, as there are fewer suspension
bridges with this type of cable and also fewer with dehumidification systems. Up till now
only three suspension bridges with main cables of helical strands have dehumidification sys-
tems installed. Flow testing was carried out on the Little Belt Suspension Bridge before the
dehumidification system was designed. The results showed that air could easily flow from the
anchor chamber through the entire 170 m long back stay. Coming into the side span trace
gas indicated that air continued to flow to the first cable band (16 m further), but not to the
next cable band. The cable bands have a very large leak area, as the entire bottom joint in the
bands was left open. While the results don’t allow a complete conclusion, they do indicate
that much air has leaked out at the first cable band. Some air was most likely still flowing
deeper inside the cable. A flow test carried out on the Högakusten Bridge, which has main
cables of parallel wire and had open bottom joints in the bands as on Little Belt Bridge,
showed that these large leaks had very little influence. We can therefore conclude that cables
of helical strands are more sensitive to large leaks than cables of parallel wires. The dehumidi-
fication systems on main cables of helical strands are however still robust, as it would require
one or more large leaks to seriously affect the system. These leaks would be indicated by the
monitoring system and as they would be large, it would be easy to locate and repair them.
Further, leaks of this size would most likely be caused by accidental actions and therefore
known to the bridge operator, meaning that they would be repaired shortly after occurring.
Experience from the Little Belt Bridge also documents the robustness of a dehumidification
system on main cables of helical strands. The dehumidification system has been in service for
10 years and the monitoring system documents that it is functioning well with no discernible
leakage development.

7.6 Sealing details

The sealing of the main cable panels with elastomeric wrap is quite straightforward to exe-
cute and very reliable. The sealing of other details, such as the covers on saddles, transitional
shrouds and cable entrance to the saddles and anchor chamber, is much more complicated.
Experience from numerous suspension bridges has shown that these types of details gener-
ally allow water or moisture intrusion with corrosion problems to follow. An upgrade of the
sealing of these details should be included in the retrofit design for dehumidification of main
cables.

NYBRIDGE_Book.indb 51 7/17/2013 8:41:46 AM

 52 Cable-supported bridges

The dehumidification system on the main cables of the Högakusten Bridge in Sweden
was installed before we had sufficient experience to realize the importance of these sealing
details. It was assumed that the transitional shrouds adjacent to the anchor chambers would
be sufficiently watertight in connection with the dehumidification system. Operation experi-
ence via the monitoring system and observations in the anchor chamber indicated that the
shrouds were not sufficiently tight and that water was still coming in through the joints in
the shrouds. The shrouds are located at the end of the flow section, where the overpressure
has fallen to about zero and is insufficient to prevent water or moisture intrusion. Therefore
it is especially important that the shrouds are fully sealed. External inspections could not
reveal where the leaks were occurring, so it was not possible to try and upgrade the sealing.
Further, the detailing of the joints was not suited for upgrading, as there were no grooves
for caulking. Therefore it was decided to remove one of the four shrouds and replace it with
a newly designed shroud. This was done in connection with the warranty works in 2010.
When the shroud was removed, discoloring and trace marks clearly indicated where water
was coming in, primarily at the connection to the cable band at the upper end and secondar-
ily through the longitudinal joints, see Figure 13. This condition made it difficult to evaluate
the effectiveness of the system in the side spans, as water was at times entering the cables just
before the monitoring point. This has however not generally affected the drying out process,
as this has been documented by monitoring data from the cables in the main span.
The new shroud completely solved the problem with water intrusion and it was decided
in 2012 to replace the remaining 3 shrouds with the new version. The design of the new
shrouds was based on experience from the design of the injection and exhaust sleeves from
earlier projects, as well as full-scale testing carried out in connection with the construc-
tion of Hardanger Bridge in Norway. Details at cables bands, the tower saddles including

Figure 13. Inside of bottom half of original shroud at left, new tight shroud at right.

Figure 14. Full-scale testing of sealing, left saddle and shroud, right cable band and transition at band.

NYBRIDGE_Book.indb 52 7/17/2013 8:41:46 AM

 Latest developments in suspension bridge main cable dehumidification 53

adjacent transitional shrouds, the cable bands and the transitional shrouds at the anchor
chamber were all tested and adjusted to achieve maximum sealing, see Figure 14. All joints
are designed with a double sealing system; an inner sealing of compressible foam neoprene
that fills out any unevenness and an outer sealing with caulk applied in a groove of a fit-
ting size.

7.7 Buffer chamber solution

In 1995 we developed the buffer chamber solution in connection with dehumidification of
the tower saddles on the Humber Bridge in England. There were problems with water ingres-
sion in the saddles and it was decided to dehumidify these as well as the main cables in the
vicinity of the saddles. This would require constant injection of dry air in the saddles, as the
dry air would flow into the main cables and eventually disappear through leaks in the existing
sealing system. It would be relatively expensive in electrical consumption if the dehumidifica-
tion unit was to run constantly, so the buffer chamber solution was developed. The air com-
ing directly from the dehumidification unit has a very low relative humidity, just over 0%,
which is much drier than necessary. Therefore a tower leg was utilized as a buffer chamber,
where the air from the dehumidification unit is mixed up with ambient air to about 40% RH
before injection in the saddles. In this manner the electrical consumption was significantly
reduced, as well as wear on the dehumidification unit. This was also the first application of
dehumidification on main cables, although the extent of the protection is not documented
and it was applied before full systems for main cables were developed and implemented. Dur-
ing the dehumidification retrofit design of the main cables in 2009 it was decided to integrate
this existing system in the complete system for the main cables and it has been operating as
part of the complete system since 2010.
Following this successful application the buffer chamber solution has been integrated in
all our main cable dehumidification projects. An existing structure is usually utilized for the
buffer chamber. The volume of the available existing structures varies, so it has not in all
cases been possible to have a chamber of optimal size. For example there are buffer chambers
in the upper cross beams of the pylons on the Högakusten Bridge, which provide an energy
savings of roughly 50%. On the same bridge there is also a buffer chamber with a much larger
volume arranged in part of the box girder, which gives an energy saving of roughly 75%.
Optimally the buffer chamber should be large enough to give maximum savings in electrical
consumption. The buffer chamber solution has a further advantage, as the dehumidification
plant is protected from the elements when it is located inside the chamber, which minimizes
maintenance requirements.

8 CONCLUSIONS

There is a vast amount of evidence that proves that traditional corrosion protection systems
for main cables do not prevent corrosion. Dehumidification is the only method that provides
such a high level of main cable corrosion protection. This can be achieved if the atmosphere
inside the cable is generally kept below the threshold level of 40% RH. Dehumidification sys-
tems for main cables are a natural extension of technology that has been successfully applied
to other steel bridge elements for over 40 years. Development of systems for main cables
began in the 1990s and the first full-scale application was in 1998. There is currently 15 years
experience with systems for main cables with 25 systems in service in 7 different countries.
Furthermore, there are projects for at least 11 more bridges under way, which will bring the
total up to 36 bridges in 12 different countries.
Corrosion protection of main cables by dehumidification is virtually a worldwide standard
and all suspension bridge owners should be planning to install systems to ensure the safety of
their bridges as well as reducing the life cycle cost.

NYBRIDGE_Book.indb 53 7/17/2013 8:41:47 AM

 54 Cable-supported bridges

REFERENCES

Bloomstine, Matthew L., 2011, Main Cable Corrosion Protection by Dehumidification—Experience,

Optimization and New Development, 6th NYC Bridge Conference, NY, USA.
Bloomstine, Matthew L. et al, 2005, State-of-the-Art Main Cable Corrosion Protection by Dehumidifi-
cation, 3rd New York City Bridge Conference, New York, USA.
Bloomstine, Matthew L. et al, 2006, Main Cable Corrosion Protection by Dehumidification, IABSE
Conference on Operation, Maintenance and Rehabilitation of Large Infrastructure Projects, Bridges
and Tunnels, Copenhagen, Denmark.
Bloomstine, Matthew L. et al, 2006, Prevention of Main Cable Corrosion by Dehumidification, 5th
International Cable Supported Bridge Operators’ Conference, New York, USA.
Bloomstine, Matthew L. et al, 2008, The Högakusten Bridge—Corrosion Protection of the Main Cables
and Maintenance of Major Components, 6th International Cable Supported Bridge Operators’ Con-
ference, Takamatsu, Japan.
Bloomstine, Matthew L. et al, 2010, Dehumidification of Suspension Bridge Main Cables, The 5th
International Conference on Bridge Maintenance, Safety and Management (IABMAS), Philadelphia,
USA.
Furuya, Kazuhiko et al, 2000, Corrosion Mechanisms and Protection Methods for Suspension Bridge
Cables, Structural Engineering International 3/2000, 189–193.
HSBE Expressway Company Ltd., The Akashi Kaikyo Bridge: Verification of Rustproof on the Cables
by the Dry Air Injection System, Newsletter on Long-Span Bridges, No. 38, July 2009.

NYBRIDGE_Book.indb 54 7/17/2013 8:41:47 AM

 Chapter 5

Design, analysis and construction of Puente Mercosur—

third bridge across the Orinoco River

K. Humpf & M. Müller

Leonhardt, Andrä und Partner Beratende Ingenieure VBI AG, Stuttgart, Germany

C. Silingardi
Odebrecht Venezuela

ABSTRACT: With the aim of developing the Southwest of Venezuela and enable its
connections with the rest of the country, a third bridge over the Orinoco River was designed
between the cities of Cabruta, State of Guárico in the north and Caicara del Orinoco in the
south. The combined highway and railroad bridge Puente Mercosur will be the third crossing
of the Orinoco with a total length of 11.125 km. The stay-cable main bridge with 360 m main
span and the adjacent approach bridges with a length of 2 × 720 m are comprised of a steel
composite truss girder with a depth of 12.0 m.

1 INTRODUCTION

1.1 General
A general comparison of Venezuela with Germany shows that the most important crude oil
exporting nation in South America has, based on a relatively low level, a strong growth in
population and as well in the gross national product (Table 1).
This growth combined with a gas price of just 0.5 USD/Gallon conducted to a very strong
increase in traffic. Supported by steadily high revenues from crude oil exports, the state initi-
ated major infrastructure projects, to improve the road network and as well the railroad con-
nections. In this context the bridges over the Orinoco have a special importance due to the
extraordinary size of the structures.

1.2 Relevancy of the 3rd crossing for the infrastructure network

The Orinoco with a total length of 2560 km (for comparison Rhine River 1320 km, Danube
2850 km) is the longest river in Venezuela and behind the Amazon and Paraná the 3rd long-
est in the Latin America (Figure 1).
Until now there are just 2 crossings of the river. The first one is a suspension bridge in
Ciudad Bolivar inaugurated in 1967 and called Puente Angostura with a total length of
1272 m and a main span of 712 m it serves for road traffic only (Lustgarten) (Figure 2).

Table 1. Comparison of Venezuela with Germany.

Population Gross national product

Area
[m2] 106 Citizen/km2 Growth [%] Per Citizen [USD] Growth [%]

Venezuela 916.500 29,3 30 1,6 12.600 ∼8

Germany 357.000 82 230 –0,05 42.000 ∼1,0
Ratio 2,57 0,36 0,13 – 0,30 8

55

NYBRIDGE_Book.indb 55 7/17/2013 8:41:48 AM

 56 Cable-supported bridges

Figure 1. Map of Venezuela.

Figure 2. Suspension Bridge.

The 2nd one downstream from the 1st one is a crossing close to Ciudad Guayana, completed
in the 2006. This system consists of two back to back cable stayed bridges with main spans of
300 m a total length of 3156 m and the deck carrying rail—and roadway traffic, known under
the official name Puente Orinoquia (Saul, Lustgarten, Humpf, et al., 2006) (Figure 3).
The Third Bridge across the Orinoco is located about 300 km upstream of the first bridge
in Ciudad Bolivar and about 400 km of the second bridge in Ciudad Guayana. The official
name will be Puente Mercosur after completion and it is a combined highway-railway bridge
which will connect the two villages of Caicara del Orinoco in the South and Cabruta in the
North.
In order to start developing the unexploited central part of the country an investment of
about one billion dollars was designated for the construction of the third bridge across the
Orinoco River, a project of the Bolivarian government with the Brazilian construction com-
pany Odebrecht. South to this structure a new city shall be developed around the industry of
producing aluminum. Then this investment in a fixed crossing will reach its importance; for
the moment and the time of completion it’s a structure in the nowhere.
Besides the construction of the river crossing building the roadway between Caicara del
Orinoco and Ciudad Bolivar and between Cabruta and San Fernando de Apure will become
a necessity. The railway connection on the northern side is currently under construction.
Construction of the bridge started in 2006.

NYBRIDGE_Book.indb 56 7/17/2013 8:41:48 AM

 Design, analysis and construction of Puente Mercosur 57

Figure 3. Cable stayed Bridges.

2 DESIGN

2.1 Description
This combined Rail and Road Bridge has a total length of 11.125 km crossing the Orinoco
River and its floodplain. The link consists of the river crossing and approach viaducts on
either side. The 2280 m long double deck river crossing is composed of an 840 m long stay-
cable main bridge and two 720 m long approach bridges (Figures 4, 5 & 6).
The deck for the combined Rail and Road Bridge comprises a steel composite truss girder
whereby the 4 lane road traffic is located on the 19.9 m wide upper deck and the rail is located
inside the truss on the lower deck. The depth of the truss is constant 12.0 m along the 2280 m
long river crossing. A cable-stayed main bridge with 360 m main span and central cable plane
is designed for the crossing of the 320 m wide navigation channel. The side spans of the cable-
stayed bridge as well as the spans of the double deck approach bridges have regular lengths
of 120 m. The cable consist of pairs of tendons made up of 41 to 61 strands d = 0.6" with a
tensile strength of 1770 MPa. The corrosion protection consists of the typical 3 independent
layers of protection and a wedge type anchoring system with highest fatigue resistances. Two
thirds of the cables are furnished with hydraulic dampers at the lower end.
The substructures were designed to resist the ship impact and earthquake loads (Saul,
Humpf, et al., 2003; Saul, Svensson 1981). The diamond shaped concrete pylons have a
height of 135 m and are founded on 39 piles with a diameter of 2.50 m. The piers of the
double deck river crossing are founded on 18 piles with a diameter of 2.00 m. Those piles
have lengths of up to 80 m.
At the transition from the river crossing to the approach viaducts for road traffic a bifurca-
tion structure is provided which separates the road and rail bridges (Figure 7).
After the bifurcation structure follows a 940 m long approach viaduct for the road traffic.
No ship impact has to be considered in this area, wherefore regular spans of 60 m could be
chosen. The composite deck is composed of two longitudinal steel girders with a depth of
4.4 m and a 30 cm thick slab. The cross girders are spaced at 3.0 m. Between the lower chords
of the longitudinal steel girders a horizontal bracing is added (Figure 8).
The pier shapes with their distinctive H-Form result from the erection process of the dou-
ble deck trusses at a lower level than finally launching the A2 m superstructure at a higher
gradient level of the road bridge A2M (Figure 9).
In the alluvial plains the approach viaducts for the road traffic consist of a concrete deck
with regular spans of 35 m. The deck of the 3885 m and 3080 m long northern and southern
approach viaducts A2 is composed of five prefabricated U-shaped channels with a depth of
2.15 m and a 18 cm thick slab. The channels are prestressed with ½" seven wire strands and
the top slab of the superstructure is cast-in-place, continuous over 140 m (Figure 10).
The 140 m sections of the approach viaducts are coupled by Shock Transmitter Bars. In
the event of a design earthquake the Shock Transmitters at each end of the 140 m bridge sec-
tions lock up. When locked, the approach viaduct becomes, in effect, a multispan continuous

NYBRIDGE_Book.indb 57 7/17/2013 8:41:49 AM

 58 Cable-supported bridges

Figure 4. Visualization of main Bridge.

Figure 5. General layout of main Bridge.

Figure 6. Superstructure, Truss girder of cable-stayed Bridge and approach Viaduct A1.

girder (hyperstatic) structure. The chosen concept allows for thermal and other service-load
movement without engaging the Shock Transmitter Bars. Modular joints and pier size can be
kept to a minimum not governed by the earthquake situation. The earthquake design of the
Shock Transmitter Bars has been performed by a non-linear time history analysis consider-
ing 10 artificial accelerograms, which are compatible with the design spectrum.
All other joints between truss girders and approach viaducts of 60 m spans are designed to
allow the free movements from seismic excitations.

2.2 Design parameters

First feasibility studies date back to 2005 and through a process of pre-qualification and
tendering the construction company Odebrecht, Brazil, qualified for signing a contract in the

NYBRIDGE_Book.indb 58 7/17/2013 8:41:49 AM

 Design, analysis and construction of Puente Mercosur 59

Figure 7. Transition from double deck to individual structures.

Figure 8. Cross section approach Viaducts A2M.

Figure 9. H-Form piles at A2M.

year 2006 with the development organization “Propatria 2000” for the design and construc-
tion of this bridge.
The basic design parameters are:
− Minimum horizontal clearance of 320 m and minimum vertical clearance of 40 m in that
navigation channel
− Impact loads due to ship collision have been considered in the design of the main structure
(Saul et al., 1981) (Saul et al., 2003)
− Varying water levels between 0 and +12.5 m
− Scour up to 12 m in the area of the CSB and Approach Viaduct A1
− Maximum slope for rail traffic 1.5%
− 4 lanes for roadway and 1 railway line
− Steel in weathering steel A-588 without coating.

NYBRIDGE_Book.indb 59 7/17/2013 8:41:50 AM

 60 Cable-supported bridges

Figure 10. Cross section Approach Viaducts A2.

The basic codes applied are:

1. AREMA (2002), Manual for Railway Engineers. American Railway Engineering and
Maintenance-of-Way Association.
2. AASHTO (LRFD) Bridge Specifications (SI UNITS 1998). American Association of
State Highway and Transportation Officials.
3. PTI Guide Specif. ‘Recommendation for Stay Cable Design, Testing & Installation’,
4th Ed.
4. COVENIN 2003-89—Norma Venezolana Acciones del Viento sobre las Construcciones.
Fondonorma, Caracas, Venezuela for the local wind loads (COVENNIN 2003).
5. Norma Venezolana para el Diseño Sismoresistente de Puentes, División de Estructuras,
M.T.C.
Train loads are given as COOPER E80 and the truck loads as HS20-44. Less than 10 trains
will pass the bridge per day. However these provide, except for the unloaded ones, generally
the design load intensity. Accidental load cases considered are one severed cable under full
service loading and seismic acceleration up to 0,21 g. All sections were sized according to
allowable stresses and for the stability checks of the plates of the box section verifications
according to International codes were provided.
In the fatigue analysis a service life of 80 years is considered. The combination of rail and
roadway traffic on one structure is not unusual in South America; however, the US codes
AREMA and AASHTO do not reflect such a combination.

2.3 Site geology and foundations

The important water depth and thick layers of soft soils make a pile foundation mandatory in
many South American rivers. Therefore, the construction of long piles with great diameters
is closely connected especially to the bridges in South America.
The site geology of the Third Bridge over the Orinoco is composed mostly by recent depo-
sition of alluvial sediments under which there are deposits of sedimentary “Las Piedras”
formation, with some very punctual and sporadic outcrops of rocks associated with the
dynamics of the Orinoco River.
Within the river bed, the first 20 meters are made up of clean sand (SP) and silty sand
(SM) with a consistency between loose and dense. In alluvial plains outside the actual river
bed, the surface soil layers are found in high plasticity clays (CH) on a stratum of low plastic-
ity clay (CL) of variable consistency, interspersed with deposits of sand.
At a depth of 20 meters, the soils have generally a higher consistency (very dense sand,
hard clay), becoming more competent in terms of strength and bearing capacity.

NYBRIDGE_Book.indb 60 7/17/2013 8:41:52 AM

 Design, analysis and construction of Puente Mercosur 61

3 CONSTRUCTION

3.1 General
From the first stages of design the potential construction methods had played an important role.
A construction site at a river with variable water depths, swamp areas with dry and with wet
seasons and in a completely remote place with very limited existing infrastructure asked for a
concept which was feasible to be built under these special conditions.
Almost 6 of 12 months of a year, piling and pile cap construction had to be interrupted
due to high water levels in the river and as the water level varies between +21 mamsl and
+33.5 mamsl. The execution of the 1141 piles started in 2007 and will take about 3½ years. By
the help of GPS systems the piles could be placed horizontally with tolerances in the range of
centimeters. Piles will have a length up to 80 m and in order to confirm the necessary length
of the piles load tests had been performed.
The bottom scaffolding of the pile caps consists of lost precast panels and the pour of the
cap has been done in two or three stages.
Only the towers will be built with climbing formwork, all piers with slip forms. The reason
for this unusual approach lies in the impact from very strong unions upon the normal produc-
tion rate slowing down the process with the climbing formwork. However the unions have to
surrender to the technical working conditions with a slip form and work here 7 days 24 hours
(Figures 11 & 12).

3.1 Assembling
Preassembled portions of the superstructure segments will be shipped from several shops in
Ciudad Guayana to a temporary assembling yard at the Orinoco River in Ciudad Guayana.
Here the main parts of the superstructure (bottom flange and top flange) are assembled by add-
ing the diagonals of the trusses to a complete superstructure segment of 60 m (Figure 13).

Figure 11. Tower construction.

NYBRIDGE_Book.indb 61 7/17/2013 8:41:52 AM

 62 Cable-supported bridges

Figure 12. Pier construction with slip forms.

Figure 13. Assembling in “Alianza”.

Upon request of the fabricators the welded connections to the top—and bottom flanges
had been changed to a bolted solution with certain advantages but major disadvantages.
However, this was as well driven by a combined welder/union problem, similar to the US.
These superstructure segments will be shipped with huge barges from Ciudad Guayana
to the construction site in Caicara at 400 km distance. In Caicara del Orinoco the segments
will be transported from the Orinoco River to the assembly yards behind the northern and
southern abutment and joined together to superstructure sections with a length of 360 m. At
low water levels (6 out of 12 months) no barge reaches Caicara del Orinoco. The delivery of
the superstructure parts has to be scheduled accordingly during the other months.

3.2 Launching
Behind the northern and southern abutment an assembling yard will be prepared to assemble
2 x 3 truss girder sections with a length of 360 m each and a dead weight of up to 9000 ton
(Figure 14).

NYBRIDGE_Book.indb 62 7/17/2013 8:41:53 AM

 Design, analysis and construction of Puente Mercosur 63

Figure 14. Launching of truss girder sections.

Figure 15. Bottom view of strand jacks.

Figure 16. Schematic view of launching system.

The truss girder sections will be slid by means of Teflon plates on launching girders into its
final position. A pair of strand jacks will be connected to each section and pull along 120 m
long static strand cables anchored to the top of forward piers to launch the deck sections
forward. After each 120 m of launch the two strand cables are winched forward to allow the
launch to continue.
For the construction of the Second Orinoco Crossing in Puerto Ordaz the same launching
procedure has been realized. A launch of 120 m was generally carried out in a single 10 hour
shift (Figures 15 & 16).

3.3 Heavy lift

After the launched truss girder section of 360 m has reached his final position the first two
pair of cables will be installed in the main and side span to compensate the cantilever defor-
mation under self weight. The concrete deck will be poured step-by-step beginning in the axis
of the pylons. After the concrete has achieved the required strength the corresponding cables
will be installed and tensioned.

NYBRIDGE_Book.indb 63 7/17/2013 8:41:54 AM

 64 Cable-supported bridges

Figure 17. Construction of cable-stayed Bridge.

The centre of the main bridge with a length of 120 m and a dead weight of 2400 ton will
be lifted from the barge by four strand jacks positioned on the existing cantilevers. In a next
step the concrete deck of the center span will be poured in various steps, the corresponding
cables installed and tensioned strand-by-strand (Figure 17).

4 CONCLUSIONS

The paper describes the design and construction of an extraordinary connection across the
Orinoco River as a pioneering investment to develop the whole area. Difficulties in the logis-
tics to construct in such a fairly remote place of the country are effectively solved within the
Design Build approach in close cooperation of designers and contractor.

ACKNOWLEDGEMENTS

Client: Ministerio de Infraestructura, Fundación Propatria 2000,

Venezuela
Construction: Construtora Norberto Odebrecht S.A., Venezuela
Preliminary & Basic Design: Leonhardt, Andrä und Partner, Stuttgart, Germany
Detailed Design: Consorcio BRAVE, Caracas, Venezuela and Sao Paulo,
Brazil
Construction Engineering: Leonhardt, Andrä und Partner, Stuttgart, Germany
Independent Checking: Leonhardt, Andrä und Partner, Stuttgart, Germany
Smoltczyk und Partner, Stuttgart, Germany (Geotechnics)

REFERENCES

Lustgarten, P.: Puente Angosturita sobre el Rio Orinoco–Boletin N° 11 de la Asociación Venezolana de

Ingeniería Estructural.
Saul, R., Lustgarten, M., Humpf, K. 2006. The Orinoco-Bridge in Ciudad Guayana/Venezuela—a
double-cable-stayed bridge with composite girders for rail and roadway traffic. Stahlbau 75: 82–92.
Saul, R., Humpf, K. et al: 2003. Innovative Protection against Ship Impact for the Rosario-Victoria
Bridge over the Paraná (Argentina). Stahlbau 72: 469–484.
Saul, R., Svensson, H. 1981: For the Pier bridge protection against ship impact, Die Bautechnik 58:
326–335, 374–388.

NYBRIDGE_Book.indb 64 7/17/2013 8:41:55 AM

 2 Bridge analysis & design

NYBRIDGE_Book.indb 65 7/17/2013 8:41:56 AM

This page intentionally left blank
 Chapter 6

Development of economical short-span steel bridge standards

K. Barth & G. Michaelson

West Virginia University, Morgantown, WV, USA

M. Barker
University of Wyoming, Laramie, WY, USA

D. Snyder
Steel Market Development Institute, Washington, DC, USA

ABSTRACT: A considerable number of bridges in the United States are either structurally
deficient or functionally obsolete. Consequently, one of the main focuses of the bridge indus-
try is on design techniques that provide a high-quality and cost-effective product in a relatively
short amount of time, with less maintenance required and a longer life span. Standardized
bridge plans significantly reduce design time, increase and provide cost-effective solutions in
the nation’s effort to repair and upgrade the infrastructure. Therefore, in this study, standard
short-span steel bridge designs were developed to create a design aid (eSPAN140) for bridge
engineers. In these designs, bridges with spans ranging from 40 feet to 140 feet in 5 foot
increments were developed for rolled beam sections, homogeneous plate girder sections and
hybrid plate girder sections. The rolled sections were designed using two design approaches:
the lightest weight possible and the lightest weight possible with a limited section depth.

1 INTRODUCTION

There are a large number of bridges in the United States that are considered structurally
deficient or functionally obsolete. In response to the deteriorating infrastructure, the Federal
Highway Association (FHWA) has introduced an initiative titled Highways for LIFE in an
effort to help in reducing these issues. This FHWA focus area promotes the development of
bridge design and construction that leads to Long lasting bridges that are Innovative, have
Fast construction times and are economically Efficient. This research, performed in conjunc-
tion with the Short Span Steel Bridge Alliance (SSSBA) of the American Iron and Steel
Institute (AISI), takes these principles and looks into methods of increasing the efficiency
of steel girder bridge design through the use of stockpiled common steel plate sizes and a
limited suite of rolled steel girders.

2 OVERVIEW OF PREVIOUS STANDARD DESIGNS

In this study, other sets of state bridge design standards were investigated. Oklahoma, Texas
and Virginia each provided sets of standard designs of steel girder bridges within the bridge
span ranges investigated.
• Oklahoma had one set of steel girder designs for bridges with span lengths between 30 and
100 feet, roadway width of 40 feet and a girder spacing of 11 ft.–10 in.
• Texas has three sets of standard girder designs with bridge span lengths between 30 and
120 feet. Each of these sets has a different overall roadway width and girder spacing:

67

NYBRIDGE_Book.indb 67 7/17/2013 8:41:56 AM

 68 Bridge analysis & design

24 foot roadway width with 7 ft. – 4 in. spacing, 28 foot roadway width with 8 ft. – 8 in.
spacing and 30 foot roadway width with 7 ft. spacing.
• Virginia had a large design aid package of pre-designed steel girder bridges that have
become outdated. This design package considered a wide variety of bridge span lengths,
girder spacings, roadway widths and bridge skew angles.
In addition, AISI published a series of standard designs for short-span steel bridges in
1994. These standards served as a benchmark for comparisons with the suite of girders
designed in this study.

3 SCOPE OF WORK

The goal of this effort has been to develop a set of standardized designs that increase the
design efficiency of short span steel bridge designs. The standardized designs were developed
based on optimized girder designs, which employ different bridge parameters and design
approaches. There are four major sets of bridge designs in this work: “limited depth” rolled
beam sections, “lightest weight” rolled beam sections, homogeneous plate girder sections
and hybrid plate girder sections. From these optimized rolled girder designs, limited suites
of rolled steel girder sections can be selected to investigate the efficiency of using stockpiled
girder sections for short span steel bridges. Also, the benefits of stockpiling common steel
plate sizes are investigated in the design of steel plate girders.
The scope of this work was to develop optimized steel girder designs for bridges with spans
between 40 and 140 feet. The girders designed to make up this wide range of bridge spans
were designed for all spans between 40 and 140 feet in 5 foot increments. To develop a wide
variety of steel girders that encompass the different bridge design parameters and practices
of practicing bridge engineers, four different girder spacings and four different girder design
approaches were investigated. Based on the designs developed for the different bridge spans,
girder spacings and design approaches, an analysis of efficiency gained from using stockpiled
common steel plate sizes and available rolled sections was performed.

4 DESIGN PROCEDURE

The short-span steel girders in this effort were designed in accordance with the 5th Edition
of the AASHTO LRFD Bridge Design Specifications (2010), and evaluated using Version
6.5 of MDX’s Line Girder Rating Software (2009), a popular steel girder design and rating
package used by many state DOTs in the United States.

4.1 Rolled beam designs

The rolled beam sections were designed using two different design approaches, termed “lim-
ited depth” and “lightest weight”. The “limited depth” rolled girder sections were designed
to meet a target L/D (Length/Depth) ratio of 25. Wide-flange sections of the given depth
were evaluated until the most economic section for the given span length and girder spacing
was found. The “lightest weight” rolled girder sections were developed in the same manner,
however without the restriction on the L/D ratio.

4.2 Plate girder designs

The plate girder sections were designed using two different material configurations: homoge-
neous and hybrid. For both material configurations the L/D ratio was used to determine the
dimensions of the web plate. The compression and tension flanges were selected to create the trial
section to begin the evaluation process. Based on the evaluation of the section, dimensions of the
flange plates were modified to find a girder section that was both adequate and economic.

NYBRIDGE_Book.indb 68 7/17/2013 8:41:56 AM

 Development of economical short-span steel bridge standards 69

In designing the steel plate girder sections, a limited selection of common steel plate
dimensions were used to take advantage of stockpiling materials. In addition, to account for
flame cutting/torching of plates, all plate depths/widths selected for design were reduced by
1/4 inches during design. The following dimensions were employed for the steel plates:
• Web plates:
− Depths: 24 in., 32 in., 40 in., 46 in., 48 in. and 54 in.
− Thicknesses: all web plates are 1/2 in. thick.
• Flange plates:
− Widths: 12 in., 14 in., 16 in., 18 in. and 20 in.
− Thicknesses: 3/4 in., 1 in., 1 1/2 in. and 2 in.
A typical girder elevation is shown in Figure 1, where L is the span length, C represents the
cross-brace spacing and the lengths of the bottom flange transitions are presented. Interior
girders were designed for the girder spacing arrangements of 6 feet, 7 feet – 6 inches, 9 feet
and 10 feet – 6 inches. In the designs, it was assumed that there were 5 girders in the bridge
system and that the bridge deck consisted of 3 lanes. The typical interior girder cross-section
layout is shown in Figure 2, and the typical bridge cross-section layout is shown in Figure 3.
Full composite action between the designed steel girder sections and the concrete slab was
assumed to develop through the use of headed shear studs.

4.3 Common design parameters

The rolled beam sections and the homogeneous plate girder sections in these designs employ
50-ksi steel. The hybrid steel plate girder sections have 50-ksi steel in the compression flange

Figure 1. Typical plate girder elevation.

Figure 2. Typical interior girder cross-section.

NYBRIDGE_Book.indb 69 7/17/2013 8:41:56 AM

 70 Bridge analysis & design

Figure 3. Typical bridge cross-section.

and web plates and 70-ksi steel in the tension flange plate. For all girder sections, excluding
the rolled beam sections of the “lightest weight” suite of girders, an L/D (Length/Depth)
ratio of 25 was assumed. The depth in this ratio includes the entire depth of the bridge super-
structure (i.e., bridge deck depth plus the concrete haunch thickness plus the girder depth).
The concrete haunch is defined as the distance from the bottom of the compression flange to
the bottom of the concrete deck.
The following parameters were assumed for each bridge girder design:
• Steel stay-in-place (SIP) formwork unit weight: 15 psf
• Future wearing surface: 25 psf
• Concrete barriers: 305 lbs/ft.
• Miscellaneous steel weight increase: 5%
• Compressive strength of concrete: 4,000 psi
• Concrete unit weight: 150 pcf
• Steel unit weight: 490 pcf
• Concrete haunch thickness: 2 in
• Constant flange width
• Constant web height

5 RESULTS OF DESIGNS

5.1 Comparison of girder designs

Figure 4 shows a comparison of the design results of the four previously mentioned design
methods (homogeneous plate, hybrid plate, limited depth, and lightest weight) for a given
girder spacing.
As shown in the higher span ranges, the economy of rolled beam solution is diminished. This
is due to the discrete number of rolled beams available; in the higher span ranges, the discrete
range of rolled beams causes the weight of the girders to increase whereas, for plate girders,
the sizes of the individual plates can be tailored to meet a given span requirement. Therefore,
in the final set of solutions selected study, rolled beam solutions are only provided for span
lengths from 40 feet to 100 feet. For plate girder solutions, homogeneous girders are provided
for span lengths from 60 feet to 140 feet and hybrid girders are provided for span lengths from
80 feet to 140 feet; these ranges were selected by the members of the SSSBA technical working
group (a group of fabricators, engineers, plate producers, service centers, and researchers within
SSSBA) to deliver the most economical solutions possible from the suite of designed girders.

5.2 Results of rolled beam designs

Tables 1 and 2 show the selected rolled beam sections for the lightest weight and limited
depth configurations, respectively. The tables provide a selected rolled shape for each 5 foot

NYBRIDGE_Book.indb 70 7/17/2013 8:41:57 AM

 Development of economical short-span steel bridge standards 71

Figure 4. Comparison of design alternatives (9'-0" girder spacing).

Table 1. Lightest weight rolled beam designs.

Girder spacing
L Selected
[ft] 6'-0" 7'-6" 9'-0" 10'-6" section

40 W21 × 62 W21 × 73 W24 × 76 W24 × 84 W24 × 84

45 W24 × 68 W21 × 101 W27 × 84 W30 × 90 W30 × 90
50 W27 × 84 W21 × 111 W30 × 99 W30 × 108 W30 × 108
55 W30 × 90 W24 × 117 W30 × 116 W33 × 118 W33 × 118
60 W30 × 108 W27 × 129 W33 × 118 W36 × 135 W36 × 135
65 W33 × 118 W30 × 132 W36 × 135 W40 × 149 W40 × 149
70 W33 × 130 W30 × 148 W40 × 149 W40 × 167 W40 × 167
75 W36 × 135 W36 × 150 W40 × 167 W36 × 182 W36 × 210
80 W40 × 149 W36 × 160 W36 × 182 W36 × 210 W36 × 210
85 W40 × 167 W36 × 182 W36 × 210 W36 × 231 W36 × 247
90 W40 × 183 W40 × 183 W40 × 211 W36 × 247 W36 × 247
95 W40 × 211 W40 × 199 W40 × 235 W40 × 249 W44 × 262
100 W44 × 230 W40 × 211 W40 × 249 W44 × 262 W44 × 262

increment in span lengths between 40 and 100 feet for each of the girder spacings employed.
Additionally, the right hand column provides a section selected to meet the requirements for
all span lengths less than that shown for all girder spacings. For example, the W30 × 108 in
Table 1 would satisfy all span lengths less than 50 feet and all girder spacings from 6 feet to
10.5 feet.
Ongoing efforts are focused on collaboration with steel mills to provide more rapid
availability of these sections, thus better insuring the success of time-sensitive projects. It
should also be noted, for example, that at the 50 foot span range for a 6 foot girder spacing a
W27 × 84 could be employed whereas the section that fits all girder spacings in the 50 ft. span
is a W30 × 108 or a per foot weight difference of 24 pounds.

NYBRIDGE_Book.indb 71 7/17/2013 8:41:57 AM

 72 Bridge analysis & design

Table 2. Limited depth rolled beam designs.

Girder spacing
L Selected
[ft] 6'-0" 7'-6" 9'-0" 10'-6" section

40 W21 × 62 W21 × 73 W21 × 83 W21 × 93 W21 × 93

45 W21 × 83 W21 × 101 W21 × 101 W21 × 111 W21 × 111
50 W21 × 111 W21 × 111 W21 × 122 W21 × 132 W21 × 132
55 W24 × 117 W24 × 117 W24 × 131 W24 × 146 W24 × 146
60 W24 × 162 W27 × 129 W24 × 146 W24 × 162 W24 × 162
65 W24 × 192 W30 × 132 W24 × 176 W24 × 192 W24 × 192
70 W27 × 194 W30 × 148 W27 × 178 W27 × 194 W27 × 194
75 W27 × 217 W36 × 150 W27 × 194 W27 × 217 W27 × 217
80 W30 × 211 W36 × 160 W30 × 211 W30 × 235 W30 × 235
85 W33 × 221 W36 × 182 W33 × 221 W33 × 241 W33 × 241
90 W33 × 241 W40 × 183 W33 × 241 W33 × 291 W33 × 291
95 W36 × 247 W40 × 199 W36 × 247 W36 × 282 W36 × 282
100 W36 × 282 W40 × 211 W36 × 262 W36 × 302 W36 × 302

Figure 5. Plate girder weights.

In a typical bridge cross-section, more girders would be required using 6 foot spacings than
wider spacings from which the W30 × 108 is derived. Thus, the actual difference in the addi-
tional cost resulting from the 24 pound/foot difference is negligible, and the most probable
scenario of having a reduced number of girder lines would likely negate perceived disadvan-
tages of a heavier girder section.

5.3 Results of plate girder designs

Previous design studies (Nagy 2008) have shown that the use of a reduced set of plate sizes,
as opposed to the use of the exhaustive set of available plates, has a minimal impact on
final girder weight. For specific dimensions of the selected plate girders the reader is referred

NYBRIDGE_Book.indb 72 7/17/2013 8:41:58 AM

 Development of economical short-span steel bridge standards 73

to Morgan (2010). A plot of the final weight versus span length for both the hybrid and
homogeneous sections for each of the girder spacings is provided in Figure 5. Several key
observations can be made from this figure:
• There is little difference, particularly in the shorter span ranges, in total girder weight as a
function of girder spacing.
• In the shorter span ranges there is little benefit provided by the use of hybrid configura-
tions. This is due to the fact that many of the sections start to be controlled as a function
of minimum allowable plate dimensions as opposed to various design limit states.
• For the longer span lengths, and particularly for the wider girder spacings the hybrid girder
configuration does provide some benefit.

6 COMPARISONS WITH OTHER STANDARDS

Figures 6 through 8 detail comparisons with the standard designs developed in this study
with those from the standards discussed earlier. It should be noted that, since these standard
designs incorporate rolled beam solutions, the comparisons in these figures are for rolled
beams only. As shown, the rolled beam solutions are competitive with other standardized
steel bridge designs.

7 INTERACTIVE WEB-BASED DESIGN TOOL

In order to ease the process of steel girder selection and provide state DOTs and owner with
a more efficient means of conducting preliminary designs of short-span steel bridges, the
authors, along with the SSSBA technical working group, the Steel Market Development Insti-
tute (SMDI), the National Association of County Engineers (NACE) Structures Committee,
FHWA, and the AASHTO T-14 Technical Committee for Structural Steel Design, have
developed eSPAN140, an interactive web-based design tool. eSPAN140 is an free, easy-to-
use application which generates a customized Solutions Book (in .pdf format) for a given
set of bridge parameters, complete with girder dimensions, cross-section information, and
associated details.
To begin to use eSPAN140, all the user has to do is go to http://www.eSPAN140.com/ and
create a free user’s account. Once an account is created, the user will have the ability to edit/

Figure 6. Comparison with AISI standard designs for a 9’-0” girder spacing.

NYBRIDGE_Book.indb 73 7/17/2013 8:41:58 AM

 74 Bridge analysis & design

Figure 7. Comparison with OklaDOT standard designs for an 11’-10” girder spacing.

Figure 8. Comparison with TxDOT standard designs for an 8’-8” girder spacing.

review/duplicate previous projects as well as to share previously-generated Solutions Books

with colleagues.
To begin a new project, the user logs into his/her eSPAN140 account. There, the user will
find a list of all of the previous projects the user has completed, along with a “Start New
Project” button. Clicking this button will open up eSPAN140’s data entry screen, where the
user inputs various parameters necessary to define a given project. Figure 9 shows a typical
data input screen in eSPAN140. In addition, eSPAN140 will display Figure 10, which defines
the range of solutions available in eSPAN140. It should be noted that eSPAN140 will also
generate corrugated steel plate solutions along with a series of fabricator and manufacturer
solutions.

NYBRIDGE_Book.indb 74 7/17/2013 8:41:58 AM

 Development of economical short-span steel bridge standards 75

Figure 9. eSPAN140 typical data input window.

Figure 10. eSPAN140 range of solutions.

A project is defined in three steps. The first step is where the user defines general project
information. Specifically, the user must input the following parameters:
• Project Name
• City/County
• State/Province
• Roadway Name
• Span Length
− It should be noted that eSPAN140 will round the span length value to the next high-
est 5 foot increment (U.S.C.S. units are listed since these are the units that eSPAN140
employs) and report the girder solution for this rounded value. For example, if the user
specifies a span length of 82 feet and 4 inches, eSPAN140 will generate a Solutions Book
containing designs for a span length of 85 feet.
− It should also be noted that, if the user specifies a span length longer than 140 feet, the
generated Solutions Book will not include steel girder designs since the girder designs
are only valid for span lengths up to 140 feet.

NYBRIDGE_Book.indb 75 7/17/2013 8:41:59 AM

 76 Bridge analysis & design

After this, the user advances to step two, where details regarding the bridge cross-section
are input. These details are described graphically in Figure 11. Specifically, the user must
input the following parameters:
• Number of Striped Traffic Lanes
• Roadway Width
• Individual Parapet Width
• Individual Deck Overhang Width
In addition, the user can specify whether sidewalks are present; the user simply has to indi-
cate the number of sidewalks and their individual widths. Once these cross-sectional param-
eters are defined, the user has to input three last parameters:
• Skew Angle
− It should be noted that, if the user specifies a skew angle larger than 20°, the generated
Solutions Book will not include steel girder designs since the girder designs are only
valid for skew angles up to 20°.
• Average Daily Traffic, selected from the following:
− “1–500”
− “501–2000”
− “Over 2000”
• Design Speed, selected from the following (it should be noted that U.S.C.S. units are listed
since these list entries are taken directly from eSPAN140):
− “0–45 mph”
− “46+ mph”
− “Don’t know”
− “Not applicable”
The user then advances to step three, where the user inputs data related to corrugated steel
plate solutions. Specifically, the user has to input the waterway area and height of cover,
or the distance from the top of the corrugated steel plate to the bottom of the layer of
pavement.
After these three steps, eSPAN140 will generate a customized Solutions Book (in .pdf
form). To generate a girder design, eSPAN140 will calculate the out-to-out width between
exterior girders (using data input for the cross-section) and then iterate even spaces between
exterior girders in order to generate a valid design (i.e., a design with a girder spacing less

Figure 11. eSPAN140 bridge cross-section view.

NYBRIDGE_Book.indb 76 7/17/2013 8:41:59 AM

 Development of economical short-span steel bridge standards 77

than 10’-6” as this is the maximum girder spacing employed in the standards) with the fewest
number of girder lines. eSPAN140 then reports the details and dimensions for the girder
designed for the next highest girder spacing. For example, if eSPAN140 calculates an interior
girder spacing of 8’-10”, it will report girder designs for a girder spacing of 9’-0”.
In addition to the details regarding girder sizes, all of the details necessary to fabricate
and erect a short-span steel bridge superstructure are included in the eSPAN140-generated
Solutions Book. These include:
• Cambers (both for steel dead weight and total dead weight)
• Stiffener sizes and spacings
• Shear stud layouts
• Individual girder weight
• Girder fabrication details, including weld sizes
• Diaphragm sizes and details
• Framing plan
• Typical cross-section details
• Rebar layout for deck design
• Elastomeric bearing pad details and steel plate sizes
• Customized manufacturer solutions and contact information for SSSBA members
The Solutions Book also provides contact information for The Bridge Technology Center.
The Bridge Technology Center is a complimentary resource available for questions specific
to standard design and detail solutions of short-span steel bridges. It is a resource provided
by West Virginia University and the University of Wyoming.

8 CONCLUSIONS

The efforts of the authors in conjunction with the AISI Short Span Steel Bridge Alliance
have great promise for improved economy and competitiveness of steel alternatives in the
short-span bridge market. This work has provided an overview of the objectives and design
process employed for the development of standard plate girder and rolled beam designs for
span lengths between 40 and 140 feet. With preselected members and details the design proc-
ess may be expedited, and a more streamlined process for shop drawing review may be cre-
ated thus eliminating many weeks in the timeline of a given bridge project, thus benefiting
current trends towards accelerated bridge construction.

REFERENCES

American Association of State Highway and Transportation Officials. 2010. AASHTO LRFD Bridge
Design Specifications, Fifth Edition. Washington, DC: AASHTO.
MDX Software, Inc. 2009. MDX Version 6.5.910. MDX Curved & Straight Steel Bridge Design Rating
Software: MDX Software, Inc.
Morgan, S.A. 2010. Development of an Optimized Short-Span Steel Bridge Design Package. Master’s
Thesis. Morgantown, WV: West Virginia University.
Nagy, G.I. 2008. Towards the Development of Efficient and Economical Short Span Modular Bridges.
Master’s Thesis. Morgantown, WV: West Virginia University.

NYBRIDGE_Book.indb 77 7/17/2013 8:42:00 AM

This page intentionally left blank
 Chapter 7

Fatigue properties of steel-concrete composite slab using

a cementitious adhesive

I. Yoshitake
Department of Civil and Environmental Engineering, Yamaguchi University, Japan

A. Ogawa
Ube Machinery Co., Ltd., Japan

Y. Fujimoto
Department of Civil and Environmental Engineering, Yamaguchi University, Japan

Y.J. Kim
Department of Civil Engineering, University of Colorado Denver, USA

ABSTRACT: A new bridge deck system, steel-concrete composite slab using a cementitious
adhesive, was developed and structural performance was examined from previous studies.
Fatigue durability is the most critical issue for bridge deck slabs subjected to wheel load, so the
property of the new slab system should be examined via experimental investigation. A moving-
wheel fatigue test was conducted using the slab system and high durability to the cyclic wheel
load was practically confirmed in our previous research. To examine the fatigue strength in
a conventional bending test, the present paper discusses fatigue properties of the slab system
in addition to the previous fatigue tests. In particular, the experimental investigation aims to
observe failure behavior of the composite slab subjected to cyclic flexural load, and to exam-
ine the flexural fatigue strength. Two and six slab specimens made by a half-scale system were
used in the static strength test and the fatigue test, respectively. The paper reports observations
in these loading tests and addresses S-N relationships of the slab system. A fatigue limit to the
2 million loading cycles of 0.7P0 is derived by the S-N relationships.

1 INTRODUCTION

In typical current practice, mechanical shear connectors, such as headed studs and L-shaped
steels, are employed in composite deck slabs for connection of steel bottom plate and con-
crete. Various shear connectors have been developed and investigated in previous studies.
Kaido & Matsui (2009) examined punching shear strength of composite slab consisting of
headed studs and lateral ribs. The composite slab is a typical composite deck slab for highway
bridges. Soty & Shima (2013) conducted shear test of L-shaped connector for composite
members. They developed a formula of the shear force and displacement for the compos-
ite structure. Choi et al. (2008) also used L-shaped shear connectors for composite bridge
decks, and they examined fatigue strength using push-out shear test. Claybaugh et al. (2004)
presented fatigue and strength properties of steel-grid bridge deck filled with concrete. To
achieve connection effectively, perfobond ribs are often used in steel-concrete composite slab.
Kim & Jeong (2006; 2009; 2010) & Jeong et al. (2009) reported structural performance of
composite decks with perfobond ribs. Their composite slab employs corrugated steel plate
to improve rigidity of the bottom plate. Lee & Han (1998) investigated fatigue behavior of
composite beam, which is an element of composite slab using pyramidal shear connectors.

79

NYBRIDGE_Book.indb 79 7/17/2013 8:42:00 AM

 80 Bridge analysis & design

Construction of such composite bridge decks may be negatively affected by a lot of shear
connectors welded to the bottom plate. Furthermore, it should be noted that these mechani-
cal connectors are a possible cause of concrete cracks at early age in addition to increases of
dead-weight and costs (Ogawa et al., 2012). Gilbert et al. (2012) measured shrinkage profile
in composite slab with wave-form steel decking, and proposed an analytical procedure of
time-dependent stresses and deformations. Larbi et al. (2007) proposed a steel-concrete com-
posite structure connected by bonding to reduce the risk of cracking, and conducted a push-
out test. As shown above, most composite systems use mechanical shear connectors and have
some issues in construction.
To improve constructability and to simplify the connecting structures, a steel-concrete
composite deck slab using a cementitious adhesive has been developed in previous studies.
Fundamental structural performance of the composite structure was described in a paper
(Yoshitake et al., 2012). Additionally, Yoshitake et al. (in press) reported a moving-wheel
fatigue test for the composite slab, and addressed fatigue response up to 60 × 106 cycles.
Though the previous report indicated sufficient high durability to wheel-load, the ultimate
fatigue strength was not observed. Fatigue durability is the most critical issue for bridge deck
slabs subjected to wheel load, so the ultimate fatigue behavior of the developed slab system
should be confirmed via a conventional fatigue test. Referring to the research by Lee & Han
(1998), the study prepared a composite beam which is an element of the developed slab. The
present study aims to observe failure behavior of the composite beam subjected to cyclic
load, and to show the flexural fatigue strength. To indicate the fatigue durability, the paper
reports the failure mode of the composite slab subjected to cyclic load and a S-N relationship
based on the conventional test.

2 MATERIALS AND TEST SPECIMEN

2.1 Cementitious adhesive

The adhesive is made with a cementitious powder material and a special emulsion (Fig. 1(a)).
The adhesive can be directly sprayed to a metallic substrate as shown in Figure 1(b)–(d).
A unique feature of the sprayable material is that it can be used for bonding fresh concrete to a
metallic substrate even after curing of the adhesive on the substrate. According to our exami-
nation, concrete can be bonded to metallic substrate after 2 weeks of curing of the adhesive.
Further details of the adhesive are available in the previous paper (Yoshitake et al., 2012).

2.2 Steel and concrete

The structural steel for the metal decking had a yield strength of 245 MPa with a nominal
elastic modulus of 200 GPa. The reinforcing bar with a nominal diameter of 13 mm had a
yield strength of 345 MPa and an elastic modulus of 200 GPa. These metallic materials are
same materials used in the previous studies (Yoshitake et al., 2012, Yoshitake et al., 2013).

Figure 1. Sprayable cementitious adhesive: (a) cementitious powder and liquid emulsion; (b) mixed
materials; (c) sprayable adhesive; (d) curing of the adhesive.

NYBRIDGE_Book.indb 80 7/17/2013 8:42:00 AM

 Fatigue properties of steel-concrete composite slab using a cementitious adhesive 81

The Japan Bridge Association recommends the use of expansive concrete for composite
deck slabs. Table 1 gives a mixture proportion of concrete. The expansive concrete was sup-
plied from the ready-mixed concrete plant used in the previous study (Yoshitake et al., 2013).
As well, it should be noted that the coarse aggregate with the maximum size of 15 mm was
used, because the concrete was used for the half-scale specimen. The concrete was placed
in the steel bottom plate after 2 weeks or longer curing of the adhesive (Figure 2). Con-
crete casting was performed twice, with 4 test specimens taken for each of the two concrete
batches. The concrete at the age of 28 days had a compressive strength of 35.2 MPa (average
of 6 cylinders tested) and a Young’s modulus of 31.3 GPa (average of 6 cylinders tested).
Though the mechanical properties of concrete made at each concrete casting were consistent
a different failure mode was observed in the static loading test (to be discussed).

2.3 Slab specimens

Ogami et al. (2012) presented flexural fatigue failure of the composite slab element, such as
a slab. The dimensions of the slab were 2000 mm long × 325 mm width × 100 mm thick. The
specimen had single rib stiffener with the bottom plate, so it was likely to indicate torsional
failure even by slight uneven load.
To prevent torsional failure, slab specimens with twin transverse rib stiffeners welded to
the bottom plate were prepared for this fatigue test. Therefore, this study employed wider slab
specimens than the previous specimen. The element specimen of composite slab is illustrated
in Figure 3 (a). This study uses one-half scale system of a composite deck slab employed for
typical twin girder bridges. Two and six slab specimens were used in the static strength test
and the fatigue test, respectively.

3 TEST PROGRAM

3.1 Static loading test

Figure 3 (b) presents the flexural test set-up. The static bending test was conducted after
the concrete had achieved a 28 day cure. The primary purpose of this test is to determine

Table 1. Mixture proportion of expansive concrete.

Unit weight (kg/m3)

w/cm water cement EXP fine agg. coarse agg. WRA

0.51 166 326 20 872 946 3.46

EXP: expansive material, WRA: water reducing agent

Figure 2. Preparing slab specimens: (a) bottom plate welded with two rib stiffeners (without the adhe-
sive); (b) bottom plate with dried adhesive; (c) concreting work.

NYBRIDGE_Book.indb 81 7/17/2013 8:42:00 AM

 82 Bridge analysis & design

Figure 3. Specimen and loading test: (a) schematic of the specimen; (b) loading method—side view.

Table 2. The applied maximum (Pu) and minimum (Pl) loads in the
fatigue test.

I.D. P0* (kN) Pu (kN) Pl (kN) N (cycles)

9010 177 166.6 17.7 72

8510 177 150.1 17.7 17538
8010a 177 147.0 17.7 1032
8010b 126 100.9 12.7 23500
7510 126 93.1 12.7 2000000
6010 126 75.5 12.7 2000000

*: load-carrying capacity in the static test.

applied load for the fatigue test. All slabs were tested in four-point bending. The applied
load was gradually increased at approximately 3 kN/min. Deflection and strain at mid-span
were measured using an LVDT and wire-strain gages, 3 mm long for steel and 60 mm long
for concrete.

3.2 Fatigue loading test

The cyclic load with speed of 1 Hz was applied to the test specimen. The maximum and
minimum loads are given in Table 2. The load ranges were determined for obtaining an S-N
relationship of the slab. To measure deflection and strain, the maximum load in each test was
statically applied at 1, 10, 102, 103, 104, 105 loadings and every 105 cycles up to 20 × 105 load-
ings. The specimens subjected to 20 × 105 loadings were tested to examine residual strength.

4 RESULTS AND DISCUSSION

4.1 Static loading test

Two specimens were tested to determine the load for the fatigue test. One slab specimen was
failed abruptly in shear at a load of 177 kN. The failure occurred at the debonding induced by
shear force after concrete crushing. The other slab failed by concrete crushing after flexural

NYBRIDGE_Book.indb 82 7/17/2013 8:42:01 AM

 Fatigue properties of steel-concrete composite slab using a cementitious adhesive 83

cracks developed. The maximum load was 126 kN, the load is significantly less than the
maximum load of the first test. The applied loads in fatigue test, therefore, were determined
employing each flexural load-carrying capacity as shown in Table 2.
Figure 4(a) shows flexural cracks observed in the second static loading test. Flexural cracks
at mid-span occurred first and several cracks developed and propagated along the shear span.
Bond failure of the adhesive was not found until the upper concrete at the mid-span failed by
crushing. Nevertheless, the bottom steel plate and rib stiffeners did not yield in the static and
cyclic loading tests (to be discussed).
Figure 5(a) shows linear responses of load—deflection at mid-span. In addition, Figure 5(b)
shows linear strain profile at mid-span at each loading, which means no-slip of the bottom
plate. The result implies the effective composite action of the adhesive. These observations
indicate similar behavior in the previous studies (Yoshitake et al., 2012, Ogami et al., 2012).

Figure 4. Flexural and shear failures: (a) static test; (b) 9010; (c) 8010a; (d) 8010b.

Figure 5. Static loading test: (a) load-deflection responses; (b) strain profile (1st test).

NYBRIDGE_Book.indb 83 7/17/2013 8:42:02 AM

 84 Bridge analysis & design

4.2 Fatigue loading test

Applied maximum loads in the fatigue test were determined as 90, 85, 80, 75 and 60% of the
flexural load-carrying capacities. Figure 4(b)–(d) illustrate failure conditions in this fatigue
test. All fatigue tests using slab specimens with 2 rib stiffeners exhibited concrete crushing or
shear failure, while torsional failure had occurred in the previous tests (Ogami et al., 2012).
The observation implies that the use of wider slab specimens is appropriate for evaluating of
fatigue life of the developed slab. As shown in Figure 4(b) and (c), a diagonal tension crack
developed in the shear span was the ultimate failure mode of the slab specimens (9010, 8010a)
subjected to relatively high load. On the other hand, the slab specimen (8010b) exhibited a
number of flexural cracks in the vicinity of the constant moment zone (Fig. 4(d)). It should
be noted that the debonding of steel plate were not observed in the fatigue tests (8010b, 7510,
6010). While the debonding is a significant failure mode of the developed composite slab,
such failure rarely occurs when the cyclic loading is less than a fatigue limit. These observa-
tions confirm the effectiveness of the sprayed adhesive in terms of enhancing bond between
the concrete and steel even under repeated loading.
Figure 6 demonstrates loading and residual deflections at mid-span. The loading deflec-
tions were measured in the static loading test after each loading cycle, and the residual
deflections were recorded after the static loading test. Figure 6(a) presents the comparison
of deflection, the load ratio (applied load/flexural load-carrying capacity) was 0.80 in both
test. The deflection in the test of 8010a increased gradually after 100 loading cycles, and the
slab specimen abruptly failed by shear cracking at 1032 cycles (Fig. 4(c)). On the other hand,
the deflection in the test of 8010b increased significantly after 10000 loading cycles. The test
specimen failed by concrete crushing at 23500 cycles (Fig. 4(d)). As well, Figure 6(b) shows
deflections in the test of 7510 and 6010. Although the loading and residual deflections in both
test increased gradually, the increases less than 1.0 mm were negligible. The results imply that
the developed composite structure has sufficient rigidity even under the cyclic loading (less
than 75% of the capacity) as well as fatigue durability of 2 million cycles or greater.

4.3 S-N curve

Figure 7(a), (b) show semi—and full-logarithmic graphs for S-N relationship of the fatigue
test. The S-N curves for each graph are given in below.

S 0 0 114 l g N + 1.00448 (1)

log S 0.02946 log
l g log1.00415 (2)

where S is defined as the ratio of applied maximum load (Pu) and the load-carrying capacity
(P0), N is the number of loading cycles.

Figure 6. Deflections at mid-span: (a) 8010a & 8010b; (b)7510 & 6010.

NYBRIDGE_Book.indb 84 7/17/2013 8:42:03 AM

 Fatigue properties of steel-concrete composite slab using a cementitious adhesive 85

Figure 7. S-N relationship: (a) semi-logarithm; (b) full-logarithm.

Yen et al. (1997) conducted bending fatigue tests using composite slab and presented
S-N curves based on the test results. They proposed each 2 curves for full and partial
composite slabs, respectively. The present study employs the 2 curves for full compos-
ites in Figure 7(a) for comparison. In addition, an S-N relationship used in a previous
study (Kaido & Matsui 2009) is given in Figure 7(b). The S-N curve was proposed by
Matsui; based on moving-wheel fatigue tests using reinforced concrete slab. Kaido &
Matsui (2009) reported the relationship is also applicable for composite slabs. It should
be noted that the relationships given above are not simply comparable with these curves
in these previous studies because the test objects and loading methods are different from
the methods used in this study. Both results indicate that coefficients for logarithm N are
significantly smaller than each coefficient of the previous curves. The comparative results
imply the fatigue life of the developed structure can be extended under an appropriate
cyclic load while the higher load (0.8P0 or greater) induces short fatigue life. According to
the S-N relationships in the fatigue test, the fatigue strength to the 2 million loading cycles
is approximately 0.7P0.

5 CONCLUSIONS

A steel-concrete composite slab system using a cementitious adhesive has been developed in
the study. The ultimate fatigue strength of the slab system was not observed, while a moving-
wheel fatigue test and conventional fatigue test were performed in the previous investigations.
The present study aimed to examine the fatigue properties of the developed composite slab.
Observations and knowledge in the experimental investigation are summarized below:
− Bond failure of the adhesive was not observed before the upper concrete at mid-span failed
by crushing in the static tests. As well, fatigue failure due to the debonding did not occur
when the cyclic loading was less than the fatigue limit. These observations indicate the
effectiveness of the adhesive in terms of enhancing bond between the concrete and steel
plate.
− All fatigue tests exhibited concrete crushing or shear failure without torsional failure. The
fatigue test using the slab specimens with 2 rib stiffeners was appropriate for the evaluation
of fatigue life of the developed slab.
− Based on the flexural fatigue test results, S-N curves were presented. The coefficients for
logarithm N in the proposed formulae are smaller than the coefficient in the previous
curves. The fatigue limit to the 2 million loading cycles is considerable as 0.7P0 for the
developed slab system.

NYBRIDGE_Book.indb 85 7/17/2013 8:42:03 AM

 86 Bridge analysis & design

The study focused on the fatigue durability of the new slab system using the adhesive.
Although the proposed composite system has shown promising test results, further struc-
tural tests, such as negative bending and impact effect, should be performed for practical use.
Future researches are necessary to examine durability against environmental effects as well
as economic performance.

REFERENCES

Choi, S., Tateishi, K., Uchida, D., Asano, K. & Kobayashi, K. 2008. Fatigue strength of angle shape
shear connector used in steel-concrete composite slab, Steel Structures, 8: 199–204.
Claybaugh, B.G., Earls, C.J. & Ahmadi, A.K, 2004. Fatigue and strength performance of concrete-filled
steel-grid bridge deck, Journal of Bridge Engineering, ASCE, 9(5): 435–443.
Gilbert, R.I., Bradford, M.A., Gholamnoseini, A. & Chang, Z.T. 2012. Effects of shrinkage on the
long-term stresses and deformations of composite concrete slabs, Engineering Structures, Elsevier,
40: 9–19.
Jeong, Y.J., Kim, H.Y. & Koo, H.B. 2009, Longitudinal shear resistance of steel-concrete composite slabs
with perfobond shear connectors, Journal of Constructional Steel Research, Elsevier, 65: 81–88.
Larbi, A.S., Ferrier, E., Jukiewiez, B. & Hamelin, P. 2007. Static behavior of steel concrete beam con-
nected by bonding, Engineering Structures, Elsevier, 29: 1034–1042.
Lee, K. & Han, J. 1998. Fatigue behavior of composite beams with pyramidal shear connectors under
repeated loading, KSCE Journal of Civil Engineering, KSCE, 2(2): 119–128.
Kaido, H. & Matsui, S. 2009. Estimation of punching shear fatigue strength for steel plate-concrete
composite decks, Steel Construction, Ernst & Sohn, 3: 181–187.
Kim, H.Y. & Jeong, Y.J. 2006. Experimental investigation on behaviour of steel-concrete composite
bridge decks with perfobond ribs, Journal of Constructional Steel Research, Elsevier, 62: 463–471.
Kim, H.Y. & Jeong, Y.J. 2009. Steel-concrete composite bridge deck slab with profiled sheeting, Journal
of Constructional Steel Research, Elsevier, 62: 463–471.
Kim, H.Y. & Jeong, Y.J. 2010. Ultimate strength of a steel-concrete composite bridge deck slab with
profiled sheeting, Engineering Structures, Elsevier, 32: 534–546.
Ogami, E., Yoshitake, I., Ogawa, A. & Fujimoto, Y. 2012. Static and fatigue strength of steel-concrete
composite beam using adhesive, Proc. of ICETCE2012, Yichang: Hubei, 1: 357–360.
Ogawa, A., Yoshitake, I., Watada, Y. & Kim, Y.J. 2012. Design concept of steel-concrete composite slab
using an adhesive, Proc.of 18th IABSE Congress, Seoul, A-0064.
Soty, R. & Shima, H. 2013. Formulation for shear force—relative displacement relationship of
L-shaped shear connector in steel-concrete composite structures, Engineering Structures, Elsevier,
46: 581–592.
Yen, J.Y.R., Lin, Y. and Lai, M.T. 1997. Composite beams subjected to static and fatigue loads, Journal
of Structural Engineering, ASCE, 123(6): 765–771.
Yoshitake, I., Ogawa, A., Kim, Y.J. & Mimura, Y. 2012. Development of a new composite slab sys-
tem using a carbon-fiber-blended cementitious adhesive, Journal of Structural Engineering, ASCE,
138(11): 1321–1330.
Yoshitake, I., Ogawa, A., Kim, Y.J. & Ogami, E. 2013. A composite deck having transverse stiffeners
bonded with a cementitious adhesive subjected to moving-wheel fatigue, Journal of Bridge Engineer-
ing, ASCE, (in press).

NYBRIDGE_Book.indb 86 7/17/2013 8:42:04 AM

 Chapter 8

Development of a shallow press-brake formed tub girder

for short-span steel bridges

K. Barth, J. Mash & G. Michaelson

West Virginia University, Morgantown, WV, USA

M. Barker
University of Wyoming, Laramie, WY, USA

D. Snyder
Steel Market Development Institute, Washington, DC, USA

ABSTRACT: This paper is focused on the development of modular shallow trapezoidal

boxes fabricated from cold-bent structural steel plate using standard mill plate widths and
thicknesses. This concept was developed by a technical working group within the Steel Market
Development Institute’s (a business unit of the American Iron and Steel Institute) Short Span
Steel Bridge Alliance (SSSBA), led by the current authors. This working group consists of all
stakeholders in the steel bridge industry, including mills, fabricators, service centers, industry
trade organizations, universities, and bridge owners. The goal was to develop innovative and
economical modular solutions for the short-span steel bridge market. The proposed system
meets the needs of current industry trends of accelerated bridge construction, while offering
an economical solution. This paper will provide an overview of experimental testing cur-
rently being conducted and further parametric analysis and design studies focused on the
refinement of cross-section dimensions.

1 INTRODUCTION

The Short Span Steel Bridge Alliance (SSSBA) is a group of bridge and culvert industry lead-
ers (including steel manufacturers, fabricators, service centers, coaters, researchers, and repre-
sentatives of related associations and government organizations) who have joined together to
provide educational information on the design and construction of short span steel bridges in
installations up to 140 feet in length. From within the SSSBA technical working group, a mod-
ular, shallow steel press-brake tub girder was developed. This girder is shown in Figure 1.
This new technology consists of cold-bending standard mill plate width and thicknesses
to form a trapezoidal box girder. The steel plate can either be weathering steel or galvanized
steel, each an economical option. Once the plate has been press-brake formed, shear studs
are then welded to the top flanges. A reinforced concrete deck is then cast on the girder in the
fabrication shop and allowed to cure, becoming a composite modular unit. The composite
tub girder is then shipped to the bridge site, allowing for accelerated construction and reduc-
ing traffic interruptions.
A key economic factor with this newly developed system is utilizing a press-brake to form
a girder from a standard-width plate, as opposed to cutting and welding plates together to
form a conventional tub girder. By employing the proposed system, the costs associated with
cutting and welding of steel plates are eliminated. Furthermore, no cross-frames are needed
(as the deck is cast compositely in the fabrication shop, providing continuous lateral support
before the girder is erected), which again reduces the overall bridge system cost. Finally, due

87

NYBRIDGE_Book.indb 87 7/17/2013 8:42:04 AM

 88 Bridge analysis & design

Figure 1. Conceptual composite press-brake tub girder.

to the methods of fabrication, increased quality control would be gained from employing this
system as the entire composite girder unit is shop-fabricated.

2 LITERATURE REVIEW

Prefabricated steel tub-girder systems have been explored as a potential design solution for
the short-span bridge market for a number of years. Many previous research efforts have
shown that these types of systems have the potential to be economical and competitive in
the short-span range. In recent years, the demands for accelerated bridge construction have
been brought to the forefront of design. Therefore, bridge systems such as the press-brake
tub girder have once again surfaced as a viable alternative to conventional bridge fabrication.
Presented in this section is a comprehensive review of previous studies focused on economical
and rapid bridge construction employing various shallow tub girder configurations.

2.1 Prefabricated press-formed steel T-box girder bridge system (Taly & Gangarao 1979)
Taly and Gangarao (1979) proposed using a press-brake to bend an A36 3/8-inch steel plate
to form a tub girder in a short-span modular bridge system. At the time of publication in
1979, The American Association of State Highway and Transportation Officials (AASHTO)
Specifications did not provide any criteria for the design of bridge members using a press-
brake to cold form the shape of girders. Therefore, the researchers checked their tub girder
design in accordance with the 1977 American Iron and Steel Institute (AISI) specifications.
In the proposed design, a prestressed concrete deck would be precast with an embedded
shear stud plate, which would, in turn, be shop welded to the steel tub girder’s flanges. The
total width of the tub girder is 3 feet, and the total width of the prestressed concrete slab is 6
feet. This reduced size permits the complete unit to be fabricated in the shop and shipped to
the construction site, greatly decreasing the amount of field labor and construction time. To
account for various bridge widths, several prefabricated tub girder units are placed adjacent
to one another and joined with a longitudinal closure pour. The system resists lateral loads
through shear keys with weld-ties placed at the junction of the prestressed concrete slab flanges.
The ends of the tub girder beams are closed off with a 3/8-inch thick steel plate diaphragm that
is welded all around the perimeter of the tub girder. To provide additional support, bearing
stiffeners are provided at the tub girder ends along with the 3/8-inch thick diaphragm.
In addition, Taly and Gangarao provided an alternative to the concrete-steel composite
tub girder bridge system which employed an orthotropic deck. To increase the longitudinal

NYBRIDGE_Book.indb 88 7/17/2013 8:42:04 AM

 Development of a shallow press-brake formed tub girder for short-span steel bridges 89

stiffness of the orthotropic deck, WT sections would be shop welded to the steel plate deck.
Like the previous design, the composite tub girder unit could be prefabricated in the shop
and shipped to the construction site. The tub girder dimensions are highly dependent on the
span length, ranging from a 2.5 foot to a 3.5 foot deep tub girder.
The researchers found the tub girder design with the composite concrete deck to be eco-
nomical for spans of 40 to 100 feet. With the all-steel configurations, the maximum span
length would be 65 feet. In addition, the authors note that the tub girders have a greater
torsional stiffness than typical I-beam sections due to their closed shape. Furthermore, 95%
of the total bridge system would be prefabricated and economy is achieved with the use of a
press-brake to cold form the members as opposed to typical fabrication procedures for steel
box girders. Also, in addition to rapid construction, the lightweight design of this system
(roughly 11 tons for a 65-foot-long girder) allows for low capacity equipment for all phases
of construction, including transportation and erection of the tub girders.

2.2 Composite girders with cold-formed steel U-sections (Nakamura 2002)

Similar to Taly and Gangarao’s proposed design, Nakamura (2002) proposed a bridge system
that utilizes a press-brake to cold form steel tub girders. Nakamura’s bridge system includes cast-
ing a prestressed concrete slab supported by twin tub girders, forming a composite modular unit.
Nakamura envisioned a continuous bridge system with multiple intermediate piers to support the
superstructure. To compensate for the potential buckling of the bottom flange at pier locations,
Nakamura designed the tub girders to be filled with concrete and prestressed by prestressed con-
crete (PC) bars, resulting in an increased strength against buckling at the support locations.
The researcher preformed several experimental bending tests on the proposed design.
These tests confirmed that the tub girder behaved as a composite beam at the center of
the span. Furthermore, at pier regions, the tub girder was shown to behave as a prestressed
beam with the prestressed concrete preventing local buckling of the bottom flange. Finally,
Nakamura concluded that this bridge system would in fact be practical and feasible since it
has adequate bending strength and deformation capacity. A drawback to Nakamura’s design
is that the tub girders require more steel than conventional plate girders. However, the cost is
offset by decreased fabrication costs, thereby resulting in an economical design.

2.3 Con-Struct prefabricated bridge system

Nelson Engineering Services has developed a cold-formed tub girder bridge system similar
to previous designs mentioned above (Tricon Precast 2008). This system, Con-Struct, incor-
porates a prefabricated composite bridge girder consisting of a shallow steel tub girder and a
concrete deck. To increase the service capacity of this system, the steel tub girders are stressed
into a camber, and the concrete deck is cast onto the girders in their stressed state. Once the
concrete is cured, the steel compressive stress is locked in to provide camber and increase the
service capacity of the structure. Employing this system, designs are valid for spans up to
60 feet (according to AASHTO LRFD Specifications).

2.4 Folded plate girders (developed at the University of Nebraska)

The University of Nebraska, Lincoln has also researched cold-bent steel tub girders and
developed a composite steel girder system utilizing folded plate girders (Burner 2010 & Glaser
2010). This system utilizes an inverted tub girder where the flanges of the girder are bent
inwards. The concrete deck is then cast on the wider center flange as opposed to previously
developed systems, where the deck is cast on the two smaller exterior flanges. An advantage
of this system is that the orientation of the girder allows maintenance and ease of inspection
of the folded plate girder. Also, the wider top flange resulting from the girder orientation
provides a safe work area for construction personal.

NYBRIDGE_Book.indb 89 7/17/2013 8:42:04 AM

 90 Bridge analysis & design

2.5 TxDOT rapid economical bridge replacement

In an effort by the Texas Department of Transportation (TxDOT) to create a more shallow
bridge superstructure, a bridge system consisting of a shallow steel tub girder was developed
(Chandar et al., 2010). Specifically, the solution was to use a tub girder that consisted of
a 5-foot-wide bottom flange width and a 3-foot-deep web. Two rows of shear studs were
welded to each top flange, and a reinforced concrete deck was cast.
An application of this concept was completed in August of 2010. The bridge consisted
of four simply supported spans of 45 feet, 100 feet, 100 feet, and 65 feet, respectively. The
total width of the bridge was 78 feet; as a result, six tub girders were utilized. It should
be noted that, in this system, while accelerated bridge construction methods were used,
conventionally-fabricated steel tub girders (as opposed to girders formed using cold bending)
were employed.

2.6 MDOT prefabricated steel box-girder systems for accelerated bridge construction
The Michigan Department of Transportation (MDOT) recognized the need for a prefabri-
cated bridge system to be shipped to the construction site where only placement and post-
tensioning were required (Burgueno & Pavlich, 2008). The goal was to create an entirely
prefabricated composite bridge which would eliminate the need for lengthy and costly road
closures for short-span bridges. In order to accomplish this, a research project was conducted
on a shallow, cold-bent tub girder system utilizing a prestressed concrete deck. Specifically,
this project focused on the design of individual units, which would be joined with longitudi-
nal deck pours. Experimental testing coupled with finite element analyses demonstrated that
this system would be competitive in the short-span bridge market.

2.7 Conclusions
Several researchers over multiple decades have researched the potential economy of prefab-
ricated bridge systems incorporating shallow steel tub girders. Many researchers have found
these technologies to be competitive in the short-span bridge market. However, while many
of the research conclusions regarding the efficiency and economy of these systems have been
promising, many of the systems were hindered by somewhat complex fabricated elements,
which would increase the total system cost. In addition, many of these systems did not have
industry-wide support, which resulted in their lack of use in mainstream construction of
short-span bridges. Therefore, a modular tub girder with simplified details, supported by all
levels of the bridge industry, would present a competitive solution for short-span bridges.

3 DEVELOPMENT OF PROPOSED SYSTEM

3.1 Background
Accelerated bridge design and construction are areas of high importance to bridge owners
due to the potential for fast, efficient, and economical bridge solutions. As a result, the bridge
industry is quickly moving towards prefabricated bridges as the preferred method of bridge
construction due to the increased quality control associated with prefabrication and the speed at
which prefabricated bridges can be erected (resulting in reduced traffic interruptions, improved
construction zone safety, decreased environmental disruptions, and improved life cycle costs).
In October of 2011, a retreat with key steel industry stakeholders was held in Chicago, IL
with the intent to develop innovative and economical modular solutions for the short-span steel
bridge market. Several solutions were developed from this meeting with a focus on press-brake
tub girders. A technical working group within the Short Span Steel Bridge Alliance (SSSBA),
led by the authors, originated the development of the press-brake tub girder at this retreat.

NYBRIDGE_Book.indb 90 7/17/2013 8:42:04 AM

 Development of a shallow press-brake formed tub girder for short-span steel bridges 91

Utilizing a standard plate width, tub girders are fabricated using a large capacity press-
brake. Plates are aligned in the press-brake, and cold bent to achieve target bend radii.
Figure 2 shows a large capacity press-brake being used to form one of the girders used for
testing (see Section 4).

3.2 Preliminary design of conceptual press-brake tub girders

Current AASHTO provisions are not specifically applicable to the design of cold-bent
tub girders. Therefore, preliminary specimen design was completed in two stages. First, a
spreadsheet was developed to compute the section properties of any configuration of tub
girder. Next, design iterations were performed based on conservative estimates of press-brake
tub girder capacity (essentially, limiting the capacity of the composite girders to the yield
moment) to assess their validity for the short-span bridge market. For this effort, two differ-
ent plate thicknesses were evaluated (7/16" and 1/2") and three different standard mill plate
widths were evaluated (72", 84" and 96"). For each variation, a design study was performed
by investigating different variations of the girder dimensions in order to obtain an optimum
girder configuration. For this study, the slope of the webs was kept at a constant 1:4 slope,
and the inside bend radii of the girders was kept at a constant value of five times the respec-
tive plate thickness. Figure 3 presents the results of these studies on an 84" × 7/16" plate.

Figure 2. Forming of a press-brake tub girder, (a) large-capacity press-brake, (b) bending of speci-
men’s top flange.

Figure 3. Preliminary design comparisons for an 84" × 7/16" standard mill plate.

NYBRIDGE_Book.indb 91 7/17/2013 8:42:04 AM

 92 Bridge analysis & design

Figure 4. Design evaluation for a press-brake tub girder utilizing a 96" × 1/2" standard mill plate.

The optimum section, in this case the highest yield moment, using an 84" × 7/16" plate was
found to have a top flange width of 6 inches and a total girder depth of 23 inches.

3.3 Preliminary design results

As previously stated, since current AASHTO provisions are not specifically applicable to
cold-bent tub girders, the girders in this study were sized based on a conservative estimate
of flexural capacity. For each of the aforementioned variations of plate widths and thick-
ness, a design evaluation was performed by computing moments according to AASHTO
Strength I and Service II load combinations for span lengths in 5-foot increments. These were
then plotted to determine the section’s validity under different span length requirements.
Figure 4 shows one of these plots for a 96" × 1/2" plate. As shown, according to conservative
assumptions of capacity, this technology would be viable for spans up to 60 feet.

4 EXPERIMENTAL TESTING PROGRAM

4.1 Overview
In order to verify the performance and capacity of this newly-developed modular tub girder,
physical flexural testing will be conducted at the Major Units Laboratory at West Virginia
University. For the experimental testing program, two plate producers, Nucor-Yamato Steel
and SSAB Americas, each donated two 84" × 7/16" × 480" HPS-50W standard mill steel
plates. These plates were then cold-bent by Greiner Industries using a commercial press-
brake. High Steel Structures, Inc. welded the shear studs and bearing plates to all four tub
girders before shipping them to the Major Units Structures Lab at West Virginia University.
Testing will consist of three composite specimens comprised of a total of four press-brake
tub girders. Using a 330-kip MTS servo-hydraulic actuator, three individual specimens will
be constructed and tested in three-point bending scenarios under both static and cyclic load-
ing (see Figure 5).

NYBRIDGE_Book.indb 92 7/17/2013 8:42:05 AM

 Development of a shallow press-brake formed tub girder for short-span steel bridges 93

Figure 5. Typical test setup.

Figure 6. Testing specimen dimensions.

4.2 Specimen design

As shown, previously (see Figure 3), the optimum section using an 84" × 7/16" plate was found
to have a top flange width of 6 inches and a total girder depth of 23 inches. Figure 6 shows
a cross-section view of the press-brake tub girder that will be used for experimental testing.
Two rows of 7/8" × 4" shear studs are placed on each top flange. End bearing plates are also
utilized to prevent potential premature bearing failure during flexural testing (see Figure 7).
A reinforced concrete deck will also be cast on the top flanges as shown.

4.3 Testing procedures

As previously stated, testing will consist of three composite specimens comprised of a total
of four press-brake tub girders. Specimen One will consist of an individual tub girder that
is 35 feet long. A reinforced concrete deck will be cast and allowed to cure for 28 days (see
Figure 8). After curing of the concrete, the specimen will be loaded to failure to assess the
section’s ultimate flexural capacity. Specimen Two will consist of a composite girder with
the same configuration as Specimen One. However, testing of Specimen Two will consist of
cyclic loading to assess the tub girder system’s fatigue performance.
Specimen Three will consist of two composite tub girders joined together by a longitudinal
closure joint comprised of looped rebar filled with ultra-high performance concrete (UHPC)

CH8_Barth et al (Tub Girder)-Final.indd 93 7/17/2013 2:27:40 PM

 94 Bridge analysis & design

Figure 7. Tub girder bearing plate.

Figure 8. Individual composite girder layout (Specimens One and Two).

Figure 9. Two-girder system layout with longitudinal closure pour (Specimen Three).

(see Figure 9). This specimen will undergo both cyclic loading (to assess fatigue performance
of the closure joint) as well as loading to failure.

5 FINITE ELEMENT ANALYSIS PROCEDURES

In addition to physical testing of the press-brake tub girders, parametric studies will be con-
ducted to assess design performance using a refined three-dimensional finite element modeling

NYBRIDGE_Book.indb 94 7/17/2013 8:42:06 AM

 Development of a shallow press-brake formed tub girder for short-span steel bridges 95

technique. Modeling will be conducted using the commercial finite element software pack-
age Abaqus/CAE (Dassault Systemes 2009). Modeling results will be benchmarked against
experimental data to assess their validity and accuracy.

5.1 Modeling procedures

Element selection for these models includes general purpose shells with reduced integration
and hourglass control. The tub girder and concrete deck are modeled using four-noded shell
elements. The bearing plates, due to their geometry, are modeled with both four-noded and
three-noded shell elements. In order to simulate composite action between the concrete deck
and the press-brake tub girder, node to node multi-point constraints are used, which restrict
the degree of freedom between the selected nodes. Boundary conditions in the finite element
model simulate a simply-supported beam (hinge-roller) condition. Figure 10 illustrates a pre-
liminary finite element mesh of a press-brake tub girder composite specimen.
The tub girders in this study are comprised of HPS-50W weathering steel, and modeled
using an elastic-plastic constitutive law including strain hardening effects. Concrete with a
compressive strength of 4 ksi is assumed for the deck and is modeled utilizing a concrete
damaged plasticity model available in Abaqus which assumes that the two main failure mech-
anisms are tensile cracking and compressive crushing of the concrete.
In order to assess the validity of these modeling techniques, experimental data from previous
laboratory experiments were employed as a benchmark. Discussed herein are the benchmark
tests utilized and results from comparisons between experimental and analytical results.

5.1.1 Benchmarking analysis #1

In 1988, Schilling and Morcos tested three steel plate girders in order to determine moment-
rotation characteristics of steel girders with ultra-compact flanges. These three girders
(denoted “S” for shallow, “M” for medium depth, and “D” for deep) were tested in three-
point bending and loaded until failure. Figure 11 shows the details of the “D” girder, which
was used for benchmarking.
A finite element model was created using the aforementioned modeling technique to model
the “D” girder. The load-deflection curve from experimental testing was plotted and com-
pared with finite element analysis results, and is shown in Figure 12. As shown, the proposed
modeling technique is efficient in capturing the nonlinear behavior of this experimental test.

5.1.2 Benchmarking analysis #2

In 2004, tests at West Virginia University were conducted on composite steel plate girder
units in order to determine the ductility of composite steel girders in positive bending
(Roberts 2004). Like the proposed experimental testing program, each specimen was loaded

Figure 10. Preliminary finite element model of typical composite press-brake tub girder specimen.

NYBRIDGE_Book.indb 95 7/17/2013 8:42:06 AM

 96 Bridge analysis & design

Figure 11. D-girder from (Schilling & Morcos 1988).

Figure 12. Comparison of Schilling and Morcos (1988) “D” girder test and finite element results.

in three-point bending using a 330-kip actuator. Figure 13 shows the details of the girder
“PL2”, which was used for benchmarking.
A finite element model was created using the aforementioned modeling technique to
model girder “PL2”. The load-deflection curve from experimental testing was plotted and
compared with finite element analysis results, and is shown in Figure 14. As shown, the pro-
posed modeling technique is efficient in capturing the nonlinear behavior of this composite
steel girder unit.

5.2 Modeling applications

A finite element technique modeling has been proposed for assessing the flexural capacity
of modular steel press-brake tub girders. The modeling technique has been benchmarked
against previously conducted flexural tests. As shown, the proposed finite element technique
is accurate in capturing the flexural behavior of steel girder specimens, both in a noncompos-
ite and composite state.

NYBRIDGE_Book.indb 96 7/17/2013 8:42:07 AM

 Development of a shallow press-brake formed tub girder for short-span steel bridges 97

Figure 13. Roberts (2004) PL2 composite girder specimen.

Figure 14. Comparison of Roberts (2004) “PL2” test and finite element results.

Testing of the previously described modular press-brake tub girders is currently being
conducted at West Virginia University. Once testing is completed, the proposed modeling
technique will be benchmarked against experimental data to assess its validity in predicting
composite press-brake tub girder capacity. Parametric studies will then be conducted in order
to more accurately assess flexural capacity. In addition, further refinement of cross-section
dimensions and proportions (according to predicted capacities) will be conducted in order to
achieve optimum efficiency of the composite system.

6 CONCLUSIONS

This paper is focused on the development of shallow steel press-brake tub girders to provide
an economical steel solution for short-span bridges. Previous research has shown that sys-
tems utilizing cold-bent steel tub girders and/or accelerated bridge construction techniques
can be competitive and economical in the short-span range. An added benefit of the system

NYBRIDGE_Book.indb 97 7/17/2013 8:42:07 AM

 98 Bridge analysis & design

proposed by the authors is the use of simplified details and industry support within the
SSSBA technical working group.
Preliminary design calculations have shown the proposed design to be a practical solu-
tion for moderate spans in the short-span range. Current testing at West Virginia University
will be used to accurately assess design performance and capacity of this newly-developed
modular system. In addition, parameteric finite element studies will be performed to further
validate design performance and improve capacity predictions.

REFERENCES

American Association of State Highway and Transportation Officials. 2010. AASHTO LRFD Bridge
Design Specifications, 5th Edition. Washington, DC: AASHTO.
Burgueno, R. & Pavlich, B. 2008. Evaluation of prefabricated composite steel box girder systems for rapid
bridge construction. Department of Civil & Environmental Engineering, Michigan State University.
Burner, K. 2010. Experimental investigation if folded plate girders and slab joints used in modular con-
struction. Civil Engineering Thesis, University of Nebraska Lincoln.
Chandar, G., Hyzak, M. & Wolf, L. 2010. Rapid, Economical Bridge Replacement. Modern Steel
Construction.
Dassault Systèmes. 2009. Abaqus/CAE Users Manual (Version 6.9). Providence, RI: Dassault Systèmes
Simulia Corp.
Glaser, L. 2010. Constructability testing of folded plate girders. Civil Engineering Thesis, University of
Nebraska Lincoln.
Nakamura, S. 2002. Bending behavior of composite girders with cold formed steel U-section. Journal of
Structural Engineering: 1169–1179.
Roberts, N. 2004. Evaluation of the ductility of composite steel I-girders in positive bending. Civil and
Environmental Engineering Thesis, West Virginia University.
Schilling, C. & Morcos, S. 1988. Moment-rotation tests of steel girders with ultra-compact flanges. Project
188 Autostress Design of Highway Bridges, American Iron and Steel Institute.
Taly, N. & Gangarao H. 1979. Prefabricated press-formed steel T-box girder bridge system. Engineering
Journal/America Institute of Steel Construction: 75–83.
Tricon Precast, 2008. Con-Struct Prefabricated Bridge System. Standard plans. Tricon Engineering
Group, Ltd.

NYBRIDGE_Book.indb 98 7/17/2013 8:42:08 AM

 Chapter 9

Load-carrying capacity of corroded end of I-section steel girder

E. Yamaguchi & T. Akagi
Kyushu Institute of Technology, Kitakyushu, Japan

ABSTRACT: One of the crucial factors that could terminate the service life of a steel
bridge is corrosion. Since many of the corrosion problems are found in the girder end, it is
quite essential to know the load-carrying capacity of the girder end for efficient maintenance.
To this end, the present study investigates the influence of corrosion at a girder end. Steel
I-section girders with various corrosion models are constructed, and placing a vertical load
at the upper flange right above the support, the girders are analyzed by nonlinear FEM. The
degradation of the load-carrying capacity of a steel girder due to corrosion at its end is thus
evaluated numerically. It is found that different corrosion patterns at a girder end can lead to
significantly different degradations of the load-carrying capacity.

1 INTRODUCTION

The maintenance of bridges is now a very important issue in Japan to keep a highway net-
work in good condition, as the number of bridges started increasing rapidly in 1960s. As for
the steel bridge, corrosion is one of the most influential phenomena on its service life: the
corrosion could reduce the load-carrying capacity of the steel bridge, threaten its safety and
eventually terminate its service. The cause for some 15% of the renewed highway steel bridges
and some 50% of the renewed railway bridges is said to be corrosion (Hung et al., 2002).
Since the replacement of a bridge requires substantial cost, it is important to evaluate the
load-carrying capacity of a corroded bridge for economic maintenance. Quite a few research
studies indeed have been carried out, but much remains to be done, since corrosion patterns are
numerous. In fact, the research is being continued: for example, the shear capacities of locally
corroded steel I-section girders have been investigated and reported in 2011 (Liu et al., 2011).
The waterproof function of an expansion joint is quite susceptible to damage due to
impact loads of moving vehicles. Therefore, water leakage from the expansion joint is often
found, leading to the corrosion environment near the girder end. As a result, more corrosion
problems occur near girder ends than in the other parts.
The girder end is subjected to a concentrated load at the bottom, a reaction force from
the bearing support, and thus the degradation of the girder end could control the service life
of a steel bridge. However, the load-carrying capacity of the girder end is yet to be studied
much: for example, a detailed case study using an actual corroded girder has been conducted
by Huang et al. (2002), but the number of such studies is quite limited and much remains to
be done.
In the present study, the load-carrying capacity of a corroded steel girder end is investi-
gated numerically. Various patterns of corrosion are considered, to this end.

2 STEEL I-GIRDER MODEL

Referring to a model bridge of Japan Bridge Association (2000), the steel I-section girder
shown in Figure 1 is employed for the present study. Young’s modulus E, Poisson’s ratio

99

NYBRIDGE_Book.indb 99 7/17/2013 8:42:08 AM

 100 Bridge analysis & design

Figure 1. Steel I-section girder model (unit: mm).

and the yield strength are 2.0 × 105 N/mm2, 0.3 and 235 N/mm2, respectively. The nonlinear
material follows the plasticity model of von Mises type. The associated flow rule and the
kinematic hardening rule are applied. The uniaxial behavior is of bilinear type with the sec-
ond slope E/100.
For the evaluation of the load-carrying capacity of a corroded girder, attention needs to
be paid to the following:
a. Local buckling in the corroded girder end above the bearing support
b. Shear buckling of the corroded web plate near the bearing support
c. Bending strength of the corroded girder
This paper focuses on local buckling in corroded girder, which is critical when a heavy truck
runs above the bearing support. Huang et al. (2002) studied the problem; and the present
study follows their investigation, but much more corrosion patterns are taken into account.
Using the design equation in Japan Road Association (2012), the design load-carrying capac-
ity of the girder end of the present steel girder is found 1735 kN when it is not corroded.

3 CORROSION MODELS

Corrosion patterns have been investigated in several studies. For example, Huang et al. (2002)
looked into the corrosion pattern of the steel girder very closely: a state of corrosion is more
sever at the lower part of and on the inner surface of the girder.
Based on the observations available in literature, corrosion models are constructed in the
pre-sent study. Those models are presented in Figure 2. Corrosion is assumed on one side of
the girder.
The locations of CM-S and CM-C are in the transverse stiffener at the support and in the
web of the cantilever part of the girder, respectively. The locations of CM-N and CM-W are
both in the web, but the widths of the corroded regions are different: CM-W is twice as wide
as CM-N. CM-S is assumed to develop on the two surfaces of the transverse stiffener since
the entire stiffener is on one side of the girder.

NYBRIDGE_Book.indb 100 7/17/2013 8:42:08 AM

 Load-carrying capacity of corroded end of I-section steel girder 101

Figure 2. Corrosion models.

Various degrees of corrosion are considered: the height h and the plate-thickness loss on
one side of a plate Δt are changed; 10%, 20% and 40% of the web height h0 are given to
h while Δt takes 2 mm, 4 mm and 6 mm. Δt is assumed uniform over each corroded region.
Since corrosion develops on both surfaces of the transverse stiffener in CM-S, the plate-
thickness loss is 2Δt, twice as much as in the other corrosion models. In the present study,
37 girders are to be analyzed, including the original (non-corroded) girder.

4 OUTLINE OF ANALYSIS

The load-carrying capacity of the steel I-section girder end is evaluated by finite element analysis
with material and geometrical nonlinearies. Initial imperfections are also taken into account: the
effects of residual stress and initial deflection are included in the analysis. To this end, the ther-
mal stress analysis and the buckling analysis (eigenvalue analysis) are conducted in advance.
All the analyses are conducted by ABAQUS (2006). 4-node shell elements are employed.
The load-carrying capacity of a girder end under loading above the support is the focus here.
Deformation is expected to occur in a localized area above the support so that the area is
modeled by finer elements. The girder is simply supported, but since it is analyzed in isola-
tion, the horizontal movement of the upper flange is also restrained so as not to incur an
inappropriate failure mode.
The localized deformation suggests that it may not be necessary to analyze the whole
girder. Therefore, a quarter-length girder model is also constructed. Assuming no corro-
sion, analyses are conducted. The numerical results of the full girder model and the quarter
girder model are shown in Figure 3. The two sets of the results are close to each other with
the difference in the load-carrying capacity being 0.1%. In the following analyses, therefore
the quarter model is to be used for saving computational cost. The finite element mesh of the
quarter girder is presented in Figure 4.

5 NUMERICAL RESULTS

Figures 5 and 6 show numerical results where Pmax and P0 are the load-carrying capacity of
a corroded girder end and that of the non-corroded girder end. P0 has been found 2542 kN
(Figure 3).
With the increase of h and Δt, the load-carrying capacity decreases. Yet the way they influ-
ence the capacity varies from corrosion model to corrosion model.
Figure 5 shows that for h/h0 = 0.1, the load-carrying capacities of CM-N and CM-W
decrease only slightly, even when Δt is large: the reductions of CM-N and CM-W are 0.1%
and 0.2%, respectively, with Δt = 6 mm. In cases of CM-S and CM-C with h/h0 = 0.1, the

NYBRIDGE_Book.indb 101 7/17/2013 8:42:08 AM

 102 Bridge analysis & design

Figure 3. Load-displacement curves at loading point.

Figure 4. Finite element mesh of quarter girder model.

Figure 5. Variation of load-carrying capacity with plate-thickness loss.

NYBRIDGE_Book.indb 102 7/17/2013 8:42:08 AM

 Load-carrying capacity of corroded end of I-section steel girder 103

reductions in the load-carrying capacity of the girder end are very little up to Δt = 2 mm.
But it decreases as Δt increases: the reductions become 20% and 13% for CM-S and CM-C,
respectively, with Δt = 6 mm. When h/h0 is larger, the reductions in the load-carrying capacity
are appreciable even with Δt = 2 mm: the reductions are 8% and 11% for CM-S and CM-C,
respectively, when h/h0 = 0.4.
The reduction changes almost linearly with the increase of Δt except for CM-S and CM-C
with h/h0 = 0.1. However, even CM-S and CM-C with h/h0 = 0.1 show the tendency of linear
change between Δt = 2 mm and Δt = 6 mm.
Figure 6 indicates that the increase of h/h0 surely reduces the load-carrying capacity of
the girder end. However, the capacity reduction is not linearly dependent on the increase of
h/h0. CM-N and CM-W show little degradation up to h/h0 = 0.l. As h/h0 increases, the load-
carrying capacity tends to decrease significantly even in CM-N and CM-W. In case of CM-S
and CM-C, the capacity reduction is small up to h/h0 = 0.1 when Δt = 2 mm, while the capacity
reduction is significant even between h/h0 = 0 and 0.1 when Δt = 4 mm and 6 mm.
The average load-carrying capacity rate Pmax/P0 of each corrosion model is found 0.855,
0.862, 0.966 and 0.945 for CM-S, CM-C, CM-N and CM-W, respectively. From this result,
it can be stated that CM-S has the largest influence while CM-N the smallest. The difference
between CM-S and CM-C and that between CM-N and CM-W are insignificant. But the
influence of CM-C is much larger than that of CM-W. Thus the corrosion models can be
classified into two groups in terms of the influence on the load-carrying capacity of a girder
end: one group consists of CM-S and CM-C; the other consists of CM-N and CM-W. The
corroded regions in CM-C and CM-S have a free boundary while CM-N and CM-W do not.
The existence of the free boundary appears essential in discussing the degradation of the
load-carrying capacity of a girder end.

Figure 6. Variation of load-carrying capacity with height of corroded region.

NYBRIDGE_Book.indb 103 7/17/2013 8:42:09 AM

 104 Bridge analysis & design

Figure 7. Deformed configuration of CM-C with Δt = 6 mm and h/h0 = 0.2.

Large deformation is observed in and around the corroded region. As an example, the
deformed configuration of CM-C with Δt = 6 mm, h/h0 = 0.2 is presented in Figure 7 where
the deformation is magnified by a factor of 3.

6 CONCLUSIONS

The present study reveals that the way corrosion influences the load-carrying capacity of a
girder end depends on the location of corrosion. If the corroded region has a free boundary,
the degradation of the load-carrying capacity of a girder end is large. The corrosion patterns
considered herein, therefore, can be classified into two groups as to the degree of the deg-
radation of the load-carrying capacity. It is also found that the load-carrying capacity of a
girder end tends to decrease linearly with the increase of the plate-thickness loss.
The observations described above can give a general idea about the seriousness of corro-
sion in terms of the degradation of the girder end. If it is found significant, measure should
be obtained such as the size of the corroded region and the plate-thickness loss as soon as
possible and a careful safety evaluation needs be done.

ACKNOWLEDGEMENTS

Financial support from the Japan Iron and Steel Federation for the present study is gratefully
acknowledged.

REFERENCES

ABAQUS 2006. User’s Manual, ABAQUS Ver. 6.6. Dassault Systemes Simulia Corp.
Hung, V.T., Nagasawa, H., Sasaki, E., Ichikawa, A. & Natori, T. 2002. An experimental and analytical
study on bearing capacity of supporting point in corroded steel bridges. Journal of JSCE 710/I-60:
141–151.
Japan Bridge Association 2000. Design Example and Explanation of Composite Girder.
Japan Road Association 2012. Specifications for Highway Bridges Part 2 Steel Bridges.
Liu, C., Miyashita, T. & Nagai, M. 2011. Analytical study on shear capacity of steel I-girder with local
corrosion nearby girder ends. Journal of Structural Engineering, JSCE 57 A: 715–723.

NYBRIDGE_Book.indb 104 7/17/2013 8:42:10 AM

 Chapter 10

Uplift bearings—selection and design considerations

G. Moor
Mageba USA, New York, USA

T. Spuler, N. Meng & C. O’Suilleabhain

Mageba SA, Bulach, Switzerland

ABSTRACT: The vertical forces exerted by bridge decks and other structures on their sup-
ports are not always downwards; uplift can occur for a variety of reasons. Upward forces
are generally transient, lasting no longer, for example, than the duration of a strong wind
or of live loading of the relevant section of the structure. Although the structure’s bearings
must typically still be capable of carrying downward forces, facilitating rotations, and resist-
ing horizontal forces and/or accommodating displacements, they must now be able to do
all of this even under uplift conditions. And if the frequency of load reversal is high, then
the uplift load condition may become fatigue-relevant, adding yet another dimension to the
design—especially if the load reversals would cause hammering of interfaces such as the slid-
ing surfaces of a sliding bearing. This paper describes key issues which must be considered in
selecting and designing bearings for uplift conditions.

1 INTRODUCTION

A structure’s bearings play a critical role in its proper functioning and performance, typically
accommodating movements and rotations while carrying loads and resisting other forces. In
doing this, they generally enable the structure to function far more efficiently than it would in
the absence of bearings, allowing bending moments and stresses to dissipate in a controlled
manner.
Most bridges require their bearings to resist downward forces, with resultant upward forces
never arising. If no horizontal forces must be resisted, the basic design of the bearing may be
relatively simple, as shown in Figure 1 for the case of a spherical bearing. If horizontal forces
are to be resisted (in the longitudinal or transverse direction, or both), this can generally be
achieved by the addition of stops or guide bars, for example as shown in Figure 2.
However, the design becomes significantly more complicated if any type of uplift force
must be considered, as described below.

Figure 1. Cross-section of a typical spherical Figure 2. Cross-section of a typical spherical

bearing (free sliding type), which must not resist bearing (guided sliding type, with guide bars on
uplift forces. sides allowing horizontal sliding movements).

105

NYBRIDGE_Book.indb 105 7/17/2013 8:42:10 AM

 106 Bridge analysis & design

2 DESIGNING BEARINGS FOR UPLIFT FORCES—KEY CONSIDERATIONS

Uplift conditions can arise for a variety of reasons, including

− wind, especially on roofs and light, narrow structures such as pedestrian bridges,
− vibrations and significant dynamic forces, such as may arise on a railway bridge,
− vertical ground acceleration (especially at near-fault locations) during earthquakes, and
− live loading, on a structure with a sensitively balanced design.
The frequency at which such uplift conditions occur is of considerable significance for the
design of the bearings which must resist these forces. If uplift is expected to occur only rarely,
then the uplift aspect of the bearing’s design may simply have to prevent upward movement
on these rare occasions. However, if uplift conditions can occur frequently, the repeated load
reversals may be fatigue relevant, and if they cause movements that can cause hammering
(e.g., at sliding interfaces that can pull apart under uplift conditions), damage to materials
can result.
The movements that must be accommodated by the bearing (if any) are also very signifi-
cant. If the bearing has to accommodate horizontal movements, by deformation or sliding,
the task of designing to resist uplift is more complicated. And if such movements are by
sliding, then the sliding interfaces require special attention to prevent damage from hammer-
ing and contamination. Of course, all other demands on the bearing must also be given due
consideration.
It must also be recognised that uplift conditions do not place demands on the structure’s
bearing alone. The uplift forces must also be safely and reliably transmitted from the bear-
ing to the connecting structures above and below, by means of suitably designed anchors in
sufficiently strong structures. And these anchorages must not only resist direct uplift forces;
they must also fully resist any horizontal forces that may arise, because friction cannot con-
tribute in the absence of pressure. In general, where a certain minimum vertical force can be
relied on to act whenever a horizontal force occurs, much or all of the horizontal force can
be resisted by friction, reducing the need for anchoring. But if uplift can occur, this benefit
is not available.
Verification of the adequacy of designs of uplift bearings is generally limited to design
proofs, perhaps including testing of components or materials as appropriate. The European
standard for bridge bearings, EN 1337, does not directly cover the design of uplift bearings
(with European Technical Approvals arising for such cases), let alone how such bearings
should be tested. Testing a bearing for uplift force is generally considered unnecessary where
the uplift force is a constant, static force or if it will only occur in extreme circumstances (e.g.
during a ULS case such as an earthquake), because the ability of a bearing to resist such
forces can generally be adequately proven by calculations and testing of materials. And test-
ing for uplift would be very costly if the uplift condition occurs frequently, with many load
reversals, requiring a dynamic testing rig. An example of how testing was carried out for such
a case, with testing limited to verifying the compressibility of bearing parts, is described in
Section 5.1 below.

3 STANDARD UPLIFT BEARINGS—WITH EXTERNAL UPLIFT PROTECTION

As noted above, if uplift is expected to occur only rarely, then the uplift aspect of the bearing’s
design may be relatively simple. The basic spherical bearing designs presented in Figures 1
and 2, for instance, may be adapted as shown in Figure 3, with uplift clamps at each side.
These can be designed to also allow horizontal movements (longitudinal or, to a degree, trans-
verse), or to prevent such movements, depending on the bridge’s requirements.
Examples of the use of bearings with such uplift-resisting capabilities are presented
below.

NYBRIDGE_Book.indb 106 7/17/2013 8:42:11 AM

 Uplift bearings—selection and design considerations 107

3.1 The uplift bearings of the Revere Beach Pedestrian Bridge, Massachusetts
The new Christina and John Markey Memorial Pedestrian Bridge (Figure 4) opened in early
2013, providing access to America’s oldest public beach from a subway station and a major
new multi-level car park at the other side of a busy road.
The structure requires two bearings at each end (Figure 5). At one end, which is designated
the fixed end, the loads are resisted, and the deck is held in place, by spherical bearings. At
the other end, which must be able to move longitudinally, the bearings are based on a slid-
ing elastomeric bearing pad and a stainless steel sliding partner. These bearings are designed
with uplift clamps to resist the significant 19 kip (84 kN) uplift force arising while facilitating
longitudinal sliding movements. The uplift clamps of one bearing allow limited sliding in the
transverse direction, making it a free sliding bearing (see Figures 6 and 7), while those of
the other bearing resist transverse forces and movements, making it a guided sliding bearing
(Figures 8 and 9).

3.2 The uplift bearings of the roof of the Kimbell Art Museum, Fort Worth, Texas
The Kimbell Art Museum in Fort Worth, Texas is a world-renowned building, and has won
wide acclaim for its design since its opening in 1972. A second building (Figure 10), designed
by world-renowned architect Renzo Piano, is scheduled to open in 2013 and will provide
space for special exhibitions, allowing the original building to showcase the permanent collec-
tion. It will also accommodate dedicated educational spaces and an approximately 300-seat
auditorium. The building design includes many striking features, including its roof, which
spans gracefully above the large exhibit areas.
To enhance its aesthetic qualities, the architect specified that the 66 bearings which sup-
port the roof and allow its movements should be designed and positioned to be very discrete,
and thus as small as possible (see Figure 11). Considering the horizontal and vertical forces
(including uplift forces of approximately 75 kips (330 kN) to be resisted by the bearings, and

Figure 3. Cross-section of a spherical bearing with uplift clamps (outlined).

Figure 4. The new Revere Beach Pedestrian Figure 5. The bearings required by the bridge’s
Bridge. design—including two uplift bearings at front.

NYBRIDGE_Book.indb 107 7/17/2013 8:42:11 AM

 108 Bridge analysis & design

Figure 6. Design of the free sliding uplift bear- Figure 7. Section through the free sliding uplift
ing, with a reinforced elastomeric pad at its core. bearing. The elastomeric pad is held in place by
The uplift clamps (bolted) at each side resist uplift keys in the steel plate at its base, and the bearing’s
but allow limited transverse sliding movement. upper plate can slide across its PTFE top surface.

Figure 8. Design of the guided sliding uplift Figure 9. Section through the guided sliding
bearing. The uplift clamps (welded) at each side uplift bearing, with the vertical plate of one uplift
are designed to also resist transverse forces. clamp removed. The horizontal upper part of the
clamp is shown, with curved lower surface allow-
ing rotations of the sliding plate it holds down.

Figure 10. The Kimbell Art Museum, Fort Worth. Figure 11. Packing of bearings for transport to
site, showing compact size.

the movements that they would have to accommodate, linear rocker bearings were proposed.
The design of these is illustrated by the renderings in Figures 12 to 15.

4 ADVANCED UPLIFT BEARINGS—WITH INTERNAL UPLIFT PROTECTION

In some cases, where loading conditions are demanding with frequent load reversals, bear-
ings with external uplift clamps as described above may be at risk of fatigue failure due to

NYBRIDGE_Book.indb 108 7/17/2013 8:42:13 AM

 Uplift bearings—selection and design considerations 109

Figure 12. Rendering of a fixed linear rocker Figure 13. Section of bearing shown in
bearing with uplift protection clamps at each side. Figure 12. The curved lower surface of the upper
The bearings feature anchor sockets for connec- element has shear key connections to the plate
tion to a concrete support structure beneath. below (as shown in Figure 15), allowing it to
rotate but not to move.

Figure 14. Rendering of a guided sliding linear Figure 15. Section of bearing shown in
rocker bearing with uplift protection clamps at Figure 14. The rocker element of the fixed bear-
each side. ing in Figures 12 and 13 is replaced by a two-part
unit, the top part of which slides across the lower
part, along one axis.

the eccentricity of the uplift clamps. This eccentricity results in moment effects and prying
action, which are demanding on the connections within the bearing. To overcome this, some
types of bearing can be designed with internal uplift protection, with uplift forces (like the
normal downward forces) flowing through the bearing’s center. Figure 16 shows a spherical
bearing with such a solution; the simple calotte of the bearing shown in Figure 3 has been
replaced by a two-part mechanism, the upper part of which is bolted through the lower part
to the concave element below.
The flow of forces through the bearing, under downward and uplift force conditions
respectively, is illustrated in Figures 17 and 18. As can be seen from Figure 18, uplift forces
are directed through the bearing’s center, resulting in only minimal eccentricity.
A current example of the use of such a design is presented below.

4.1 The uplift bearings of the landing area of the Sochi Ski Ramp
The all-season resort city of Sochi on the north-eastern shores of the Black Sea is currently
being prepared to host the 2014 Winter Olympics. For the engineers who must design and
construct the extensive new facilities, the project has presented many challenges, such as the
high seismicity of the area. One of the new facilities is a ski jumping area (Figure 19), with
two jumps and a landing area with seating for spectators. Adjacent to and part of this land-
ing and viewing area is a ski-out area—which will also serve as the start and finish zone
for the Alpine Combination—which constitutes a composite steel-concrete bridge deck with
multiple season-dependent purposes. The design of the structure is such, that when the view-
ing area is subjected to the weight of thousands of spectators, uplift force conditions result

NYBRIDGE_Book.indb 109 7/17/2013 8:42:15 AM

 110 Bridge analysis & design

Figure 16. Cross-section of a spherical bear- Figure 17. The flow of downward forces through
ing with internal uplift protection—avoiding the bearing shown in Figure 16.
eccentricity of forces and reducing the effects of
loading.

Figure 18. The flow of uplift forces through the Figure 19. The new ski jump facility at Sochi.
bearing shown in Figure 16.

Figure 20. The design of a free sliding uplift Figure 21. A free sliding uplift bearing during
bearing (allowing both longitudinal and limited assembly, before placing of the sliding plate on
transverse movements). top.

at several of the structure’s support bearings. At other times, however, load distributions will
be very different, resulting in a number of load cases with uplift forces acting for prolonged
periods of time.
Spherical bearings with internal uplift protection were selected to address this challenge,
with free sliding, guided sliding and fixed varieties required. The design and manufacture of
these bearings are illustrated in Figures 20 to 25.

NYBRIDGE_Book.indb 110 7/17/2013 8:42:17 AM

 Uplift bearings—selection and design considerations 111

Figure 22. A guided sliding uplift bearing (allow- Figure 23. The guided sliding uplift bearing
ing longitudinal movements, resisting transverse shown in Figure 22, following placing of the slid-
forces), during assembly (before placing of sliding ing plate on top but before fixing of the guide bar
plate on top or guide bars at each side). at one side.

Figure 24. A fixed uplift bearing during assem- Figure 25. The main uplift-resisting element of
bly (before placing of top plate). the fixed uplift bearing shown in Figure 24.

5 EXCEPTIONAL UPLIFT BEARINGS FOR SPECIAL REQUIREMENTS

Occasionally, uplift bearings, which are already quite special in their own right, must also sat-
isfy very special additional demands. In such cases, the bearing’s primary function of resist-
ing uplift as well as downward forces is likely to play a key role in developing a solution,
perhaps presenting the greatest challenge to be overcome. This is illustrated by the following
example.

5.1 The special uplift bearings of the Golden Ears Bridge, Vancouver
The Golden Ears Bridge (Figure 26), near Vancouver, British Columbia, was opened to traf-
fic in 2009 and features an unconventional hybrid cable-stayed design which allows the bridge
deck, rather unusually, to rise and fall under the influence of traffic alone. The bridge has
three main spans of 794 feet (242 m) each and end spans of 397 feet (121 m), and the move-
ments of the entire 3175 foot (968-meter) length are accommodated by expansion joints and
bearings at the two ends only. The design of the bridge resulted in the following very demand-
ing combination of requirements for each bearing (Spuler et al., 2010):
− longitudinal movement of 122 inches (3,100 mm)
− transverse movement of 2 inches (50 mm)
− rotation of 0.039 radians
− downward bearing capacity of 1,034 kipf (4,600 kN)

NYBRIDGE_Book.indb 111 7/17/2013 8:42:19 AM

 112 Bridge analysis & design

− uplift capacity of 881 kipf (3,920 kN)

− and all of this with frequent changes between downward and uplift force conditions, many
times a day.
A solution based on the spherical bearing type shown in Figure 16 could have been devel-
oped to satisfy all force and movement requirements, but the final requirement, for the bear-
ing to be designed to withstand frequent changes between downward and upward force
conditions, was defining in this challenge. Since vertical movements, however minute, of one
part relative to another could not be ruled out (due to fabrication tolerances and material
deformations in particular), it was concluded that the materials used at the sliding interfaces
could with time have suffered from the hammering which would result each time the verti-
cal force on the bearing changed from upward to downward. To prevent such hammering,
it became clear that the bearing design had to ensure that the sliding interfaces were in a
constant state of compression.
A standard bearing type fulfilling these requirements was not known to the bearing sup-
plier, so a special design had to be developed. The final design features a long part which is
bolted to the bridge deck and a shorter part which is anchored to the concrete pier below.
These parts interact by virtue of their overlapping horizontal plates, which are separated by
large, sliding elastomeric bearings (see Figures 27 and 28). Each bearing was constructed
with a designed pre-compression force of 337 kips (1,500 kN), which would ensure that no
gaping joint occurs at any of the bearing’s sliding surfaces under Serviceability Limit State
conditions.
Production of bearings of such dimensions presented many challenges not normally
encountered in the manufacture of bridge bearings, requiring special measures to be utilized.
For example, to ensure the parallelism of critical bearing elements, which could have been
compromised by welding distortions, it was decided to bolt, rather than weld, the critical
plates together. And the compressibility of each elastomeric bearing was tested after manu-
facture to confirm theoretical predictions about its performance. Such measures enabled con-
fidence to be gained that these special bearings would perform as required in service.
The fully fabricated bearings (Figures 29 and 30), each weighing 37,000 lb (17,000 kg),
attest to the fact that suitably qualified and experienced engineers can develop a solution to
almost any uplift bearing supply challenge.

6 CONCLUSIONS

The requirement to facilitate uplift conditions can increase the challenge to design and manu-
facture structural bearings considerably. Although relatively simple external uplift clamps

Figure 27. Cross section of pre-compressed

uplift bearing. The elastomeric pads have
PTFE surfaces (bottom of upper pads, top
Figure 26. The Golden Ears Bridge, during of lower pad) to allow sliding of the upper
construction. part across the lower part.

NYBRIDGE_Book.indb 112 7/17/2013 8:42:21 AM

 Uplift bearings—selection and design considerations 113

Figure 28. Longitudinal section of bearing—facilitating sliding movements of +/–5 ft (+/–1550 mm).

Figure 29. One bearing being securing on a Figure 30. A bearing during installation on the
truck for transport to site. Golden Ears Bridge.

may suffice in many cases, when load reversals are infrequent, the structural performance of
some types of uplift bearing can be significantly enhanced by locating the uplift prevention
feature at the bearing’s centre, largely avoiding force eccentricities that can result in unwanted
moment forces and prying action. But even this advanced solution may not suffice in some
instances, e.g. when a bearing is subjected to frequent load reversals, which could result in
hammering at sliding interfaces which are not pre-compressed. It is thus clear that the nature
and frequency of the uplift condition must be assessed and understood to enable a suitable
bearing solution to be proposed and implemented.

REFERENCES

Spuler, T., Moor, G. & O’Suilleabhain, C. 2010. Supporting innovative bridge design—The bearings and
expansion joints of the Golden Ears Bridge. Proc. 3rd FIB Int. Congress, Washington D.C.

NYBRIDGE_Book.indb 113 7/17/2013 8:42:21 AM

This page intentionally left blank
 Chapter 11

Design of Idemli Bridges in landslide areas

C. Ozkaya, G. Cetin & F. Tulumtas

Yuksel Proje Uluslararası Co., Ankara, Turkey

ABSTRACT: In this paper, design of a highway viaduct located in a landslide area is

studied. The depth of landslide is expected to be up to 20 meters from the surface. At the
time of construction, since retrospective inclinometer data was not available, movement
rate of the mobilized soil was unknown during harsh climatic conditions etc. One alter-
native is to design a foundation system capable of resisting all forces and displacements
resulting from mobilized soil for even the greatest movement rate that could be expected.
Another alternative is to keep span lengths of the structure as large as possible with a foun-
dation system having technically feasible small dimensions in order to not to be affected
by the mobilized soil as much as possible. The second alternative was adopted with 165 cm
diameter composite piles having a thick steel tube and steel fiber reinforced concrete in the
foundation system.

1 INTRODUCTION

A new highway connects city of Samsun to town of Gerze in northern part of Turkey. Old
highway is a narrow one crossing mountainous region with a curved profile. Increases in
tourism revenues necessitated a new modern highway.
The new highway is 40 km long including many bridges, viaducts and tunnels due to topog-
raphy of the site. The route of the site mainly follows the narrow area between the mountains
and North Sea due to economical reasons.
Unfortunately, deep and massive landslides are expected at some parts of the route up to
20 meters deep from the surface. The view of the Idemli bridge site is presented in Figure 1.

Figure 1. General view of the site-Idemli Viaducts.

115

NYBRIDGE_Book.indb 115 7/17/2013 8:42:22 AM

 116 Bridge analysis & design

2 LITERATURE REVIEW ON DESIGN OF BRIDGES ON LANDSLIDE AREAS

The number of bridges designed and built on landslide regions is quite small since routes are
normally designed to eliminate destructive effects of landslides. Therefore, in literature, solu-
tions and design guidelines for bridges on landslide areas are scarce.
In Croatia (Nossan et al., 2009), a viaduct on landslide area is designed and constructed
utilizing a foundation system consisting of diaphragm walls under a pile cap. The diaphragm
walls are positively connected to the pile cap, forming a typical pile foundation system. The
diaphragm walls are socketed into the firm ground below movable soil. Soil layers that are
prone to landslide consist of medium to highly plastic clay. The main idea of designing such a
laterally rigid foundation system is to resist full thrust resulting from a possible landslide. The
diaphragm walls are designed to sustain the lateral thrust of the sliding soil mass approxi-
mately equal to three times the theoretical passive earth force. The view of this system is
presented in Figure 2.
Inclinometer readings taken at various positions near the bridge indicates stabilization of
the mobilized soil after completion of the diaphragm wall construction. The initial yearly
movement of 15–20 mm reduced to 1–7 mm after construction of the diaphragm walls
(Nossan et al., 2009).

3 GEOTECHNICAL PROPERTIES AND SEISMICITY OF THE IDEMLI SITE

Seismicity of the site is quite low and peak ground acceleration for 475-year return period
earthquake is assumed as 0.1 g. Therefore, seismically induced landslide is not a major issue
for the bridge site.
At the time of structural design, inclinometer readings were available for a 3 month time,
indicating 1–3 mm horizontal movement. On the other hand, the data was collected during
autumn, without heavy rainfall. It is well known in engineering community that heavy rain-
fall triggers landslides. For this reason, foresight was necessary during structural design in
order to find out optimum solution. Structural design processed in parallel with construction
due to tight schedule.
At Idemli Site, inclinometer data implied horizontal soil movement up to 20 meters from
free-field. There are some underground water streams as indicated by Figure 3. There are
two threats in landslide areas: (1) Whole ground can slip and move as one. (2) Debris may
roll down the surface and therefore hit the deck, which is above ground level. In scope of this
project, only the first threat is taken into account.
The movable soil ranges from loose sand to rock particles. As shown in Figure 1, over the
bridge, a steep valley is present indicating a major risk for the bridge, due to huge amount of
potentially movable soil.

Figure 2. Application for design of a bridge foundation in a landslide area- (P1, P2, P3, DIK5, DIK6,
IK4 indicates inclinometer positions) (Nossan et al., 2009).

NYBRIDGE_Book.indb 116 7/17/2013 8:42:23 AM

 Design of Idemli Bridges in landslide areas 117

Figure 3. View of the movable soil and an underground stream.

4 ADOPTED SOLUTION

There are two alternatives for foundation design of bridges located on landslides. First
solution as adopted by Nossan et al., is to design a laterally rigid system capable of resisting
full lateral thrust applied by moving soil. According to authors, this solution requires well
documentation of site and geotechnical features. Moreover, soil movement rates should be
known for a long period of time.
Second solution implies minimal interference of the foundations with movable soil. This
solution necessitates long span bridges with piles that are founded at a level below the pos-
sible landslide depth, and strong enough to resist the force arising during a landslide. Instead
of diaphragm walls, piles with circular cross-section are preferable in this solution since cir-
cular cross-sections exhibit Omni-directional properties, being independent of the direction
of the landslide, at least in a cross-sectional basis without considering pile group effect. In
case of diaphragm walls, the direction of the landslide should be known exactly in order
to place short dimension of the wall perpendicular to landslide so as to reduce total lateral
thrust applied by the moving soil and to increase lateral rigidity of the foundation system.
In Idemli bridges, second alternative is adopted due to uncertainties in character and extent
of the landslide expected at the bridge. The foundations of the bridges are located on rela-
tively shallow landslide prone regions, as presented in Figure 4. The piles are embedded into
firm ground although not presented in Figure 4. Embedment lengths are in the range of
7–10 meters. The symmetrical arrangement of the span lengths necessitated some founda-
tions to be located on relatively deeper movable soils as compared to others.
At Idemli Viaducts, a composite superstructure consisting of four steel I girders and cast
in situ R/C slab with a maximum span length of 75 meters is selected. The superstructure
is continuous to be able to absorb lateral movements resulting from mobilized soil without
unseating of girders, if occurs. The span length is limited by girder height since profile of
the bridge is low. The girder height should be kept optimum/minimum so as to provide a
free flow path for the mobilized soil. Accumulation of soil at the bridge level may lead to
sweeping of the superstructure toward sea. The uphill soil as well as soil at the bridge site
consists of moveable soil, whereas difference in movement rates and restraints provided by
bridge etc. may result in accumulation of soil debris. The cost of constructing a barrier to
inhibit movement of uphill soils is prohibitively high, implying many rows of piles with
large diameter.

NYBRIDGE_Book.indb 117 7/17/2013 8:42:23 AM

 118 Bridge analysis & design

Figure 4. Foundations of the Idemli-2 Viaduct are located on relatively shallow landslide prone
regions.

Figure 5. View of Idemli Viaduct.

In foundation system CFT (Concrete Filled Steel Tubes) piles with a diameter of 165 cm are
selected. The outer steel shell is 20 mm thick and it is made up of S355 JR steel. At the inner
periphery of the pile, eight steel T profiles are used as longitudinal reinforcement. The concrete
inside of the steel tube is steel fiber reinforced concrete with steel fiber density of 30 kg/m3.
The steel fiber is utilized so as to increase ductility of the concrete. The depth of pile ranges
from 20 meters to 30 meters. The piles are embedded in base rock about 8 to 10 meters. There
are 16 piles in a pier foundation.
The views of the bridge, a pile and pile system are presented in Figure 5, Figure 6 and
Figure 7, respectively.
The pile system is capable of resisting thrusts applied by mobilized soil within its elastic
range. The expected thrust at a pile system is presented in Figure 8.
The construction of the viaducts has been completed and they are in use since November
2012. Up to now, no problem is observed due to lateral soil movement. Unfournately, incli-
nometers were removed following completion of the construction hindering a quantative
verification of this application.

5 CONCLUSIONS

Design of bridges located on movable soils is a special task due to uncertainties associated
with rate of movement and lateral thrust applied by mobilized soil.

NYBRIDGE_Book.indb 118 7/17/2013 8:42:24 AM

 Design of Idemli Bridges in landslide areas 119

Figure 6. View of a CFT pile.

Figure 7. View of pile system.

At the site of Idemli Viaducts, depth of landslide can reach up to 20 meters and character
of movable soil ranges from loose soil to rock particles. In design of these viaducts, maximum
span length of 75 meters was adopted which minimizes the risk of soil accumulation due to
very high girder depths and also minimizes number of piers in movable soils. In the founda-
tion system, concrete filled steel tubes (CFT) are utilized so as to minimize dimension of

NYBRIDGE_Book.indb 119 7/17/2013 8:42:25 AM

 120 Bridge analysis & design

Figure 8. Representation of the thrusts applied by the movable soils to the pile system.

the system prone to lateral soil movement, while maximizing resistance of the cross-section.
Sixteen steel CFT piles are utilized at each pier foundation.

REFERENCE

Nossan, S.A., Nossan, S.V., Stanic, B., Mihaljevic, I. 2009. A Bridge Foundation Resisting Sliding Soil Mass.
Proceedings of the 17th International Conference on Soil Mechanics and Geotechnical Engineering.

CH11_O_zkaya et al (Landslide)-Final.indd 120 7/17/2013 2:28:42 PM

 3 Bridge rehabilitation and retrofit

NYBRIDGE_Book.indb 121 7/17/2013 8:42:27 AM

This page intentionally left blank
 Chapter 12

Precast concrete filled steel grid panels for deck replacement

at the Newburgh-Beacon Bridge—South Span

W. Moreau
New York State Bridge Authority, Highland, New York, USA

S. Abdou
Ammann & Whitney, New York, NY, USA

ABSTRACT: Over 440,000 square feet of deck on the Newburgh Beacon Bridge/South
Span, which spans the Hudson River and carries three eastbound lanes of I-84, will be
replaced with precast, high performance, concrete filled, galvanized steel grid panels for the
entire 7,789 foot length of the structure. The deck replacement will be constructed in stages.
Over 1,900 panels, topped with a 1" polyester concrete overlay will replace the existing con-
crete deck. An innovative aspect of the design involves the use of steel haunches to support
the deck on the existing girders/stringers. The top flanges of the existing stringers will be
spliced, providing continuity at 28 relief joints. The project also includes the modifications
of existing finger joints, major electrical work, new sign structures, new drainage systems
and 15,680 linear feet of new precast concrete barriers. This paper describes the bridge and
its history, the development of alternatives and selection of deck type, innovative details and
construction stages.

1 INTRODUCTON

The Newburgh-Beacon Bridge, officially the Hamilton Fish Newburgh-Beacon Bridge,

spans the Hudson River carrying Interstate 84 in New York State between the Town of
Newburgh in Orange County to the west and the City of Beacon in Dutchess County to
the east. The Newburgh-Beacon Bridge is one of six bridges over the Hudson River run
by the New York State Bridge Authority. The bridge consists of two parallel spans with
similar appearance consisting of deck and cantilever trusses. The original north span was
awarded the 1965 AISC “Most Beautiful Bridge” award for long span bridges. The bridge
is the world’s 21st longest cantilever bridge and with 25 million crossings per year, it is the
most travelled of all spans maintained by the New York State Bridge Authority. Although
the newer south span looks very similar, major differences include wider piers, trusses, and
roadway and the addition of a pedestrian walkway, while the south span utilizes weathering
steel instead of carbon steel.
The north span was completed and opened to traffic on November 2, 1963 with one lane
in each direction. Due to increasing traffic volumes, the second span was constructed to the
south and opened to traffic seventeen years later on November 1, 1980. Following the open-
ing of the south span, the original structure was reconfigured by widening the roadway to
provide three-12-foot wide lanes with no shoulder and used for westbound traffic only. The
south span structure features a 53'-5" wide curb-to-curb roadway accommodating 3–12 foot
wide eastbound lanes, a 3-foot wide left shoulder, a 14-foot wide right shoulder and a 9-foot
wide pedestrian sidewalk along the south side.

123

NYBRIDGE_Book.indb 123 7/17/2013 8:42:27 AM

 124 Bridge rehabilitation and retrofit

2 BACKGROUND

Although the south span of the Newburgh-Beacon Bridge is the younger of the two structures,
wear and tear has taken its toll on the roadway deck, which was originally constructed with-
out an overlay. Early in 2012, the New York State Bridge Authority awarded a design con-
tract to Ammann & Whitney Consulting Engineers, P.C. of New York City for the design of
the South Span Deck Replacement on the Newburgh-Beacon Bridge. Ammann & Whitney’s
scope of work included performing an evaluation of various modular deck replacement sys-
tems for an accelerated construction contract for the south span of the Newburgh-Beacon
Bridge and the preliminary and final designs and design services during construction. The
program is funded by NYSBA’s five year, 160 million dollar capital program for all facili-
ties. A toll increase in effect since January 2012 will fund approximately $60 million of the
program.
The work undertaken by Ammann & Whitney was expedited to meet the November 2012
deadline for project advertisement, allowing a construction start by early spring of 2013. The
goal was that customers could see immediate improvements resulting from the toll increase,
the first in over 10 years. In addition to the redecking of the south span, the painting of the
north span is also currently underway. On January 22, 2013, the bids for this project were
opened and the low bid of 93.1 million dollars was accepted. This project will be the largest
self-financed capital project ever undertaken by the New York State Bridge Authority.

3 STRUCTURE DESCRIPTION

The north and south spans are similar in design and appearance with multi-span girder and
deck trusses along the approaches and a 3-span cantilevered through-truss spanning the
navigational channel. All spans are supported on concrete piers supported on caissons for
cantilever span piers and pile foundations for deck truss spans.
Superstructures: The south span is 7,788'-9" in length between abutments. The total length
of the cantilever channel span is 2,203'–6" with a 1,000 foot long main span. The west and
east approaches consist of girder spans from the abutments, with a total length of 541'–3" for
the west spans and 377'–6" for the east spans. Girder spans are composed of seven (7) built-
up welded plate girders, typically spaced at 8'-10", supporting the deck slab.
Deck truss spans: Connecting the main cantilever channel spans to the girder spans on
both the west and east sides are multi-span deck trusses. The deck truss spans have a total
length of 1,228'-3" on the west side and 3,438'-3" on the east side. The deck truss spans con-
sist of 2 trusses spaced at 33'-4" connected with lateral bracing along the top and bottom
chords and cross bracing at alternating panel points along the truss. Typical deck truss spans
consist of 13 panels at 34'-0" each, with the exception of the deck truss spans immediately
adjacent to the girder spans which have 10 panels, also spaced at 34'-0".
The floor system and deck framing consists of stringers between floorbeams located at
panel points. The roadway deck and barrier parapets in the deck truss spans are supported
on nine (9) W27 × 84 stringers spaced at 6'-8" resting on top of floorbeams. Floorbeams are
built-up welded plate members with web depth of 60" (see Figure 1).
Cantilever truss channel spans: The 3 spans comprising the channel spans include the
1,000 foot long main span and two 601'-9" long side spans. The side spans consist of 15 pan-
els at 40'-0" spacing, while the main span consists of 24 panels at 41'-8" spacing.
Hinges in the main span of the cantilever truss provide for expansion. For the cantilever chan-
nel truss spans, the deck is supported on nine (9) W30 × 99 stringers spaced at 6'-8" resting on top
of floor-beams. Floor-beams are built-up welded plate members with web depth of 72".
Roadway deck: The existing 53'-5" wide roadway deck, on the south span is the original
deck placed during construction of the bridge in 1980. It is a 7-3/4" thick, cast-in-place,
reinforced concrete deck with a 3/4" thick NovaChip wearing surface overlay added in 2005.

NYBRIDGE_Book.indb 124 7/17/2013 8:42:28 AM

 Precast concrete filled steel grid panels for deck replacement 125

Figure 1. Existing cross section at deck truss span.

The deck is non-composite with the stringers along the cantilever and deck truss portions of
the span. However, the deck acts in composite action with girders in the girder spans.
Bridge joints: There are a total of 18 expansion joints and 28 deck relief joints along the
entire structure. A majority of the deck relief joints and some of the expansion joints have
been modified or replaced since original construction. The original expansion joints in the
girder spans, which were elastomeric joints, have since been replaced with asphalt plug joints.
A majority of the original deck relief armored joints, along the deck truss and channel can-
tilever spans were replaced with asphalt plug joints. All expansion joints along the deck truss
and channel cantilever spans are open finger joints, with the exception of one multi-cells
modular joint at the east deck truss span.
The west girder spans are two and three span continuous over the piers and are separated
by expansion joints, in addition to the expansion joint at west abutment. The east girder
spans are three-span continuous with expansion joints at east abutment. The two expansion
joints between deck truss spans and girder spans are also finger joints.
The west and east eleven (11) deck truss spans have nine (9) intermediate finger joints at
the pin/hangers, two (2) finger joints at both ends of channel truss spans and twenty (20)
relief joints. In addition, there are two intermediate finger joints at the pin/hanger and eight
(8) relief joints at the channel truss spans.
Although the existing finger joints are generally in good condition, moisture below the
joints has led to corrosion and deterioration of floor system members and joint support
members. A majority of finger joint hold-down bolts are either missing or deteriorated
and vertical misalignment of the finger joint at one panel (cantilever truss main span) have
resulted in a noticeable bump for passing vehicles.
Barriers: Existing barrier parapets are 2'-10" Jersey shape concrete barriers with a steel
railing bolted to the top of the barrier on both sides of the roadway. The north barrier has
a one foot tall tube rail attached to the top of the barrier for the entire length of the span.
The south barrier is similar to the north barrier along the channel spans. For the deck truss
and girder spans, a two foot tall fence is attached to the top of the south barrier to provide
separation from the pedestrian walkway.

4 DECK REPLACEMENT

4.1 Deck replacement evaluation

Many factors played a role in determining the most cost effective deck type for this structure.
These factors included the construction cost and the need for both accelerated construction
and traffic maintenance during peak hours. It was immediately clear that modular deck types
would satisfy these project objectives.

NYBRIDGE_Book.indb 125 7/17/2013 8:42:28 AM

 126 Bridge rehabilitation and retrofit

Since the deck has been in service for 32 years, the team only needed to confirm the
extensive deterioration that had been noted in previous inspections. The most recent bien-
nial inspection, performed in 2011, identified multiple locations with potholes in the asphalt
wearing surface as well as in the concrete slab, with exposed reinforcing steel. The biennial
inspection reports also identified deficiencies in the truss superstructure, joints and other
critical components. Repairs for these deficiencies were included under the deck replacement
project.
The established history along with confirmatory investigation led to a recommendation to
replace the existing roadway on the South Span to ensure structural integrity, safety and rider
comfort along this important crossing of the Hudson River.
Three major deck replacement alternatives were evaluated:
Alternative 1: Precast concrete high performance (HP) deck panels
Alternative 2: HP concrete-filled steel grid deck panels (5" steel grid deck) with 1.5" over-
fill (see Figures 2 and 3)
Alternative 3: Steel orthotropic deck (open rib with ¾" steel deck plate, supported/welded
to WT shape subfloor beams)

Figure 2. Typical cross section at girder spans.

Figure 3. New cross section at channel spans.

NYBRIDGE_Book.indb 126 7/17/2013 8:42:28 AM

 Precast concrete filled steel grid panels for deck replacement 127

Each alternative was modular and met construction staging requirements. For each
alternative, options were investigated and load rating calculations were performed for
stringers/girders and the truss system to determine applicability.
Options that were considered include:
• Use of normal-weight concrete vs. lightweight;
• Use of 9-1/2" monolithic slab vs. 8" slab and 1-1/2" wearing surface overlay for precast
concrete deck alternative;
• Use of 1" wearing surface overlay with 1-1/2" concrete overfill vs. 2" concrete overfill for
steel grid deck alternative;
• Use of thin vs. thick wearing surface overlays.
Different methods and details for installation in the field were also investigated for each
alternative and corresponding option. Items studied included panel-to-panel connections
in the longitudinal and transverse direction and panel to existing stringer/girder connec-
tions. Other options that were investigated included utilizing steel haunch connections, high
early strength (HES) concrete, rapid setting concrete, and ultra-high performance concrete
(UHPC) for closure pours.
Major options for each alternative included the following:
Precast concrete panels alternative (Hallmark et al., 2012): The use of ultra-high perform-
ance concrete (UHPC), grouted bars, post-tensioning and grouted curved threaded bars
were investigated for both the longitudinal and transverse panel to panel connections.
Panel to existing steel connections: The use of UHPC haunches with shear studs and
bolted steel angle haunches were examined and investigated for panel to existing steel
connection.
Precast concrete full filled steel grid alternative (Claybaugh et al., 2004): For the panel to
panel and existing stringer connections: options investigated included rapid setting con-
crete in haunches concrete haunches with threaded shear studs welded to existing string-
ers, and couplers to splice reinforcing bars and trim steel plates at end of each panel at the
longitudinal joints over stringers. The steel haunch option included the use of the rapid
setting concrete within the deck and the use of a longitudinal steel haunch to attach the
deck to existing stringers/girders. For the transverse panel to panel joints, options included
rapid setting concrete with welded distribution bars and alternating welded studs to the
steel grid main bars.
Orthotropic deck options (Higgins 2003; Tsakopoulos and Fisher 2003): The investigation
of the orthotropic deck included closed and open ribs and providing composite action
and increased structural rigidity by connecting the deck plate to the truss vertical mem-
bers (shear connector between the deck and trusses). We determined that the use of a ¾"
open rib plate, welded (fillet weld) to the WT subfloor beams, spanning between existing
stringers/girders, would satisfy load and fatigue requirements. Full penetration field deck
welds for the panel to panel connections and bolted splices for the sub-floor beam connec-
tions were investigated.
One of the primary considerations in the selection of the deck panel attachment meth-
ods was the need to accommodate overnight and off-peak hour construction staging limita-
tions which limited lane closures to off-peak periods (9:30 AM to 3:00 PM and 6:30 PM to
6:00 AM). Of particular importance was the need to adequately support new panels on the
existing floor system while also providing support for existing deck sections adjacent to the
newly installed deck and meeting peak hour traffic demands. This would require longitudinal
stringers or girders to simultaneously support new and existing deck sections with an open
joint and appropriate overlay transitioning to allow for traffic. Similar considerations were
addressed for transverse joints between new and existing deck sections.
Conceptual designs for each deck alternative were prepared. The alternatives were evalu-
ated to determine the most desirable deck alternative to proceed through to preliminary and

NYBRIDGE_Book.indb 127 7/17/2013 8:42:29 AM

 128 Bridge rehabilitation and retrofit

final design. Each alternative was evaluated while considering constructability, maintenance
and protection of traffic, construction staging, environmental issues, service life, maintain-
ability, initial construction cost, maintenance costs, and life cycle costs. The suitability of
each alternative for overnight construction work was a pass/fail evaluation factor.
All alternatives would result in a weight reduction when compared to the existing rein-
forced concrete deck and therefore would allow for an increase in live load capacity of the
existing steel superstructure without providing modification. Load ratings of all alternates
found that the redecked existing structure would exceed HS25 loading capacity.

4.2 Life Cycle Cost Analysis (LCCA)

Life Cycle Cost Analysis (LCCA) was provided for the preferred option of each of the three
alternatives. Since each scheme yielded a different service life, cost comparison based solely
on initial cost would not be appropriate. To accurately assess the live cycle costs for each
selected alternative, a present worth analysis of each alternate was performed.
Factors that affect life cycle costs include initial construction cost, maintenance frequency and
costs, expected service life of the structure, discount rate utilized and the planning horizon.
The expected service life of the structure for the three selected alternatives was based on
prior projects and experience. The planning horizon was 75 years and a discount rate of
3.5% was used for the analysis. The construction cost was estimated using 3.5% escalation,
construction beginning toward the end of 2012 and lasting 34 months.

4.3 Selection of deck type system

Factors that played a major role in the selection of the most desirable alternative included:
• Construction cost and duration,
• Use of steel haunches to accelerate construction which would allow traffic during peak-
hours and facilitate future inspection and maintenance,
• Use of fast setting concrete (one hour cure) to allow for construction during limited work
windows,
• Overnight/off-peak hour construction for the deck replacement,
• Use of high performance materials (concrete and grout) for longer service life,
• Use of overlays with waterproofing properties to protect the new deck,
• Service life of deck, grout, joints and overlays,
• Life cycle costs.
Although the deck replacement investigation indicated that all three of the above described
alternatives (precast concrete panels, precast concrete filled steel grid and orthotropic deck) were
viable, the construction costs proved to be the principle deciding factor. This led Ammann &
Whitney to recommend a precast high-performance (HP) normal-weight concrete full-filled
galvanized steel grid for the deck replacement on the South Span of the Newburgh-Beacon
Bridge. The deck itself consists of 5" WT section main bars. In addition, a 1-1/2" concrete
overfill and 1" thick wearing surface overlay was also recommended for a total deck thickness
of 7-1/2". Deck panels will be attached and supported on the existing stringers and girders
using steel angle haunches. All deck relief joints on the truss spans will be eliminated to reduce
water infiltration and future deterioration on the steel superstructure.
The concrete-filled steel grid deck includes the use of high-friction granite aggregate con-
crete mix in the panels and polyester concrete for the wearing surface overlay due to their
superior durability and successful past performance. Particular attention was paid to the
concrete mix and overlay, as it is important to provide materials with long lasting service life
and to minimize maintenance. The specified concrete mix will reduce alkali–silica reaction
(ASR) and the use of the high friction granite aggregate will minimize future deterioration
of the concrete.

NYBRIDGE_Book.indb 128 7/17/2013 8:42:29 AM

 Precast concrete filled steel grid panels for deck replacement 129

4.4 Deck design, criteria and details

The recommended steel grid incorporates 5" WT sections as main bars with secondary and
round bars in the perpendicular direction. The use of a steel grid has the advantage of allow-
ing splices between adjacent panels to achieve continuity. The steel grid and reinforcing bars
will be galvanized for corrosion protection. The use of the steel angle haunches will allow the
deck panels to be installed and adequately support traffic at the completion of off-peak hour
construction periods. The precast concrete steel grid deck will provide an estimated 75 year
service life, while the polyester concrete overlay has a service life of 30 years.
The following design criteria were used during final design of the steel grid deck:
• WT section (main bars) galvanized steel grid with steel haunches/connection was detailed
and checked for fatigue to connect the deck to the existing weathering steel stringers/gird-
ers (during construction stages and the permanent conditions),
• 1.5" concrete overfill and 1" polyester concrete overlay we used,
• Designs were provided for 6'-8" stringer spacing for the truss spans and 8'-10" for the girder
spans,
• The steel haunch system was designed for composite action of steel girder/stringer with
steel grid deck. The bolted steel angle haunch system was checked for both bearing and
horizontal shear transfer, and to satisfy slip-critical design criteria,
• Bolted connections (3/4" HS bolts—ASTM A 490) were used to connect the steel grid trim
plates to the steel haunch angles at the end supports of the panel and to connect the upper
haunch angles (welded to the main bar) to the lower haunch angles welded (intermittent
filet welds) to the top flanges of the existing steel stringers/girders,
• Fatigue analysis utilizing 3 times the fatigue truck axle load on dual axles,
• Category D detail was used to check the welded connections (Fatigue limit = 7 ksi),
• Design of the new deck considered the elimination of relief joints and splicing the existing
stringers,
• Steel deck design was based on AASHTO LRFD Section 6.6 which considers magnified
fatigue stress in the deck (AASHTO 2012),
• Modifications to finger expansion joints for the truss span deck were provided and the
existing modular joints will be replaced with a new finger joints,
• New armorless expansion joints with elastomeric concrete headers were designed for the
girder spans,
• High-performance normal-weight concrete mix (F’ c = 5 ksi) with high friction granite
aggregate is used in the design of the precast concrete,
• Rapid setting concrete is used for the closure pours (one hour cure) to allow deck replace-
ment during off peak hours,
• Galvanized reinforcing bars were used,
• Wire mesh at negative moment areas was detailed for crack control (at floorbeam &
stringer areas) to eliminate potential cracking at the negative moment areas,
• The location of the transverse joints between panels will be optimized to minimize crack-
ing potential by locating the joints at mid span of the stringers and away from the floor-
beams. This will minimize deterioration to the floorbeams due to water seepage and
provide stringer continuity at the intermediate floorbeams.

4.5 Steel haunch details and panel to panel connections

Panel-to-panel connections along longitudinal joints will be achieved by providing a bolted
connection at the trim plates and using couplers for splicing the reinforcing bars. The trans-
verse joint connection will be established by field welding secondary bars. The reinforcing
bars may be welded or substituted with studs welded to the web of the main bars at the clo-
sure pour. Rapid setting concrete will be used for the closure pours. The detail was developed
for longitudinal joint to provide an adequate support to the panel ends to accommodate

NYBRIDGE_Book.indb 129 7/17/2013 8:42:29 AM

 130 Bridge rehabilitation and retrofit

the traffic during construction stages, where the existing deck is still in place adjacent to a
new panel (see Figure 4 & Figure 6). Intermediate panel support will consist of two angles;
one shop welded to the grid deck panel main bars and the other angle to the top of existing
stringers/girders, which will then be bolted together during final installation (see Figure 5).
Oversized holes will be provided to allow adjustment during placement. The steel haunch/
deck panel supports provide superb versatility as they allow newly installed deck panels to be
adequately supported without need for concrete haunches and curing periods. Steel haunches
also allow new deck panels to be attached to exposed portions of stringers that are support-
ing the existing concrete deck (Slade 1989; Steinberg 1997; Zoghi et al., 2008).

4.6 Modifications to existing finger joints

Modifications to the existing finger joints resulted in two major advantages; cost effective-ness
due to eliminating the cost of new finger joints (15 joints) and the need for temporary steel plat-
ing during construction. Temporary steel plating is always a concern since they can result in an
unsafe deck condition under the heavy truck loading. The joint modification will be provided
in two phases. In the first phase, during deck replacement, the existing deck will be removed 2'

Figure 4. Steel haunch longitudinal joint between new panel and existing deck.

Figure 5. Steel haunch at intermediate support and fascia stringer.

NYBRIDGE_Book.indb 130 7/17/2013 8:42:29 AM

 Precast concrete filled steel grid panels for deck replacement 131

Figure 6. Longitudinal precast steel grid panel details at deck truss and channel truss spans.

Figure 7. Precast barrier detail.

from the limits of the finger joints and the overfill for the new deck panels will be tapered by ½"
to meet the elevation of the existing finger joints. After the completion of the deck replacement,
shim plates will be used to raise/reset the existing finger joints by 1.5" to match the elevations
of the top of new overlay. Rapid setting concrete will be used for the new joint headers after the
field installation of the new steel channels, angles, studs and reinforcing bars.

4.7 Precast barrier details

Based on the NYSDOT Bridge Manual, a design testing level of TL-4 for the new barriers is
appropriate. Per Section 6 of the Manual: “TL-4 (PL-2)–Taken to be generally acceptable for
the majority of applications on high-speed highways, expressways and interstate highways
with a mixture of trucks and heavy vehicles.” Consequently, the types of barriers which can be
specified are: 3'-6" Single-slope, 3'-6" F-Shape and 3'-6" Vertical Parapet. It is expected that a
single metal rail, similar to the existing, will be installed atop the new barrier on the north side.
The existing metal fence that currently sits atop the barrier on the south side (adjacent to the
right shoulder) will be removed, galvanized and reinstalled on the new barrier.

NYBRIDGE_Book.indb 131 7/17/2013 8:42:30 AM

 132 Bridge rehabilitation and retrofit

Due to speed of construction and staging, it may be advantageous to utilize precast barri-
ers in lieu of cast in place (see Figure 7). The concrete barrier can by constructed by one of
three methods: cast-in place, slip formed or precast. As per the NYSDOT Bridge Manual, if
the precast method is chosen, the contractor must use one of the preapproved precast con-
crete barrier systems.

5 CONSTRUCTION STAGING

Maintenance and Protection of Traffic options have been developed in order to construct the
new deck without significant impacts to bridge traffic. Numerous factors affecting the stag-
ing have and will continue to be taken into account, including the following:
• Maintain the present number of lanes (3) during all peak traffic periods.
• A minimum lane widths of 11'-6" during all peak traffic periods.
• Avoid the use of “cattle chutes”.
• Construction activities should not impact toll operations.
• Temporary traffic barrier lengths should be limited to the expected extent of deck removal/
replacement for a 1-week period of time.
• Transitions between old and new deck should be smooth and uniform. The use of
temporary steel deck plates should be avoided.
• Provide adequately sized work zones for the Contractor.
It is desirable for the expedited construction stages to include the following stages/phases,
for peak and off peak hours:
Stage 1 - Phase A: Install temporary fencing 2-foot from the southern roadway barrier,
within the pedestrian walkway to separate construction activities from pedestrian traffic.
Place temporary concrete barrier 15'-4" from existing roadway barrier, providing 3–12'
travel lanes. Remove existing roadway barrier fence. Remove existing southern barrier and
deck during peak hours. Clean top of steel and install lower haunch angles (see Figure 8).
Stage 1 - Phase B: Off-Peak Hours: Close southern-most lane for a length of roadway equaling
the expected night’s operations. Place temporary concrete barriers approximately 26'-8" from
southern roadway barrier during off-peak activities. Remove 13'-4" of existing deck, install
remaining haunch angles and replace the deck removed in phases A and B (see Figure 9).
Stage 2 - Phase A: Place temporary concrete barrier 13'-4" from existing roadway barrier,
providing 3–11'- 6" travel lanes. Remove existing roadway barrier and deck. Clean top of
existing steel and install lower haunch angles.

Figure 8. Stage 1A—Peak.

NYBRIDGE_Book.indb 132 7/17/2013 8:42:30 AM

 Precast concrete filled steel grid panels for deck replacement 133

Figure 9. Stage 1B—Off-peak.

Stage 2 - Phase B: Off-Peak Hours: Staging area along the right shoulder (southern por-
tion of the roadway) will be vacated. Close the northern-most lane adjacent to staging area
for a length of roadway equaling the expected night’s operations. Place temporary concrete
barriers approximately 26'-8" from northern roadway barrier during off-peak activities.
Remove remaining portions of existing deck, install remaining haunch angles and replace
the deck removed in stage 2—Phases A and B.
The dimension of each panel in the transverse direction is half the width of the roadway
and approximately 8–10' in the longitudinal direction. Minimizing the longitudinal joints
to only one in the middle of the structure significantly contributed to containing cost and
accelerating the construction.

6 CONCLUSIONS

The current roadway of the south span of the Newburgh-Beacon Bridge is the original deck dating
from time the original construction in 1980. The deck consists of 7-3/4" thick cast-in-place rein-
forced concrete. The roadway exhibits extensive deterioration as noted by previous inspections.
It is necessary to replace the existing CIP deteriorated deck to ensure continued and safe
operation of the span, and to provide additional live load capacity to handle current heavy
truck loads. During 2012, Ammann & Whitney Consulting Engineers performed investiga-
tions and evaluations of various panelized/modular deck replacement systems for use on the
south span of the Newburgh-Beacon Bridge. These included precast concrete deck, concrete-
filled steel grid deck and steel orthotropic deck alternatives. Each alternative also had options
that were investigated and utilized for load rating calculations for stringers/girders and truss
system to determine applicability of the alternative.
Each alternative was evaluated based on feasibility and ease of construction, maintenance
and protection of traffic, construction staging, environmental concerns, service life, ease of
maintenance, initial construction cost, maintenance costs and life cycle costs with particular
emphasis on the suitability of each alternative for overnight construction work.
The recommended alternative was the precast high-performance (HP) normal-weight con-
crete full-filled galvanized steel grid. The use of steel haunches to connect panels to existing
stringers/girders and of rapid setting concrete was key to allowing the construction to pro-
ceed during off peak hour periods. A polyester concrete overlay will be installed after all deck
sections are placed to protect the deck against deterioration and extend its service life.

NYBRIDGE_Book.indb 133 7/17/2013 8:42:30 AM

 134 Bridge rehabilitation and retrofit

The total cost of construction which includes the demolition and disposal of the existing
deck and fabrication and installation of the new deck, barriers, drainage, electrical system,
and modifications of existing finger joints, while taking into account construction staging
and placing the deck in two pieces with a single longitudinal joint in the middle of the road-
way, was bid at $93.1 million with a construction schedule of 34 months.

REFERENCES

AASHTO. (2012). AASHTO LRFD Bridge Design Specifications, Customary U.S. Units, the American
Association of State Highway and Transportation Officials, Washington, D.C.
Claybaugh, B., Earls, C., and Ahmadi, A. (2004). “Fatigue and Strength Performance of Concrete-
Filled Steel-Grid Bridge Deck.” Journal of Bridge Engineering, 9(5), 435–443.
HÃllmark, R., White, H., and Collin, P. (2012). “Prefabricated Bridge Construction across Europe and
America.” Practice Periodical on Structural Design and Construction, 17(3), 82–92.
Higgins, C. (2003). “LRFD Orthotropic Plate Model for Live Load Moment in Filled Grid Decks.”
Journal of Bridge Engineering, 8(1), 20–28.
Slade, L. (1989). “Haunches and Hangers.” Civil Engineering—ASCE, 59(3), 64–67.
Steinberg, E. (1997). “Reliability of Structural Steel Haunch Connections for Prestressed Concrete.”
Journal of Structural Engineering, 123(10), 1382–1389.
Tsakopoulos, P., and Fisher, J. (2003). “Full-Scale Fatigue Tests of Steel Orthotropic Decks for the
Williamsburg Bridge.” Journal of Bridge Engineering, 8(5), 323–333.
Zoghi, M., Farhey, D., and Gawandi, A. (2008). “Influence of Haunches on Performance of Precast-
Concrete, Short-Span, Skewed Bridges with Integral Abutment Walls.” Journal of Performance of
Constructed Facilities, 22(2), 101–107.

NYBRIDGE_Book.indb 134 7/17/2013 8:42:31 AM

 Chapter 13

Blatnik and Bong Bridges wire rope replacement: Maintaining

Minnesota’s unique bridge inventory

L. Amundson
Parsons Brinckerhoff, Minneapolis, Minnesota, USA

R. Haight
Parsons Brinckerhoff, New York, New York, USA

E. Lutgen
Minnesota Department of Transportation, St. Paul, Minnesota, USA

ABSTRACT: Previous bridge inspections identified corrosion and other condition issues
with the suspenders that support the main span bridge decks of the John A. Blatnik and
Richard I. Bong Memorial Bridges located between Duluth, Minnesota, and Superior,
Wisconsin. The Blatnik crossing is 7,975 feet long with the main span consisting of a 600-foot
arch truss with a 480-foot suspended deck. The Bong Bridge is an approximately 11,800-foot
crossing with a main span consisting of a steel tied arch with a 500-foot span. With major
bridge preservation efforts programmed for both bridges in 2012, it was determined that a
detailed assessment of the condition of the wire ropes and sockets on the Blatnik Bridge and
the helical wire strand and sockets on the Bong Bridge was warranted. This would provide
Mn/DOT and WisDOT information to determine if replacement of the wire ropes is needed
to extend the service life of these two major and unique structures. A construction contract
was executed to remove two suspenders from each bridge and have them tested destructively
to determine their modulus of elasticity, ultimate strength, and overall rope and socket con-
dition. This presentation will discuss the configuration and existing condition of the sus-
pended deck system, the removal and replacement procedure, the results of the testing, and
the maintenance recommended to extend the service life of the two bridges.

1 INTRODUCTION

On August 1, 2007, the State of Minnesota experienced a disaster resulting in a loss of life
and infrastructure with the collapse of the I-35 W Bridge. This tragedy triggered a mandate
within and outside of the state for the inspection of bridges focusing on the state’s fracture
critical inventory. On August 5, 2007, a team lead by Parsons Brinckerhoff had bridge inspec-
tion staff in Minnesota begin a fracture critical inspection of the John A. Blatnik Bridge
(Blatnik) connecting Duluth, Minnesota, and Superior, Wisconsin. It was during this inspec-
tion that the condition of the wire ropes supporting the deck was examined and provided
to the Minnesota Department of Transportation (Mn/DOT). Since an additional border
structure with shared ownership by Mn/DOT and Wisconsin Department of Transporta-
tion (WisDOT)—the Richard I. Bong Bridge (Bong)—also shared a similar suspended deck
feature, inspectors familiar with suspender rope condition evaluation assessed the condition
of the Bong Bridge as well. As a result of those inspections, it was recommended to perform
further destructive testing on a selected sample of suspenders from both bridges. Destructive
testing of wire rope and strand is a conventional method for assessing the ultimate strength
of these bridge members, as a visual inspection and condition assessment applies only to the

135

NYBRIDGE_Book.indb 135 7/17/2013 8:42:31 AM

 136 Bridge rehabilitation and retrofit

exterior wires in the cross section, which make up a small percentage of the cross section of
the wire strand or rope.
These two bridges are among eight bridges in the Mn/DOT inventory that contain sus-
pended decks but, due to their significant length over the St. Louis Bay, they represent
60 percent of the area of suspended deck in the state. That said, bridges with suspended
decks represent a very small portion of the state’s total inventory as a measure of deck area
and, as a result, the state looked for assistance in developing a strategy to assess the condition
and make recommendations for long-term maintenance and possible replacement.
Mn/DOT and WisDOT have proactively maintained these structures with Mn/DOT per-
forming a major rehabilitation of the Blatnik Bridge that was completed in 1994. At that
time, wrapped seizing wires were removed from suspenders where they passed through the tie
at the lower sockets. In addition to removal of the wire seizings, the ropes were cleaned and
painted and the cut-out where the ropes pass through to the lower sockets was enlarged to
mitigate the abrasion that occurred.

2 BRIDGE LOCATION, SUSPENDER CONFIGURATION, AND GENERAL

CONDITION

The Blatnik Bridge (Figure 1) crosses St. Louis Bay in Duluth, Minnesota, and opened to
traffic in 1961. The bridge underwent a major rehabilitation that was completed in 1994.
The crossing is 7,975 feet long and includes an arch truss with main span of 600 feet. The
suspended deck portion of the arch truss is 480 feet. The Bong Bridge (Figure 3) also crosses
St. Louis Bay and opened to traffic in 1985. That crossing is 11,800 feet and includes a
500-foot suspended span-tied arch.
Each suspender location on the Blatnik Bridge consists of four ASTM A603 galvanized
structural bridge rope suspenders, socketed top and bottom, and positioned in a square con-
figuration that straddles the center line of the floor beam. Original drawings and specifica-
tions indicate that the suspenders are 1-5/8-inch diameter wire rope with Class A coating and
with original test breaking strengths of 275.0 kips, 281.2 kips, and 282.4 kips. The minimum
breaking strength with a Class A coating is 123.0 tons (246 kips).
From the 2007 fracture critical inspection, it was noted that there were numerous wire
breaks found on the Blatnik Bridge wire rope suspenders. These were generally located where

Figure 1. John A. Blatnik Bridge.

NYBRIDGE_Book.indb 136 7/17/2013 8:42:31 AM

 Blatnik and Bong Bridges wire rope replacement 137

Figure 2. Blatnik Bridge elevation view, showing recommended suspender removal locations.

Figure 3. Richard I. Bong Bridge.

the suspender ropes pass through the top of the anchor brackets approximately 2 to 3 feet
above the lower suspender socket. The wire breaks were caused by abrasion of the ropes
against the steel truss member due to live load vibration and wind action. It was evident dur-
ing the 2007 inspection that the cut-out for the suspender ropes in the anchor brackets had
been enlarged during the 1990s bridge rehabilitation to avoid further abrasion. The Blatnik
suspenders have seizing wires and the original openings in the truss near the upper sockets
(no enlargements in structural steel adjacent to the suspenders) and seizing wires removed
with enlarged openings near the lower sockets at the bottom brackets.
Each suspender location on the Bong Bridge consists of two ASTM A586 helical wire sus-
penders, socketed top and bottom, and positioned to straddle the center line of the tie girder.
Original drawings and specifications indicate that the suspenders are 2-9/16-inch diameter hel-
ical wire strand with Class A coating on the inner wires and Class B coating on the outer wires.
The minimum breaking strength with this coating configuration is 386.0 tons (772 kips).
From the 2007 fracture critical inspection, it was noted that the condition of the strands
did not vary significantly across the bridge. There was no paint on the suspenders and all
ropes had zinc oxide from corrosion of the galvanizing. There were some locations with signs
of abrasion where the rope passed through the box tie girder. There was also corrosion on
the shim plates at the lower socket. The condition of the strand under the shim plates was
not visible.

NYBRIDGE_Book.indb 137 7/17/2013 8:42:32 AM

 138 Bridge rehabilitation and retrofit

3 SELECTION OF LOCATIONS FOR REMOVAL AND TESTING

OF SUSPENDERS

In an effort to better understand the range of suspender condition, one suspender was
selected to be removed from each bridge as the test candidate in worst condition and one was
selected to be the test candidate in average condition. Suspenders were selected from the same
side of each bridge for ease of traffic maintenance in their removal. The shortest suspenders
were also avoided due to the stiffness of the truss at these locations and the steeper incline of
the upper chord making their removal more difficult. In addition, these locations would have
yielded a shorter suspender length for sample testing. See Figures 2 and 4 for each bridge
general elevation and suspender locations.

3.1 Blatnik Bridge

For the worst location, the suspender rope at the West Truss, Panel Point 17N, the Northwest
suspender, was removed and replaced. This rope was found to have six broken wires on two
different strands that comprise the rope. This suspender rope is 53 feet 7-11/16 inches long
under full dead load. For the worst-case location sample, it was first recommended on the
contract drawings to remove and replace the suspender at Panel Point 16S, the Northwest
suspender. See Figure 5 for the general condition of this suspender. During a preliminary
site walk-through by the contractor and Mn/DOT, it was noted that there was a riveted truss
splice connection at this location and at every even-numbered panel point where the suspend-
ers pass through the tie girder. Using this location would result in a large number of rivet
removals and replacements to access the socket area. As a result, the 2007 inspection notes
were reviewed, and another location with similar suspender conditions was found at Panel
Point 17N, Northwest suspender.
For a typical suspender condition, the suspender rope at the West Truss, Panel Point 15N, the
Southeast suspender, was removed and replaced. This rope was found to have spots of Stage 3
corrosion on the bottom 12 feet and spots of Stage 4 at the roadway level. Wire corrosion is
categorized visually by corrosion stages 1 through 4 in accordance with NCHRP Report 534
(TRB 2004). Stage 3 corrosion exhibits spots of brown rust covering up to 30% of the surface
of a 3-inch to 6-inch length of wire. Stage 4 corrosion exhibits brown rust covering more than
30% of the surface of a 3-inch to 6-inch length of wire. This rope is 47 feet 4-3/16 inches long
under full dead load. See Figure 6 for the general condition of this suspender.

3.2 Bong Bridge

For the worst case location on the Bong Bridge, removal and replacement of suspender
strands at Panel Point 7R, East Strand, was performed. This location was found to have
fretting rust at the box chord penetration and one broken wire. See Figure 7 for the general
condition of this suspender.

Figure 4. Bong Bridge elevation view, showing recommended suspender removal locations.

NYBRIDGE_Book.indb 138 7/17/2013 8:42:42 AM

 Blatnik and Bong Bridges wire rope replacement 139

Figure 5. Blatnik Bridge: west truss panel point

16S northwest rope (Six broken wires, three broken Figure 6. Blatnik Bridge: West truss
wires on two different strands at the anchor panel point 15N southeast rope.
bracket. West Truss Panel Point 17N was in similar
condition.)

Figure 7. Bong Bridge: panel point 7R east Figure 8. Bong Bridge: panel point 3R east
strand (Corrosion on shim plates and on cable strand (Corrosion on shim plates and on cable
adjacent to shims. Broken wire at socket. 70% adjacent to shims.)
zinc corrosion.)

For a typical suspender condition, removal and replacement of suspender strands at Panel
Point 3R, East Strand, was performed. This location was found to have Stage 1 and Stage 2
corrosion with pack rust in the shims at the lower socket. In accordance with NCHRP Report
534, Stage 1 corrosion exhibits spots of zinc oxidation on the wires and Stage 2 exhibits zinc
oxidation on the entire wire surface. See Figure 8 for the general condition of this suspender.
Both of these suspender strands are located on the designated Right (R) (or South) Arch.
Both strands were also selected from the south side of the bridge for ease of traffic mainte-
nance and because this side of the bridge has a sidewalk/bikeway. The suspender strand at
Panel Point 3R is 59 feet 11-7/16 inches long and at Panel Point 7R, the suspender strand is
77 feet 9 inches long, both under full dead load.

3.3 Summary of suspenders recommended for removal and destructive testing

Blatnik Bridge Worst case: West Truss, Panel Point 17N, the NW rope
Average case: West Truss, Panel Point 15N, the SE rope
Bong Bridge Worst case: Panel Point 7R East Strand
Average case: Panel Point 3R East Strand

NYBRIDGE_Book.indb 139 7/17/2013 8:42:42 AM

 140 Bridge rehabilitation and retrofit

3.4 Suspender testing results

The suspender removal and replacement was performed by Piasecki Steel Construction
Company based in Castleton, New York. The suspenders were removed from the Bong
Bridge during the week of October 11–15, 2010, and from the Blatnik Bridge the following
week. The suspenders were cut into lengths that provided 20 feet at each socket end for labo-
ratory testing and ease of shipment.
Field work involved using a jacking frame to jack the suspender against the truss tie girder
(in the case of Bong Bridge) and against the floorbeam (in the case of Blatnik Bridge) using
a clamping plate attached to the suspender to be removed. See Figure 9 for details of the
jacking frame. This system relieves the load on the suspender between the clamping plates
and the lower suspender socket. The lower suspender socket is then removed from its support

Figure 9. Conceptual jacking frame for suspender removal and replacement (Bong Bridge shown,
Blatnik Bridge similar).

NYBRIDGE_Book.indb 140 7/17/2013 8:42:44 AM

 Blatnik and Bong Bridges wire rope replacement 141

point, the jacking load is then relieved, and the suspender is removed. A new suspender
is then inserted in its upper socket support, the clamping plates are reattached to the new
suspender, and the process is reversed, with the jacking frame jacked to the same load as
for removal. If there is a gap between the suspender and its support point, this gap is filled
with shims. Using this method, the new suspender takes the same load as the removed sus-
pender. Laser vibration measurement was considered for determining the tension in the old
suspender prior to removal as well as after installation of the new suspender, to ensure that
the proper load was being applied to the new suspender. Although this method was used on
the Cass Street Bridge in Wisconsin, which had been recently constructed at the time, DOT
staff deemed that the reuse of new suspenders of the same lengths as existing ones, as well as
the tensioning of the new suspenders to the same tension as those removed, would provide
acceptable accuracy without the use of laser vibration measurement.
The removed suspender samples were shipped to WireCo WorldGroup (WireCo) in
Carrollton, Missouri. Suspender testing occurred in December 2010. The suspender sam-
ples used the existing socket ends of each suspender, an 18-foot test length, and a fabricated
socket at the opposite end of the test length. The samples were tested for modulus of elastic-
ity and maximum load at failure, visual examination before and after testing, and splitting
and examination of existing socket cones. The modulus of elasticity and maximum load
at failure data are summarized in Table 1 below. While corrosion-fatigue is an important
aspect of deterioration of suspender ropes, under this investigation, no fatigue tests were
performed. Only tension tests were performed under this investigation.

3.5 Blatnik Bridge

Through documentation received from Mn/DOT, the original suspender ropes for the Blatnik
Bridge were 1-5/8-inch diameter, ASTM A603 with Class A coating, gross metallic area
1.270 square inches, and a minimum breaking strength of 123.0 tons (246 kips). All samples
tested above this minimum breaking strength except for the Panel Point 17N bottom sample
at 228.8 kips (worst case location). This suspender’s top sample, Panel Point 17N top, also
had the highest breaking strength of the four samples for this bridge, indicating that the cor-
roded and broken wires at the cutout approximately 6 feet above the bottom socket location

Table 1. Maximum load at failure and modulus of elasticity test data for Blatnik and Bong Bridges.

Blatnik Bridge Bong Bridge

Maximum load Modulus Maximum load Modulus

at failure of elasticity at failure of elasticity

pounds psi pounds psi

Panel Point 263,859 24,846,894 Panel Point 759,431 24,365,482

15N top 3R top
Panel Point 270,943 24,496,938 Panel Point 765,725 24,445,632
15N bottom 3R bottom
Panel Point 288,205 24,496,938 Panel Point 757,305 24,283,166
17N top 7R top
Panel Point 228,844 24,649,093 Panel Point 754,864 24,283,166
17N bottom 7R bottom
High 288,205 24,846,894 High 765,725 24,445,632
Low 228,844 24,496,938 Low 754,864 24,283,166
Average 262,963 24,622,466 Average 759,331 24,344,362
Standard 24,938 165,923 Standard 4,653 77,871
deviation deviation

NYBRIDGE_Book.indb 141 7/17/2013 8:42:45 AM

 142 Bridge rehabilitation and retrofit

(that were noted in the 2007 inspection) have affected the overall strength of this suspender.
While the cutouts at the anchor bracket shown in Figure 5 were enlarged in the 1990s to
avoid abrasion, the loss in strength between top and bottom of the Panel Point 17N sample is
approximately 20 percent. See Figure 10 for the general condition of the Blatnik suspenders.
The breaking strength of the Panel Point 17N bottom sample is below the 123.0-ton (246-
kip) minimum acceptable breaking strength for 1-5/8-inch ASTM A603 wire rope. The break-
ing strength of the Panel Point 17N bottom sample is also below the 95 percent of minimum
strength rejection limit for new wire rope. The governing dead and AASHTO Standard
Specification (AASHTO 2002) live load for this bridge was computed to be 315 kips per
suspender panel for one truss. Divided over four suspenders per truss panel location yields an
average of 78 kips per suspender rope. The average safety factor for the four samples tested is
approximately 3.4, while the safety factor for the lowest breaking strength at Panel Point 17N
bottom is approximately 2.9. A rule of thumb for designing new bridges is generally for sus-
pender safety factors to be a minimum of 3.0. While the broken wires in the suspenders are
of concern and the broken wires have contributed to this reduction in the safety factor of the
suspenders, Mn/DOT took action in the 1990s to alleviate the cause of abrasion between the
suspenders and the bottom anchor brackets and slow rate of deterioration to help preserve
the life of this bridge component. If there is a suspender loss, there are three other suspenders
at a panel on the Blatnik Bridge so the deck does not span between two panels if a suspender
is lost. The analysis undertaken for this project did not consider the load carrying capability
of the remaining suspenders if there was a loss of one suspender.
Some wire break failures noted in WireCo’s report for the Blatnik Bridge indicated shear
wire failures in some of the outer wire layers where wire steel section loss had occurred,
rather than the typical cup-and-cone ductile wire failure where no steel section loss was evi-
dent. Socket cone splits showed no signs of corrosion, no air gaps around wires that would
indicate poor socketing during original construction, nor any wire pullout. The shear wire
failures indicate wire brittleness, with virtually no yielding.
Recommendations for maintenance of the suspenders was provided to Mn/DOT, includ-
ing cleaning the exterior of the suspenders using SSPC-SP2 (hand tool cleaning) or careful

Figure 10. Blatnik rope shown (Prior to testing, the areas of corrosion and damage were noted. The
damage shown here is located 6 feet from the bottom socket.)

NYBRIDGE_Book.indb 142 7/17/2013 8:42:46 AM

 Blatnik and Bong Bridges wire rope replacement 143

use of SSPC-SP3 (power tool cleaning) and applying a protective three-coat zinc-rich paint
system or a spray galvanizing protective system until a more thorough protective coating can
be applied. This is particularly important for ferrous corrosion in the splash zone from road-
way overspray down to the bottom sockets in order to slow down and possibly arrest cor-
rosion deterioration. Special emphasis of suspender inspection during biennial inspections
was also recommended to drive targeted maintenance painting and to identify suspenders
that may need to be replaced due to wire breakage or advanced steel wire section loss. It was
recommended that inspectors note any overall deterioration of the zinc galvanizing, general
bulging, or individual wire bulging in suspenders that may indicate interior wire corrosion or
localized deterioration at the suspender/socket interface, so timely preventive maintenance
can be performed. It should also be noted that the contract documents specified that the
jacking frames for both bridges remain the property of Mn/DOT. In addition, the contractor
provided Mn/DOT with extra suspender material for each bridge so that they can replace
suspenders and continue to perform testing to monitor the overall suspender condition.

3.6 Bong Bridge

Through documentation received from Mn/DOT, the original suspender ropes for the Bong
Bridge were 2-9/16-inch diameter, ASTM A586 with Class A coating on the inner wires and
Class B coating on the outer wires and a minimum breaking strength of 386 tons (772 kips).
All four samples were in good condition visually, including the zinc galvanizing, except for
some localized zinc and ferrous corrosion at the guide casting location at the top of the box
tie girder (where some abrasion can take place between the steel guide casting and galvanized
wires), and at the suspender/socket interface. These locations exhibited some Stage 3 condi-
tions, rated per NCHRP Report 534. All four samples tested just slightly below this minimum
breaking strength, with an average sample breaking strength of 759.3 kips, or a reduction of
1.7 percent from the ASTM A586 minimum specified strength.
Wire break failures noted in WireCo’s report for the Bong Bridge suspenders indicated
ductile wire failures. See Figure 11 showing the typical cup and cone failure. Socket cone
splits showed no signs of corrosion, no air gaps around wires, or wire pullout. See Figure 12
for the general condition of the zinc cone.
While the reduction in breaking strength is minor for the years of service, it should be
noted that the computed dead load plus live load at an individual suspender for this bridge is
263 kips, yielding an average suspender safety factor of 2.9. Even at the ASTM A586 mini-
mum breaking strength of 772 kips, the computed suspender safety factor when new was only
slightly greater than 2.9. A rule of thumb for designing new bridges is generally for suspender
safety factors to be minimum 3.0. If there is a suspender loss, there is at least one additional

Figure 11. Bong rope shown (The wire failure shown was ductile in nature showing the typical cup and
cone and the failure point.)

NYBRIDGE_Book.indb 143 7/17/2013 8:42:46 AM

 144 Bridge rehabilitation and retrofit

Figure 12. Bong zinc cone shown (The zinc cone is shown removed from the production body and cut
in half to reveal full zinc penetration throughout.)

suspender at a panel on the Bong Bridge so the deck does not span between two panels if a
suspender is lost. The analysis undertaken for this project did not consider the load carrying
capability of the remaining suspender if there was a loss of one suspender.
It was recommended that the suspenders for the Bong Bridge be targeted for special empha-
sis during inspections. With two suspenders per location, and each suspender near the mini-
mum generally accepted safety factor for new bridges, it is important to monitor the condition
of the suspenders. Inspections should note any overall deterioration of the zinc galvanizing at
the upper and lower guide castings, general bulging, or individual wire bulging in suspenders
that may indicate interior wire corrosion or localized deterioration at the suspender/socket
interface, so timely preventive maintenance can be performed. As some suspender lower
socket shims show active corrosion, this condition should also be monitored during future
inspections. It was suggested that if pack rust between shims occurs that alters the height of
the shim pack in its bearing area, consideration is given to replacing the affected shim packs.
Recommended maintenance of the suspenders includes localized cleaning of the exterior
of the suspenders where there is ferrous corrosion. This can be accomplished using SSPC-
SP2 (hand tool cleaning) or careful use of SSPC-SP3 (power tool cleaning). The cleaned
wires should then be protected by applying a three-coat zinc-rich paint system or a spray gal-
vanizing protective system until a more thorough protective coating can be applied. Special
inspection emphasis of suspenders during biennial inspections is also recommended to
− drive targeted maintenance painting
− identify suspenders that may need to be replaced due to advanced wire steel section loss
− identify shim packs that may need to be replaced due to pack rust in the bearing area of
the shim
If targeted and aggressive preventive maintenance is carried out on a timely basis in the
identified vulnerable suspender areas near the guide casting and suspender/socket interface,
the suspenders can be expected to approach or reach their design life of 50 years before need-
ing replacement.
Another area of focus for future rehabilitation is the guide casting detail for the suspend-
ers on the Bong Bridge These castings guide the suspenders through the cutouts in the box
tie girder and arch box. The suspender/casting faying surfaces are currently galvanized wire
against steel casting, and this has caused abrasion of the zinc protective coating on the sus-
penders. Rehabilitation of the detail to include a zinc-lined casting would avoid contact
between dissimilar metals and avoid untimely abrasion of the protective coating from the
suspender wires in these localized areas.

NYBRIDGE_Book.indb 144 7/17/2013 8:42:47 AM

 Blatnik and Bong Bridges wire rope replacement 145

4 CONCLUSIONS

While bridges like the Blatnik and Bong are more common to the east and west coast, they are
relatively unique structures within the Mn/DOT and the WisDOT inventories. Their location
in a cold, snowy, and windy climate places demands on these structures and particularly the
suspenders with snow removal abrasion and the resultant corrosion occurring in the splash
zone typically from November through March. Mn/DOT has taken a proactive approach to
understand the condition of and maintenance for these structures to provide continued serv-
ice to residents and businesses of the Arrowhead Region. Because visual inspection does not
reveal the internal condition of suspender ropes, other bridge owners with similar suspended
deck spans can benefit from Mn/DOT’s experience and consider a removal, replacement and
testing program to ascertain the condition of their suspender ropes. With an experienced
contractor, the removal and replacement process was undertaken very quickly, both bridges
were complete in eight working days.
While the status of the Blatnik Bridge is subject to Mn/DOT’s plans for possible replace-
ment, the condition assessment of the suspender ropes is important to maintaining the condi-
tion of this unique aspect of Mn/DOT’s inventory. With an understanding of the typical and
worst case conditions of the suspender ropes of each bridge and with the contractor-supplied
jacking frame and additional suspender material for each bridge, Mn/DOT can monitor and
maintain the condition of the remaining suspenders to provide a robust structure for years to
come until a determination is made regarding future rehabilitation or replacement.

REFERENCES

AASHTO (American Association of State Highway and Transportation Officials), 2002. AASHTO
Standard Specification for Highway Bridges, 17th Edition.
TRB (Transportation Research Board), 2004. Guidelines for Inspection and Strength Evaluation of Sus-
pension Bridge Parallel Wire Cables, Washington, DC, NCHRP Report 534.

NYBRIDGE_Book.indb 145 7/17/2013 8:42:48 AM

This page intentionally left blank
 Chapter 14

Rehabilitation of the Alexander Hamilton Bridge and Highbridge

Interchange Ramps

M. Tughral Shaikh & H.-P. Ho

Jacobs Engineering, New York, NY, USA

T. Bashir
New York State Department of Transportation, New York, NY, USA

ABSTRACT: The Alexander Hamilton Bridge (AHB) has not undergone major repairs
since it opened to traffic on January 15, 1963. Currently, the estimated traffic on the bridge
is in excess of 200,000 vehicles per day. With a construction cost tag of $407 Million, the
Alexander Hamilton Bridge Rehabilitation Project is the largest single-contract construction
project in the history of the New York State Department of Transportation (NYSDOT).
Construction on the project began in the spring of 2009 and is scheduled to continue until
the end of 2013. The scope of work for this project is the strengthening, widening, and com-
plete deck replacement of the AHB main steel arch span including its two (2) West approach
spans and seven (7) East approach spans, and the rehabilitation and/or replacement of eight
(8) other existing bridge structures that would be directly impacted by the construction of
AHB. The project also includes the rehabilitation of on-grade approach roadways for all
nine (9) bridges. The project limits extend from Amsterdam Avenue in New York County
to Undercliff Avenue in Bronx County, New York City. Constructing a new state of the art
skateboard park, redesigning and improving the parklands, and reconstructing the existing
playgrounds within the project limits are also part of this project. This paper presents the
complexities in the designs of the permanent as well as the temporary structures to suc-
cessfully complete the proposed rehabilitation work for the heavily travelled AHB and its
associated Highbridge Interchange Ramps, development of methodologies for the staged
demolitions/constructions, multi-agency coordination to minimize interruptions to their
operations, constructability issues due to limited clearances and restricted work-hour win-
dows, and challenges and measures taken during construction to complete the rehabilitation
works and to meet the ambitious construction schedule.

1 INTRODUCTION

The Alexander Hamilton Bridge and Highbridge Interchange Ramps rehabilitation/

replacement project includes a total of nine (9) existing bridge structures on or spanning the
Cross Bronx Expressway (CBE) from Amsterdam Avenue in New York County to Undercliff
Avenue in Bronx County, New York City. The project is under the jurisdiction of NYSDOT
Region 11. The nine (9) bridges in this construction package include the Alexander Hamilton
Bridge, Ramp “TE” over the CBE, the Highbridge Interchange Connector Ramps included:
three (3) bridges on “Ramp D”, two (2) bridges on “Ramp A”, and a Ramp “B & L” bridge,
and the Undercliff Avenue Bridge over the CBE. The on-grade roadways included the seg-
ments of Ramps “D” and “A” roadways between the bridge structures, and the segments of
AHB, Ramps “TE”, Ramp “B” and Ramp “F” approach roadways. See Figure 1 for project
site key map and Figure 2, which shows the overall view of the project including AHB. The
entire project is done in metric units.

147

NYBRIDGE_Book.indb 147 7/17/2013 8:42:48 AM

 148 Bridge rehabilitation and retrofit

1.1 Bridge 1—Alexander Hamilton Bridge, (AHB)

Alexander Hamilton Bridge was constructed in 1962. The superstructure for the Main Arch
Span consists of a total of four (4) steel box arch ribs supporting the spandrel columns
that support the four lines of longitudinal spandrel girders. The ends of each rigid steel
box arch rib are anchored into the massive concrete anchorage located at either side of the

Figure 1. Project site key map—Alexander Hamilton Bridge and Highbridge Interchange Ramps.

Figure 2. Alexander Hamilton Bridge and Highbridge Interchange Ramps.

NYBRIDGE_Book.indb 148 7/17/2013 8:42:48 AM

 Rehabilitation of the Alexander Hamilton Bridge and Highbridge Interchange Ramps 149

Harlem River. The longitudinal spandrel girders in the Main Arch Span support the trans-
verse floorbeams and cantilevered brackets which support the stringers. The stringers sup-
port the concrete deck slab of the AHB. There are two (2) West Approach spans and seven (7)
East Approach spans at the ends of the Main Arch Span. The existing superstructure for two
(2) West approach spans and six (6) East approach spans of AHB consists of non-redundant
steel girders, floorbeams and stringers framing supporting the concrete deck slab. Only the
last span in East approach, 7E, consists of redundant multi-girders framing supporting the
concrete slab. Figure 3 shows AHB south elevation.
The original AHB consists of two (2) independent structures divided by a longitudinal deck
joint along the center median and carries six (6) through traffic lanes of Cross Bronx Express-
way, I-95, three (3) northbound lanes and three (3) southbound lanes in each half of the bridge,
and an additional acceleration lane for On-Ramp “A” on the southbound side of the bridge and
a deceleration lane for Off-Ramp “D” on the northbound side of the bridge. The bridge spans
(total length = 452.7 m; 1,485.4 ft.) over the Highbridge Park and the Harlem River Drive in
Manhattan, the Harlem River, and the Metro-North Commuter Railroad tracks, On-Ramp
“A”, Off-Ramp “D”, the Major Deegan Expressway and Sedgwick Avenue in the Bronx.

1.2 Bridge 2—Ramp TE

Ramp “TE” over I-95 is located at the west end of the AHB. It was constructed in 1952 and
carries one lane of Off-Ramp traffic from the northbound Cross Bronx Expressway, I-95, to
the eastbound Washington Bridge (181st Street Bridge or Route US1). Ramp “TE” spans
over the Highbridge Park, Trans Manhattan Expressway Connector Ramps, and the Cross
Bronx Expressway, I-95.
The segment of north approach roadway that links between Ramp “TE” and the Washington
Bridge is on-grade. This segment of roadway will also be rehabilitated as part of this project.

1.3 Highbridge Interchange Ramp – Bridge 3, 4 & 5 – Ramp “D”

These three bridges are segments of the circular Ramp “D” which was constructed in 1962
and carry two lanes of Off-Ramp traffic from the northbound Cross Bronx Expressway, I-95,
to the Major Deegan Expressway, I-87, both northbound and southbound. The bridges span
over the Major Deegan Expressway and Sedgwick Avenue.
The two segments of Ramp “D” between Bridge 3 and Bridge 4, and between Bridge 4 and
Bridge 5, and a segment of Ramp “F” (north approach roadway of Bridge 4) are on-grade
roadways. These segments of roadway will also be rehabilitated as part of this project.

1.4 Highbridge Interchange Ramp – Bridge 6 & 7 – Ramp “A”

These two bridges are segments of the circular Ramp “A” which was constructed in 1962 and
carry two lanes of On-Ramp traffic from the southbound and northbound Major Deegan

Figure 3. Alexander Hamilton Bridge south elevation.

NYBRIDGE_Book.indb 149 7/17/2013 8:42:49 AM

 150 Bridge rehabilitation and retrofit

Expressway, I-87, to the southbound Cross Bronx Expressway, I-95. The bridges span over
the Major Deegan Expressway and Sedgwick Avenue.
The segment of Ramp “A” between the two structures is on-grade roadway. This segment
of roadway will also be rehabilitated as part of this project.

1.5 Highbridge Interchange Ramp – Bridge 8 – Ramp “B & L”

This structure, a segment of Ramps “B & L”, was constructed in 1962 and carries two lanes
of traffic from the northbound Major Deegan Expressway, I-87, to the northbound Cross
Bronx Expressway, I-95 and southbound Cross Bronx Expressway, I-95. The bridge spans
over Sedgwick Avenue.
The segment of Ramp “B” (south approach roadway of Bridge 8) is on-grade roadway.
This segment of roadway will also be rehabilitated as part of this project.

1.6 Bridge 9—Undercliff Avenue Bridge

This structure carries the Undercliff Avenue over I-95 was constructed in 1962 and has one
lane of one direction of vehicular traffic, two parking lanes and two wide sidewalks for
pedestrian traffic.

2 PROJECT OBJECTIVE

This is a bridge rehabilitation project and the primary objective of this project is to upgrade
the existing deteriorating bridge structures to current standards and to extend their service
life for an additional fifty years or to replace the existing bridge structure entirely where fea-
sible. Included in the project is also the rehabilitation of the associated segments of on-grade
approach roadways for the bridges. The need for improvements on this section of the Cross
Bronx Expressway (CBE) is evidenced by the structural deficiencies caused by age and dete-
rioration of the existing structures as found in the in-depth inspections performed in 1993
(NYSDOT 1995–2004 and NYSDOT 2004). If this project were not implemented, the bridges
would have continued to deteriorate at an increasing rate, until load posting would have been
required, and subsequently may have led to bridge closures.
Subsequent to the in-dept inspections performed in 1993 for all the bridge structures on or
spanning the CBE and within the Highbridge Interchange an Interim Repairs Contract was
let in 1998 by NYSDOT to address the immediate repairs needed for the bridge deteriora-
tions. During construction of the Interim Repair Contract (NYSDOT 1998), it was discov-
ered that the conditions of the existing concrete deck slabs and some critical steel members of
the AHB and its connector ramps are in much worse conditions than anticipated. They were
in need of immediate replacement or repairs. Furthermore, the existing bridges do not meet
the current seismic codes (AASHTO 2004) and must be adequately strengthened. Therefore,
NYSDOT decided and initiated the final rehabilitation designs project for those selected few
bridge structures.
However, for the rehabilitation of AHB two other bridges at both ends of AHB must also
be involved due to the presence of their piers. A pier column of Ramp “TE” at the western
end of AHB and the center pier of the Undercliff Avenue Bridge to the east of AHB both fall
in the median of CBE, I-95. Removal of these piers were needed to accomplish the planned
staged construction and shifting of traffic lanes on AHB, and, therefore, it is reasonable to
include the planned rehabilitation works for these two structures in the same construction
contract with AHB and its associated ramp structures.
Similarly, since construction on the bridges noted above will require a coordinated mainte-
nance of traffic plan to minimize disruption and further limit the disruption to one construc-
tion project, it was agreed that it would be advantageous to also include the six Interchange

NYBRIDGE_Book.indb 150 7/17/2013 8:42:49 AM

 Rehabilitation of the Alexander Hamilton Bridge and Highbridge Interchange Ramps 151

ramp bridges and seven segments of on grade roadways which are directly impacted by the
construction on AHB in the same construction contract.

3 DESIGN APPROACH AND CHALLENGES

Rehabilitation of the Alexander Hamilton Bridge (the main steel arch span and its approach
spans), six (6) Interchange ramp structures, Ramp “TE”, and Undercliff Avenue Bridge is
one of the most challenging and complex design projects in the history of NYSDOT. One of
the major design constraints was to maintain and not to interrupt the entire interchange traf-
fic (approximately AADT 200,000 vehicles) (NYSDOT 2004) during construction.
The first Final Designs began in the winter of 2002 and completed at the end of 2006.
NYSDOT solicited construction bids in 2007 but no contractor came forward because of
their unwillingness to accept the challenge to design the complex temporary supports for the
AHB and the temporary ramp structures while reconstruction Ramps “A” and “D”.
As per NYSDOT directive, Jacobs took the challenge to provide the designs of all the
temporary supports and ramp structures (AASHTO 1995 and OSHA) which is usually the
contractor’s job to design during construction. Upon completion of all the temporary sup-
ports and structures designs, NYSDOT re-issued the construction bid on January 22, 2009.
This time, five bids were received and the joint venture team CCA Civil/Halmar International
was selected based on their lowest bid of $407 Million.
Five (5) alternatives were originally considered for each of the nine (9) bridge structures
(NYSDOT 1995–2004 and NYSDOT 2004).
Alternative 1: No-Build and Alternative 2: Deck Rehabilitation with Monolithic Overlay
alternative did not meet any of the project objectives; therefore, both alternatives were elimi-
nated. The only three (3) remaining feasible alternatives considered for each of the nine (9)
project bridges are as follows:
• Alternative 3: Rehabilitation of Superstructure with Deck Replacement
• Alternative 4: Replacement of Superstructure and Deck
• Alternative 5: Complete Replacement of Bridge
For each of the nine (9) bridges, several design alternatives were investigated and the rec-
ommended design alternatives were selected for each bridge. The recommended alternative
for Alexander Hamilton Bridge and the six (6) Highbridge Interchange Ramp bridges is
Rehabilitation of Superstructure with Deck Replacement including retrofit substructure to
meet current seismic requirements and the recommended alternative for Ramp “TE” and
Undercliff Avenue Bridge is Complete Replacement of the Bridge.

3.1 Alexander Hamilton Bridge (AHB)—Steel Arch Span Bridge and approach spans
The rehabilitation designs of the Main Arch Span and its approach spans took into consid-
eration the difficulty of staged deck replacement as well as 3.353 m (11.0 ft) widening (each
side) to introduce standard shoulders and modifications to the existing superstructure fram-
ing including the retrofit of existing substructure to meet the current seismic requirements
(AASHTO 2004) while maintaining all through lanes of traffic in each direction during con-
struction. The feasibility of using prefabricated components was also investigated intensively.
See Figures 4 thru 7.
In order to minimize the duration of construction and the number of traffic detours, it was
determined that new longer cantilever brackets in the Main Arch Span and addition of new
fascia girders in the approach spans would be the most structurally sound and economical
method to increase the existing roadway width for maintaining and shifting of existing traf-
fic lanes. Based on the existing and new superstructure framing configurations, the proposed
work on AHB can be minimized to six (6) construction stages. Furthermore, based on the

NYBRIDGE_Book.indb 151 7/17/2013 8:42:49 AM

 152 Bridge rehabilitation and retrofit

Figure 4. Alexander Hamilton Bridge—existing and proposed arch span section.

reduced number of construction stages, it was estimated that due to the irregular framing and
geometry for the proposed staged construction, using prefabricated components for the new
construction offered no advantages as compared to the conventional construction. In fact,
the cost of using prefabricated components would be higher than the conventional construc-
tion. Consequently, cast-in-place concrete deck slab was selected for the proposed deck slab
replacement for AHB. In addition, stainless steel rebar is being used to acquire longer life
and to reduce the thickness of roadway deck slab to optimize superstructure and reduce the
seismic responses.
Due to the complexities of the existing and proposed Alexander Hamilton Bridge struc-
tures, detailed computer models, sophisticated structural analyses, numerous design studies,
and intensive evaluations were developed prior to selecting the final design scheme for the
rehabilitated new structure (NYSDOT 2004).
For the Main Arch Span with new longer cantilever brackets, several structural alterna-
tives were studied to strengthen the existing steel arches, spandrel columns and girders, as
well as floorbeams while eliminating the existing longitudinal deck joints and combining
the two existing separate bridges together into one structure as shown in Figure 4. The final
design required the addition of new median struts between intermediate arch ribs for tying all
four arch ribs together, strengthening of existing spandrel columns and arch ribs by adding
steel reinforcement plates, and reinforcing the critical connections. Repair or replacement of
several floorbeams and stringers was also necessary for the final design. The effects of each
staged constructions with complicated temporary loading cases as well as the thermal behav-
ior of the modified structure (temporary and final cases) were carefully studied to optimize
the final framing in the Main Arch Span.

NYBRIDGE_Book.indb 152 7/17/2013 8:42:50 AM

 Rehabilitation of the Alexander Hamilton Bridge and Highbridge Interchange Ramps 153

For the approach spans, the additions of new fascia girders and interior retrofit girders
were needed for the new widened bridge deck and to provide additional redundancy for
the new bridge structure. See Figures 5 and 6 for existing and proposed approach span
section. Furthermore, all the existing single span girders were spliced at alternate piers and
made two spans continuous resulting in the elimination of 50% of the existing transverse
deck joints. Similar to the Main Arch Span the elimination of the existing longitudi-
nal deck joint along the center of the existing two separate bridges required the replace-
ment of existing two (2) median cantilever brackets with one (1) piece of new median
floorbeams.
Numerous seismic alternatives were investigated during the final design phase to achieve
the optimum responses for the existing bridge substructures (abutments, anchorages and
piers) and to isolate the required strengthening to only few existing substructures.
• The new bridge deck is widened on each side throughout its entire length; therefore, add-
ing considerable amount of dead load; consequently, additional seismic response load.
• The existing East approach spans consist of tall rigid concrete frames founded on pile-
footings, with the exception of the East abutment on spread footing, and Pier 7E solid wall
on caissons.
• The basic design concept was to strategically modify the existing superstructure bearing
layout to limit the required strengthening for only 2–3 sets of pier foundations.
• All existing fixed bearings at both West and East abutments were replaced with expansion
bearings to reduce the overturning forces for the abutments.

Figure 5. Alexander Hamilton Bridge—existing east approach span section (west approach similar).

Figure 6. Alexander Hamilton Bridge—proposed east approach span section (west approach
similar).

NYBRIDGE_Book.indb 153 7/17/2013 8:42:51 AM

 154 Bridge rehabilitation and retrofit

• All existing simply supported superstructure both in West and East approach spans were
spliced in between piers and made continuous.
• Fixed bearings were selected at the massive existing Pier 1W for the West approach spans
as well as Pier 1E in the East approach spans to resist the seismic loading.
• Similarly, Pier 2E, 4E and 5E were strengthened for the new seismic loading cases:
added piles for 2E & 4E and thickened the pier column bases and footing for shear. See
Figure 7.
• Due to the rigidity of existing Pier 7E, it was found that no strengthening was required.
• Pier 3E support the widest portion of AHB and the two connector ramp structures:
RAMP A & D, therefore, additional piles were added for Pier 3E footing due to additional
vertical loads.

3.2 Ramp TE & Undercliff Avenue Bridge

The existing Ramp TE bridge superstructure consists of a reinforced concrete hollow box girder
with 200 mm (8 in) thick concrete deck supported by nine (9) pier columns, Figure 8. The piers
consist of single reinforced concrete circular cantilever columns enclosed in steel shell of 2 m
(6.5 ft) diameter. The Undercliff Avenue Bridge consists of a 2-span reinforced concrete rigid
frame structure with the center rigid wall running along the centerline of CBE, I-95.
The design of the new Ramp TE and Undecliff Avenue Bridge took into consideration
the difficulty of replacing the existing bridge structure while maintaining AHB Traffic below.
A pier column of existing Ramp TE penetrates through the median of AHB at west approach
spans, and the center pier of the Undercliff Avenue Bridge is located in the median of the
CBE, a short distance from the east end of the AHB approach slab. These two piers inter-
fere with the implementation of staged shifting of traffic required to maintain the existing
number of lanes of CBE, I-95, traffic in each direction on AHB during its rehabilitation, and
as a result, the removal of these obstructions is essential. Consequently, due to the removal

Figure 7. Alexander Hamilton Bridge—Proposed pier 2E (Pier 4E similar) with new cap beam and
new added piles and pile caps.

NYBRIDGE_Book.indb 154 7/17/2013 8:42:51 AM

 Rehabilitation of the Alexander Hamilton Bridge and Highbridge Interchange Ramps 155

Figure 8. Ramp TE—existing aerial view.—Dated 2004.

of piers, the span lengths of the existing two bridges over the CBE mainline are essentially
doubled. Therefore, for Ramp TE and the Undercliff Avenue Bridge, the only feasible solu-
tions were to replace both existing bridge structures. It should be noted that for the Ramp
TE, some existing substructure elements are reused.
In addition to the removal of one Ramp TE pier column that penetrates through the
median of Alexander Hamilton Bridge, two other pier columns immediately adjacent to the
existing AHB fascias had to be relocated to accommodate the new widened AHB. Only two
existing piers were to be rehabilitated and reused to accommodate the new steel box girder
span lengths. The new bridge ramp was designed to meet current NYSDOT bridge design
standards and ASSHTO seismic criteria. The new designs also eliminated or improved the
nonstandard features of the existing bridge. The Ramp TE was constructed with long-term
full closure of the ramp utilizing an alternate detour route, and the Undercliff Avenue Bridge
was constructed in two (2) stages while maintaining one (1) lane of traffic on the bridge. Fig-
ures 9 and 10 show existing and proposed elevations while Figures 11 and 12 show existing
and proposed cross-section of Ramp TE respectively.

3.3 Highbridge Interchange Ramps

Rehabilitation of superstructure included cast-in-place deck replacement, widen deck by
0.533 m (1.75 ft) each side of Ramp “A” & “D”, bearing replacement, substructure concrete
repair, retrofit substructure to meet current seismic requirement is also applicable to all six (6)
interchange ramp structures. Unlike the Alexander Hamilton Bridge, the Interchange ramps
have limited deck slab widths and their geometries have either sharp skews or sharp curva-
tures. The physical constraints of these structures impede the proposed deck replacement to
be performed by closing the entire ramp. A complete closure of these structures is the only
choice to complete the required deck replacement work.
The deck replacement work will have to be performed during full bridge closures using
temporary bridge structure for traffic detours. Full width deck replacement in the work area

NYBRIDGE_Book.indb 155 7/17/2013 8:42:52 AM

 156 Bridge rehabilitation and retrofit

Figure 9. Ramp TE—existing elevation.

Figure 10. Ramp TE—proposed elevation.

Figure 11. Ramp TE—existing cross-section.

is required for these bridges due to the variable spacing of the support framing and sharp
curvatures of the existing ramps.
Temporary Bridges were designed for both Ramps “A” and “D” and built adjacent to both
existing ramps.

NYBRIDGE_Book.indb 156 7/17/2013 8:42:53 AM

 Rehabilitation of the Alexander Hamilton Bridge and Highbridge Interchange Ramps 157

Figure 12. Ramp TE—proposed cross-section.

4 CONSTRUCTION

Construction on the project began in the spring of 2009, and is scheduled to continue until
the end of 2013. Currently the construction is on schedule and under budget.

4.1 Site constraints

Construction of this project was estimated to take several years to complete and would
severely affect the daily lives of motorists traveling through or with an origin or destination
in the CBE (I-95) corridor. Since the CBE operates at or near capacity during peak hours,
the existing three Expressway lanes of through traffic in each direction must be maintained
for Alexander Hamilton Bridge during construction. Temporary lane/lane closures/closures
will be implemented only within the permissible off-peak hours. Accordingly, the Alexander
Hamilton Bridge will have to be widened and the median removed to enable shifting of traf-
fic to maintain the same number of travel lanes on the bridge during construction. As for
the access ramp work (other than Ramp TE) and for the Undercliff Avenue Bridge work,
short-term closures with reasonable detours and staged construction are being employed,
respectively. Ramp TE was closed to traffic for the duration of replacement construction.
Two (2) of the nine (9) bridge structures included in this bridge rehabilitation/bridge
replacement project: Ramp TE and the west approach spans of Alexander Hamilton
Bridge (AHB) are within parkland (Highbridge Park in New York County and Washington
Bridge Park in Bronx County). The proposed bridge rehabilitation/bridge replacement of
both structures involved both temporary and permanent property impacts. The temporary

NYBRIDGE_Book.indb 157 7/17/2013 8:42:54 AM

 158 Bridge rehabilitation and retrofit

and additional permanent easements in Highbridge Park and Washington Bridge Park for
the construction of AHB and Ramp TE has required the preparation of a Programmatic
Section 4(f) (NYSDOT 2004) Evaluation and close coordination with the NYC Department
of Parks and Recreation (NYCDPR) in order to minimize disturbances and to convert the
Section 4(f) properties for the use of the proposed bridge rehabilitation/replacement work.
The proposed widening of the existing AHB deck on each side will have permanent
impacts to the existing R.O.W. of the Cross Bronx Expressway. The need for additional width
on the Alexander Hamilton Bridge is for providing standard shoulders on the bridge and
shifting and maintaining the existing number of travel lanes during construction. The per-
manent impact to the parkland is limited to minor permanent easement takings of the 4(f)
property (NYSDOT 2004). However, the proposed replacement of Ramp “TE” will have
no permanent impacts to the 4(f) property. The new bridge structure will follow the existing
bridge alignment and will be within its existing right-of-way and an additional easement is
not required for this ramp structure. Four of the existing Ramp “TE” pier columns within the
4(f) property will be eliminated for the proposed structure.

4.2 Construction stages

There are six (6) stages for the AHB:
Stage 1—Modification of the existing median
Stage 2—Rehabilitation and widening of north portion of the bridge
Stage 3—Rehabilitation and widening of the south portion of the bridge
Stage 4—Rehabilitation of the south inner portion of the bridge
Stage 5—Rehabilitation of the north inner portion of the bridge
Stage 6—Rehabilitation of the median portion of the bridge

4.3 Design challenges during construction

Agencies Coordination:
PANY&NJ—Truck traffic restrictions for GWB lower deck and closures of GWB
NYC Parks—Limited disturbance to parklands under the bridge constructions
NYCDEP—Coordination with the on-going DEP project within the project limits
NYCDOT—Restricted lane closures
MetroNorth Railroad—Coordination of construction works over the railroad tracks
Miscellaneous Field Issues:
• Encountered unexpected utility lines and features that interfere with the proposed
constructions.
• Field Survey of existing structures do not meet the As-Built and Contract Plans.
• Top of existing steels vary from the theoretical values shown on the As-Built and Contract
Plans requiring adjustments of haunches.
• Existing drainage pipes system either non-existing or no-functioning.
• Existing structural details differ from the details shown on the As-Built and Contract Plans.

4.4 Temporary structures

AHB:
• Temporary supports for the replacement of bearings and repairs/strengthening of existing
pier bents placed on existing piles footing cap to prevent settlement.
• Temporary braces to the existing pier columns for lateral stabilities.
• Limited spaces/heights and presence of existing structures, roadways or streets for the con-
struction of temporary supports.

NYBRIDGE_Book.indb 158 7/17/2013 8:42:55 AM

 Rehabilitation of the Alexander Hamilton Bridge and Highbridge Interchange Ramps 159

Figure 13. Plan of Temporary Widening of Highbridge Ramps.

RAMPS:
• Independent temporary bridge structures were constructed adjacent and parallel to the
existing ramp structures, Figure 13.
• Temporary spread footings were used for the support of all temporary bridge structures.
• Reinforced earth walls were used extensively for temporary structures/roadways.

5 CONCLUSIONS

The Alexander Hamilton Bridge Rehabilitation Project is the largest single-contract con-
struction project in the history of the New York State Department of Transportation. To
complete a project of this size and complexity within both the planned schedule and the
available budget is a great challenge. At the beginning of the project, numerous unanticipated
field conditions resulted in the project falling behind schedule by many months. However, as
the construction progressed, the project has been able to recoup much of this lost time and is
currently back on the original schedule. The ability to achieve this was primarily due to the
open and constant lines of communication and cooperation between the NYSDOT person-
nel, contractor, inspection team, and designer. With all of them are working together, they
are able resolve the difficulties presented in the field in both a time and cost efficient manner.
Based on recent projections, this complex and technically challenging project is expected to
be completed both on time and within budget. Extensive Inter-agency coordination, incen-
tive and disincentive provisions in the construction contract, radio ad campaign and the
outreach efforts ahead of every construction stage, all resulted in significantly alleviating the
impacts of construction on the traffic flow.

NYBRIDGE_Book.indb 159 7/17/2013 8:42:55 AM

 160 Bridge rehabilitation and retrofit

REFERENCES

American Association of State Highway and Transportation Officials (AASHTO) 1995, Guide Design
Specifications for Bridge Temporary Works, 1995.
American Association of State Highway and Transportation Officials (AASHTO) 2004, Bridge Design
Specifications, LRFD design Method for Superstructure and ASD Method for Substructure. 2004.
New York State Department of Transportation (NYSDOT) 1995–2004, Bridge Rehabilitation Project
Reports (9 Reports), Rehabilitation of 16 Bridges on or Spanning The Cross Bronx Expressway (I-95)
Corridor Between Amsterdam Avenue in New York County and Undercliff Avenue in Bronx County,
NYSDOT, 1995–2004.
New York State Department of Transportation (NYSDOT) 1998, Interim Repair of 31 Bridges on or
Spanning The Cross Bronx Expressway (I-95) and 2 Bridges on Bruckner Expressway (I-278), NYS-
DOT 1998.
New York State Department of Transportation (NYSDOT) 2004, Design Report-Alexander Hamilton
Bridge and High Bridge Interchange Ramps Rehabilitation, I-95 Corridor Between Amsterdam Avenue
in New York County and Undercliff Avenue in Bronx County, NYSDOT, 2004.
Occupational Safety and Health Administration (OSHA), Health and Safety Standards.

NYBRIDGE_Book.indb 160 7/17/2013 8:42:56 AM

 Chapter 15

Jointless ‘floating’ deck for the Walt Whitman Bridge

R. Prior & D. Griffith

AECOM, Philadelphia, PA, USA

Q. Ye
Weidlinger Associates, New York, NY, USA

V. Pandya
Delaware River Port Authority, Camden, NJ, USA

ABSTRACT: Due to water leakage at deck relief joints, the stringers and the top chord
of the floor trusses were heavily corroded. The poor condition of the framing, coupled with
the advanced age and the condition of the existing deck, resulted in the Delaware River Port
Authority electing to replace the deck system in 2006. Several deck replacement options,
including a steel orthotropic deck and a concrete-filled steel grid deck, were studied. The grid
deck option was selected for its lower initial construction cost and the feasibility of making
it jointless. To make the grid deck jointless in each span, it had to be either rigidly connected
to the stiffening trusses through deck shear keys along the length of the span, or allowed
to move freely, i.e. “float”, relative to the stiffening trusses. The “floating” deck option was
determined to be more economical and have better performance, thus advanced through final
design and construction.

1 INTRODUCTION

The purpose of this paper is to give insight into the decision making and evaluation process used
to determine the chosen deck alternative for the suspended spans on the Walt Whitman Bridge
(Figure 1), to elaborate on the key elements of the design of the jointless “floating” system and
to highlight some of the items to consider in construction of this type of deck system.
The Walt Whitman Bridge was opened to traffic in early 1957, and has become the busiest
of the DRPA’s four bridges, carrying an average of 140,000 vehicles per day (vpd). The sus-
pended structure features a 2,000 foot (ft) main span, two 770 ft back spans and a minimum
navigational vertical clearance of 150 ft. The bridge carries seven lanes of traffic in 79 ft
width (curb-to-curb). The seven lanes include a reversible center lane, which can be switched
to accommodate peak traffic demands by relocating the moveable median barrier.
The suspended structure is composed of two stiffening trusses connected to the main sup-
port cables and suspenders, and transverse floor beam trusses spaced at approximately 20 ft
on center. The existing deck is supported on stringers (W18 × 50). An important characteris-
tic of the existing deck in the suspended spans is the presence of relief joints located approxi-
mately every 121 ft (Figure 2). These joints have been the source of corrosion affecting the
stringer ends and the top chords of the floor beam trusses. This situation, combined with the
design of the stringer to floor beam connection, and the overall lateral-torsional behavior
of the suspended structure, has introduced areas of stress concentrations at the relief joints.
The loss of section due to corrosion coupled with the stress concentrations has combined to
create fatigue cracks in the stringer webs. The elimination of deck joints was an important
factor in the deck alternatives evaluation.

161

NYBRIDGE_Book.indb 161 7/17/2013 8:42:56 AM

 162 Bridge rehabilitation and retrofit

Figure 1. Walt Whitman Bridge.

Figure 2. Deck, stringers and top laterals.

Between 2004 and 2005, the DRPA had several studies commissioned to evaluate the
condition of the suspended portions of the bridge. These studies included a detailed cable
evaluation and a deck condition study. The cable evaluation concluded that the overall cable
strength had been reduced by nearly fifteen percent due to wire corrosion. The reduction in
cable strength together with the additional weight from the moveable barrier resulted in a
reduction in the cable factor of safety from 2.74 to 2.33. Even though a safety factor of 2.33
was acceptable, the benefit of reducing the overall dead load on the cables was considered an
important factor in the evaluation of deck alternatives.
The deck condition study revealed significant corrosion of the grid deck steel pan forms,
main bars and transverse reinforcing bars, and cracking and deterioration of the concrete
fill. The pan forms were severely deteriorated over eighty-percent of the total deck area, with
nearly ten-percent missing, primarily in the areas of the deck relief joints. Concrete cores
revealed expansive corrosion on the reinforcing bars as the cause of the concrete cracking. The
concrete cores also revealed a high chloride content, nearly twice the threshold limit for the
onset of corrosion. The extent of the deterioration of the stringers at the relief joints was also
of particular concern. The results of the study also indicated that significant deterioration of

NYBRIDGE_Book.indb 162 7/17/2013 8:42:56 AM

 Jointless ‘floating’ deck for the Walt Whitman Bridge 163

the deck had occurred since the previous study was conducted in 1987. Based on the results
of these studies, the DRPA determined that a total replacement of the existing deck and
stringers was in order.

2 EVALUATION OF DECK ALTERNATIVES

Two primary deck replacement alternatives, a steel orthotropic deck and a concrete-filled grid,
were evaluated. Both alternatives were studied in detail with plans developed to a twenty-five
percent level for each alternative. The deck alternatives were evaluated based on initial and
life-cycle costs, construction schedule, overall constructability, construction staging and traf-
fic impacts, fabrication and overall weight reduction.

2.1 Orthotropic
The use of an orthotropic deck as a replacement alternative offers advantages in weight
reduction, design life, maintainability and an increase in the global structural performance of
the bridge. The deck plate proposed for the Walt Whitman Bridge was 3/4" thick, with 5/16"
rib plates, and a rib spacing of 24 inches. Preliminary design of the orthotropic deck alterna-
tive produced a design with a weight of 58 pounds per square-foot (psf), approximately 8 psf
less then the original grid deck.
In addition, the use of an orthotropic deck allowed for the removal of the stringers and
laterals, further reducing the dead load. This benefit, assuming all retrofitting was completed,
reduced the overall deck weight by 15 psf and reduced the cable stress by 5.5 percent relative
to the original allowable design stress of 80 kips per square-inch (ksi) and would increase the
cable safety factor from 2.33 to 2.47.
Orthotropic decks also have their disadvantages. Initial fabrication costs for orthotropic
decks are typically higher than other deck types due to the increased steel fabrication
demands, and the need to apply a high degree of quality control to welded connections and
details. During the design phase, it was estimated that the Owner’s requirement to include
“Buy America” contract provisions would increase the cost of fabrication by approximately
$11.5 million were such a provision included. In addition to cost considerations, the Owner’s
experience with orthotropic deck maintenance on their other suspension bridge facility
(the Benjamin Franklin Bridge) was not positive, particularly with respect to the wearing
surface.

2.2 Filled grid

There were two grid deck concepts initially investigated for replacing the existing filled grid
deck. One was to replace it with a system that had deck joints similar to the existing, but
with improved stringer to floor beam connections. The other was to construct a jointless
grid deck. The relatively low costs of filled grid fabrication and construction came with both
options. The latter concept required a more in-depth analysis to confirm that large relative
displacements between the deck and support framing could be accommodated. Both options
had the main grid bars running transversely to the supporting stringers, which in turn sit on
elastomeric bearings that increase the flexibility and hence the performance of stringer-floor
truss connections. The benefit of a design with deck joints is that it allows for the differential
temperature and live load displacement of the deck floor system relative to the floor trusses
and the longitudinal stiffening truss. The deck joints allow for this displacement without
introducing large forces into the deck system.
Preliminary design for the grid deck using normal weight concrete resulted in a design heav-
ier than the existing deck. The new deck including the grid and concrete (flush-filled) weighed
about 68.07 psf, as opposed to 66.02 psf for the existing deck. Grid deck manufacturers

CH15_Prior et al-Final.indd 163 7/17/2013 2:31:20 PM

 164 Bridge rehabilitation and retrofit

consulted during the preliminary design process recommended using a half-filled grid with
an overfill design (Figure 3), to provide more concrete cover on the grid bars.
This type of design provides better protection against corrosion, better structural perform-
ance and provides a better substrate for the wearing surface.
The initial concept for a jointless grid deck was to install steel shear connectors between
the deck and the stiffening trusses, so that the relative movements between them would be
restrained. After the initial study was completed, it was concluded that in order to transfer
the shear forces, special panels of steel orthotropic deck or cast steel grid deck would be
required between the stiffening truss and the regular grid deck. The additional cost for the
special deck panels was estimated to be around $9 million.

2.3 Initial recommendations and re-evaluation

The preliminary deck design information was summarized in a draft preliminary design
report that was submitted to the DRPA in 2007. The initial AECOM recommendation was
to replace the existing deck with an orthotropic steel deck. The decision was based heav-
ily on the advantages of weight reduction and enhanced global structural performance of
the main spans. However, DRPA expressed trepidation with an orthotropic deck based on
three primary factors: (1)—their bad experience with maintaining a wearing surface on the
Ben Franklin Bridge (BFB) orthotropic deck; (2) the initial cost of orthotropic, particularly
given the agency’s strong preference for domestically fabricated steel; (3) concerns regarding
the ability to control weld quality. Based on these concerns, DRPA asked the design team
to revisit the initial recommendation in order to develop a grid option that would provide
some of the same advantages as an orthotropic deck (weight reduction, low maintenance,
improved structural performance).
Additional design and investigation was performed to determine if an economical grid
deck option was available that provided some of the same advantages as orthotropic. In an
effort to reduce the weight, a design using lightweight (120 pcf) concrete (LWC) was inves-
tigated. Though lightweight concrete has been used in bridge decks since the 1930’s, little
information was available for a lightweight concrete grid system on a suspended span bridge.
Since examples of long-term performance of lightweight concrete grid decks on suspended
spans was not found, a reduced service life was estimated for the lightweight concrete grid
deck option. In the life-cycle cost analysis performed to evaluate the proposed deck alterna-
tives, a sixty-five year service life was assumed for the normal weight concrete grid deck and
fifty year service life was assumed for the lightweight concrete grid deck.

Figure 3. Typical half-filled grid.

CH15_Prior et al-Final.indd 164 7/17/2013 2:31:20 PM

 Jointless ‘floating’ deck for the Walt Whitman Bridge 165

Since elastomeric bearings had already been proposed for the grid deck with deck joints, it
was logical to take advantage of their flexibility to develop a jointless concept that would elim-
inate the need for expensive shear panels by permitting the deck to move or “float” independ-
ently from support framing. The bearings for the “floating” deck option would be thicker, i.e.
more flexible, than those for the deck with deck joints, and would allow relative movement
between the deck and the floor beam trusses by deflecting longitudinally. The flexibility of the
bearings would also eliminate the need for hard steel bearings, a feature associated with the
original bridge, which contributed to the previously mentioned widespread fatigue cracking.
Based on the details developed (Figure 4) above, the total weight of the lightweight con-
crete “floating” deck system was recalculated and compared to the existing deck, as well
to a new lightweight concrete grid deck with joints. The “floating” deck weight, including
the stringers, bearings, laterals and wearing surface was calculated as 96.55 psf, which was
slightly higher than the proposed grid deck with joints (92.70 psf), but still less than 105.19
psf of the existing deck.

2.4 Decision process and final recommendations

The new lightweight grid options provided the DRPA with four deck replacement options
that would reduce suspended span dead load. The four deck type options presented were the
orthotropic deck with foreign fabrication, orthotropic deck with domestic fabrication, light-
weight filled grid with joints and lightweight filled grid without joints. Recognizing that each
option had unique advantages and disadvantages, AECOM hosted a series of workshops to
solicit input as to the optimal deck replacement alternative. Workshop participants were pro-
vided a decision matrix which would allow for each of the four deck types to be compared,
contrasted and ranked based on agreed-upon categories. Based on DRPA input, evaluation
categories included: dead load of the deck system, initial cost, life-cycle cost, fabrication and
constructability. The goal of the decision matrix was to allow each stakeholder to independ-
ently apply an importance or weighting factor based on what they felt was most important.
For example, one person may feel weight reduction is of primary importance and therefore
give the category of dead load an importance factor of 3, while another stakeholder may
consider initial construction cost as most important and give that category an importance
factor of 3 (See Figure 5). The product of the importance factor and the rating factor pro-
duced a total score for each alternative and provided the DRPA with a quantitative method
to compare the various deck alternatives with the highest total indicating the preferred deck
alternative.

Figure 4. New grid deck, stringers, and elastomeric bearings.

CH15_Prior et al-Final.indd 165 7/17/2013 2:31:21 PM

 166 Bridge rehabilitation and retrofit

Figure 5. Sample deck decision matrix – Stakeholder “A”.

Based on aggregate scoring, the lightweight concrete-filled grid deck “jointless” option
was selected because it offered a moderate yet meaningful amount of dead load reduc-
tion at a significantly reduced construction cost, with no increase in life-cycle cost. The
grid deck could also be domestically fabricated and supplied, and is a system familiar
to a broad array of contractors. In summary, the DRPA felt that a modern lightweight
concrete-filled grid deck, which incorporates a new stringer support system designed
to correct the long-standing lateral-torsional stringer cracking inherent in the original
design, together with a long-term cable protection program, was the most prudent course
for the DRPA to take.

3 FINAL DESIGN AND DETAILING

3.1 Finite element model

A global model was developed using the ANSYS program for the bridge with the “floating”
deck (Ye et al., 2010). This model included all the main structural elements, such as cables,
suspenders, towers, decks, floor trusses, main stiffening trusses, etc. The towers were fixed at
the base of the tower legs. The anchorages were not explicitly modeled, since it was believed
they had no appreciable impacts on the results of this study.
The concrete decks were modeled by shell elements. Each stringer and bearings, as well
as the transverse stringer bracing, were explicitly modeled. The bearing models consist of
matrix elements with 6 degree-of-freedoms (DOFs). Both the upper-and lower-bound elas-
tomer properties were used for the bearing elements, so the worst case scenarios for the dif-
ferent members were captured.

3.2 Loads and load cases

The types of loads used for the study were dead loads; live loads; thermal loads—uniform
and vertical gradients; and wind loads. For most of the study, AASHTO LRFD method and
load combinations were used.
The ASCE live loads, consisting of both uniform and concentrated loads, were used for
global effects, such as deflections or deformations of the entire spans. Their recommended
transverse load distribution is shown in Figure 6. In order to capture the governing global
effects, four load cases were created, for which uniform loads were applied respectively on:
the entire length of the bridge; one side span only; main span only; half of main span. The
concentrated loads were moved longitudinally to different locations on the bridge, and their
effects were combined with those from the uniform loads.

CH15_Prior et al-Final.indd 166 7/17/2013 2:31:22 PM

 Jointless ‘floating’ deck for the Walt Whitman Bridge 167

Figure 6. Transverse distribution of ASCE loading.

For the load rating of floor trusses, the PHL-93 loading, which is a PennDOT modified
version of AASHTO HL-93, was used. Similar to AASHTO HL-93 loading, PHL-93 load-
ing is a lane loading combined with a HS20 truck or a tandem axle, except that the tandem
axle weight total is 62.5 kips, as opposed to 50 kips in the HL-93 loading. A dynamic factor
of 1.33 is applied to the HS-20 trucks and tandem axles. In addition, a multiple presence fac-
tor of 0.65 is applied to both the truck and lane loading.
Thermal Loads: Temperature rises were additive to the live load effects, so a temperature
rise of 70°F was used. In addition, a thermal gradient of the bridge caused by direct sun light
was determined by AECOM during the field inspection and was included in the analysis.
Wind Loads: The original design wind loads were used for the analysis and they are:
• a total of 625 plf on stiffening trusses;
• a total of 115 plf on cables and suspenders.

3.3 Stringers and bearings

In the proposed details (Figure 7), there is a 5 ¾" tall bearing underneath each W14 stringer,
which is about 4" shorter than the existing W18 stringer. The grid deck is 5 3/16" deep with
1 ½" overfill, on top of which a 1 ½" Rosphalt 50 LT wearing surface will be applied. The
proposed bearing details are shown in Figure 8. To transfer the shear force, the elastomer is
vulcanized to both the sole plate and the base plate.
The elastomeric bearings are the key elements in this “floating” deck system. The benefit of
the bearings are twofold: they will deflect longitudinally to allow relative movements between
the deck and the floor trusses, so that the latter will not be overstressed; at the same time, the
shear stiffness of the elastomer will reduce these relative movements so that the bearing shear
deformations are within the allowable value. The maximum bearing shear deformation due
to the service limit state is 1.2 inches, which is less than the 1.5" allowable for a 3" thick elas-
tomer. The maximum live load rotation on the bearings is about 0.002 radians. This rotation
was combined with an additional rotation of 0.005 radians specified by AASHTO for erec-
tion tolerance to determine the design rotation.
Due to the lightness of the deck, the dead load reaction on each bearing is around 9 kips
and the maximum live load reaction is around 36 kips under the PHL-93 tandem axles. At
the service limit state, one edge of the bearing might be subject to uplift. The 2009 AASHTO
LRFD interim provides new provisions for checking the elastomer under tension and it
was found that the maximum tensile stresses in these bearings are well within the allowable
limits.
After the initial design of the jointless grid deck with bonded elastomeric bearings was
completed, the design was further refined to use a combination of bonded and sliding bear-
ings: bonded bearings in the middle portions of the main and side spans; and sliding bearings
near the towers and anchorages (Figure 9).

CH15_Prior et al-Final.indd 167 7/17/2013 2:31:23 PM

 168 Bridge rehabilitation and retrofit

Figure 7. Comparison of deck, stringer and bearing details.

Figure 8. Proposed bearing details.

Figure 9. Bearing layout.

3.4 Grid deck longitudinal splice details

One common design challenge for grid decks is in the detailing of longitudinal deck splices,
which are required for staged construction. The longitudinal deck splice detail that was used
for the Walt Whitman Bridge is shown in Figure 10. The size of splicing rebars is #5, and
rebar couplers are used to achieve the continuity of resisting the negative bending moments
at the supporting stringers.
Precast deck panels were specified in the design stage, and each longitudinal closure pour
was about 2'–4" in length. Prior to the construction, the Contractor proposed to use the cast-
in-place method and was approved, as discussed hereafter.

CH15_Prior et al-Final.indd 168 7/17/2013 2:31:24 PM

 Jointless ‘floating’ deck for the Walt Whitman Bridge 169

Figure 10. Grid deck longitudinal splice details.

3.5 Load rating of floor trusses

The load ratings of floor truss members for the jointless deck option were performed
using the AASHTO LRFR method. In these ratings, bi-axial moments, as well as axial
forces were calculated. Ratings revealed that a few diagonal members in the floor trusses
under the deck relief joints, which have the smallest member sizes among all types of floor
trusses, needed strengthening in order to satisfy the rating of the PHL-93 tandem. These
members were strengthened as an element of the advanced work performed prior to the
redecking.

4 CONSTRUCTION STAGING AND MAINTENANCE AND PROTECTION

OF TRAFFIC

The Walt Whitman’s diverse and demanding operational climate makes full use of the bridge’s
79 ft. cartway. During weekday periods, the bridge functions as a typical commuter facility,
with distinct peak periods in the westbound direction in the morning and eastbound direc-
tion in the afternoon/evening. Peak demand periods are extended and compounded during
the summer months (Memorial Day to Labor Day), with the heaviest travel occurring on
Friday evenings in the eastbound direction and on Sunday evenings and Monday mornings
in the westbound direction. Demand also increases during heavily attended events at the
sports complex in South Philadelphia and the entertainment center in Camden, with peaks
occurring prior to and immediately after events.
A capacity analysis of the proposed construction staging was performed using Highway
Capacity Manual reduction factors. The analysis assumed a minimum lane width of 10 ft for
interior lanes, 11 ft for curb lanes, and 12 ft for cattle chute lanes. This analysis determined
a per lane capacity of 1740 vph. This capacity was compared to peak volumes calculated
for the anticipated construction period by applying a two percent growth factor. This com-
parison indicated that four lanes of traffic would be required to accommodate peak period
volumes, assuming no diversion of traffic existed.

CH15_Prior et al-Final.indd 169 7/17/2013 2:31:27 PM

 170 Bridge rehabilitation and retrofit

The DRPA’s experience during the redecking of approach spans, which occurred between
1995 and 1999, was that three lanes of traffic were sufficient to carry peak traffic in the east-
bound direction without causing delays. Using yearly traffic data obtained from the DRPA’s
annual reports, a broader analysis on the regional traffic patterns during this period was
performed to evaluate any changes to WWB traffic or regional traffic patterns that may have
helped reduce eastbound peak hour demand at the WWB. The analysis revealed that demand
on the WWB was reduced by approximately fifteen percent in the eastbound direction during
the peak hour, and that corresponded to an increase of the same amount in the eastbound
peak hour traffic at the Benjamin Franklin Bridge (BFB). Westbound diversion to the BFB
would not have occurred, since this would have been ineffective in reducing travel time due to
the bottleneck conditions that exist at the western terminus of the BFB.
Based on this analysis, it was reasonable to expect that a minimum of fifteen percent of
peak hour eastbound volume would be diverted to the BFB during construction. With this
diversion of traffic, an hour-by-hour comparison of volume/capacity revealed that three lanes
of traffic were sufficient to handle peak hour traffic volumes without causing delays in the
eastbound direction. Westbound traffic was evaluated in a similar manner to the eastbound
traffic and free flow conditions existed with four lanes but delays would be experienced if
only 3 lanes were provided.
In order to accommodate the anticipated traffic volumes during the construction without
causing delays, it was recommended that four lanes be provided in the westbound direction
and two lanes in the eastbound direction during the morning peak period, and that the move-
able barrier be shifted to provide three lanes in each direction during the evening peak period
and at all other times. This maintenance-of-traffic scheme would require that the construc-
tion be performed in seven long-term stages. The seven stages were identical to that which
was successfully used during the approach reconstruction in the late nineties.
It should be noted that during the fourth stage (Stage 4), the work zone was in the center
of the bridge, making it impossible to shift the moveable barrier to accommodate peak hour
traffic. The typical section is shown in Figure 11. During this stage, three lanes of traffic
were provided in each direction during both morning and evening peak periods, which had
the potential to result in westbound morning delays. It was recommended that the DRPA
engage in an aggressive media and public outreach effort in the weeks leading up to Stage 4,
in order to encourage alternative travel patterns during this period. In addition, the DRPA

Figure 11. Stage 4. Three lanes in each direction.

CH15_Prior et al-Final.indd 170 7/17/2013 2:31:28 PM

 Jointless ‘floating’ deck for the Walt Whitman Bridge 171

included a monetary incentive for Stage 4, if the contractor could complete the stage work
in less than 120 calendar days. The contractor completed Stage 4 in 90 days, earning the full
monetary incentive.
In order to achieve the minimum lane widths used in the traffic evaluation, it was recom-
mended that the existing cartway be widened from the existing 79'–0" to 80'–6" (9" widening
on each side). The widening would provide a benefit to the construction staging, but it would
also provide a narrow 9" shoulder, which could be used to store/collect storm water runoff.
The nine inch widening resulted in a reduced width of the maintenance walkway behind the
roadway barrier (2'–9" to 2'–0"), however it was agreed that this reduction would not impact
future maintenance access.
It was also recommended to replace the existing 24" wide concrete moveable barrier with a
13" wide steel barrier. The narrower barrier would not only maximize the width available to
traffic during construction, but it would also provide better long-term performance, requir-
ing less routine maintenance. The steel barrier also performed better during impact testing,
deflecting 28 inches, as opposed to 53 inches for the concrete barrier.

4.1 Construction barrier

The seven (7) stage construction phasing requires that temporary concrete barrier be installed
adjacent to the construction work zone. The design team desired to limit and if possible
eliminate the need for deck penetrations into newly placed deck sections to install temporary
barrier. In the design phase, the AECOM investigated barrier from other states and agencies,
however all examples included some form of direct anchoring into newly placed deck. In an
effort to minimize the number of drilled holes, the barrier detail recommended by AECOM
utilized anchor spacing of 4 ft, the maximum permitted by PADOT. Calculations confirmed
that this spacing would be sufficient to resist TL-4 crash loading. TL-4 loading was utilized,
since this was considered to be the appropriate test level for the permanent barrier on the
bridge.
During Stage 1 of the construction, the Contractor submitted a request to use a NYSDOT
approved and NCHRP crash tested (TL-3) stiffened temporary barrier for Stages 2 thru 7.
Since the contactor’s proposal utilized an unanchored “stiffened” barrier system that con-
formed to NCHRP (TL-3), the primary consideration in evaluating this system was whether

Figure 12. Stage 5. Looking West with the moveable barrier in the afternoon position to accommodate
3 lanes in each direction. The barrier is shifted to the left to provide 4 lanes of inbound traffic for the
AM morning traffic period.

CH15_Prior et al-Final.indd 171 7/17/2013 2:31:30 PM

 172 Bridge rehabilitation and retrofit

Figure 13. Stiffened construction barrier on WWB.

this criterion was appropriate for this project. As a starting point in our evaluation, AECOM
considered the standards and practices of other agencies. In addition to looking closely at
PADOT and NJDOT standards (DRPA’s host state agencies), AECOM also investigated
the standard temporary barriers approved for use in a number of other States including
GADOT, WVDOT, TXDOT, VADOT, INDOT, SCDOT, OHDOT, RIDOT, NYSDOT and
CTDOT. All DOT’s that we investigated require a pinned or bolted barrier for work on
bridges except for NYSDOT which, as noted, had recently approved the stiffened (unan-
chored) barrier system. In general, many states indicate that temporary concrete barrier shall
conform to NCHRP 350 Test Level 3 (TL-3). However, many DOT’s are also silent on the
Test Level required. The NYSDOT barrier was tested in 2008 (Stolle 2008). With the excep-
tion of PADOT, we did not find any DOT with a barrier taller than 32" and we did not find
any DOT that specifically stated that a test level 4 (TL-4) is required for temporary construc-
tion barrier.
In addition, our evaluation of the stiffened NYSDOT barrier considered that the tangent
geometry thru the work zone reduces the potential for side angle impacts to the construction
barrier, and that the Contractor’s means and methods limits exposure to open hole condi-
tions across a 250 ft length for no longer than 72 hours. Based upon a consideration of all
above factors, the project team concluded that the proposed barrier system could be consid-
ered as an alternative suitable for use on the Walt Whitman Bridge project.

5 CONSTRUCTION CONSIDERATIONS

As with many elements on the project, the need to maintain six (6) lanes of traffic during
peak periods and the proposed seven (7) stages of construction increased the difficulty of
the construction operations over what would be expected with a redecking project with fewer
stages. The Contractor utilized a number of methods to improve construction efficiency,

CH15_Prior et al-Final.indd 172 7/17/2013 2:31:30 PM

 Jointless ‘floating’ deck for the Walt Whitman Bridge 173

including the design and installation of a construction gantry crane in each span and the use
of an under deck work platform beneath the full length and width of the suspended spans.
Both the gantry crane and the under deck platform were constructed as part of the advanced
work on the project. The use of a construction gantry crane proved essential in efficiently
removing demolished deck components and in installing new stringer and deck elements.
While AECOM made provisions in the design phase to allow the use of the gantry crane, the
gantry cranes were a contractor designed option. The under deck platform was a SafeSpan
corrugated metal deck platform system hung from the floor truss members. The platform is
dual rated as a working platform with a live load capacity of 25 psf or a shield with a maxi-
mum ultimate capacity of 100 psf.
There were a number of elements of the proposed “floating” grid deck system that presented
constructability challenges and required thorough investigation to insure a final product that
met the intent of the design. Among these elements were the installation and adjustment of
the stringers and elastomeric bearings, installation of the grid panels and casting of the con-
crete deck in the short work windows that were required by the bid documents.

5.1 Installation and adjustment of stringers and bearings

Each exterior stage (stages 1 and 7) required the installation of four lines of roadway string-
ers and approximately 680 elastomeric bearings per stage and each interior stage (stages 2
thru 6) required the installation of two lines of stringers and approximately 340 elastomeric
bearings per stage. The seven stages of work progressed from the South side of the bridge to
the North side as previously described and was dictated by the approved maintenance and
protection of traffic staging.
Restrictions were placed in the Contract documents as to the maximum length of deck
that could be removed at any one time, where the replacement work should begin and how
it should progress along the 3,500 ft length of the bridge. The limitations on the maximum
length of deck to be removed (500 ft in center span) were primarily put in place to limit
overall bridge rotation as the side of the bridge with the deck removed would elevate with the
removal of the existing deadload. While the amount of anticipated bridge rotation was not
considered detrimental to the structure, there was a potential concern that it would be percep-
tible to the users of the bridge. This proved to not be a significant issue since the Contractor’s
operation of continually demolishing portions of existing deck and stringers was balanced
with the operation of continually installing the new framing and deck structure.
The design documents specified that the demolition and deck replacement work start at the
center of the main span and progress toward the towers. However, the Contractor requested
a change so that he could work in the direction of traffic (i.e., work from PA anchorage to NJ

Figure 14. Gantry crane. Figure 15. Under deck platform.

CH15_Prior et al-Final.indd 173 7/17/2013 2:31:31 PM

 174 Bridge rehabilitation and retrofit

anchorage in stages 1 thru 4 and work from the NJ anchorage to the PA anchorage in stages
5 thru 7). This was reviewed and found to be acceptable as long as the approved stringer and
bearing installation procedures, as specified in the contract plans, were followed.
The stringer erection procedure resulted in the need for longitudinal jacking of some
stringers to complete the bolting of the stringer splice plates which in turn induced shear
deformation of the bearings. The shear deformation during installation was not found to
be excessive. It is important to note that keeper bars were installed on all sliding bearings so
that the sliding bearings would behave as bonded bearings during the course of construction.
A concern was the potential for inconsistent deformation of the bearings along a common
floor truss bearing line since the environmental conditions varied significantly between many
of the stages. For example, stage 2 bearings and stringers were installed in the winter of 2012
while stage 4 bearings and stringers were installed in the summer of 2012. Despite the dras-
tic change in ambient air temperature between stages, only minor stringer jacking has been
required to complete stringer field splices and to maintain proper bearing alignment.

5.2 Grid deck concrete placement

The staged construction and short windows available for deck casting were key considera-
tions in the design and construction of the grid deck. The proposed grid deck was a precast
grid deck with cast-in-place closure pours. Precast concrete deck panels were chosen over
cast-in-place grid panels in the design phase since it was expected that the limited windows
available for casting were not conducive to casting the entire deck as cast-in-place. Also,
since work would proceed through cold weather winter conditions it was felt by the design
team that reducing the amount of cast-in-place work to only closure pours would result in
a superior deck over what could be achieved by proceeding with a 100% cast-in-place deck.
The Contractor submitted a request to provide a 100% cast-in-place concrete filled grid deck,
in lieu of the precast filled grid deck that was presented in the Contract bid documents. The
Contractor identified the following advantages of a cast-in-place deck:
1. Less rigid deck panels allowing for better conformance with stringer profile.
2. Reduced need for shimming which would improve the schedule.
3. Reduced cold joints which should produce better long-term construction quality.
The design team recognized the advantages of the cast-in-place deck and approved pro-
ceeding with this method but due to concerns of the impact of traffic vibration on the quality
of the concrete, additional vibration monitoring and early age concrete testing was specified.
Specifically the design team recommended that the Contractor monitor vibrations of the
concrete deck prior to casting, during casting and for a period of 48 hours after the deck
is cast and compare measured peak particle velocities (PPV) against acceptable threshold
levels.
While a number of reports have found little evidence to indicate that the performance of a
concrete deck is degraded by traffic induced vibrations during placement or curing (NCHRP
Synthesis 86), the studies were performed on conventionally reinforced concrete decks with
stiffer superstructures than proposed for the Walt Whitman Bridge. Thus, there was a concern
that the large pours proposed for the WWB and the relative flexibility of the deck may make
the concrete more susceptible to vibration damage during and immediately following casting.
A number of agencies and DOT’s have published recommended peak particle velocities for
concrete at various ages and the following threshold velocities were initially proposed:

Concrete Age Allowable Maximum PPV

(Hours) (Inch/second)

Less than 12 0.2

12 to 24 1.0
24 to 48 2.0

CH15_Prior et al-Final.indd 174 7/17/2013 2:39:43 PM

 Jointless ‘floating’ deck for the Walt Whitman Bridge 175

Figure 16. Concrete placement—Stage 1 shown.

The vibration testing was performed using a Blastmate III monitor from Instantel with
5 monitors on the bridge spaced in the work zone in the area of the proposed concrete pour.
Velocity sampling started prior to Stage 2 concrete casting and was performed with traffic
in the lanes on both sides of the Stage 2 work zone (lanes 5 and 7). It became immediately
apparent that measured velocities were significantly higher than the threshold values with
multiple occurrences of velocities over 2.0 inch/second recorded each hour with values as
high as 3.16 inch/second. A reduction in the number of occurrences of values over 2.0 inch/
second was experienced with the closure of an adjacent lane (either lane 5 or 7) but there were
still random peak particle velocity spikes above 2.0 inch/second.
Based on the information gathered in Stage 2 and verifying that no deck cracking was
experienced, the following recommendations were made for concrete placement:
• Close the adjacent lane to traffic that is positively connected to the new deck.
• Keep traffic off of the adjacent lane for a minimum of 15 hours or until the concrete
achieves a strength of greater than 1,000 psi.
• Temporarily clamp the grid to the stringers to reduce and dampen the unfilled grid.
• Reduce speeds on the bridge during the 36 hours following the pour.
• Smooth the pavement at the existing relief joints.
Through Stage 6, the deck has not experienced any cracking.

6 CONCLUSIONS

Selecting the preferred deck alternative for the deck replacement of the Walt Whitman Bridge
was a collaborative effort between the design team and the owner. The studies conducted by
AECOM and Weidlinger provided the engineering and knowledge for the owner to perform
their own internal review and objectively evaluate each of the proposed alternatives. The use
of the decision matrix provided the owner with a tool to quantitatively compare the alterna-
tives based on five key categories.
This project has demonstrated that a lightweight concrete filled grid deck without joints
(floating deck) is a feasible deck replacement option for owners looking to replace aging
decks on suspended type structures. The jointless grid deck has an advantage over ortho-
tropic decks in initial costs and is comparable to orthotropic in life-cycle costs.

CH15_Prior et al-Final.indd 175 7/17/2013 2:31:32 PM

 176 Bridge rehabilitation and retrofit

A floating deck system requires that stringer and bearing installation be tightly monitored
and controlled to insure proper bearing deformation within construction stages and
throughout variations in construction seasons. As demonstrated by the WWB, this can be
achieved even with construction being performed throughout significantly differing seasonal
conditions.
Both precast and cast-in-place grid decks have advantages. A 100% cast-in-place deck will
reduce the number of cold joints while simplifying the connections to adjacent previously
constructed deck panels.
The newly constructed deck has shown no signs of cracking despite experiencing peak
particle velocities above threshold levels. Moderate restrictions to minimize vibrations—
including reducing travel speeds and closing traffic lanes adjacent to the work zone during
initial curing-have so far been effective in controlling deck cracking.
A temporary concrete barrier that is stiffened in lieu of a barrier that is anchored through the
deck is a potential construction barrier option where a TL-3 crash test level is appropriate.
The use of a gantry crane system and under-bridge platform will overcome many of the
constructability and access limitations associated with long term staged construction.

REFERENCES

Stolle, C., 2008. Evaluation of Box Beam Stiffening of Unanchored Temporary Concrete Barriers,
MwRSF Research Report No. TRP-03-202-08.
Ye, Q., Griffith, D., Pandya, V. and Fanjiang, G., 2010. Floating Jointless Steel Grid Deck for Walt
Whitman Bridge, International Bridge Conference. Pittsburgh: Pennsylvania.

CH15_Prior et al-Final.indd 176 7/17/2013 2:31:33 PM

 Chapter 16

Brooklyn Bridge orthotropic deck installation

R. Leo & P. Reynolds

Greenman-Pedersen Inc., New York, NY, USA

ABSTRACT: The Franklin Square Arch Structure roadway of the iconic Brooklyn Bridge
was recently retrofitted with orthotropic deck panels. The advantages of orthotropic deck
panels include rapid installation, reduction of dead weight loading, longer service life and
minimal maintenance. The main disadvantage is cost of fabrication. During installation, the
orthotropic deck panels were aligned before being welded in the flat and overhead positions.
Multiple alignment tolerances needed to be met, including: vertical alignment, root open-
ings, diaphragm offsets and panel elevations. Welding processes which were used include sub-
merged arc (SAW) and shielded metal arc (SMAW). Also, the air carbon arc cutting process
was used to achieve a U-groove on top of the panels. The work was performed during over-
night closures without disruption of daily rush hour and peak traffic flow.

1 INTRODUCTION

Franklin Square Structure is located in Manhattan on the Brooklyn Bridge Manhattan

Approach, spanning Pearl Street. The location is very historic, with the east abutment once
being the location of the first presidential mansion in which George Washington, his family,
and his household staff lived from April 23, 1789 to February 23, 1790. The location looks
quite different from how it looked prior to 1950, when there was an elevated transit track
with a station located directly under the structure.
The Franklin Square Structure is a truss bridge that utilizes pin connections with eyebar
members. It is a trapezoidal shape; with the shortest span about 169’ and the longest span at
about 212’ in length. See Figure 1.
There are theoretically no bending stresses throughout the structure due to the freely rotat-
ing pinned connections. The trusses were constructed from wrought iron, which has low
carbon content in contrast to steel. Wrought iron is tough, ductile, and has good resistance
to corrosion, which is a contributing factor as to why the structure is still in good condition.
The pins are made of carbon steel, which is considered a frontrunner to A36 steel used in

Figure 1. Franklin square structure.

177

NYBRIDGE_Book.indb 177 7/17/2013 8:43:12 AM

 178 Bridge rehabilitation and retrofit

construction today. In 1999, after it was found that the existing pins were failing, six steel
arches were added below each of the existing trusses. Arches were chosen so that the roadway
could remain open, and also to keep the appearance consistent with the original historic
structure. Increased dead load throughout the years was the cause of the pins failing.

2 BACKGROUND

An orthotropic deck system consists of a flat, thin steel plate, stiffened by a series of closely
spaced longitudinal ribs and supported by orthogonal transverse floorbeams. This is short-
ened from the original term orthogonal-anisotropic. Orthogonal means that it is pertaining
to or involving right angles or perpendiculars. Anisotropic means that it has the property of
being dependent directionally. These terms apply because the orthotropic deck has a series
of longitudinal ribs that are at right angles with the supporting diaphragms and floorbeams,
and deck also has a completely different set of properties longitudinally and transversely.
There are many advantages to implementing an orthotropic deck system, and sometimes the
system is necessary because of construction time restraints and span length requirements.
Compared to traditional reinforced concrete decks, orthotropic decks are lighter, have a shal-
lower deck depth, and can be installed much more rapidly. The lighter weight allows longer
spans to be more efficiently designed. Also, orthotropic decks are often used in moveable
bridges due to the lower weight. The disadvantages of orthotropic decks include a high cost
of fabrication due to the amount of welding required and less flexibility in the field due to
the fact that the panels must be completely pre-fabricated.

3 CONSTRUCTION

3.1 Floorbeam replacement

The first step in the Brooklyn Bridge Project at the Franklin Square Structure was to replace
the floorbeams that the orthotropic deck would eventually be supported by. To replace the
floorbeams, a new network of temporary floorbeams was installed connecting to the trusses,
and the existing deck was lifted to remove the load from the existing floorbeams.
The floorbeams were then cut into sections and removed, which was followed by the instal-
lation of new floorbeams into place by a crane from above. On the Brooklyn Bound side of
Franklin Square Structure, 22 of 41 floorbeams were replaced, and on the Manhattan Bound
side, 24 of 45 floorbeams were replaced. See Figure 2.

Figure 2. Floorbeam installation.

NYBRIDGE_Book.indb 178 7/17/2013 8:43:13 AM

 Brooklyn Bridge orthotropic deck installation 179

3.2 Concrete deck removal

After the floorbeams were replaced, the next step was to remove the existing reinforced
concrete deck, including the underlying supporting steel. For the Franklin Square Bridge, this
was done utilizing a crane during nightly traffic closures. After removal of the existing deck
and supporting steel, a temporary steel deck panel system was installed. The roadway could
then be opened to traffic for the following morning. A total of 32 temporary deck panels were
installed on the Brooklyn Bound roadway. When the orthotropic deck panels were ready to be
installed, the temporary steel deck panels were removed with a crane. See Figure 3.

3.3 Panel delivery and installation

Upon delivery to the project site, the orthotropic deck panels were inspected for damage by
American Welding Society Certified Welding Inspectors. The new orthotropic deck panels
were then transported to the roadway installation location on flatbed trailers during nightly
closures. Each orthotropic deck panel weighs between 20,000 pounds and 50,000 pounds,
ranging in size from 200 square feet to 600 square feet. Temporary ¾" lifting lugs were welded
to the top face of the orthotropic deck panels for lifting purposes. The panels were transferred
from the flatbed trailers and into place on to the newly installed floorbeams using dual 40 ton
cranes. Licensed surveyors ensured panels were set at proper elevation and offset from the
existing truss. When the panel elevations and truss offset requirements were satisfied, 13/16"
diameter holes were drilled in the orthotropic deck ribs/splice plates for temporary ¾" diam-
eter bolts, and 11/16" diameter holes were drilled through the orthotropic deck diaphragms/
connections angles for temporary 5/8" diameter bolts (New York State Steel Construction
Manual 2008). The temporary bolts maintained the orthotropic deck panel positions at the
proper elevation. With the temporary bolts in place, the temporary lifting lugs were removed
for daytime traffic. The temporary bolts were replaced with permanent 1" and 7/8" diameter
high strength galvanized bolts after the panels were completely welded. The temporary lifting
lug weld locations were Magnetic Particle inspected for soundness by a technician. A total of
24 orthotropic deck panels were installed on the Manhattan and Brooklyn bound lanes at the
Franklin Square Structure. See Figure 4.

3.4 Panel alignment

When all twelve orthotropic deck panel pairs were installed at either the Brooklyn bound
or Manhattan bound side of Franklin Square Arch, the certified welding inspector (CWI)
verified the diaphragm/floorbeam centerline offsets, root openings and vertical alignment
parameters for all the orthotropic deck panels. This allowed the contractor to make necessary

Figure 3. Concrete deck removal.

NYBRIDGE_Book.indb 179 7/17/2013 8:43:13 AM

 180 Bridge rehabilitation and retrofit

Figure 4. Panel installation.

minor adjustments required by Contract specifications prior to shielded metal arc welding and
submerged are welding of adjacent panels. Once the final fit-up conditions were established,
the orthotropic deck panels were welded in the longitudinal and transverse directions.
The Contract specifications required the adjacent orthotropic deck panel vertical align-
ment to be within 1/16". The maximum allowable diaphragm-floorbeam web offset was
7/8". Locations with diaphragm-floorbeam web offsets greater than ½" require additional
floorbeam stiffeners to be installed. The Contract specifications also required the difference
between the maximum and minimum root opening along any given joint to be 1/8" or less.
The root opening in the longitudinal and transverse direction was determined by the Con-
tractor’s approved Welding Procedure Specification (WPS). The approved WPS for the lon-
gitudinal joint required a root opening between 0" to 3/16". The allowable root opening for
the transverse joint was 3/16" to ½". These strict tolerances had to be adhered to prior to the
start of production welding (American Welding Society 2008).

3.5 Panel welding

The longitudinal joint of the adjacent ¾" thick orthotropic deck panels consisted of a square
groove at the top and a V-groove at the bottom. The transverse joints of adjacent orthotropic
deck panels consisted of a V-groove and a permanent backing bar. After all orthotropic
deck panels were aligned in their final position, the V-groove of the longitudinal joints was
SMAW welded in the overhead position from below. This overhead weld was ground flush
after SAW welding was completed from the top of the panel (New York State Steel Construc-
tion Manual 2008). Permanent backing bars were not allowed in the longitudinal joint per
Contract specifications (American Welding Society 2008). Overhead welding on the bottom
face of the longitudinal joint served as a backing bar for the submerged arc weld on the top.
The approved WPS for overhead shielded metal arc welding required 5/32" diameter E7018
electrodes to be used per the following weld parameters:
i. 130–210 Amps;
ii. 22.3–25.7 Volts;
iii. 4–11 inches per minute travel speed;
All electrodes were transported to the project site in hermetically sealed containers. The
electrodes were capable of being used directly from the hermetically sealed containers.
Once removed from the containers the electrodes are kept in a storage oven set at 250° F
until used. E7018 electrodes not used within 4 hours after removal from sealed containers
or storage ovens were re-dried for a minimum of 2 hours at temperature between 450° F
and 550° F or discarded. When the relative humidity was greater than 70%, the out of stor-
age time for E7018 electrodes was decreased from 4 hours to 2 hours. Wet electrodes were
discarded. Electrodes were allowed to be re-dried once (New York State Steel Construction
Manual 2008).

NYBRIDGE_Book.indb 180 7/17/2013 8:43:14 AM

 Brooklyn Bridge orthotropic deck installation 181

Once all longitudinal joint overhead welding was complete, air carbon arc electrodes (½"
diameter) were used to achieve a U-groove on the top of the longitudinal joints. A 150° F
preheat was required prior to air carbon arc gouging. Acceptable preheat was achieved
with open torches from above. A straight uniform U-groove was achieved with an Airarc
N7500 mounted on a Bug-O track system. Welding shields were used to protect pedestrians
on the active footwalk from arc flash. Sound blankets were also used to minimize the noise
created from the air carbon arc gouging equipment. Once all slag created from the air carbon
arc gouging process was removed, the joint was MP inspected by a licensed MT technician
for soundness. When discontinuities such as slag were discovered, they were removed and
repaired per approved WPS before proceeding to the submerged arc weld process.
Before SAW welding begins, the orthotropic deck panel diaphragm and rib splice bolts
were fully tightened on one panel and loosened on the adjacent panel to allow the free panel
to move in response to weld shrinkage forces. Contract specifications required weld distor-
tion control measurements to be taken. Shrinkage and angular deformation/cusping meas-
urements were taken before and after submerged arc welding at two feet intervals along the
joint. Contract specifications allowed a maximum shrinkage of 1/16". The shrinkage was
measured by taking a measurement between two points located 2 inches from the edge of
adjacent panels before and after submerged arc welding. Per Contract specifications, the
cusping was not allowed to exceed 1/8" for 90% of the measurements and not to exceed 3/16"
for the remaining 10% of the measurements for any given joint. Cusping measurements were
taken after all strongbacks, clamps and wedges were removed. Cusping was measured 1" on
either side of the joint centerline using a 2 foot straight edge centered over the joint.
The Contractor submitted a Welding Procedure Qualification Record (WPQR) for
approval. The WPQR included all weld parameters the Contractor planned on adhering to
during field welding. A welder operated the submerged arc welding machine on a test plate
according to the weld parameters identified on the WPQR. The test plate was sent to a
laboratory to perform tension tests, side bend tests, charpy impact tests, reduced section
tension tests, radiograph tests, macroetch tests and chemical analysis (New York State Steel
Construction Manual 2008). Once the test results were accepted, the Contractor submit-
ted a WPS with the same parameters as the WPQR for final approval. Acceptance of the
WPQR test results qualified the welder operator that completed the test plate to operate
the submerged arc welding machine. Additional welder operators were required to operate
the submerged arc welding machine using the same welding parameters utilized on previous
test plates. The additional test plates were radiographed for welder operator approval.
The SAW welding of the longitudinal joint required a preheat of 250° F and a maximum
interpass temperature of 450° F. Preheat was achieved with torches from above. The pre-
heat and interpass temperatures were checked with temperature indicating crayons. The sub-
merged arc welding of longitudinal joints was initiated on the downhill side of the joint and
proceeded uphill. See Figure 5. The approved submerged arc welding WPS for longitudinal
joints required the following operating parameters:
i. 504–600 Amps;
ii. 26–30 Volts;
iii. 18.7–25.3 inches per minute travel speed;
Lincolnweld 960 flux and Lincolnweld L61 1/8" diameter electrodes were used for the
SAW process. The flux was stored indoors at a relative humidity less than 70% and an ambi-
ent temperature between 40° F and 122° F. The flux was protected from condensation.
According to the manufacturer, the flux was able to be used directly from the bag. Humid
conditions required the flux to be heated in an oven at a temperature between 500° F and
900° F. Wet flux was not allowed for use. Flux that fused with the weld could not be reused.
Flux that did not melt during welding was able to be reused after recovery. Recovered flux
was passed through a filter screen and magnet to remove all unwanted particles before return-
ing to the flux supply system. Flux that was not recovered from the weldment surface within

NYBRIDGE_Book.indb 181 7/17/2013 8:43:14 AM

 182 Bridge rehabilitation and retrofit

Figure 5. Submerged arc welding panels.

one hour of being deposited was discarded and not used. The percentage of new to recycled
flux was a minimum of 33%.
After completing all the longitudinal joint welds, the transverse joints were welded.
The transverse joint was slightly different than the longitudinal joint. A permanent backing
bar was used in lieu of the overhead welding. The permanent backing bar was tack welded in
place in preparation for the SAW weld. The transverse joints were preheated to 250° F with
torches. The interpass temperature was kept between 250° F and 450° F. After the last sub-
merged arc weld pass was made, the joint was post heated for one hour at 250° F. The temper-
ature of the joint was then gradually lowered by 100° F each hour until ambient temperature
was achieved. The weld reinforcement for both the longitudinal and transverse joints was
ground flush (New York State Steel Construction Manual 2008). After grinding the joints
flush an UT technician checked the welds for discontinuities. Discontinuities, such as minor
porosity and slag, were repaired according to an approved repair WPS. Repair locations were
then tested by the UT technician. A waterproofing membrane was applied to the orthotropic
deck panels after completion of all welding, followed by the placement of an asphalt wearing
surface.

4 CONCLUSIONS

The key to assuring the highest quality welded splices was in the caliber of the quality assur-
ance inspectors. Only formally trained (AWS certified) inspectors with field weld experience
possessed the requisite technical knowledge to work with the Contractor’s alignment and
welding crew. Project specific reporting forms were created for the Brooklyn Bridge project.
All requirements and specifications were embedded within the inspection forms. Successful
completion of the orthotropic deck installation was due to the dedication and persistence of
many individuals from NYCDOT, Designers, Contractors and the Resident Engineering and
Inspection team.

REFERENCES

American Welding Society, 2008. AWS D1.5—Bridge Welding Code.

New York State Steel Construction Manual 3rd Edition, 2008. Department of Transportation, Office
of Structures.

NYBRIDGE_Book.indb 182 7/17/2013 8:43:14 AM

 4 Seismic analysis & design

NYBRIDGE_Book.indb 183 7/17/2013 8:43:15 AM

This page intentionally left blank
 Chapter 17

Case studies on several isolated bridge configurations adopting

a performance based approach

D. Forcellini
Università della Repubblica di San Marino, San Marino Republic

M. Banfi
Alga S.p.A., Milano, Italy

ABSTRACT: The most recent development of engineering approach is based on concept

of design consisting in prescribed performance rather than the more traditional prescriptive
approaches. The paper aims to assess the effects of isolation devices on a benchmark bridge
in reducing the column displacements and repair costs adopting various configurations of
abutments and pier connections. The data were derived to relate response to damage of indi-
vidual components within the structure and tied to explicit repair procedures that could be
used for cost estimation and repair effort necessary to return the bridge to its original func-
tionality. The contribution of the isolation technique is assessed evaluating the resistance
effects applied to Peak Ground Acceleration levels in terms of cost and time repair quantities
with particular attention to the bridge-abutment configuration.

1 INTRODUCTION

Performance-Based Earthquake Engineering (PBEE) aims to quantify the seismic per-

formance and risk of engineered facilities using metrics that are of immediate use to both
engineers and stakeholdlers. It is a new methodology based on the concept of design for
prescribed performance rather than the more traditional prescriptive approaches, developed
by the Pacific Earthquake Engineering Research (PEER) Center (http://peer.berkeley.edu).
The main contribution consists of theoretical development, applications in academia and
industry and also to the inclusion of precepts into the next generation of building/design
codes (Cornell and Krawinkler 2000). In particular, the present work is based on a recently
developed bridge PBEE user interface (Lu et al., 2010; Mackie et al., 2010b), thanks to the
effort born from the above earlier developments.
Isolation technique aims mainly at the uncoupling the structure from the damaging effects
of earthquake motions generally in the longitudinal direction. Its effectiveness on bridges
has been studied in experimental, analytical and numerical papers (Constantinou et al., 1993,
Eftekhari and Tehrani Zadeh, 1996 and Haque et al., 2010), while relatively few attempts
have been proposed taking into account performance based criteria.
In this regards, the presented work is focused on the longitudinal behaviour that might be
provided by several isolation device models, focusing on several aspects such as the suitabil-
ity of a particular arrangement and the type of isolation system. The beneficial effects are
studied comparing several configurations of abutments and column connections with isola-
tion devices at the top of the column and at the abutment supports, in reducing the bridge
repair cost and time by exercising the underlying analysis framework. Three types of isola-
tion devices (elastomeric bearings, frictional/sliding bearings and roller bearings) developed
and used in many countries all over the world (Kunde and Jangid, 2003) were performed.

185

NYBRIDGE_Book.indb 185 7/17/2013 8:43:15 AM

 186 Seismic analysis & design

In particular, the response is assessed in terms of repair cost and time quantities such as
Crew Working Days (CWD) and the total Repair Cost Ratio (RCR) defined as the ratio
between the cost of repair and the cost of the new construction. In the sections 2 and 3 the
employed PBEE methodology is described. The study and its results are presented respec-
tively in sections 4 and 5, comparing bridge response for the different configurations. Conclu-
sions are then presented in section 6.

2 PBEE METHODOLOGY

The methodology (Mackie et al., 2008, 2010a) is subdivided to achieve performance objectives
stated in terms of the probability of exceeding threshold values of socio-economic decision
variables (DVs) in the seismic hazard environment. The PEER PBEE framework is funda-
mentally based on the application of the total probability theorem to disaggregate the prob-
lem into several intermediate probabilistic models that involves intermediate variables, such
as repair items or quantities (Qs), damage measures (DMs), engineering demand parameters
(EDPs), and seismic hazard intensity measures (IMs). Then, the EDPs are computed directly
from the ensemble of time history analyses performed and automatically associated with the
PGs and the DVs. The data used to populate the relationships that associate EDPs to DMs
and DMs to Qs is presented in Mackie et al. (2008) and the numerical implementation of the
methodology inside the interface (Lu et al., 2011) is described in Mackie et al. (2010a).
Consequently, the applications of this methodology are very general, since the possible
variables that can be taken into consideration depend on the objectives that the decision
makers choose to refer to.

3 PBEE ANALYSIS

PBEE analysis can be organized into the major components: definition of Ground Motion
Input, bridge-ground Finite Element Model and definition of performance groups, respec-
tively shown in the next subsections.

3.1 Ground input motion

The framework can employ any kind of user-specified input motion. In particular, in this
study the motions are taken from the PEER NGA database. They consist of 10 selected
ground motions to be representative of seismicity in typical regions of California. They take
into account two main characteristics: moment magnitude (Mw) and epicentre distance (R).
For more details, see Mackie, Lu and Elgamal (2010b).

3.2 FEM model

The investigate model, representative of the prevalent ordinary construction types for new
California bridges, is a 90 m long, 2-span structure, supported on one circular column (1.22 m
diameter) 12 m long, 6.71 m above grade (Figure 1). The deck is 11.9 m wide and 1.83 m deep,
and the weight is 130.30 kN/m. Each abutment is 25 m long with 30000 kN as total weight.
The reinforced concrete column is modelled with nonlinear forced-based beam-column ele-
ments (nonlinear Beam Column, Mazzoni et al., 2009 and Lu et al., 2010) and fiber cross
section, with 0.2 rad/m as the maximum curvature value at 11900 kN compression axial
load. The deck is modelled with five separate elastic beam-column (BC) elements (with cross
Area of 5.72 m2, transversal inertia 2.81 m4 and vertical inertia 53.9 m4). For more details see
Lu et al. 2010. The approach ramps make the connection with the longitudinal boundaries.
The deck-column and deck-abutments connections consist of several isolated configurations

NYBRIDGE_Book.indb 186 7/17/2013 8:43:15 AM

 Case studies on several isolated bridge configurations 187

Figure 1. The investigate model.

Figure 2. F.E.M. definition.

that are descripted and compared in section 4. In order to concentrates on the structural
response only, a very hard soil with a sufficient high stiffness (if compared to isolator stiff-
ness, as shown in Tongaonkar and Jangid, 2003) that simulates fixed conditions is considered
and thus, influence of soil-structure interaction was neglected. For more details see also as
Elgamal et al. (2011). Figure 2 shows the performed 3D 200 m × 200 m mesh model.

3.3 Performance Groups

PBEE methodology assesses the damage and repair grouping the various structural and non-
structural components into Performance Groups (PGs) that are used in PBEE analysis. Each
major bridge component is grouped into a PG. Each PG contains a collection of components
that reflect global-level indicators of structural performance and that contribute significantly
to repair-level decisions. The notion of a PG allows grouping several components for related
repair work; therefore PGs are not necessarily the same as the individual load-resisting struc-
tural components. The interface is built with 11 PGs (Mackie, Lu and Elgamal, 2010b), repre-
sentative of typical bridge schemes and used in this study, as shown in Table 1.
For each performance group, discrete damage states (DSs) are defined, and each of these
has a subset of different repair quantities (Qs), associated for a given scenario. Once the
Qs have been established (damage to different PGs), the total repair costs can be generated
through a unit cost function that is based on the Qs. Finally, for each Q, an estimate of the
repair effort can be obtained through a production rate. More information on the derivation
of the default DSs, Qs, unit costs, and production rates can be found in Mackie et al. 2008,
2009, 2010b, Elgamal et al. 2009, 2011, 2012 and Forcellini et al. 2012.

NYBRIDGE_Book.indb 187 7/17/2013 8:43:15 AM

 188 Seismic analysis & design

Table 1. Performance Groups.

Performance group (PG) Performance group Names

1 Max Column drift ratio

2 Residual column drift ratio
3 Max rel. deck-end/abutment displ. (left)
4 Max rel. deck-end/abutment displ. (right)
5 Max bridge-abutment bearing displ. (left)
6 Max. bridge-abutment bearing displ. (right)
7 Approach residual vertical displ. (left)
8 Approach residual vertical displ. (right)
9 Abutment residual pile cap displ. (left)
10 Abutment residual pile cap displ. (right)
11 Column residual pile displ. at ground surface

4 ISOLATION CONFIGURATIONS

In order to evaluate the longitudinal resistance and behaviour, several isolated configuration
were performed. First of all, it was considered an original model (F-01) where abutments
were isolated with two HDS 650 × 337 s. They consist of two Soft Damping Rubber Bearing
isolators, of soft compound with modulus of elasticity G = 0.4 Mpa and equivalent viscous
damping ξ = 10% (Marioni 2006). They are modelled with 2 simple elastic springs (730 kN/m
each), roughly representative of real high damping rubber bearing response commonly used
in professional bridge engineering applications.
This model is compared with the theoretical case of a simple roller link connections between
the deck and the abutments (F-02) that provides no resistance (simple roller device).
The original fixed connection between the top of the column and the deck is then modelled
with two different models of isolators.
In the first model (HDRB-L), the isolation devices are two HDN 650 × 337 s, consisting
of two Normal Damping Rubber Bearing isolators, of normal compound with modulus of
elasticity G = 0.8 Mpa and equivalent viscous damping ξ = 10% (Marioni 2006).They are
modelled with one simple elastic spring (2920 kN/m; 2 × 1460 kN/m).
The second model (FP-NL) consists of two sliding pendulum devices on the top of the
column and one on each abutment. They are made of the sliding materials HOTSLIDE
series, according to EN 15129 (European Standard on Antiseismic Devices) requirements
and tested at the Eucentre laboratory at Pavia University. They are modelled with a sim-
plified two-spring model (Kelly, 1997) that can take into considerations both non-linear
behaviour and buckling response using an explicit force—deformation relations. In partic-
ular, the stability theory resembles the linearized theory of an elastic column, but accounts
for shear deformation by considering rotation of the cross section, which is independent
of the lateral deflection (Koh & Kelly 1987). Also predicted by stability analysis, the multi-
layer bearing under simultaneous lateral and axial loading undergoes an additional vertical
displacement beyond that due to material axial flexibility. The two-spring model (Figure 3)
is composed by two rigid elements connected by moment springs across hinges at the top
and bottom and by shear springs and frictionless rollers at mid-height. The kinematics
of the model is described by two DOF (degree of freedom): the shear displacements and
the relative rotation (for more details, see Kelly 1997 and 2003). In the implementation,
the linear shear spring has been replaced by a bilinear spring to represent the nonlinear
behaviour observed in elastomeric and lead-rubber bearings (Figure 4). The behaviour of
the nonlinear spring is controlled by two initial bending stiffness k1/2, the yield strength
Fyo and the post-yield stiffness k2o. Figure 5 and Table 2 summarize the configurations
adopted in the study.

NYBRIDGE_Book.indb 188 7/17/2013 8:43:16 AM

 Case studies on several isolated bridge configurations 189

Figure 3. Simplified two-spring model (I-NL) element (http://opensees.berkeley.edu/OpenSees/

manuals).

Figure 4. Simplified two-spring model (I-NL) behaviour (http://opensees.berkeley.edu/OpenSees/

manuals).

Figure 5. Schematic plan of the model.

Table 2. Data for model stiffness.

Model KLA KLC Isolation

F-01 730 × 2 fixed On abutment

F-02 0 fixed No Isolation
HDRB-L 730 × 2 2920 Full Isolation
FP-NL NL NL Full Isolation

NYBRIDGE_Book.indb 189 7/17/2013 8:43:16 AM

 190 Seismic analysis & design

5 RESULTS

The developed user interface (Lu et al., 2010, 2011, Mackie et al., 2010b) can compute directly
several figures that summarize the different disaggregated quantities. Due to the assembly-
based (vector) nature of the method applied and according to the total probability theorem,
it is possible to disaggregate the results (repair costs and time) into individual contributions.
In particular, the main contribution to the Total Repair Cost and Time is shown to be the
longitudinal drift ratio (PG1), the relative longitudinal displacement between the deck end
and the abutment (PG3), representing the column and the abutments damage respectively.
This section illustrates the results in terms of Total Repair Cost Ratio (RCR) and Repair
Time in crew working days (CWD) for the four configurations (F-01, F-02, HDRB-L and
FP-NL), neglecting Soil Structure Interaction, as mentioned.

5.1 PG1 column damage

Figure 6 shows that the cost connected to column deformability (PG1-Longitudinal drift
ratio) is activated with increasing PGA values. It is seen that the isolated configurations
(HDRB-L and FP-NL) have very low repair cost if compared with the two non-isolated
models, as expected by the design of the isolation devices. This is due to the isolation effec-
tiveness, since the bridge and the soil have independent behaviour. In F-01 and F-02 models
the main deformation of the bridge is seen to be a longitudinal rigid translation that affects
the column.
The results first of all confirm that the isolation devices chosen for the pilot study solve
to their function and that they were correctly designed. Secondly, Figure 6 shows the effec-
tiveness of sliding pendulum compared to HDRB-L especially for high values of PGA. For
example, if a PGA equal to 0.8 g is considered, the costs connected to HDRB-L configura-
tion is 20% more than the FP-NL configuration.

5.2 PG3 abutment damage

Figure 7 shows abutment damage (PG3-the relative longitudinal displacement between the
deck end and the abutment) at increasing values of PGA. Comparing F-01 and F-02, it is
possible to assess the importance of abutment isolation in order to reduce the costs connected
to abutment damage. HDBR-L model response is close to the F-01: column isolation does
not seem effective for the abutments, cause the linear behaviour of the abutment connection

Figure 6. Maximum longitudinal drift ratio (column)—PG1 results.

NYBRIDGE_Book.indb 190 7/17/2013 8:43:17 AM

 Case studies on several isolated bridge configurations 191

Figure 7. Max longitudinal relative deck end—abutment displacement—PG3 results.

Figure 8. Total repair cost ratio (%) results.

in both cases. On the contrary, FP-NL curve starts to increase for higher values of PGA if
compared with other models responses (0.4 g instead of 0.25 g). This reflects the influence
of friction pendulum devices used in correspondence of the abutments. Non linearity can
reduce the deformations (and consequently the damage) transferred to the abutments in the
other three cases. This is true for a big range of PGA (between 0.45 g to 0.70 g).

5.3 Total damage

Figures 8 and 9 show the results in terms of total Repair Cost Ratio (RCR) defined as the
ratio between the cost of repair and the cost of the new construction and repair time, such
as Crew Working Days (CWD). Comparing the cost connected to PG1 (column damage)
and PG3 (abutment damage), it is possible to assess that the total damage is mainly affected
by the abutment contribution (PG3). In particular, in F-01 and F-02 cases the displacement

NYBRIDGE_Book.indb 191 7/17/2013 8:43:18 AM

 192 Seismic analysis & design

Figure 9. Total repair time (CWD).

and consequently column damage affects the total response, since the column isolation is
not present. Anyway, this contribute is small, if compared with the abutments damage. In
details, PG1 has its maximum value at around $35000,00 while PG3 maximum value seems
to be around $180000,00. For the isolated models instead, the costs associated to the column
damage are less important, thanks to the presence of the isolation. For instance, PG1 has its
maximum value at around $12000,00 versus PG3 maximum around $180000,00. In particu-
lar, HDBR-L model response shows that column isolation has no benefits on the total repair
costs and time. Instead for FP-NL model, all the costs and time are connected to abutments
damage and, as shown in the previous paragraph and non-linearity mainly affects the range
of PGA between 0.45 g to 0.70 g.

6 CONCLUSIONS

The study conducted in this paper may be viewed as a pilot investigation to assess isola-
tion technique applying the overall PBEE analysis framework for a simple single-column
bridge-abutment configuration. This study, neglecting ground soil-structure interaction
effects, allows for estimating the bridge model response with particular attention to the
bridge-abutment configuration.
In this regards, four isolated configurations were intentionally chosen in order to highlight
the potential beneficial effect of base isolation technique. The comparison allows to focus on
the bridge-abutment interaction that permits the abutment to provide lateral and transverse
resistance to the bridge deck displacement, reducing the seismic demand on the bridge col-
umn. Several main conclusions can be underlined.
The paper shows that the damage is mainly due to displacement both to the column and
to the abutments. The results shows how isolation effectiveness can reduce column damages
especially for high values of PGA. Then, comparing the original configuration with the two
isolated configurations, it can be deduced that saving the column does not necessarily means
reducing the total costs and time for the bridge. This is due to the fact that the damage is
moved to the abutments. The paper shows how this damage can be reduced equipping the
abutments with a high non-linear isolators.
Further analysis will aim to reproduce more realistic responses and to assess more properly
the best isolated configurations. First of all the transverse behaviour has to be taken into
consideration because it can significantly modify the response of the bridge. Secondly, soil

NYBRIDGE_Book.indb 192 7/17/2013 8:43:19 AM

 Case studies on several isolated bridge configurations 193

structure interaction has to be taken into account in order to reproduce more realistically the
abutments influence on the entire bridge—soil system.

ACKNOWLEDGEMENT

The study was possible thanks to professor Ahmed Elgamal and doct. Jinchi Lu from
University of California, San Diego who helped the authors to perform, develop the interface
and introduce the isolator devices models inside the platform. The authors want to thank also
professor James Kelly from University of California, Berkeley. His assistance in the isolators
properties modelling gave determinant contributions to this paper. The reported research
was supported by Fondazione Ente Cassa di Faetano and Banca di San Marino, under the
Fondo per l’Eccellenza Sammarinese Award. This support is gratefully acknowledged.

REFERENCES

Constantinou, M.C., Tsopelas, P., Kim, Y.S. and Okamoto, S. 1993. NCEER-Taseis Corporation
Reaserach program on sliding seismic isolation systems for bridges: Experimental and analytical
study of friction pendulum system (EPS) Technical Report NCEER-93-0020.
Cornell, C.A., and Krawinkler, H. 2000. Progress and challenges in seismic performance assessment.
PEER Center News 3:2.
Eftekhari, M. and Tehrani Zadeh, M. 1996. Effect of seismic isolation systems on dynamic behavior of
bridges under earthquake loading. 1996, 11th World conference on earthquake Engineering.
Elgamal, A., Yang, Z., Parra, E. and Ragheb, A. 2003. Modeling of cyclic mobility in saturated cohe-
sionless soils. International Journal Plasticity 9:6,883–905.
Elgamal, A., Yan, L., Conte, J.P. 2008. Three-dimensional seismic response of Humboldt bay bridge-
foundation-ground system. Journal of Structural Engineering 134:7,1165–1176.
Elgamal, A., Lu, J., Forcellini, D. 2009. Mitigation of Liquefaction-Induced lateral deformation in slop-
ing stratum: Three-dimensional Numerical Simulation. Journal of Geotechnical and Geoenvironmen-
tal Engineering 135:11,1672–1682.
Elgamal, A., Forcellini, D., Lu, J., Mackie, K.R. and Tarantino, A.M. 2011. Seismic assessment of a
simple bridge-abutment model adopting a Performance-Based Earthquake Engineering approach.
(N.1129) XIV Convegno ANIDIS L’ingegneria Sismica in Italia, September 18–22, Bari, Italy.
Elgamal, A., Forcellini, D., Lu, J., Mackie, K.R. and Tarantino, A.M. 2012. A parametric study on
several bridge-abutment configurations adopting a performance-based earthquake engineering
methodology. II International Conference on Performance-Based Design in earthquake geotechnical
engineering, May 28–30, Taormina, Italy.
Forcellini, D., Tarantino, A.M., Elgamal, A., Lu, J. and Mackie, K. 2012. Seismic Assessment of
Isolated Bridge configurations subjected to liquefaction adopting a PBEE methodology. Proceed-
ings (N.260) of the 15th World Conference on Earthquake Engineering, September 24–28, Lisbon,
Portugal.
Haque, M.N., Bhuiyan, A.R. and Alam, M.J. 2010. Seismic response analysis of base isolated high-
way bridge: Effectiveness of using laminated rubber bearings. IABSE-JSCE Joint Conference on
Advances in Bridge Engineering-II, August 8–10, 2010, Dhaka, Bangladesh.
Kelly, J.M. 1997. Earthquake-Resistant Design with Rubber. Springer-Verlag.
Kelly, J., M., 2003. Tension buckling in multilayer elastomeric bearings, Journal of Engineering Mechan-
ics, ASCE, 129(12):1363–1368.
Koh, C.-G. and Kelly, J.M. 1987. Effects of axial load on elastomeric isolation bearings, Rep. No. UCB/
EERC-86/12. Earthquake Engineering Research Center, University of California, Berkeley.
Kunde, M., C., Jangid R., S. 2003. Seismic behavior of isolated bridges: a state of the art review. Elec-
tronic Journal of Structural Engineering 3: 140–170.
Lu, J., Mackie, K., and Elgamal, A. 2010. Performance-based earthquake evaluation of a single column
bridge system, Joint conference proceedings, 7th International Conference on Urban Earthquake
Engineering (7CUEE) & 5th International Conference on Earthquake Engineering (5ICEE), March
3–5, 2010, Tokyo Institute of Technology, Tokyo, Japan.

NYBRIDGE_Book.indb 193 7/17/2013 8:43:20 AM

 194 Seismic analysis & design

Mackie, K.R., Wong, J.-M., and Stojadinovic, B. 2008. Integrated Probabilistic Performance-Based
Evaluation of Benchmark Reinforced Concrete Bridges. PEER Center, University of California,
Berkeley Report No. 2007:09.
Mackie, K.R., Wong, J.-M., and Stojadinovic, B. 2010. Post-earthquake bridge repair cost and repair
time estimation methodology. Earthquake Engineering & Structural Dynamics 39:3,281–301.
Mackie, K.R., Lu, J. and Elgamal, A. 2010. User interface for performance-based earthquake engineer-
ing: a single bent bridge pilot investigation. 9th US National and 10th Canadian Conference on Earth-
quake Engineering: Reaching Beyond Borders. July 25–29, Toronto, Canada.
Marioni, A. 2006. Sistemi di isolamento sismico innovativi prodotti dalla Società ALGA (http://www.
alga.it).
Mazzoni, S., McKenna, F., Scott, M.H., Fenves, G.L., et al. 2009. Open System for Earthquake Engi-
neering Simulation, User Command-Language Manual, Pacific Earthquake Engineering Research
Center, University of California, Berkeley, OpenSees version 2.0.
Tongaonkar, N.P., Jangid, R.S. Seismic response of isolated bridges with soil-structure interaction Soil
Dynamics and Earthquake Engineering. 2003,23(4),287–302.

NYBRIDGE_Book.indb 194 7/17/2013 8:43:20 AM

 Chapter 18

Assessment and design of seismically isolated bridges in proximity

of major faults and located on liquefiable soils

C. Ozkaya, G. Cetin & F. Tulumtas

Yuksel Proje Uluslararası Co., Ankara, Turkey

ABSTRACT: In this paper, seismic isolation design of one highway viaduct and one high-
speed railway viaduct located within few hundred meters of North Anatolian fault in Turkey
is studied. For the highway viaduct, a hybrid seismic isolation system is chosen for the bridge
consisting of four natural and two lead-rubber bearings supporting the six steel-girders at
each substructure. Lead cores in the lead-rubber bearings are designed to have an overall
characteristic strength equal to 10% of the superstructure weight. This large characteristic
strength is intentionally chosen to limit the isolator displacements under near-field effects.
Furthermore, the supplementary four natural rubber bearings over each substructure pro-
vided additional stiffness. For the railway viaduct, the same design philosophy was followed
with considering the serviceability issues related to high-speed railway bridges. Curved sur-
face slider bearings having a dynamic friction coefficient of 12% with 2.54 mm yield displace-
ment are selected.

1 INTRODUCTION

Bozuyuk-Mekece State Highway is connecting the cities in the Mediterranean region and
internal Anatolia to Istanbul, Turkey. Upgrading the highway was planned due to heavy traf-
fic and accidents. Total length of the project is about 85 km and the project includes many tun-
nels, bridges and overpasses. Sakarya-II viaduct is the most important structure of this project
since half of the project is at one side of Sakarya River and other half at the other side.
Sakarya-II viaduct is the key element connecting the halves. Highspeed railway line between
Ankara and Istanbul follows almost the same route with Bozuyuk-Mekece Highway.
Turkey is an earthquake prone country and about 95% of the population is susceptible
to earthquakes. Most of the earthquakes in Turkey take place on North Anatolian fault.
North Anatolian fault is about 1500 km long and crosses the country from east to west. The
fault is a strike-slip fault having characteristics similar to San Andreas fault. Destructive
earthquakes up to Mw = 7.5 occurred on the fault-line resulting in many life and economical
losses. Recent earthquakes on the North Anatolian fault are 17 August 1999 Izmit (Mw = 7.4)
and 12 November 1999 Duzce (Mw = 7.2) earthquakes.
The Sakarya-2 viaduct and high-speed railway viaduct VK-14A are very close to each
other as shown in Figure 1.

2 SEISMICITY AND GEOTECHNICAL PROPERTIES OF THE SITE

North Anatolian fault has two branches at Marmara region. The northern branch is more
active compared to southern one and 17 August earthquake occurred on this branch.
Northern branch is very close to TEM Highway and Istanbul. On the other hand, part of
the southern branch between Geyve-Iznik is not active for centuries. Therefore, this portion
of the fault may be classified as seismic gap. Sakarya-II Viaduct is about 300 meters far from

195

NYBRIDGE_Book.indb 195 7/17/2013 8:43:20 AM

 196 Seismic analysis & design

Figure 1. General view of the Highway and Railway Viaducts at Pamukova—Turkey.

Figure 2. Seismicity of the Marmara Region and bridge site (1900–2006) (Cetin, 2008).

the southern branch of the North Anatolian fault. Historical seismicity of the region and
bridge site is presented in Figure 2. In Figure 2, earthquakes with magnitudes larger than 4.0
are highlighted.
Site-specific studies have indicated that scenario earthquake having a magnitude of
Mw = 7.5 may be expected at the site during the economic life of the bridge. Return period of
the maximum credible earthquake was selected as 1000 years. By using the attenuation rela-
tionships, peak ground acceleration at the bedrock level is calculated as 0.96g (Cetin, 2008).
However, there is little seismic data for sites that are closer than 3 km to a major fault line.
Due to close proximity of the bridge to the fault, strong directivity effects are expected at
the bridge site during a major earthquake. Near-fault earthquakes contain significant wave
pulses. For strike-slip faults, they dominate the horizontal motion and may appear as single
or double pulses with single or double-sided amplitudes. The period of the main pulse is in
between 0.5 sec-5 sec (Bolt, 2004). These pulses result in very high displacements and forces.
Elastic response spectrum analysis cannot capture strong nonlinear characteristics of near-
field earthquakes. Therefore, nonlinear time history analyses were utilized in design of the
bridge. Seven near-field earthquake records were selected and scaled according to AASHTO
Guide Spec (AASHTO, 1999). Records (Cetin, 2008) are presented in Table 1.
Conformance of ensemble spectrum formed by taking the average of the SRSS spectra
for the individual earthquakes and 1.3 × Code spectra is presented in Figure 3. It is observed

NYBRIDGE_Book.indb 196 7/17/2013 8:43:20 AM

 Assessment and design of seismically isolated bridges in proximity of major faults 197

Table 1. Earthquake Record Utilized In Analysis (Cetin, 2008).

Earthquake Station code Mw PGA EW PGA NS PGV EW PGV NS

1976–Gazli-Uzbekistan Karakyr 6.8 1.74 2.05 187 205

1994–Northridge-USA Jensen 6.7 1.01 1.82 135 119
1994–Northridge-USA La Sepulveda 6.7 2.74 3.41 285 277
1995–Kobe, Japan Takatori 6.9 0.49 0.50 103 98
1999–Duzce, Turkey Bolu 7.2 2.05 2.31 158 175
Northridge–01 Sylmar 6.69 0.76 1.11 146 126
1999–Chi Chi, Taiwan TCU 065 7.62 1.00 0.74 155 97

* PGA in g, PGV in cm/sec.

Figure 3. Code spectrum and average spectrum of 7 earthquakes.

Figure 4. Code spectrum and SRSS of selected earthquake records.

from the figure that average of the earthquakes match the code spectrum at intermediate
periods corresponding to isolation period. Scaling factors are selected for each record so as to
match the code spectra in between 1.5 sec-2 sec. On the other hand, as presented in Figure 4,
at short periods, spectra of some records do not fit well to the code spectra. Since short peri-
ods represent vibration periods of the substructure, this unfit may result in overestimation or
underestimation of substructure forces.
Upper 30 meters of the soil at the bridge site consists of sand, gravel and silt. SPT values
increase as depth increases. SPT values (N1,60) in between 15–30 are common at the upper
15–20 meters. Moreover, level of water table is close to the free field since the viaducts are on
Sakarya River. Liquefaction triggering is a function of ground motion intensity and char-
acteristics as well as geotechnical parameters (Kramer, 1996). Procedure proposed by Seed

NYBRIDGE_Book.indb 197 7/17/2013 8:43:23 AM

 198 Seismic analysis & design

Figure 5. Equivalent shear modulus during expected earthquake.

et al. (Kramer, 1996) was utilized for establishing the liquefaction susceptibility of the site.
Results have revealed that the site is susceptible to liquefaction during a major earthquake
due to high level of seismic input.
In order to verify susceptibility of the bridge site to liquefaction, equivalent linear dynamic
analyses were performed by using software Proshake (EduPro, 2003) Scaled earthquake
records and geotechnical parameters of the site were set as input. Analyses results have indi-
cated that lateral stiffness of the soil strata between 0–10 meters from the top decreases to
almost zero during a major earthquake as presented in Figure 5.

3 SAKARYA-2 VIADUCT

Sakarya-II viaduct consists of two twin bridges. Two bridges have very similar span lengths
and substructures. The bridges have nine spans with a total length of 384 meters for the
longer one and 365 meters for the shorter one. Longest span length is 52 meters and shortest
span length is 32 meters. The viaduct has a horizontal curvature near one of the abutments
as presented in Figure 6.

3.1 Superstructure
The superstructure is slab on steel girder with six continuous steel girders. Steel girders are
selected in order to provide a lighter superstructure compared to reinforced concrete super-
structure. A view of the superstructure cross-section is presented in Figure 7.
The depths of the girders are variable with minimum depth of 1.75 meters and maximum
depth of 2.75 meters. Center-to-center spacing between steel girders is 2.30 meters. Expan-
sion joints are only present at abutments. By selecting a continuous superstructure, risk of
unseating due to rotation of the substructure and unexpected displacements is eliminated to
an important extent.

3.2 Substructure
Piers are hammerhead type with an oblong shaped column. Past experience has indicated
that seismic performance of bridges with piers having multiple columns is better compared to
those having only one column. In case of piers with single column, stability problems in the
transverse direction and large vertical accelerations at the cap beam level due to transverse

NYBRIDGE_Book.indb 198 7/17/2013 8:43:24 AM

 Assessment and design of seismically isolated bridges in proximity of major faults 199

Figure 6. View of the Sakarya-2 Viaduct.

Figure 7. Cross-section of superstructure.

displacement may be the reasons of such an advantage. Unfortunately, a pier with double
column could not be selected for this bridge due to space problems since the bridge crosses
the present railway line, highway line and the highspeed railway line that is recently in con-
struction. The bridge has 20° skew angle. The dimensions of the oblong shaped column are
6.00 meters × 2.00 meters.
Cast in place reinforced concrete piles with a diameter of 120 cm were utilized in the bridge
foundations. There are 20 piles at a pier and 25 at an abutment. The piles are embedded to the
bedrock. One pier at each bridge has a spread footing since a very hard rock (conglomerate)
was encountered at the surface during construction.

3.3 Seismic Isolation Bearings

A hybrid seismic isolation system consisting of natural rubber bearings and lead rubber bear-
ings were selected for this viaduct. At each substructure, four natural rubber bearings and
two lead rubber bearings support six steel I girders.
The four natural rubber bearings over each substructure was intended to provide addi-
tional stiffness to achieve a small post elastic period to deflect the earthquake input energy
transferred through the underlying soft soil and to limit the isolator displacements under
near-field effects.
Lead core in lead rubber bearings increases damping of the system by its elasto-plastic behav-
ior under lateral loads. Structural deformations and forces diminish with energy dissipation
capacity. Overall characteristic strength of lead rubber bearings is selected as 10% of the super-
structure weight in order to limit the isolator displacements. On the other hand, higher charac-
teristic strength may result in high heat generation at lead core. Therefore, heating up of the lead
core may result in reduced energy dissipation capacities under repeating loading cycles.

NYBRIDGE_Book.indb 199 7/17/2013 8:43:25 AM

 200 Seismic analysis & design

3.4 Structural model

A 3D model of the viaduct is built and analyzed using the software Larsa 4D (Larsa Inc.).
The structural model is capable of simulating the non-linear behavior of piers, seismic isola-
tion bearings and soil-structure interaction effects.
Lead rubber bearings are modeled using hysteretic translational springs. An equivalent
pile diameter was found by performing a non-linear static analysis on the foundation system.
Equivalent single pile has nearly identical lateral response with that of the foundation system.
Utilizing a single pile in analyses drastically reduced computation time. Soil-structure inter-
action is modeled using linear viscous dampers representing effects of soil damping and radi-
ation damping. P-y curves are utilized in analyses. Dimensions of the soil column should be
selected sufficiently large in order to be able to change vibration characteristic of pile group.
Shear moduli of the soil layers are taken from dynamic analyses results of the soil site.

3.5 Analysis results

Non-linear analyses with near-field earthquake records resulted in very high substructure
forces and isolator displacements. Effective isolation period is found out to be around 1.5 sec.
This relatively short isolation period was intentionally selected in order to reduce the sensitiv-
ity of the bridge to long-period velocity pulses and high dominant site period due to soft soil
and liquefaction. Lateral design displacement for the lead rubber bearings and elastomeric
bearings are ±850 mm which resulted in very large isolation bearings. Maximum isolator
displacements for each record are presented in Table 2 (in cm).

3.6 Tests of seismic isolation bearings

Natural rubber bearings and lead rubber bearings at 9 of 10 axes have a diameter of 1100 mm.
Overall heights of the bearings are 580 mm. Diameter of the bearings utilized in the pier with
spread footing is 1400 mm with an overall height of 912 mm. Shear moduli of the rubber
bearings is in between 0.42 MPa-0.915 MPa. Bearings with 1400 mm diameter were manu-
factured in two parts and then connected by using pins and welding.
Test velocity of the lead rubber bearings was calculated as 2.2 m/sec. Combined with lateral
design displacement of ±850 mm, testing became the critical issue since there are few labora-
tories in the world capable of performing such tests. Bearing manufacturer Alga Spa selected
EUCenter Laboratory for prototype tests. However, lateral stroke capacity of the testing
equipment in EUCenter is limited to ±580 mm. Therefore, prototype tests were divided into
two parts as static prototype tests and dynamic prototype tests.
Static prototype tests were performed at Alga spa laboratory with a velocity of 1.6 mm/s
and maximum lateral displacement of approximately 1060 mm. Lateral displacement capac-
ity and stability of the bearings were verified during static prototype tests. Nevertheless, no
heating in the lead core was observed due to very low testing velocity.
Dynamic prototype tests started with a velocity of 2.2 m/s. However, lateral accelera-
tions in the movable base plate increased to 4 g at tests having low displacement amplitudes

Table 2. Isolator displacements (cm).

Earthquake Long disp. Trans disp.

Karakyr 50 100
Jensen 40 58
La Sepulveda 117 89
Takatori 24 60
Bolu 41 80
Sylmar 32 80

NYBRIDGE_Book.indb 200 7/17/2013 8:43:27 AM

 Assessment and design of seismically isolated bridges in proximity of major faults 201

Figure 8. Lead rubber bearing after dynamic prototype test.

Table 3. Damping ratios in seismic loading cycles (Test Velocity 1.4 m/s) (EuCentre 2009).

Test name S1-3 S1-4 S1-5 S1-6 S2-7 S1-7B SP-9

cycle 1 16.12 19.57 22.59 23.94 24.39 22.72 24.10

cycle 2 14.99 18.53 21.89 24.30 23.91 22.87 23.09
cycle 3 14.06 17.76 21.61 23.82 23.39 22.43 22.49
cycle 4 13.36 17.45 21.42 23.44 22.82 21.95 22.05
cycle 5 12.94 17.27 21.25 23.15
cycle 6 12.17 17.02 20.88 22.77

(EuCentre 2009). Lateral inertial force at the base plate accounted for approximately 50%
of the total lateral force. Moreover, stability of the equipment was endangered due to very
low axial load levels. For these reasons, tests with 1.4 m/sec, 1.6 m/sec and 1.8 m/sec were
performed as confirmation tests. Unfortunately, heating-up of the hydraulic oil did not allow
successive cycles without interruption. On the other hand, tested bearings were not damaged
during tests as presented in Figure 8. Performance level of the bearing was verified by these
tests. Average damping ratios during seismic loading cycles are presented in Table 3.

4 VK14 A VIADUCT

VK14 A viaduct is a 1020 meter long structure with 33 meter spans. There is a gentle hori-
zontal curvature on the viaduct. The viaduct is a part of Ankara-Istanbul high-speed railway
line. A general view of the viaduct is presented in Figure 9.

4.1 Superstructure
The superstructure consists of 10 precast prestressed I girders with a 30 cm cast-in-place
R/C slab. The view of the superstructure is presented in Figure 10. Superstructure is made
continuous by cast in place diaphragm beams connected with I girders by means of U bars
emerging from I girders, as presented in Figure 11.
The rails are continuous UIC 60 type rails. In the superstructure, expansion joints are
present at every 3 or 4 spans in order to limit additional rail stresses due to train acceleration,
deceleration and temperature changes.

4.2 Substructure and seismic isolation system

The selected bearing system consists of curved surface sliding bearings with 12% dynamic
friction coefficient. There are two bearings at each pier located on diaphragm beams. View

NYBRIDGE_Book.indb 201 7/17/2013 8:43:28 AM

 202 Seismic analysis & design

Figure 9. View of VK14 A Viaduct.

Figure 10. Cross-section of VK14 A Viaduct.

of a curved surface sliding bearing during installation is presented in Figure 12. Relatively
high friction coefficient is selected in order to fulfill stringent service load requirements of
railway bridges. Radius of curvature of the bearings is selected as 2000 mm in order to limit
post-elastic period of the structure, which is critical for near-field structures. The bearing
supplier is Alga Spa.
At intermediate expansion joints, steel shear keys allowing free longitudinal movement
and restraining transverse movement during service load conditions are selected in order to
prevent differential transverse movement between modules which can damage rails. During
an earthquake, the connection of the steel pipe to its anchorage plate fails and the seismic iso-
lation system becomes fully active. Configuration of a shear key is presented in Figure 13.
In the longitudinal direction, viscous lock-up devices were utilized in order to connect
modules to provide continuity. The devices are force-limiting type, meaning that after a force

NYBRIDGE_Book.indb 202 7/17/2013 8:43:28 AM

 Assessment and design of seismically isolated bridges in proximity of major faults 203

Figure 11. Reinforcement cage of I Girders.

Figure 12. Curved surface sliding bearing installation.

threshold the control valves open and the devices act as viscous damper. Total stroke of the
device is ±200 mm.
Double column piers are utilized with a pile foundation. Pile diameter is 165 cm. Since,
piers heights are low and they are mostly embedded in soil, the substructure is designed to
behave essentially elastic during a major earthquake.

4.3 Structural model and analysis results

The structural model of this viaduct is similar to that of Sakarya-2 viaduct. Design isolator
displacement is selected as 500 mm. According to Eurocode 15129 (Eurocode 2009), restrain-
ing rings should not be utilized at sliding bearings so as to prevent seismic pounding, which
may result in bearing damage and transmission of very high pounding forces to substructure.
Sectional view of the utilized bearing is presented in Figure 14.

4.4 Tests of seismic isolation bearings

Tests of seismic isolation systems for railway viaducts is more severe compared to other struc-
tures due to stringent service load requirements for these structures in addition to seismic

NYBRIDGE_Book.indb 203 7/17/2013 8:43:29 AM

 204 Seismic analysis & design

Figure 13. Shear key detail (dimensions in mm).

Figure 14. Sectional view of the selected curved surface sliding bearings.

load requirements. Wearing is an important issue for sliding bearings and as expected the
effect of wearing increases as the friction coefficient increases.
Wearing resistance of sliding material utilized for this viaduct should be very high since
selected friction coefficient is 12% and the superstructure is located directly on bearings.
Wearing tests are performed for a total travel path of 1.5 miles. Results of wearing tests are
presented in Figure 15. The red broken line in Figure 15 indicates friction coefficient at room
temperature, at slow moving rates representing daily temperature changes.
As seen in Figure 15, wearing resistance of the sliding material is quite high which has a
favorable effect on rail-structure interaction. Temperature sensitiveness of friction coefficient
implies that friction coefficient shall not be considered as a definite value but its variation
with temperature and wearing shall be considered during structural design. Behavior of the

NYBRIDGE_Book.indb 204 7/17/2013 8:43:30 AM

 Assessment and design of seismically isolated bridges in proximity of major faults 205

Figure 15. Results of wearing tests (Politecnico Di Milano 2011).

Figure 16. Results of dynamic prototype tests (Politecnico Di Milano 2011).

Table 4. Test Results for VK14 A Bearings (Politecnico Di Milano 2011).

NYBRIDGE_Book.indb 205 7/17/2013 8:43:33 AM

 206 Seismic analysis & design

bearings during dynamic prototype tests was quite satisfactory and degradation in energy
dissipation capacity was quite low under repeating load cycles, as presented in Figure 16
although test velocities were as high as 1 m/s. Energy dissipation capacity (EDC), effective
lateral stiffness (Keff) and friction coefficient (μ) during Prototype Test 3 are presented in
Table 4.

5 CONCLUSIONS

• The two seismically isolated viaducts were designed to withstand an MCE Mw = 7.5
earthquake.
• Seismic isolation period in near-field structures can be kept relatively small in order to
provide high restoring force and shift isolation period from that of high intensity velocity
pulses.
• Soil structure interaction shall be considered for structures on soft soils.
• Testing of seismic isolation systems for near-field structures require special attention.
• For railway bridges, requirements for bearing systems during service load conditions are as
important as seismic load conditions.

REFERENCES

AASHTO. 1999. Guide Specification for Seismic Isolation Design.

Bolt, B.A. 2004. Seismic Input Motions for Nonlinear Structural Analysis. ISET Journal of Earthquake
Technology.
Cetin, K.O. 2008. Probabilistic Seismic Risk Report for Sakarya-II Viaduct. Middle East Technical
University.
EduPro Civil Systems Inc. 2003. Proshake.
EuCentre. 2009. Pre-characterization and Characterization Tests of LRB Algasism 1100 × 58.
Pavia-Italy.
Eurocode. 2009. EN 15129: Anti-Seismic Devices.
Kramer, S.L. 1996. Geotechnical Earthquake Engineering. Prentice-Hall.
Larsa Inc. Larsa 2000.
Politecnico Di Milano. 2011. Wear Test on Special Sliding Material Used In Friction Pendulum Bear-
ings for VK12 & VK14A Viaducts at Eskisehir-Kosekoy High-speed Railway Bridges. Milano-Italy.
Politecnico Di Milano. 2011. Qualification Tests on Anti-Seismic Sliding Isolators for VK12 & VK14A
Viaducts. Milano-Italy.

NYBRIDGE_Book.indb 206 7/17/2013 8:43:35 AM

 Chapter 19

Seismic retrofitting of curved highway viaducts

G. Moor & C. Mendez

Mageba USA, New York, USA

T. Spuler
Mageba SA, Bulach, Switzerland

ABSTRACT: In recent years, curved highway bridges, which are more prone to seismic
damage than straight ones, have become an important component in modern highway sys-
tems, as the most viable option at complicated interchanges or river crossings where geomet-
ric restrictions apply. Curved structures may sustain severe seismic damage owing to rotation
of the superstructure or displacement toward the outside of the curve line. A commonly
adopted earthquake protection strategy consists of replacing the vulnerable conventional
bearings with seismic isolation devices. This paper describes, with reference to a current
project, how seismic isolation, in the form of Lead Rubber Bearings, can be retrofitted to
curved highway viaducts. As a result of such works, the constant serviceability of the struc-
ture is ensured even after the occurrence of a strong earthquake, facilitating the passage of
emergency vehicles and contributing to the safety of the population.

1 INTRODUCTION

Increasing awareness of the threats posed by seismic events to critical transport infrastruc-
ture has led to the demand to seismically retrofit highway viaducts and other bridges to
improve their ability to withstand a strong earthquake. Continually evolving technology and
the improving evaluation and design abilities of practitioners have also contributed to the
demand—as have, of course, increasingly stringent national design standards.
In recent years, curved highway bridges (Figure 1) have become more widely used, as the
most viable option at complicated interchanges or river crossings. Curved structures are more
prone to seismic damage than straight ones, and may sustain severe seismic damage owing to
rotation of the superstructure or displacement toward the outside of the curve line due to the
complex vibrations that arise during strong earthquake ground motions.
A commonly adopted earthquake protection strategy consists of replacing the vulnerable
conventional bearings with seismic isolation devices such as lead rubber bearings (Figure 2),
as described below.

Figure 1. A curved highway viaduct. Figure 2. A lead rubber bearing in a bridge.

207

NYBRIDGE_Book.indb 207 7/17/2013 8:43:35 AM

 208 Seismic analysis & design

2 SEISMIC ISOLATION

The main objective of a seismic isolation system is to increase the natural period of a structure.
However, rather than simply increasing the natural period to a high value, an efficient seismic
design also considers how energy dissipation capability can be increased and how lateral
forces can be distributed to as many substructures as possible. Bridges, and viaducts in par-
ticular, are ideal candidates for the adoption of such a seismic isolation approach due to the
ability to distribute lateral forces among multiple supports, and thanks also to the general
ease of installation and inspection of isolation devices.
An isolation system placed between the bridge superstructure and its supporting substruc-
ture is generally capable of increasing both flexibility and energy dissipation. Flexibility in the
horizontal plane will lower the frequency of the bridge, decreasing earthquake-induced accelera-
tion, while the energy-dissipating capacity of the seismic isolators will considerably reduce the
damaging energy exerted to the bridge piers. Moreover, when isolation devices are installed at
the tops of a bridge’s piers, the lateral force from the superstructure during a seismic event can be
distributed among all piers, avoiding the concentration of lateral forces at specific locations.
Among the different seismic isolation devices available, elastomeric isolators have found
wide application in bridge structures. This is due to their simplicity and their combining of
isolation and energy dissipation functions in a single compact unit. They provide a high level of
damping—a crucial aspect of seismic protection—to minimise the seismic energy flow to the
superstructure and to limit the horizontal displacements of the isolators (Mendez et al., 2011).
A particularly popular type of elastomeric isolator among seismic engineers is described
below: the Lead Rubber Bearing.

3 LEAD RUBBER BEARINGS (LRB)

As shown by Figures 3 to 8, a Lead Rubber Bearing (LRB) is similar to a normal elastomeric

bearing with anchor plates, but made using specially chosen elastomer and with a lead plug at

Figure 3. 3D illustration of a multi-directional

Figure 4. Cut-out view of a multi-directional
LRB, designed to allow movement in any hori-
LRB, showing the lead core at the centre of the
zontal direction.
elastomeric pad.

Figure 5. 3D illustration of a guided LRB, Figure 6. Section through a guided LRB, show-
designed to allow movements along one axis and ing guide-bars at each side of the elastomeric
resist transverse forces. pad, and the pad’s lead core.

NYBRIDGE_Book.indb 208 7/17/2013 8:43:36 AM

 Seismic retrofitting of curved highway viaducts 209

Figure 7. Cut-out view of a guided Lead Rubber Bearing (LRB), showing inner construction (including
lead core at centre of the elastomeric pad).

Figure 8. A typical rectangular Lead Rubber Figure 9. Force—Displacement hysteresis loop

Bearing, ready for installation. of a LASTO®LRB Lead Rubber Bearing.

Figure 10. Effective secant stiffness of a Figure 11. Equivalent damping ratio of a
LASTO®LRB Lead Rubber Bearing. LASTO®LRB Lead Rubber Bearing.

its centre. The lead plug deforms plastically during an earthquake, dissipating energy through
hysteretic damping as indicated by the hysteresis loop graph in Figure 9.
The effective secant stiffness and equivalent damping ratio are two of the most important
characteristics of an LRB for use in the design of a seismic isolation system. Examples are
shown in Figures 10 and 11.
This type of seismic isolator was among the first to be tested in accordance with the new
European standard for anti-seismic devices, EN 15129 (CEN 2009, Moor et al., 2011), see
Figure 12. The testing, carried out at the European Centre for Training and Research in

NYBRIDGE_Book.indb 209 7/17/2013 8:43:36 AM

 210 Seismic analysis & design

Figure 12. Testing of LRB in accordance with EN 15129, with displacements of +/- 250 mm at speeds
of up to 1.634 m/s, while subjected to vertical loads of up to 3,450 kN.

Earthquake Engineering (EUCENTRE) in Pavia, Italy, paved the way for certification with
the CE label, verifying conformance with the standard. The tested LRB isolators each had
a diameter of 500 mm and a total height of 286 mm, and were designed for a maximum dis-
placement of 250 mm and a maximum vertical load of 3,450 kN. The samples were subjected
to 14 different tests, most of them including dynamic conditions, with a total of 37 cycles,
and with frequency and amplitude varying from one test to the next. The velocities reached
during the testing also varied, with values of up to 1.634 m/s. For all dynamic testing, the
applied vertical pressure of 6 MPa represented a vertical load on the samples of 1,131 kN.
The use of Lead Rubber Bearings in retrofitting seismic isolation to a curved highway
viaduct is described below.

4 THE RETROFITTING OF SEISMIC ISOLATION TO THE VIADUC

DE CHILLON

The Viaduc de Chillon is an important viaduct on the A9 highway near the Swiss city of
Montreux. When constructed between 1966 and 1969, the structure was designed in accord-
ance with the design standards of the day, but significant advances have been made in these
standards in the intervening period—most notably in relation to seismic safety. The current
Swiss design code SIA 262, published in 2003, defines the effects and influences for which
structures must be designed, and includes details of the seismic actions and the braking forces
that structures must be designed to withstand, depending on location and intensity of traf-
fic. A review of the viaduct’s design concluded that substantial changes would be required in
order to bring the structure up to the safety standard defined by the code.
The viaduct has a length of 2,100 m, and consists of two structures, side by side, each
12 m wide. It has a total of 23 spans of prefabricated concrete, each of approximately 100 m
length. Views of the structure are presented in Figures 13 and 14.

4.1 The viaduct’s existing design for movements and rotations

The twin slender pillars at the ends of each span support the deck above by means of simple
concrete hinges, which allow small rotations but no sliding movements. As a consequence
of this, horizontal deck movements are facilitated by deflection of the piers alone. The piers
beneath each carriageway are 0.8 m wide and 5 m long, and vary greatly in height, from 1 m
to 45 m, resulting in greatly differing stiffnesses. See Figures 15 to 18.

NYBRIDGE_Book.indb 210 7/17/2013 8:43:37 AM

 Seismic retrofitting of curved highway viaducts 211

Figure 13. View of the viaduct, showing its loca- Figure 14. View of twin structures of viaduct,
tion on the side of a hill. showing typical arrangement of two slender pil-
lars per structure at each end of each span.

Figure 15. The simple existing connections of Figure 16. A concrete hinge, which prevents
the deck to the support pillars, without any type horizontal movement of the deck but permits
of mechanical bearing. small longitudinal rotations.

The longitudinal and transverse movements of the deck that would arise in a design earth-
quake could not be facilitated by deflection of the piers alone, especially considering their
greatly varying stiffnesses, both longitudinal and transverse. A solution which would apply
modern seismic engineering methods to this decades-old structure was required.

4.2 The solution using Lead Rubber Bearings

In order to make the viaduct capable of withstanding a seismic event as defined by current
design standards, it was determined that the concrete hinge connections between pier and
deck at the short piers and the abutments should be replaced by seismic isolation devices.
The isolators required for this project would be required to fulfil the following requirements:
1. facilitate movements of up to 40 mm in each direction in most cases (or 80 mm in the lon-
gitudinal direction at the abutments)—thus countering the relative stiffness of the short
piers and preventing them from failing;
2. provide damping of up to 28%;
3. dissipate hysteretic energy;
4. ensure re-centering following an earthquake;

NYBRIDGE_Book.indb 211 7/17/2013 8:43:38 AM

 212 Seismic analysis & design

Figure 17. Some pillars, with lengths of up to Figure 18. Some pillars, with a length of just
45 m, can facilitate significant longitudinal deck one metre, are too short and stiff to facilitate any
movements through pillar deflection. deck movements by deflection.

Figure 19. Sections through pillar (parallel (left) and transverse (right) to span of bridge), showing first
stage of pier extension and bearing replacement works. The existing concrete hinge can be seen in each
section. Holes are drilled through the pillar for the connection of concrete extensions at each side.

5. increase the period of the deck of the bridge to more than 2 seconds; and
6. be able to transmit horizontal loads of up to 400 kN.
Lead rubber bearings (LRB) were chosen to fulfil these needs. Having determined the
exact properties they should possess, in terms of stiffness, damping characteristics and so on,
the bearings were designed to achieve those properties, with particular attention to dimen-
sions and design of elastomeric mix.
Before replacing the existing concrete hinges with the new seismic bearings, it was first neces-
sary to adapt the main structure to suit the new isolators. In particular, the individual pillars
which would support the new bearings had to be extended, to enable them to support the
hydraulic jacks used to hold the bridge during the installation of the new isolator bearings.
Since the works have not yet been completed, the approach is illustrated below by details from
the design drawings. Figures 19 and 20 show the first and final stages respectively of the extend-
ing of individual pillars and the installation of a new LRB to replace each existing concrete
hinge.

NYBRIDGE_Book.indb 212 7/17/2013 8:43:40 AM

 Seismic retrofitting of curved highway viaducts 213

Figure 20. Sections through pillar (parallel (left) and transverse (right) to span of bridge), showing
final stage of extension and bearing replacement works. After cutting off the top of the pillar to make
space for the new bearing, the pillar has been extended with new concrete at each side, and a new Lead
Rubber Bearing has been placed with the aid of hydraulic jacks placed on the new concrete extensions.

Figure 21. Section (parallel to span of bridge, Section 1 in Figure 22) through both pillars of one
structure at one pier location, showing steel bracing placed between the pillars.

In addition, strengthening of the piers was required, with new bracing installed between
the individual pillars at each location, as illustrated by Figures 21 and 22.
Initial images from the ongoing construction work on site are presented in Figures 23 and
24. The project is scheduled for completion in 2014.

5 CONCLUSIONS

The use of a seismic isolation strategy in this case—and in particular, one based on Lead Rubber
Bearings—has already proven to be a sensible approach to the challenge presented by the need to
make this important highway viaduct seismically safe in accordance with current seismic design
standards. By providing an alternative to conventional earthquake resistance design measures,
it saves the major strengthening works which would otherwise be required, if the benefits of
energy dissipation and damping were not incorporated in the design. The project thus demon-
strates the potential such an approach, and such devices, have to significantly reduce seismic risk
without compromising the safety, reliability, and economy of bridge structures.
As a result of these works, the constant serviceability of the structure is ensured even after
the occurrence of a strong earthquake, facilitating the passage of emergency vehicles and
contributing to the safety of the population.

NYBRIDGE_Book.indb 213 7/17/2013 8:43:42 AM

 214 Seismic analysis & design

Figure 22. Plan view of both pillars of one structure at one pier location, showing steel bracing placed
between the pillars.

Figure 23. One pillar at one location after drilling Figure 24. Addition of reinforcing steel for new
of holes for connection of concrete extensions. concrete extensions.

REFERENCES

CEN (European Committee for Standardization). 2009. EN 15129 – Anti-seismic devices. Brussels,
Belgium.
Mendez Galindo, C, Spuler, T, Moor, G, & Stirnimann, F. 2012. Design, full-scale testing and CE-
certification of anti-seismic devices according to the new European norm EN 15129: Elastomeric
Isolators. Proc. 15th World Conference on Earthquake Engineering. Lisbon, Portugal.
Moor, G, Mendez, C, and O’Suilleabhain, C. 2011. Measuring Up, Bridge Design & Engineering, Issue
63. London, England.

NYBRIDGE_Book.indb 214 7/17/2013 8:43:43 AM

 Chapter 20

Retrofit procedures of bridge networks against seismic

actions in Italy

A. Pipinato
Department of Civil, Architectural and Environmental Engineering, University of Padova, Italy

ABSTRACT: Existing bridges represent a relevant category of the international cultural

heritage, being the evidence of the modern industrial technology. Many of these structures
require particular rehabilitation due to design defects, basic elements deterioration, variation
of use or intensity change of the imposed loads. With regard to Italy, the historical heritage
is rich of significant structures, which played an essential role in the growth of the industrial
civilization: the most part of this heritage is represented by bridges, and a large part of the
Italian bridges has more than one hundred years. In this paper, a comprehensive overview
on practical issues on the assessment of bridge networks are presented in order to evaluate
the structural integrity of historical and deteriorated bridges, incorporating code aspects,
analytical, mechanical and structural topics.

1 INTRODUCTION

The risk associated to the seismic vulnerability of bridges and infrastructure in general is a
relevant issue to guarantee standard safety and security of citizens in everyday life and more-
over in the case of disaster. In Italy, tragic events have kicked the nation also recently (Aquila
2009 and Emilia 2012), and in the past—in the second half of the 20th century (Friuli 1976
and Irpinia 1980)—even if the transportation infrastructure has not suffered significantly. In
the Apennine mountains crossing of the A16 highway the bridges did undergo some damage,
mainly due to the inadequacy of the bearing devices, but this was promptly remedied by the
owner through the systematic adoption of the seismic isolation. The delay in the apprecia-
tion of the risk is not exclusive to Italy, but also in US, the Federal Highway Administration
(FHWA) published a first document titled “Retrofitting guidelines for Highway Bridges”
some years after (FHWA-ATC 1983) the San Fernando (1971) earthquake. Still, in 1989,
despite of the large retrofit program developed, the Loma Prieta earthquake exposed sub-
stantial deficiencies in bridges in California (Pinto et al., 2011). A relevant number of bridges
and viaducts of medium and large span are present in the national infrastructural system,
Figures 1 and 2. The situation as briefly outlined above is sufficient to understand that the
state of the art on seismic assessment and retrofit of bridges still needs to be advanced in sev-
eral areas. The areas considered to be of priority interest were assessment methods, retrofit
criteria and intervention techniques. Although, a relevant amount of work on existing bridge
network has been performed in the past, a comprehensive overview of recent studies should
help in the identification of the state of knowledge of the matter.

2 SEISMIC ACTIVITY IN ITALY

In the last two thousand years, Italy has been hit by over thousands of medium to strong
earthquakes measuring more than grade V on the Mercalli scale, and by hundreds events of
an intensity equal to or higher than grade VIII on the Mercalli scale. Earthquakes with 1,000

215

NYBRIDGE_Book.indb 215 7/17/2013 8:43:44 AM

 216 Seismic analysis & design

Figure 1. Aglio Viaduct, Milano-Roma-Napoli Highway.

Figure 2. Italia Viaduct, Salerno-RC Highway.

or more deaths since 1900 (USGS 2013) have been reported in Table 1, while selected earth-
quakes of historic interest in Italy (USGS, 2013) have been reported in Table 2: as could be
inferred from the analysis of available data, in the twentieth century alone, 7 earthquakes had
a magnitude of 6.5 or more (grade X and XI on the Mercalli scale). Disastrous earthquakes
like the long seismic period in 1783 in Calabria (which peaked at grade XI on the Mer-
calli scale), have deeply scarred the land and left recognizable signs of recovery and recon-
struction (Protezione Civile 2013). In the last forty years, the economic damage caused by

NYBRIDGE_Book.indb 216 7/17/2013 8:43:44 AM

 Retrofit procedures of bridge networks against seismic actions in Italy 217

Table 1. Earthquakes with 1,000 or more deaths since 1900 (USGS, 2013).

Date UTC Location Deaths Magnitude Comments

1908/12/28 Messina, 72000 7.2 Over 40% of the population of Messina and
Italy more than 25% of Reggio di Calabria killed
38.15 N 15.68E by the earthquake and tsunami, as well as by
fires in some parts of Messina. Casualty toll is
based on census data 1901–1911, some estimates
are as high as 110,000. Severe damage in large
parts of Calabria and Sicily. Felt throughout
Sicily and north to Naples and Campobasso.
Also felt on Malta, in Montenegro and Albania
and on the Ionian Islands. Tsunami heights
of 6–12 m (20–39 ft) observed on the coast of
Sicily south of Messina and heights of 6–10 m
(20–33 ft) observed along the coast of Calabria.
Aftershocks continued into 1913.
1915/01/13 Avezzano, Italy 32610 7.0 Severe damage in the Avezzano-Pescina area.
41.98 N 13.65E An estimated 3,000 more people died in the
next few months from indirect effects of the
earthquake. Felt throughout Central Italy from
Veneto to Basilicata.
1980/11/23 Southern Italy 2735 6.5 According to official statistics, 2,735 people
40.9 N 15.3E were killed, about 9,000 were injured, about
394,000 homeless and extensive damage
(maximum intensity X) in Basilicata, Campania
and parts of Puglia. Castelnuovo di Conza,
Conza della Campania, Laviano, Lioni,
Sant’Angelo dei Lombardi and Santomenna
were almost completely destroyed. In Basilicata
and Campania, more than 77,000 homes were
destroyed and 755,000 were damaged. Landslides
caused many houses to collapse and ground
cracks were observed in the area. The earthquake
was felt from Sicily to the Po Valley.
1930/07/23 Irpinia, Italy 1404 6.5 Most of the damage was in the Ariano
41.05 N 15.37E Irpino-Melfi area of Avellino, Potenza and
Foggia Provinces. Damage occurred as far away
as Napoli (Naples). The quake was felt from the
Po Valley to Catanzaro and Lecce Provinces.
Earthquake lights were reported in the epicentral
area.
1976/05/06 Northeastern 1000 6.5 1,000 reported killed, at least 1,700 injured, and
Italy extensive damage in the epicentral area. The
46.4 N 13.3E quake was reported felt throughout Europe.
A magnitude 4.6 foreshock preceded the main
shock by about 1 minute and 7 seconds. The main
shock was followed by a number of aftershocks,
at least one reaching a magnitude of 5, that
caused additional damage and injuries.

seismic events has been assessed at around 80 billion euro, to which must be added damage
to historical, artistic and monumental cultural heritage. In Italy, the relationship between
the damage caused by earthquakes and the energy released during the events is much higher
than in other countries with high seismicity, such as California or Japan. For example, the
earthquake in 1997 in Umbria and Marche was responsible for damage (homeless: 32,000;

NYBRIDGE_Book.indb 217 7/17/2013 8:43:45 AM

 218 Seismic analysis & design

Table 2. Selected earthquakes of general historic interest in Italy (USGS, 2013).

Date-UTC Location Magnitude Fatalities

2012 05 20 North-east Italy M 6.0 27

2009 04 06 Central Italy M 6.3 295
2002 11 01 Southern Italy M 5.8 –
2002 10 31 Southern Italy M 5.9 29
2002 09 06 Southern Italy M 6.0 2
1997 09 26 Central Italy M 6.4 11
1980 11 23 Southern Italy M 6.5 3,000
1976 05 06 Northeastern Italy M 6.5 1,000
1968 01 15 Western Sicily M 6.5 231
1930 07 23 Irpinia, Italy M 6.5 1,404
1920 09 07 Toscana, Italy M 6.4 171
1919 06 29 Mugello, Tuscany M 6.3 100
1915 01 13 Avezzano, Italy M 7.0 32,610
1905 09 08 Calabria, Italy M 7.9 557
1908 12 28 Messina, Italy M 7.2 70,000
1857 12 16 Naples, Italy M 6.9 11,000
1783 02 04 Calabria, Italy – 50,000
1693 01 11 Sicily, Italy M 7.5 60,000

economic damage: approximately 10 billion euro) to be compared with that caused by the
1989 earthquake in California (14.5 billion dollars), despite it having around 30 times less
energy. This is mainly due to the high population density and to the considerable fragility of
our constructions (Protezione Civile 2013).
Italy has a medium-high seismic hazard (due to the frequency and intensity of phenom-
ena), very high vulnerability (due to the fragility of building, infrastructural, industrial, pro-
duction and service assets) and an extremely high exposure (due to population density and
its historical, artistic and monumental heritage that is unique in the world). This territory
therefore has a high seismic risk, in terms of victims, damage to buildings and direct and
indirect costs expected after an earthquake (Protezione Civile 2013).
The precise timing of the seismic classification has been precisely reported by (Protezione
Civile 2013):
• up to 2003 the national territory was classified in three seismic categories with different
forces;
• in the past, Ministerial Laws between 1981 and 1984 classified “seismic” totally 2,965
italian municipalities on the total of 8,102, that correspond to the 45% of the national ter-
ritory, in which the 40% of the population lives;
• new criteria for seismic classification were published in 2003: they are based on recent
studies and processing regarding seismic dangerousness of the territory, i.e. the analysis
of the likelihood that a territory may be affected, during a given time interval—generally
50 years—by an event that exceeds a given intensity or magnitude threshold;
• current Technical Regulation for Buildings (Ministerial Decree of 14 January 2008), has
indeed modified the role that seismic classification had for planning purposes: for each
zone—and thus municipal territory—a value of peak acceleration, and consequently a
spectrum of elastic response, was previously supplied to calculate seismic actions (Figure 3).
As of 1 July 2009, 2008 Technical Regulations for Buildings came into force: each building
has its own acceleration, according to geographical coordinates of the project area and
to the nominal design life of a building: the degree of core dangerousness, then, can be
defined for each point of the national territory, within an area of 5 sq. meters, regardless
of local administrative borders. Seismic classification (which seismic zone a municipality

NYBRIDGE_Book.indb 218 7/17/2013 8:43:46 AM

 Retrofit procedures of bridge networks against seismic actions in Italy 219

Figure 3. Italian seismic classification (Protezione Civile, 2013).

belongs to) is thus useful only for planning management and territorial control by relevant
boards (Region, Genio, Municipal Agency etc.).
Along the analysis of the aforementioned data, one principal observation could be inferred:
no earthquake over the threshold of M = 6.5 has been recently observed, as the last one of
this magnitude refer to 1915. At the same time, the national infrastructure are not so old,
being the large part of railway being built from the 1920, and concerning highway from the
1960. Since a majority of the bridges were built before explicit code requirements for seismic
design, it is expected that widespread damage to the bridge infrastructure might occur in the
case of highest magnitude earthquakes. Many owners have recognized the existing vulner-
ability and have begun to perform bridge retrofits along the nation. However, funding con-
straints, the nature of the hazard, and the lack of clear understanding of the seismic behavior
of bridges have led to very few bridges having undergone seismic retrofit. Moreover, a wide-
spread understanding of the retrofits types that are available for addressing the common
vulnerabilities of deficient bridges is not available. This paper provides a detailed account of
the common bridge assessment and retrofit practices, with the goal of providing critical and
timely information for the more relevant Italian managing agencies (Figures 4 &5) that are
considering options for reducing the seismic vulnerability of their bridges.

NYBRIDGE_Book.indb 219 7/17/2013 8:43:46 AM

 220 Seismic analysis & design

Figure 4. Italian highway network (Autostrade per l’Italia, 2013).

Figure 5. Italian railway network (RFI, 2013).

3 CODE AND STANDARDS FOR SEISMIC DESIGN

3.1 Eurocode
Eurocodes system includes a normative document for the seismic design of new bridges,
which is at least partially based on the recent concepts of performance-based design:
Eurocode 8 Part 2 (CEN 2005a). This document, however, is not matched by a specific
document on existing bridges, while a document is available in the form of Eurocode
8 Part 3 (CEN 2005b) for existing buildings. The scope of Eurocode 8 is defined in EN
1998-1:2004, 1.1.1 and the scope of this Standard is defined in 1.1.1. Additional parts of
Eurocode 8 are indicated in EN 1998-1:2004, 1.1.3. Within the framework of the scope set
forth in EN 1998-1:2004, the Standard contains the particular Performance Requirements,

NYBRIDGE_Book.indb 220 7/17/2013 8:43:47 AM

 Retrofit procedures of bridge networks against seismic actions in Italy 221

Compliance Criteria and Application Rules applicable to the design of earthquake resistant
bridges. This primarily covers the seismic design of bridges in which the horizontal seismic
actions are mainly resisted through bending of the piers or at the abutments; i.e. of bridges
composed of vertical or nearly vertical pier systems supporting the traffic deck superstruc-
ture. It is also applicable to the seismic design of cable-stayed and arched bridges, although
its provisions should not be considered as fully covering these cases. Suspension bridges,
timber and masonry bridges, moveable bridges and floating bridges are not included in the
scope of the code, but contains only those provisions that, in addition to other relevant
Eurocodes or relevant Parts of EN 1998, should be observed for the design of bridges in
seismic regions. In cases of low seismicity, simplified design criteria may be established (see
EC8, 2.3.7-1). The following topics are dealt with in the text of the code: basic requirements
and compliance criteria, seismic action, analysis, strength verification, detailing. The code
also includes a special section on seismic isolation with provisions covering the application
of this method of seismic protection to bridges. Annex G contains rules for the calcula-
tion of capacity design effects. Annex J contains rules regarding the variation of design
properties of seismic isolator units and how such variation may be taken into account in
design. The basic principle of Eurocode 8 lies on the no-collapse requirement (ultimate
limit state) and on the minimization of damage (serviceability limit state). According to the
first criteria, after the occurrence of the design seismic event, the bridge should retain its
structural integrity and adequate residual resistance, although at some parts of the bridge
considerable damage may occur: the bridge should be damage-tolerant i.e. those parts of the
bridges susceptible to damage, by their contribution to energy dissipation during the design
seismic event, should be designed in such a manner as to ensure that, following the seismic
event, the structure can sustain the actions from emergency traffic, and inspections/repair
can be performed easily; according to the minimization of damage criteria, only second-
ary components and those parts of the bridge intended to contribute to energy dissipation
during the design life of the bridge should incur minor damage during earthquakes with a
high probability of occurrence. The non-collapse requirement for bridges under the design
seismic event is more stringent than the relevant requirement for buildings, as it contains the
continuation of emergency traffic. According to EC8, the bridge shall be designed so that
its behavior under the design seismic action is either ductile, or limited ductile/essentially
elastic, depending on the seismicity of the site, on whether seismic isolation is adopted for its
design, or any other constraints which may prevail (Figure 6). This behavior (ductile or lim-
ited ductile) is characterized by the global force-displacement relationship of the structure,
shown schematically in the following Figure 6. In regions of moderate to high seismicity it
is usually preferable, both for economic and safety reasons, to design a bridge for ductile
behavior, i.e. to provide it with reliable means to dissipate a significant amount of the input
energy under severe earthquakes. This is accomplished by providing for the formation of
an intended configuration of flexural plastic hinges or by using isolating devices in accord-
ance with Section 7 of the code. The part of this sub-clause that follows refers to ductile
behavior achieved by flexural plastic hinges: bridges of ductile behavior shall be designed so
that a dependably stable partial or full mechanism can develop in the structure through the
formation of flexural plastic hinges. These hinges normally form in the piers and act as the
primary energy dissipating components. As far as is reasonably practicable, the location of
plastic hinges should be selected at points accessible for inspection and repair. The bridge
deck shall remain within the elastic range. However, formation of plastic hinges (in bend-
ing about the transverse axis) is allowed in flexible ductile concrete slabs providing top slab
continuity between adjacent simply-supported precast concrete girder spans. Plastic hinges
shall not be formed in reinforced concrete sections where the normalized axial force defined
in EC8-5.3 exceeds 0,6.
The code does not contain rules for provision of ductility in pre-stressed or post-tensioned
members: consequently such members should be protected from formation of plastic hinges
under the design seismic action. Flexural plastic hinges need not necessarily form in all piers.

NYBRIDGE_Book.indb 221 7/17/2013 8:43:49 AM

 222 Seismic analysis & design

Figure 6. Seismic behavior.

However the optimum post-elastic seismic behavior of a bridge is achieved if plastic hinges
develop approximately simultaneously in as many piers as possible. The capability of the struc-
ture to form flexural hinges is necessary, in order to ensure energy dissipation and consequently
ductile behavior (see 4.1.6(2)). The deformation of bridges supported exclusively by simple low
damping elastomeric bearings is predominantly elastic and does not lead in general to ductile
behavior (see 4.1.6(11)P). The global force-displacement relationship should exhibit a significant
force plateau at yield and should ensure hysteretic energy dissipation over at least five inelastic
deformation cycles. Elastomeric bearings used over some supports in combination with mono-
lithic support on other piers, may cause the resisting force to increase with increasing displace-
ments, after plastic hinges have formed in the other supporting members. However, the rate of
increase of the resisting force should be appreciably reduced after the formation of plastic hinges.
Supporting members (piers or abutments) connected to the deck through sliding or flexible
mountings (sliding bearings or flexible elastomeric bearings) should, in general, remain within
the elastic range. For bridges of ductile behavior, capacity design shall be used to ensure that an
appropriate hierarchy of resistance exists within the various structural components. This is to
ensure that the intended configuration of plastic hinges will form and that brittle failure modes
are avoided. The fulfillment of the ductility shall be achieved by designing all members intended
to remain elastic against all brittle modes of failure, using “capacity design effects”. Such effects
result from equilibrium conditions at the intended plastic mechanism, when all flexural hinges
have developed an upper fractile of their flexural resistance (over strength). For bridges of lim-
ited ductile behavior the application of the capacity design procedure is not required.

3.2 Italian code

In 2003 Italian codes started a change in the direction of an harmonization with Eurocode 8.
Now a specific seismic chapter of the current Eurocodes-aligned national design code pro-
duced by the “Ministero delle Infrastrutture” (DM2008) is currently used in structural design.
The need for a document dealing with the problem of assessing and retrofitting bridges in seis-
mic areas dates back to 2003, when the update of the seismic design code came together with
the obligation of assessing into five years all the strategic structures and infrastructures in Italy
for a code obligation (Pinto et al. 2011). Adhering to this code obligation and with reference to
bridges, with funding from the Civil Protection Department, the national infrastructure Com-
mitting Agencies launched a program for the assessment of all existing bridge structures.

NYBRIDGE_Book.indb 222 7/17/2013 8:43:49 AM

 Retrofit procedures of bridge networks against seismic actions in Italy 223

4 INVENTORY AND CHARACTERISTICS OF BRIDGES IN ITALY

Structural typologies characterizing the most of the existing highways are quite different.
Highways built after 1980, includes rather uniform typologies: a) precast segmental box
girder bridges with pier heights up to 90 m and span lengths between 40 and 100 m, b) girder
bridges in concrete and in composite steel-concrete with pier heights between 5 and 30 m
and span lengths between 20 and 80 m. To exemplify, representative bridges are: the Bor-
gone viaduct (20 + 26 × 40 + 20 m), the Ramat viaduct (50 + 9 × 100 + 50 m and tall piers),
Bardonecchia bridge (7 × 42 m) and the Millaures bridge (6 × 80 m, composite steel-concrete).
Older highways, built from 1960shows a greater typological variability, which can be reduced,
however, to a few homogeneous sets. Representative bridges are: the Borgotaro viaduct (slab
bridge with several interconnections), the Narboreto bridge (4 × 30 m), the Rio Verde viaduct
(2 × 65 + 6 × 95 + 76 m and very tall piers, h = 150 m) and the Roccaprebalza South viaduct
(13 × 45 m and tall piers). As specificed before, older bridges were not designed against seismic
actions, while most recent infrastructures are mainly designed also according to Eurocodes.

5 STRUCTURAL DESIGN AND RETROFIT

5.1 Superstructure retrofits

The primary failure leading to collapse identified after the observations of the recent major
earthquakes, like San Fernando (1971) was the unseating of superstructures at the expansion
joints and abutments, a problem shared by both types of superstructures, the weak cross-
bracing and/or diaphragms. As during the motion adjacent spans could vibrate out of phase,
girders could fall down. Depending on the type of girder external restraints, there are several
ways of retrofitting.
• Simply supported steel girders and/or precast concrete girders: cable restrainers could be
adopted in order to avoid unseating; this measure could be combined with a cap seat exten-
sion if expected longitudinal displacements are larger than the available seat width; for
steel girders, a more efficient retrofit lies on the modification of the structural longitudinal
scheme, making continuous adjacent girders; this last method could not be applicable for
curved bridges or in bridges with particular seating details.
• Continuous girders: in this case, specific interventions are often needed at the supports, oth-
erwise specific interventions are needed at the internal connections; weak points like cross-
bracing or diaphragms may have to be improved. Bracing systems may tend to buckle in
presence of the seismic action and, if channel diaphragms are not full depth of the girder,
the webs could cripple. In general, the ideal solution is to introduce additional sets of brac-
ing, stiffeners, and/or full-depth channel diaphragms as close to the bearings as physically
possible (Keady et al., 2000).

5.2 Seismic isolation

There are three main objectives when adding seismic isolation bearings to bridges: to shift
the natural frequency of the structure out of the region of dominant earthquake energy,
to increase damping in the structure, and last, to lessen the dynamic reactions between
the bridge superstructure and substructure (Wendichansky et al., 1995; Mayes et al., 1984,
1994; Wright et al., 2011). An elastomeric isolation bearing consists of a number of rub-
ber layers and steel shims, bonded in alternating layers, to produce a vertically stiff but
horizontally flexible isolator. The alternating steel and rubber layers act to restrain the
rubber layer from bulging laterally. This kind of bearings can provide for flexibility and
hysteretic/viscous damping forces. They can be either low damping or high damping bear-
ings. The insertion of a lead plug in an elastomeric isolator provides energy dissipation for

NYBRIDGE_Book.indb 223 7/17/2013 8:43:50 AM

 224 Seismic analysis & design

seismic response and stiffness for static loads. They can be grouped in (i) Natural Rubber
Bearings, (ii) High Damping Rubber Bearings (HDRB) and (iii) Lead Rubber Bearings
(LRB). Sliding devices consists of sliding supports providing for frictional damping forces.
Modern sliding bearings consist of a sliding interface and a rotational element needed for
maintaining the full contact at the sliding interface. The rotational element may take vari-
ous forms such as in the pot bearing, the spherical bearing, the disc bearing, the articulated
slider in the friction pendulum bearing or an elastomeric bearing. The type of material at
the slider interface may be: un-lubricated PTFE, where the un-lubricated interfaces consist
of highly polished austenitic stainless steel in contact with PTFE or similar composites (as
those used in FPS bearings); lubricated PTFE, where lubricated interfaces consist of highly
polished austenitic stainless steel in contact with unfilled PTFE, and lubrication is applied
by grease stored in dimples; bimetallic interfaces: interfaces consisting of stainless steel in
contact with bronze or similar metals impregnated with a lubricant such as lead, PTFE or
graphite. The additional issues that this kind of interface may lead to those already related
to the previous categories make it critical for use. Metallic and friction dampers are relatively
economic, and are used when a control is needed on the level of the provided force, when
an increased initial structural stiffness is needed, and/or when the main concern is to reduce
displacement as opposed to acceleration. Friction Dampers dissipate the seismic energy by
friction developing between two solid bodies sliding relatively one to the another. Typical
examples of these devices are: slotted-bolted connections; pall devices; Sumitomo devices.
Metallic dampers take the advantage of hysteretic behavior of metals when deformed into
the post-elastic range. A wide variety of different types of devices have been developed,
with basic shapes cut from thick steel plates, among these: C/E-shaped Hysteretic Dampers;
EDU device; ADAS and TADAS Elements; Lead Extrusion Devices; torsional beams, bell
dampers, steel tubes, etc. Some recently developed damper systems incorporate re-centering
capabilities, characterized with the so-called flag-shaped hysteretic loop, among these: shape
Memory Alloys Dampers (SMA); Energy Dissipating Restrain (EDR); Friction Spring
Seismic Dampers; Post-Tensioned Energy Dissipating (PTED) steel connections. The main
advantage of the self-centering behavior consists in reducing permanent offsets when the
structure deforms in-elastically. Magneto-rheological Dampers (DMPs) typically consist of
hydraulic cylinders containing micron-sized magnetically polarizable particles suspended
within a fluid. With a strong magnetic field, the particles polarize and offer an increased
resistance to flow. By varying the magnetic field, the mechanical behavior of the MRD can
be modulated: MR fluids can be changed from a viscous fluid to a yielding solid within mil-
lisecond and the resulting damping force can be considerably large with a low-power sup-
ply. Electro-rheological Dampers (ERDs) are the electric analogue ones. ER fluid contains
micro-sized dielectric particles and its behavior can be controlled by subjecting the fluid to
an electric field. Magneto-rheological fluids are an alternative solution to Electro rheologi-
cal ones when very compact devices are needed, as the rheological behavior is similar to the
ER-fluids but with higher yield stress. In the case of steady fully developed flow, the shear
resistance of MR/ER fluids may be modeled as having a friction component augmented by
a Newtonian viscosity component. MR/ER Dampers can be placed between the chevron-
brace and the rigid diaphragm or beam. The most common seismic isolation interventions
performed in Italy, is reported in a bridge database analysis of 160 bridges performed by
the author, presented in the following Figure 7. The seismic isolation has been successfully
adopted in several countries mainly using the elastomeric and sliding bearings. A total of
255 isolated bridges had been built (Priestley et al., 1996). Lead-rubber bearings were used
in a very large variety of bridges, showing clearly that this is the preferred choice, except in
some cases where dampers are coupled with traditional sliding support.
Seismic isolation and energy dissipation devices can also be used in retrofitting the bridges
(Buckle and Mayes 1989) and (Penzien 2001). These are used to replace vulnerable support
bearings by which the bridges system’s flexibility can be increased considerably, lengthen-
ing the fundamental periods resulting in reduced horizontal seismic forces but increasing

NYBRIDGE_Book.indb 224 7/17/2013 8:43:50 AM

 Retrofit procedures of bridge networks against seismic actions in Italy 225

Figure 7. Bridge and viaducts with seismic devices: HD-hysteretic damper; LRB lead rubber bearing;
RB rubber bearing; SHD slider with hysteretic damper; STU-shock transmission unit; VD-viscous
damper; EL-Elastic; EP-elastic plastic; OL-hydraulic dampers, EP equivalent; OP-hydraulic damper;
SL-sliding support; ST-shock transmitter associated with SL RB-rubber bearings; LRB-lead rubber
bearings; hydraulic ST-hydraulic shock transmitter. Other abbreviations:RC-reinforced concrete; PCB-
pre stressed concrete beams; UB-universal beam; PSC-pre-stressed concrete.

superstructure displacements. As reported by (Kunde 2001) and (Kelly et al., 1984) studied
the retrofit of an existing freeway overpass undertaken to improve earthquake performance
by the installation of lead-rubber bearings between the superstructure and the supporting
columns. Before the retrofit, the columns of the bridge were capable of resisting approxi-
mately one-quarter of the design site earthquake but the lead-rubber bearings are shown to
improve this performance; (Parducci and Mezzi 1992) discussed great number of highway
bridges in Italy provided with seismic isolating devices. Some applications of seismic isola-
tion in Italy are given in Table 3.

5.3 Interventions on railway bridges

A special case of seismic retrofit is related to railway bridges, where the serviceability is not
related to the structural damage, but principally to track displacements due to vibrations and
relative displacements in dynamically excited bridges: these are very dangerous problems, as
also very small displacement of the track could lead to tilting moment in trains, and eventu-
ally derailment. If compared to the way and highway bridges, the seismic assessment of a
railway bridge is much more complicated by the analysis of the train-track-bridge complex,
considering the non-linear behavior of piers, bearings, ballast, trains etc.

6 CURRENT RESEARCH AND REAL SCALE TESTING EXPERIENCE

Not satisfactory performance of older reinforced concrete bridges in major earthquakes is

because of the fact that they were not designed for earthquake effects; as such, the design

NYBRIDGE_Book.indb 225 7/17/2013 8:43:50 AM

 226 Seismic analysis & design

Table 3. Applications of seismic isolation in Italy.

Bridge Description Other data

Carboj viaduct Structural intervention on the Carboj

viaduct (longitudinal non-linear viscous
dampers; transversal rubbearing).
Changed the static scheme, from
isostatic to iperstatic, continuous deck
after the intervention. The devices are
activated only at the seismic event
(OP series); in correspondence of
the deck interruption, they allow thermal
displacement (OTP series) (Castellano
and Infanti, 2007).
Rio Gozzo and Structural intervention on the Rio Gozzo
San Lorenzo (108 EL) and San Lorenzo (shear keys)
viaducts. viaducts for the seismic isolation
(Marioni, 2011).

Agrifoglio Structural intervention on the Agrifoglio

viaduct. viaduct for the seismic isolation
(elasto-plastic sliding isolators, with
U-shaped steel plates as energy dissipating
elements). To allow thermal variations,
the elasto-plastic devices are coupled
with the shock transmitters (Dolce, 2001).
SS647 Biferno Structural intervention on the SS647
viaduct Biferno viaduct (longitudinal non-linear
viscous dampers; transversal hysteretic
steel dampers). Changed the static scheme,
from isostatic to iperstatic, continuous
deck after the intervention (Castellano,
Infanti 2007).

details do not meet current seismic design criteria, and are vulnerable to seismic deformation
demands. In the 1994 Northridge earthquake, the primary cause of collapse and major dam-
age was summarized as insufficient ductility in the bridge structural frames (Zelinski 1999). It
was reported that in the 1994 Northridge earthquake, as long as the steel and fiber reinforced
polymer (FRP) composite shells confined the concrete and longitudinal reinforcement, the
columns of rehabilitated bridges could handle large displacements beyond yield, while car-
rying the dead load (Yashinsky 1998). In the 1995 Hyogo-ken Nanbu earthquake, columns
exhibited shear failure because the volumetric steel ratio of the ties was low, the tie bars were
not arranged effectively for confinement or shear resistance, and a plastic hinge could not
form (Taylor 1999). For tested RC bridge piers in which posttensioning of the beam was
combined with glass FRP wraps for the beam and column; the performance of the rehabili-
tated piers was satisfactory (Sexsmith et al., 1997). The feasibility and advantages of applying
steel jacket rehabilitation measures developed for single-column pier bridges to multicolumn
pier bridges were investigated analytically (Zhang et al., 1999). It is clear that seismic rehabili-
tation is a politically convenient and technically effective and practical means to meet current
seismic performance objectives.

NYBRIDGE_Book.indb 226 7/17/2013 8:43:51 AM

 Retrofit procedures of bridge networks against seismic actions in Italy 227

A recent extensive research has been performed in Italy by a group of Universities (Pinto
et al., 2011). Half scaled bridge piers were designed with an insufficient overlapping length
of the longitudinal bars across the critical zone that should lead to an early loss of the lateral
strength due to bar slippage. The built specimens have the following characteristics: hollow-
core rectangular cross-section with external dimensions 800 × 1500 mm and wall thickness
of 150 mm; pier height of 6 m (aspect ratio equals 4); longitudinal reinforcement: 80 ϕ 10
(ρL = 1.05%) with an overlapping length equal to 20 diameters (200 mm) at the base of
the pier; transversal reinforcement: stirrups Φ6/150 mm (ρV = 0.38%); axial load equal to
1000 kN (ν = 4.3%) or 2000 kN (ν = 8.6%); concrete Rck400; steel FeB44 K. The new stress
path created using longitudinal FRP strips applied to the overlapping region is expected to
cause the plastic hinge shift upwards where the longitudinal steel is well anchored allowing
for an efficient energy dissipation. During the design phase different possible solutions have
been considered concerning the retrofit materials (carbon, aramid or glass FRP), the retrofit
geometry (width and length of the region to be retrofitted), the techniques to be used for
the anchoring of the FRP strips to the foundation. This was possible employing a numerical
FE model developed to predict/reproduce the tests results. Regarding the materials, the final
choice was to use carbon FRP (C-FRP): the analyses indicated that this material is the only
one able to sustain the acting tension forces. Too many FRP layers would have been needed
to carry the same force using glass or aramid fibers, affecting the effectiveness of the retrofit
intervention. For what concern the geometry of the retrofit intervention, the final solution
was to apply longitudinally two C-FRP layers on the four sides of the specimen. As far as
the exploitation of the material strength is concerned this choice appears to be questionable
since the fibers applied to the pier sides parallel to the imposed motion will not have the
same stress as those on the other two sides, but the adopted solution seemed to be the only
possibility to assure the maximum stress diffusion across the pier section. It is worth men-
tioning that even though anchoring 1500 kN force to the foundation of the scaled specimen
would have been probably feasible using a steel collar fixed to the foundation with some high-
strength steel bars taking advantage of the deep foundation of the specimen, moving back
to real structures the anchoring to the foundation of the tensile force induced in the FRP by
a seismic excitation would have been much more difficult, if not unfeasible. Spreading the
tensile force on the four sides of the pier, the anchoring is clearly easier. Between different
possibilities initially considered to anchor such force, final choice was to use an anchoring
system realized with FRP too. The idea was to employ aramid connectors, normally used to
transfer shear stresses. If this solution will be found to be effective, as it seems from its design,
multiple advantaged will arise both on the economic and technologic sides. Due to external
constraints only two piers have been tested within the duration of the project, those without
the FRP retrofit in the lap-splice region. The tests confirmed that, as expected, lap-splice
with an overlapping length equal to 20 times the diameter of the spliced bars is insufficient
to assure the anchoring of the bars. The tests also underlined that the effectiveness of the
lap-splice decreases while the axial load increase: that is because of the higher stresses and
damages (such as partial concrete spalling) in the overlapping region. Large-scale tests on
framed piers have been also undertaken. This typology, characteristic of many old viaducts
of the Italian highway system, has been chosen for its high seismic vulnerability.

7 CONCLUSIONS

Existing bridges represent a relevant category of the international cultural heritage, being the
evidence of the modern industrial technology, in particular the Italian bridge network built in
the nineteenth century, both in the railway and in the roadway lines. Many of these structures
require particular retrofitting interventions in relation to the seismic issue. A comprehensive
overview on the assessment of bridges is presented in order to evaluate the structural integrity

NYBRIDGE_Book.indb 227 7/17/2013 8:43:52 AM

 228 Seismic analysis & design

of historical and deteriorated bridges, incorporating code aspects, analytical, mechanical and
structural topics. Concerning available codes, EC8 has been presented especially with refer-
ence to the lack in this code of a specific section dealing with existing bridges. Structural
seismic isolation/dissipation has been deepen in detail, with a comprehensive overview on
retrofitting interventions on 160 bridges developed by the author.

REFERENCES

Buckle, I.G. and Mayes, R.L. 1989. “The application of seismic isolation to bridges”, Structures Con-
gress ‘89: Seismic Engineering: Research and Practice, ASCE, NY, 633–642.
Castellano M.G., Infanti S. 2007. “Seismic retrofit of bridge and viaducts with seismic isolation”, NT
1700, Fip Industrial, Italy.
CEN. 2005a. “Eurocode 8 Part 2: Seismic design of bridges” European Committee for Standardization,
Brussels, Belgium.
CEN. 2005b. “Eurocode 8 Part 3: Assessment and retrofitting of existing structures” European
Committee.
DM 2008. “Nuove norme tecniche per le costruzioni” Decreto Ministeriale del Ministero delle Infras-
trutture 14/1/2008.
Dolce M. 2001. Remarkable design examples concerning recent application of innovative anti-seismic
tecniques to bridge and viaducts in Europe. 7th International seminar on seismic isolation, passive
energy dissipation and active control of vibrations of structures, Assisi, Italy, October 2–5, 2001.
Keady, K.I., Alameddine, F., Sardo, T.E. 2000. “Seismic Retrofit Technology”. Bridge Engineering
Handbook. Ed. Wai-Fah Chen and Lian Duan. Boca Raton: CRC Press, 2000.
Kelly, T.E., Jones, L.R. and Mayes, R.L. 1984. “Seismic retrofit of bridges utilizing ductile base isola-
tion concepts”, Proc. of the 8th World Conf. on Earthquake Engineering, New Jersey, California,
Vol. 1, 651–658.
Kunde M.C. 2001. Seismic behavior of isolated bridges: A-state-of-the-art review. Electronic Journal of
Structural Engineering, 3.
Marioni P. 2011. I viadotti Rio Gozzo e San Lorenzo sull’Autostrada dei Fiori. Strade e autostrade,
4/2011, ed. Edi-Cem, Milano.
Mayes, R.L., Choudhury, D., Crooks, R.S., Jones, D.M., and Knight, R.P. 1994. Seismic isolation ret-
rofit of existing bridges, ASCE, Reston, Va.
Mayes, R.L., Jones, L.R., Kelly, T.E., and Button, M.R. 1984. Base isolation concepts for seismic bridge
retrofit, ASCE, San Francisco.
Parducci, A. and Mezzi, M. 1992. “Seismic isolation of bridges in Italy”, Bulletin of the New Zealand
National Society for Earthquake Engineering, Vol. 25, 193–202.
Parducci A. 2007. Nuovi sistemi per la protezione sismica dei ponti: il viadotto Coltano. Corridoio V,
EDA, n. 2/2007, by Alessio Pipinato, Il Prato Edition, Padova, Italy.
Penzien, J. 2001. “Earthquake engineering for transportation structures-Past, Present and Future”,
Earthquake Spectra, Vol. 17, 1–34.
Pinto P.E. et al. 2011. Seismic assessment and retrofit of existing bridges, Proceedings of The state of
Earthquake Engineering Research in Italy, 111–140, © 2009 Doppiavoce, Napoli, Italy.
Priestley, M.J.N., Seible, F. and Calvi, G.M. 1996. “Seismic design and retrofit of bridges”, John Wiley
and Sons, New York.
Wendichansky, D.A., Chen, S.S., and Mander, J.B. 1995. “In-situ performance of rubber bearing retro-
fits.” National Seismic Conf. on Bridges and Highways, Progress in Research and Practice, FHWA,
San Diego, Calif.
Wright T., DesRoches R., Padgett E.J. 2011. Bridge Seismic Retrofitting Practices in the Central and
Southeastern United States. Journal of Bridge Engineering, Vol. 16, No. 1.

WEB REFERENCES

• Autostrade per L’italia. 2013. Official web-site, www.autostrade.it

• Protezione Civile. 2013. Official web-site, www.protezionecivile.it
• RFI. 2013. Rete Ferroviaria Italiana. Official web-site, www.rfi.it

NYBRIDGE_Book.indb 228 7/17/2013 8:43:52 AM

 5 Bridge monitoring

NYBRIDGE_Book.indb 229 7/17/2013 8:43:52 AM

This page intentionally left blank
 Chapter 21

Investigation of feasibility of deterministic, theoretical predictive

models underpinning visual condition monitoring outcomes

A. Sagar, S. de Silva & S. Setunge

RMIT University, Melbourne, Australia

ABSTRACT: This paper focuses on the viability of theoretical predictive models in

recognizing the most appropriate time for maintenance intervention in critical structural ele-
ments of bridge assets. The key objective is in providing deterministic, theoretical under-
pinning to visually inspected condition data. It is proposed to achieve that objective by
conducting a real-time theoretical simulation of flexural stresses, taking into account the age
dependent behavior of shrinkage, creep (under sustained loads) and imposed loads. In doing
so, crack widths can be theoretically estimated over the life cycle of the bridge girders. The
paper proposes a Theoretical Age of Intervention (TAoI), based on crack depth reaching a
critical value, which may not be otherwise conceivable to human eye. The paper demonstrates
that such a theoretically predicted time of intervention, when calibrated using the condition
monitoring data, would provide asset owners with a useful tool for maintenance planning.

1 BACKGROUND

1.1 Introduction
Bridges are considered critical links in transportation infrastructure and provide significant
contribution towards the growth of a nation. In US, transportation sector accounts for 20% of
gross nation product and almost 800,000 jobs (Qader et al., 2003). First bridge in Australia was
built around 1788 over Tank Stream, in New South Wales (NSW), a state of Australia. It is evi-
dent that most of the developed metropolitans’ road bridges are aging. As per Stewart and Val
(1999), 50% of bridges are 50 years old in Australia and United States. Most of these bridges
were constructed in the US in between 1930s and 1970s (Estes and Frangopol 2003). Their
expected maintenance cost was estimated around US$10.6 billion per year in next ten years
(Estes and Frangopol 2003). Similarly, an estimated sum of US$300 million was required to
maintain the deteriorating bridges in NSW (Stewart 2001). In addition, VicRoads (a state road
authority of Australia, for Victoria state) and Victorian councils spent over AUD$30 million
dollars each, during 2009 and 2011 on bridge maintenance and monitoring (Report 2011).

1.2 Comparison of current inspection criteria in Australia and USA

Both USA (Graybeal et al., 2003) and Australia (VicRoads 2011) adopt condition monitor-
ing criteria of bridges based on visual inspections. Types of inspection practiced in Australia
and USA have been studied and discussed comparatively under Tables 1–2. American inspec-
tion guidelines are described under AASHTO (2011), whereas, Australian inspection proce-
dures are set by the respective state road authorities. Henceforth, inspection types described
in VicRoads manual have been discussed as an Australian example.
Above mentioned inspections are performed to assess the condition of superstruc-
tures and substructures of bridges. According to AASHTO (2011) and VicRoads (2011),
superstructures include steel and reinforced concrete beams and girders and box sections,

231

NYBRIDGE_Book.indb 231 7/17/2013 8:43:52 AM

 232 Bridge monitoring

Table 1. Current inspection types in Victoria, Australia (VicRoads 2011).

Types of inspection Frequency

Level 1 – Routine maintenance inspection: A brief inspection Twice yearly frequency with a
of all structural components and associated infrastructure to maximum interval of 6 months
assess and report any visible signs of damage and distress.
Level 2 – Road structures condition inspection: Managed 2 to 5 year frequency depending
on a statewide basis to provide a consistent visual assessment upon condition of components and
of the condition of each structure. estimated rates of deterioration.
Level 3 – Detailed engineering investigation: Such an Initiated when structures are initially
investigation may be one of a number of field and analytical reported in Poor Condition in Level
investigations to provide detailed assessment of the structural 2 inspection.
component.

Table 2. Current inspection types in America (AASHTO 2011).

Types of Inspection Frequency

Initial inspections: Initial inspection of a bridge, as it becomes a part

of the bridge file.
Routine inspections: Regularly scheduled inspections comprising of
observations, measurements or both, needed to determine the physical
and functional condition of bridge, to identify any changes from “initial”
or previously recorded conditions. Each bridge should be
Damage inspections: An unscheduled inspection to assess structural inspected at regular
damage resulting from environmental factors or human actions. intervals not to exceed
In-depth inspections: It is a close-up, hands-on inspection of one or more 24 months or at longer
members above or below the water level to identify any deficiencies not intervals for certain
readily detectable using Routine Inspections procedure. bridges where such
action is justified by
Fracture critical inspections: Inspection of steel bridges should include past reports and
the identification of fracture-critical members (FCM) and the performance history
development of a plan for inspecting such members. and analysis
Underwater inspections: Underwater inspection involves locating of
deterioration of substructure and undermining, diving to visually
inspect and measure bridge components, or some combination thereof.
Special inspections: Inspection scheduled at the direction of the bridge
owner or the responsible agency.

while substructures consist of abutments, retaining walls and piers. This paper limits its dis-
cussion to reinforced concrete (R/C) bridge girders.
Typical condition monitoring procedure includes a set of inspections as described in above
Tables, where each inspection is dependent on the outcome of preceding inspections except pri-
mary inspections, i.e. Level 1 and Initial inspections. Since a structure is not expected to experi-
ence major degradations under serviceability conditions in early service life, routine and Level
2 inspections are given higher prominence (Phares et al., 2004) than Level 1. But, these inspec-
tions are reckoned as labor-intensive, costly and highly qualitative (Caner et al., 2008; Koh and
Dyke 2007; Phares et al., 2004), where reliability is hard to measure. A limited reliability is also
noticed in most of the visual inspection reports gathered by the research team after consulting
several councils. Visually inspected conditions are documented in qualitative terms (as stated
in Tables 3–4) and serve as a strong evidence for bridge owners to take future maintenance and
monitoring decisions (Alampalli 2010). In addition, this subjective assessment is associated
with high degree of variability due to continuous rotation of inspectors (Qader et al., 2003).

NYBRIDGE_Book.indb 232 7/17/2013 8:43:52 AM

 Investigation of feasibility of deterministic, theoretical predictive models 233

Conditions recording of visual inspections prescribed in respective manuals are done in

two distinctive ways. For example, USA prescribes a condition rating (FHWA 1995), whereas
Australia prescribes condition states, as given in Tables 3–4. FHWA manual provides a brief
discussion on the condition ratings of individual structural components of a bridge. How-
ever, condition states are well delineated for each structural element in Australian practices.
Table 4 illustrates the delineated condition states of a R/C girder.

Table 3. Condition ratings, USA (FHWA 1995).

Condition rating Description

N Not applicable
9 Excellent condition
8 Very good condition: no problems noted
7 Good condition: some minor problems
6 Satisfactory condition: structural elements show minor deterioration
5 Fair condition: all primary structural elements are sound but may have
minor section loss, cracking, spalling, or scour.
4 Poor condition: advanced section loss, deterioration, spalling or scour.
3 Serious condition: loss of section, deterioration, spalling, or scour have
seriously affected primary structural components. Local failures are
possible. Fatigue cracks in steel or shear cracks in concrete may be present.
2 Critical condition: advanced deterioration of primary structural elements.
Fatigue cracks in steel or shear cracks in concrete may be present or scour
may have removed substructure support. Unless closely monitored it may
be necessary to close the bridge until corrective action is taken.
1 Immediate failure condition: major deterioration or section loss present in
critical structural components, or obvious vertical or horizontal movement
affecting structural stability. Bridge is closed to traffic but corrective action
may put bridge back in light service.
0 Failed condition: out of service: beyond corrective action

Table 4. Condition state for R/C girder, Victorian bridges, Australia (VicRoads 2011).

Condition state Description

1 The beams are in good condition with only very minor fine cracking due to
corroding reinforcement, shrinkage or lack of curing.
2 Flexural cracking and cracking due to reinforcement corrosion is fine though
no rust staining is visible in the cracks. A few minor spalls may be present. Fine
cracking may have occurred at bearing areas of the beam.
3 Flexural cracking has increased with cracks in the medium size. Cracking
due to reinforcement corrosion is medium and a number of spalls may
have occurred, and/or heavy rust staining is present at the cracks. Where
reinforcement is closer together, some delamination of the concrete may have
occurred. Loss of section of any corroding reinforcement is less than 20%.
Medium cracking may have occurred at the ends of the beam affecting the
bearing area of unit.
4 Corrosion of reinforcement is well advanced with loss of bar section greater
than 20%. Flexural cracking in the member may be heavy with the beam
noticeably deflecting under load. Severe spalling may have occurred due to
corroding reinforcement or at the ends of the beam at the bearing areas.

NYBRIDGE_Book.indb 233 7/17/2013 8:43:52 AM

 234 Bridge monitoring

The extent of each condition state affecting reinforced concrete girder shall be quantified
as a percentage of the whole component, i.e. the percentage in each condition state (1, 2, 3
and 4) must add up to 100% of whole component observed at site (VicRoads 2011).
Both manuals advise to record the location of cracks, their widths and primary cause of
cracking in order to determine the condition of bridge girders (AASHTO 2011). Generally,
cracks can be categorized as flexural and shear cracks. This indicates that the recording of
characteristic crack patterns is one of the integral criteria to determine the condition of
a girder. In order to make visual condition monitoring more effective, VicRoads manual
suggests capturing of the photographs of the deteriorated element, such as, cracks appear-
ing on girders. These photographs are then compared with reference photos, provided in
manual, to determine the condition of element. As per Phares et al. (2004), maintaining an
inventory of photographs also provides a strong basis in determining the rate of deteriora-
tion and describing the full condition of bridge components in conjunction with visually
recorded condition ratings. On the other hand, Jahanshahi and Masri (2011) proposes that,
in reality a constant camera object distance cannot be maintained under the inspection.
Therefore, it will be unreliable to cross check the photographic evidences with reference
images because, the focal length and resolution of those images is unknown too.
According to Hearn and Shim (1998), condition states and ratings indicate the severity but
not the extent of damage.

1.3 Brief description of paper

This paper demonstrates a methodology using a theoretical approach, which uses ‘as built’
conditions of the bridge girders to quantitatively determine the condition at any given point
of time. Proposed methodology introduces a new tool for bridge owners, which can theo-
retically predict the age of intervention for maintenance purposes, supplementing and cross
examining the visual inspection outcomes. Crack widths are theoretically predicted as a
function of age, taking into account flexure under varying loads, shrinkage and creep. In
addition, paper also demonstrates the importance of deterministic approach in informed
and proactive decision-making. This hypothesis is illustrated through a bridge currently in-
service, and has been in service since 1964. This 48 year old bridge is located 300 km North-
West of Melbourne, Australia, represents over 60% of the bridge assets in the region. Further
information is discussed under section 3.2.

2 HYPOTHESIS

2.1 Theoretical age of intervention (TAoI)

Theoretical Age of Intervention (TAoI) is a term and mechanism that the authors introduce
in this paper. TAoI is defined here as, theoretically identifiable age of intervention which pre-
dicts cracks within a girder reaching a structurally critical stage, which may not be necessarily
visible to human eye. TAoI would provide a reasonable lead time for the asset manager as
a sufficient buffer for proactive and informed renewal forecasting. Here, structurally critical
stage means, cracks are wide enough to set corrosion setting, triggering a vicious cycle of
stiffness and strength degradation, without timely intervention. Visual condition monitoring
data can be supplant and examined by this deterministic theoretical simulation. As men-
tioned before, inspection of the soffit of bridge girders is a challenging task especially from
a distance. However, as per Graybeal et al. (2001), sometimes locations of cracks and severe
deteriorations do not get recorded due to physical constrains prevails at the site, which is also
observed by the authors on the several site inspections. Therefore, to verify any subjectiveness
of data recorded, the proposed methodology can be used to examine location and severity of
cracks. Concrete cracks with age due to shrinkage and creep effects and flexural stresses. Once
theoretically predicted crack width value crosses the threshold figure, i.e. structurally critical
crack width, bridge can be tagged for in-detail visual inspection or cost effective protective

NYBRIDGE_Book.indb 234 7/17/2013 8:43:52 AM

 Investigation of feasibility of deterministic, theoretical predictive models 235

measures, to avoid any future deterioration. EuroCode2 (Mosley et al., 2007) and Gilbert
and Ranzi (2011) suggest 0.3 mm as the critical crack width. Whereas, American Concrete
Institute (ACI 2001) specify 0.41 mm as critical crack width.
Recommended mechanism is expected to provide assistance to asset owners in precaution-
ary and cost effective amelioration techniques, such as, protective coatings. It empowers the
owners to predict the crack width of bridge girder at time economically, without completely
relying on the visual condition inspectors to detect them. In addition, it helps in identifying the
present condition of aging structure in-service, whose preceding inspection inventory is not
well maintained. Methodology to find Theoretical Age of Intervention is demonstrated using
theoretical time-series behaviour of a rectangular R/C section, representing a bridge girder.

2.2 Age dependent influences & cracking

Assessment of the theoretical time-dependent cracking behaviour of an R/C bridge girder
is a complex phenomenon and has always been the foremost concern for researchers and
designers in application. Flexural cracking within the tension zone of a beam is obtained
when serviceability is checked during design (Allam et al., 2012; Chowdhury and Loo 2001).
Whereas, age dependent cracking due to creep and shrinkage, are unlikely to get computed
during in-service life span. Most condition monitoring reports indicate that cracks usually
appear at the one third of span.
Cracks in R/C developed from an early age, due to shrinkage and handling during construc-
tion, hardly visible to eye. But, over the time as concrete matures and loads imposed on the
girders, crack widths increase and become visible to human eye. Such cracking takes place due
to sustained applied dead loads, cyclic traffic loading and creep and shrinkage effects. Moreo-
ver, creep and shrinkage are challenging processes, which contain various uncertainties origi-
nating due to variations in material characteristics (Yang 2007). Reduction in girder strength,
change in environmental conditions and growth in traffic volumes (Qader et al., 2003), also
contributes to crack propagation at certain age of the girder. Shrinkage is constrained by
bonded reinforcements due to which tensile stresses on concrete develops with age and result
in time-dependent cracking and widening of initial cracks (Gilbert 2001). Crack widening can
result in loss of load-bearing capacity (Dawood and Marzouk 2010), surface deterioration
(Chowdhury and Loo 2001) and quick diffusion of corrosion factors (Allam et al., 2012).
Besides above influencing parameters, regular increment in live load, above designed con-
ditions, result in development of residual tensile steel stress which in effect, increases crack
widths. Truck load, which is considered as design live load during designing of structures,
has increased significantly with time (Qader et al., 2003), as shown in Figure 1. In Aus-

Figure 1. Variation in design live load over last 60 years in Australia.

NYBRIDGE_Book.indb 235 7/17/2013 8:43:52 AM

 236 Bridge monitoring

tralia, during 1960’s live load of 9.3 kN/m, H20-S16 truck load as per NAASRA (1958) was
used for design purposes. However it has increased to 24 kN/m, SM1600 truck load as per
AS5100.2 (2010) and is currently used for designing purposes.
Theoretical consequences of above mentioned time influential factors haven’t been studied
on real girders and is yet in scarcity. This paper demonstrates a methodology considering
simply supported and singly R/C beam as a real girder exposed to time influential factors.

3 METHODOLOGY

3.1 Time series behavior of R/C beams and age dependent crack widths
As already mentioned, cracking in R/C members due to shrinkage, creep and flexural stresses
are time dependent. Current body of knowledge in this field is well advanced. Several empiri-
cally and theoretically developed methods can predict the time dependent behaviour of crack-
ing and reinforcement detailing for crack controlling (Allam et al., 2012; Chowdhury and Loo
2001; Dawood and Marzouk 2010; Frosch 1999; Gilbert and Nejadi 2004; Makhlouf and Mal-
has 1996). Majority of formulae proposed in assessing crack widths are empirically based and
have been performed under controlled environments. An experimental study of flexural crack-
ing in reinforced concrete members under sustained loads is conducted by Gilbert and Nejadi
(2004) to understand the behaviour of concrete cracking. Formula proposed by (Gilbert and
Ranzi 2011) is adopted in this paper to predict the crack widths. The reasons for selecting
Gilbert formula in this work is based on the fact that it is more applicable under Australian
conditions and the formula takes into account, not only the flexural stresses induced by variable
imposed loading, but also the stresses due to shrinkage and creep. It is also important to note
that empirical formulae are sensitive to unit conversion; hence this paper uses the SI units.

3.2 Bridge case study

A bridge constructed in 1964, is chosen from one of the rural city council in Victoria, to
demonstrate the methodology. The bridge girders are considered typical to that era. Also
material properties, structural configurations and construction methods are considered to
provide a representation of major proportion of existing bridge assets. As per the documents
provided by the council, simply supported reinforced concrete I beams, as shown in Fig-
ure 2, are used in this bridge to support the traffic loads. Bridge was designed in 1960s, using
the working stress method. Original material properties of I girder, are given in Table 5. To
simplify the computations, “I” girder cross-section is transformed to a rectangular section,
while maintaining effective depth and the second moment of area, identical to the existing
girder. Table 6 shows the bridge dimensions and structural configuration. Information of
transformed equivalent rectangular bridge girder is provided in Figure 2. Figures 5–6 show
the current condition of the selected bridge.

3.3 Constitutive equations

An empirical expression proposed by Gilbert for calculating crack width is shown below. This
expression is used for a given flexural member.

*⎡ ⎤
T τ b s*
(w )
*
soffit
=
kcov
cov eer s
Es ⎣ Ast
⎢ −
db
( ne * tc ) ( ssh * E s )⎥ (1)
⎦

where, (w*)soffit = final maximum crack width at the member soffit, kcover = term to account for
the dependence of crack width on the clear concrete cover, c and can be calculated from the
following equation

NYBRIDGE_Book.indb 236 7/17/2013 8:43:53 AM

 Investigation of feasibility of deterministic, theoretical predictive models 237

Figure 2. Cross-section view of redesigned R/C rectangular beam (all dimensions are in mm).

Table 5. Material properties of bridge girder.

Material properties 1960’s

Concrete compressive strength, f ’c (MPa) 25

Elastic modulus of steel, Es (MPa) 200,000
Yield strength of steel, fsy (MPa) 260 ‘Y Bars’

Table 6. Bridge dimensions.

Year of construction 1964

Total length 55.8 m

Span 5
Overall width 8.1 m
Traffic width 7.1 m

⎛ D dn ⎞ ⎛ 5c ⎞
kcov er = ⎜ ⎟ *⎜ ⎟ (2)
⎝ d d n ⎠ ⎝ ( D d n ) − 2 d b⎠

where, D = overall depth of the beam, d = depth to resultant force in tensile steel, dn = depth
to neutral axis in a cross-section, measured from extreme compressive fiber and db = nominal
diameter of a bar.
T = resultant tensile force estimated using Ast = area of steel in tensile zone and Ms = in
service bending moment. τb = bond shear stress between the steel and the surrounding tensile
concrete and ρtc = reinforcement ratio of the tension chord and may be taken as
Asst
ρtc = (3)
Acct

Act = area of concrete between cracks in tension chord. εsh = shrinkage strain which may
be computed using clause 3.1.7 of AS3600 (2009) and is a time dependent value. ne = Es/Ee
effective modular ratio

NYBRIDGE_Book.indb 237 7/17/2013 8:43:54 AM

 238 Bridge monitoring

Ec
Ee = (4)
( + φ (t,τ )))

Ec = elastic modulus of concrete and φ(t,τ) = creep coefficient. Both parameters are time
dependent and can be calculated using clause 3.1.2 and 3.1.8 of AS3600 (2009), respectively.

3.4 Results
Crack widths calculated based on equation (1), usually occur within tensile zone of the girder
and move towards neutral axis as bridge ages. Primary cracks, cracks developed at initial stage,
have a tendency of opening and closing due to variability in loading and plasticity of con-
crete. But, after certain time, cracks start widening because of many factors such as, gradual
reduction in plasticity and applied load exceeding permissible cracking load. Such a behavior
of concrete has been highlighted in Figure 3. Graph clearly illustrates age dependent crack
widths. In Figure 3, early age micro cracking phase is ignored as macro cracking takes over in
later years. It exceeds the critical crack width value, i.e. 0.3 mm on 17,000 days or 46.5 years.
Hence, 46 years can be considered as the TAoI, which is depicted in Figure 3 by an arrow.
Figure 3 shows the theoretical time dependent crack width behaviour of bridge girder,
where shrinkage, creep coefficient and effective modular ratio acted as a time dependent vari-
ables. Shrinkage was a function of concrete compressive strength, fc,’ hypothetical thickness
of beam, th and exposed environment. Creep coefficient was a function of initial steel stress
and elastic strain. Both shrinkage and creep increased with time and resulted in the incre-
ment of crack widths with age. Also, crack width computation of an R/C section is directly
proportional to the tensile steel stress (as shown in Figure 4) and concrete cover. Concrete
cover module is considered extremely significant for crack width calculations. As given in
equation (1), crack width is directly proportional to the concrete cover thickness.
Figure 4 gives the stress variation within tensile reinforcement of the girder over a 60 year
period. Steel stress provided in Figure 4 has also included additional stresses resulting from
increasing traffic load, as given Figure 1. It is interesting to note that crack widths indicated
in Figure 4 are under Serviceability Limit State (SLS) of loads.
In essence time dependency of the crack width is therefore proportional to and influenced by;
• Design loads induced flexural stresses.
• Shrinkage induced steel stress.
• Creep induced steel stress.

Figure 3. Crack widths with time under Serviceability Limit State (SLS).

NYBRIDGE_Book.indb 238 7/17/2013 8:43:56 AM

 Investigation of feasibility of deterministic, theoretical predictive models 239

Figure 4. Crack widths versus varying in steel stress.

Another contributing variable is the change in elastic modulus of concrete Ec over life span.
It was noticed that the value of Ec increased at a higher rate during initial stages whereas,
reduced afterwards. As per elastic modulus formula mentioned in AS3600 (2009), Ec value
varies depending on the type of cement is used.
It is important to note, results given in Figures 3–4 bears following assumptions
• Constant concrete compressive strength, f ’c, throughout life.
• fsy is taken as 260 MPa representing 1960 applications.
• Concrete cover is kept constant, i.e. 30 mm.
Effects of varying concrete cover are also scheduled to be included in future work along with
area of steel and concrete compressive strength. A time dependent reliability approach can also
be developed by incorporating variability of material properties and environmental conditions.
Based on theoretically predicted outcomes given in Figures 3–4, it can be effectively concluded
that girder exceeds the critical crack width value of 0.3 mm after 17,000 days or 46.57 years. As
a result of above quantitative theoretical prediction, 46 years of age can be tagged as TAoI for
this particular bridge. At this TAoI, a detailed inspection of the structure is required.

4 VALIDATION OF THEORETICAL APPROACH

We have used two methods to validate the theoretical outcomes of the proposed method.
Firstly, we have compared the theoretically estimated crack widths against the recently
obtained Level 2 condition monitoring inspection reports. Secondly, authors have conducted
an independent inspection by visiting the site with trained field inspectors from the council.
Condition monitoring Level 2 inspection was conducted in August 2012 and was prepared
by experienced bridge inspector based on VicRoads inspection manual specifications. In
the Level 2 inspection report of the bridge includes photographic evidences showing visible
cracks. Authors have also collected further photographic evidence, given in Figure 6, during
the site visit conducted in February 2013. Material properties and structural dimensions of
the bridge girder provided in Tables 5–6 were also clarified in these site visits.

4.1 Validation of proposed theoretical model

Most inspection guidelines put major emphasis on inspecting the soffit of deck due to the
fact the bridge girders are considered structurally critical elements of a bridge, both, in
terms of functionality and safety. Experienced bridge inspectors in these councils agree that

NYBRIDGE_Book.indb 239 7/17/2013 8:43:57 AM

 240 Bridge monitoring

Figure 5. Side view of bridge from the left embankment looking towards right embankment.

Figure 6. Crack widths appearing outside on bridge girders.

60–70% of Level 3 interventions involve bridge girders. Asset managers depend solely on
condition monitoring data, in planning their renewal forecasting and capital works programs.
This approach appears rather reactive, considering budgetary constraints. Priority setting
lacks lead-up time available for action. Moreover, maintenance budgets are very limited and
need careful planning.

NYBRIDGE_Book.indb 240 7/17/2013 8:43:58 AM

 Investigation of feasibility of deterministic, theoretical predictive models 241

Bridge was inspected in 2006 prior to the last inspection in 2012. Condition monitoring
inspection report was thorough, and credit to the inspector, all possible signs of deterioration
were diligently recorded. However, that report was not explicit on the width or extent of the
cracks on the soffit of the girders. As a result, it was reported as condition state 1 with a 100%
condition rating. That means, bridge girders are in very healthy condition with fine cracking.
Refer Table 4 for exact description of stage 1. VicRoads inspection manual classify cracks
between 0.1 to 0.3 mm as fine cracking. Figure 6 below illustrates the photographic evidence
of the cracks taken during authors’ recent site visit, arguably greater than what would be
classified as fine cracking.
We have also established that most flexural cracks were at one-third-span and the shear
cracks were not existent. These flexural cracks were then compared with crack identification
guidelines as given in Figure 7.
From our visual comparison between Figures 6–7, it can be drawn that most of the crack
widths vary between 0.3 to 0.7 mm and bridge girder falls under the category of condition
state 3, where cracking is considered medium size; between 0.3 to 0.7 mm. As per Figure 6,
most of the cracks are emerging at the soffit region in tensile zone and some seem to appear
on the web. Bottom left and right images in Figure 6 are zoomed in snap shots of the girder.
Closer photos were taken to inspect and measure the crack widths. By visually comparing
these cracks with reference crack sizes, their widths seem to be falling around 0.6 mm, which
is almost double the critical crack width value. Most of the cracks are vertical or near vertical
indicating they are flexural cracks due to imposed loads, creep occurred over 48 years and
the shrinkage. It is interesting to note that condition report noted visible signs of corrosion
initiation. These visible signs of corrosion initiation also evidently prove that cracking in
bridge girders has passed its critical limit, setting corrosion in motion. As per the guidelines,
structure is clearly reaching Level 3 conditions. It can be stated with confidence that above
bridge girders have already crossed the critical crack width values, i.e. 0.3 mm, but based on
condition report, asset owner would not have identified that the bridge has passed renewal
phase but may needed to be included in the capital works program. This confronts many of
the small-scale bridge asset owners.
Such subjectivity in data can be supplanted by the proposed quantitative methodology in
a very cost effective manner. As demonstrated in section 3, Figure 3, the proposed predictive
model returns theoretical intervention period, TAoI = 46.5 years based on critical crack width
of 0.3 mm. However, as mentioned visual inspection reports the situation as fine cracking.
The adoption of the proposed method would have provided the asset manager a buffer zone
well over 10 years. Alternatively, a two coating of protective paint could have arrested and
retarded the deterioration process. On the other hand photographic evidence and proposed
quantitative analysis is in agreement that girders demand in-detail structural evaluation, i.e.
Level 3 inspection. This appears to highlight the strength of the proposed approach, that
deterministic, theoretical predictive model can supplant and underpin the visual condition
monitoring data.

Figure 7. Reference Crack Sizes (VicRoads 2011).

NYBRIDGE_Book.indb 241 7/17/2013 8:44:01 AM

 242 Bridge monitoring

As already mentioned, authors strongly recommend using protective coating in tension

zones of R/C bridge girders as a general periodic maintenance practice, similar to buildings,
regardless of age. But, in bridge maintenance this practice is not well recognized. If such
coating is applied when crack widths are between 0.1 to 0.2 mm that could extend life expect-
ancy in a very cost effective way. In our theoretical prediction model (Figure 3), that age is
around 25 years. That could be considered as a proactive preserving measure.
Proposed methodology is easy to use, cost effective tool which has a scientific basis. Pro-
posed procedure can be tailored for many variations of R/C girders. Proposed theoretical
evaluation can certainly be adopted for Steel girders, by replacing time dependent behavior
of creep and shrinkage with time dependent behavior of corrosion. It has been mentioned
above, that, asset owners rely heavily on condition monitoring reports. Such tool can simply,
yet accurately underpin the visually inspected outcomes. In addition, it can also be used to
predict and identify the present conditions of old structures, just by using some basic as
built structural details. This tool can also be used in conjunction with visual condition moni-
toring data to develop a cost effective method to forebode the future conditions of bridge
components.
Theoretical Age of Intervention (TAoI) is an indicator which can be used as a cautionary
sign for bridges.

5 CONCLUSIONS

This paper discusses the validity and application of theoretical predictive modeling of R/C
girders as a tool of bridge asset management, which can underpin the visual condition moni-
toring data. A thorough comparison was drawn between available inspection techniques and
condition rating criterion. An example of Australian and US practices was discussed to illus-
trate different inspection techniques between nations. In relation with condition rating, their
recording patterns were also highlighted. Concept of theoretical age of intervention TAoI
was introduced, with demonstrating the viability and strength of the concept. TAoI was cal-
culated on the basis of time series behavior of concrete. Time dependent crack widths result-
ing from shrinkage and creep effects and flexural stresses were used as the basis in predicting
time of intervention in a typical situation of an aging bridge. Methodology to calculate TAoI
was demonstrated in detail and validated by evaluating a real bridge girder qualitatively as
well as quantitatively. Crack width of 0.3 mm, the threshold of corrosion setting, was con-
sidered as the prime indicator to identify the right time of intervention in bridge monitoring
and maintenance. This feasibility study provides a handy, yet cost effective quantitative tool
to asset owners for informed maintenance and monitoring decisions.

REFERENCES

AASHTO (2011). The manual for bridge evaluation: American association of state highway and trans-
portation officials.
ACI (2001). Control of cracking in concrete structures (ACI 224R-01), ACI manual of concrete prac-
tice, American concrete institute.
Alampalli, S (2010). Special issue on bridge inspection and evaluation. Journal of bridge engineering 15:
349–351.
Allam, S.M, Shoukry, M.S, Rashad, G.E and Hassan, A.S (2012). Crack width evaluation for flexural
RC members. Alexandria engineering journal 51: 211–220.
AS3600 (2009). Australian Standard.
AS5100.2 (2010). Australian standard – bridge design – design loads.
Caner, A, Yanmaz, A.M, Yakut, A, Avsar, O and Yilmaz, T (2008). Service life assessment of existing
highway bridges with no planned regular inspections. Journal of performance of constructed facilities
22: 108–114.

NYBRIDGE_Book.indb 242 7/17/2013 8:44:01 AM

 Investigation of feasibility of deterministic, theoretical predictive models 243

Chowdhury, S.H. and Loo, Y.C (2001). A new formula for prediciton of crack widths in reinforced and
partially prestressed concrete beams. Advances in structural engineering 4(2): 101–110.
Dawood, N and Marzouk, H (2010). An analytical model for crack spacing of thick reinforced concrete
plates. Engineering structures 32: 472–482.
Estes, A.C and Frangopol, D.M (2003). Updating bridge reliability based on bridge management sys-
tems visual inspection results Journal of bridge engineering 8: 374–382.
FHWA (1995). Recording and coding guide for the structure inventory and appraisal of the nation’s
bridges, U.S department of transportation.
Frosch, R.J (1999). Another look at cracking and crack control in reinforced concrete. ACI structural
journal 96(3): 437–442.
Gilbert, R.I (2001). Shrinkage, cracking and deflection—the serviceability of concrete structures. Elec-
tronic journal of structural engineering 1(1): 2–14.
Gilbert, R.I and Nejadi, S (2004). An experimental study of flexural cracking in reinforced concrete
members under sustained loads, University of New South Wales, Sydney, Australia.
Gilbert, R.I and Ranzi, G (2011). Time-dependent behaviour of concrete structures, Oxon, Spon press.
Graybeal, B.A, Phares, B.M, Rolander, D.D, Moore, M and Washer, G (2003). Visual inspection of
highway bridges. Journal of nondestructive evaluation 21(3): 67–83.
Graybeal, B.A, Rolander, D.D, Phares, B.M, Moore, M.E and Washer, G.A (2001). Reliability and
accuracy of in-depth inspections of highway bridges. Transportation research record: Journal of the
transportation research board 1749: 93–99.
Hearn, G and Shim, H.S (1998). Integration of bridge management systems and nondestructive evalu-
ations. Journal of infrastructure systems 4: 49–55.
Jahanshahi, M.R and Masri, S.F (2011). A novel crack detection appraoch for condition assessment of
structures. Computing in civil engineering: 388–395.
Koh, B.H and Dyke, S.J (2007). Structural health monitoring of flexible bridge structures using correla-
tion and sensitivity of modal data. Computers and structures 85: 117–130.
Makhlouf, H.M and Malhas, F.A (1996). The effect of thick concrete cover on the maximum flexural
crack width under service load. ACI structural journal 93(3): 257–265.
Mosley, B, Bungey, J and Hulse, R (2007). Reinforced concrete design to Eurocode 2. NewYork, Palgrave
Macmillan.
NAASRA (1958). National association of Australia state road authorities—bridge design
specification.
Phares, B.M, Washer, G.A, Rolander, D.D, Graybeal, B.A and Moore, M (2004). Routine highway
bridge inspection condition documentation accuracy and reliability. Journal of bridge engineering
9: 403–413.
Qader, I.A, Abudayyeh, O and Kelly, M.E (2003). Analysis of edge-detection techniques for crack iden-
tification in bridges. Journal of computing in civil engineering 17: 255–263.
Report, Victorian Auditor-General’s (2011). Management of road bridges. 18.
Stewart, M.G (2001). Reliability-based assessment of ageing bridges using risk ranking and life cycle
cost decision analyses. Reliability engineering and system safety 74: 263–273.
Stewart, M.G and Val, D.V (1999). Role of load history in reliability-based decision analysis of aging
bridges Journal of structural engineering 125: 776–783.
VicRoads (2011). Road structures inspection manual. Victoria, Australia.
Yang, I.H (2007). Prediction of time-dependent effects in concrete structures using early measurement
data. Engineering structures 29: 2701–2710.

NYBRIDGE_Book.indb 243 7/17/2013 8:44:02 AM

This page intentionally left blank
 Chapter 22

GPR testing of reinforced concrete viaduct beams

D. Beben, W. Anigacz & A. Mordak

Faculty of Civil Engineering, Opole University of Technology, Opole, Poland

ABSTRACT: The paper presents the Ground Penetrating Radar (GPR) tests to determine
the parameters of reinforced concrete (RC) viaduct beams. GPR method is designed for non-
destructive investigations of the shallow subsurface of soil, structural elements, roads and
bridges. The subject of this study is a three-span RC road viaduct located over railway lines,
which is planned to modernization. In relation to the adopted reinforcement method and
the lack of design documentation for the existing viaduct, it was necessary to obtain basic
information on the beams occurring in the object. Determine the geometric parameters of
beams and location of reinforcing bars was the main purpose of the GPR tests. The distances
between particular reinforcing bars and the depth of their position were also determined. Two
selected RC beams were thoroughly tested (the first one located in the central span and the
second in the left span).

1 INTRODUCTION

Many reinforced concrete (RC) bridges require fast and effective repair or strengthening. In
some cases, there is a lack of basic information on the cross-section of the bridge. The choice
of modernization method requires knowledge of the load carrying capacity of the exist-
ing structures, which may be obtained using destructive and/or non-destructive methods of
testing. Taking into consideration the advantages and disadvantages of non-destructive test
methods for bridges presented by McCann & Forde (2001), the Ground Penetrating Radar
(GPR) technique was applied in this research.
GPR is a high-resolution electromagnetic (EM) technique, which originated in the 1970s. It
is designed for non-destructive investigations of the shallow subsurface of soil, structural ele-
ments, roads, and bridges. Annan (2001), Hugenschmidt (2002) state that the non-destructive
assessment of the condition of engineering structures using GPR has lately become quite an
attractive and effective method of testing.
The principle of EM wave scattering is used in the GPR method. The fundamental princi-
ples and theory of GPR operation have evolved over many years, resulting from the develop-
ment of electrical engineering, and geophysical and seismic testing. In favorable conditions,
the GPR technique can provide precise information on the nature of underground facilities,
the depth of their location, the occurrence of structural anomalies such as voids in the testing
object, etc.
The main purpose of the experimental tests using the GPR technique was to determine
the geometric parameters of beams and reinforcing bars within them. The distances between
particular reinforcing bars and the depth of their position were also to be determined. Two
selected RC beams were thoroughly tested (the first one located in the central span and the
second in the left span). The velocity of the wave propagation during the GPR test was esti-
mated based on a preliminary test and literature review.

245

NYBRIDGE_Book.indb 245 7/17/2013 8:44:02 AM

 246 Bridge monitoring

2 SHORT DESCRIPTION OF TESTED VIADUCT

The road viaduct consists of three spans made of simply supported RC beams. Each span
consists of 25 main beams with unknown geometrical and strength parameters. All spans
have the same length of 11.50 m. The total object length is 39.00 m and its width is 13.00 m.
The object intersection angle in relation to the railway line axis under the viaduct is approxi-
mately 70º. The main beams are directly situated on the piers (without bearings) and are
likely anchored to the heads of piers.
The viaduct repair project foresaw the change of the static scheme by adding a new layer of
continuous RC slab on the whole object. This change was planned to permit the movement
of heavy trucks weighing 150 kN during normal service of the viaduct. However, the owner
of the viaduct decided to increase its carrying capacity to 300 kN. For this reason, detailed
testing of existing main beams was necessary. During the repair works, the asphalt layers and
leveling concrete were removed to uncover the upper surfaces of the existing main beams.
This was done to minimize the influence of other elements which could disturb the GPR
measurement. The tested surface was dry, and the atmospheric conditions did not influence
the accuracy of the results.

3 DESCRIPTION OF EQUIPMENT AND METHODOLOGY OF GPR TESTS

3.1 Applied the GPR equipment

Each GPR apparatus consists of a few essential components. The main component of the
whole measuring system is a control unit. The basic tasks of this unit are the following: con-
trol of the antenna, signal receiving and converting (from analog to digital), and cooperation
with a computer. The second component of the GPR configuration is the antenna, which is
especially important from the user’s point of view. GPR antennas consist of suitably prepared
steel plates called dipoles. Antennas differ from one another primarily in the operation fre-
quency. The differences in frequencies also result in variations in antenna dimensions. From a
practical point of view, the antenna’s operation frequency demonstrates physical parameters
of the EM wave. The measurement accuracy and the depth of penetration directly depend on
the operation frequency of the antenna. The frequency range of GPR antennas is from more
than 10 to 2000 MHz. Low frequency antennas are usually large and allow deep penetration
but with low accuracy (resolution). In turn, the highest-frequency antennas (approximately
2000 MHz) are considerably smaller and allow high-accuracy testing, but at shallow depth.
Middle-frequency antennas (100–1000 MHz) are the most universal and are usually applied,
among other things, to the detection of underground infrastructure.

3.2 Test description

Prior to testing, the velocity of EM wave propagation in the analyzed medium should be
determined. This may be done experimentally. The time of EM wave propagation through an
element of known thickness (measured with accuracy of ±5 mm) should be estimated. This
study should be carried out on an unreinforced structural element. In the case of the viaduct,
because of the difficulties in accessing the non-reinforced beam element, the traditional pre-
tests of calibration of the EM wave propagation velocity were not conducted. However, a few
preliminary GPR tests using various wave velocities (v(ω) = 0.08 m/ns, 0.10 m/ns, 0.12 m/ns,
and 0.15 m/ns) were carried out. Generally, concrete is not a homogeneous material and
should be considered as a composite material composed of aggregate, cement paste (often
with some mineral additives, e.g., fly ash), and some air and water (mainly inside the pores).
As it turned out, the best results were achieved for the wave velocity of 0.10 m/ns. Therefore,
the velocity of wave propagation in concrete was adopted from the preliminary test and from
a literature review. McCann & Forde (2001), Drobiec et al. (2010) propose the velocities of

NYBRIDGE_Book.indb 246 7/17/2013 8:44:02 AM

 GPR testing of reinforced concrete viaduct beams 247

the electromagnetic wave propagation in various materials. Additionally, it should be noted

that at the test preparation stage, the beam cross-section and the physical and chemical
parameters of the concrete were unknown.
Two main beams placed in the longitudinal axis of the viaduct were chosen for testing, one
from the central span of the viaduct and the other from the left span (Figure 1). Each of these
beams was a separate quarter of the coordinate system, where the longer side was the T-axis
and the shorter was the L-axis. Parallel to the T-axis, longitudinal scanning (L) was executed,
while parallel to the L-axis, transverse scanning (T) was carried out (Figure 2).
After uncovering the upper part of the beams, it turned out that their width was approxi-
mately 0.45 m and length approximately 11.50 m. It was decided to scan a band of width
0.50 m and length 12.00 m, because it was very hard to accurately define the beginning and
the end of individual beams and there was a lack of information approximately the shape of
the beam cross-sections.
Individual scans were executed in increments of 0.05 m to obtain the best accuracy of
the radargrams. Eleven longitudinal profiles (L1–L11) with length of 12.00 m were gener-
ated for the first measurement. These profiles fully covered beam no. 1. Additionally, two
sections of length 1.00 m were chosen, of which 42 transverse scans (T) were made in incre-
ments of 0.05 m. The selected sections were located between 0.00 and 1.00 m (T12–T32) and
between 6.00 and 7.00 m (T33–T53) along the beam length. Due to the partial exposure of
beam no. 2, the measurements were shortened to a length of approximately 6.00 m. In the
next step, eleven longitudinal profiles of length 6.00 m were executed, once again in incre-
ments of 0.05 m. As for the first beam, two sections of length 1.00 m were chosen (between
0.00 and 1.00 m and between 4.00 and 5.00 m along the beam length). Both sections were
scanned in the transverse direction (T12–T53) in increments of 0.05 m (Figure 2). Generally,
the longitudinal scans (L) were used to find the position of the reinforcement bars (stirrups)
placed in the transverse direction of the beam. In turn, the transverse scans (T) were used to
locate the main longitudinal reinforcement bars and possible anomalies in the beams.

Figure 1. Top view on the viaduct beams (bolded lines mean the tested beams).

Figure 2. Inspection traces on the tested viaduct beams.

NYBRIDGE_Book.indb 247 7/17/2013 8:44:02 AM

 248 Bridge monitoring

4 RESULTS OF THE GPR TESTS AND THEIR ANALYSIS

4.1 RC beam no. 1

The selected results of the radargrams for the longitudinal scanning (L) are presented in
Figure 3. All longitudinal radargrams (L) show that the first level of reinforcing bars (stirrups)
is on a similar depth, i.e. 0.055 m from the upper edge of the beam. Possible differences could
be the result of unevenness of the tested surface (the leveling concrete layer); in these places, the
stirrups are situated a little bit deeper. In order to define finally the position of the stirrups, it
was necessary to analyze the beginning fragments of each radargram in detail. In these places,
the tested beam was fully uncovered and the stirrups were clearly visible on the radargram.
Additionally, the layer at a depth of 0.12 m, which is clearly visible in Figure 3, turned out to
be the upper edge of an internal hole in the beam. In the same figure, the layer at a depth of
approximately 0.40 m is also visible and is likely the bottom edge of the internal hole. Moreover,
all the radargrams show the layer at a depth of 0.56 m, which is the bottom edge of the main
beam. A few anomalies were detected, e.g. reflections from an undefined object (element) are
visible at a depth of 0.30 m (marked with circles in Figure 3). This type of anomaly is likely
connected with the internal hole existing in the beam and could be due to reflections of the EM
waves from the edges or bends in the internal hole. On the basis of the results obtained, the
initial cross-section of the main beam of the viaduct was established (Figure 4).
The distances between the reinforcing bars (stirrups) were estimated based on the part
of the radargram where they were most visible, i.e. on the length of 2 m of each scanning.
In these places, the main beam was fully uncovered (Figure 3). In the remaining part of the
beam, only single stirrups were visible and it was difficult to find the repeated their locations
on the radargrams. The average distance between the stirrups was 0.17 m; the largest distance
was 0.20 m (in the middle part of the beam) and the smallest was 0.13 m (at both ends of the
beam). Figure 5 shows the top view of the stirrup arrangement inside the beam.
As mentioned earlier, transverse scans (T) were also executed. These scans confirmed that
the first level of reinforcing bars and the location of the internal hole were at the same depth.
Moreover, transverse scanning allowed us to estimate the range of the internal hole inside the
beam. Considering the small length of scanning and numerous reflections, the visualization
color was changed in order to obtain improved sharpness. Figure 6 shows an example of a
transverse radargram.
In the transverse scans (T), two reinforcing bars were detected in the upper part of the
beam at a depth of 0.06 m. These rebars existed in the corners of the main beam. Moreover,
the edges of the hole inside the beam were visible on all radargrams, the upper edge at a
depth of 0.12 m and the lower edge at a depth of 0.40 m. The beginning of the hole is visible
approximately 0.18 m from the beginning of the scan, and it ends at 0.37 m of the wide scan.

Figure 3. The processed radargram (L1).

NYBRIDGE_Book.indb 248 7/17/2013 8:44:02 AM

 GPR testing of reinforced concrete viaduct beams 249

Figure 4. The cross-section of the viaduct beam based on the radargrams (longitudinal profiles).

Figure 5. Top view of the beam with marked longitudinal profiles and stirrups.

In the majority of radargrams, reflection on the beginning of the scan is visible at a depth of
0.055 m. These are likely the stirrups situated at a similar depth to the main reinforcing bars.
The bottom level of the main reinforcing bars could not be determined accurately. However,
at a depth of 0.50 m, some reflections of the EM waves were observed, likely due to existing
reinforcement bars.

4.2 RC beam no. 2

The results obtained for beam no. 2 confirm the results obtained for beam no. 1. The distance
between the stirrups was estimated based on radargram L4 (Figure 7), where the fragment
of the rebar group is most visible. The average distance between stirrups was approximately
0.18 m. Larger spacing was detected in the middle part and smaller distances at the ends of the
beam. Besides, the stirrups at a depth of 0.055 m were noted. An anomaly was also detected at
a depth of 0.07 m (falling down). These are likely the upper longitudinal reinforcing bars.
In the case of the second beam, the rebar level and edges of the internal hole were
more visible on transverse radargrams (T). Two reinforcing bars were also detected for
this beam at a depth of approximately 0.06 m. The first was 0.06 m and the second 0.40 m
from the beginning of the scanning area. Moreover, the internal hole at a depth of 0.12 m
is quite well visible. The internal hole reaches a depth of 0.40 m, where the image struc-
ture changes visibly. The second level of rebars is also visible (unfortunately, the number
of rebars cannot be distinguished, likely due to the large number of rebars merging in
the image). The bottom edge of the beam is also clearly visible at a depth of 0.56 m
(Figure 8).

NYBRIDGE_Book.indb 249 7/17/2013 8:44:03 AM

 250 Bridge monitoring

Figure 6. The processed transverse radargram (T22) of beam no. 1.

Figure 7. The L4 radargram of beam no. 2 with well visible stirrups.

4.3 Discussion
Based on the GPR test it was determined that the main upper reinforcing bars were situated at
a depth of approximately 0.06 m from the top edges of beams. The possible differences in their
positions could result from inequalities of the test surface on which the measurement antenna
was moved. Spacing of the main upper longitudinal reinforcing bars was determined quite accu-
rately. After detailed analyses of both considered beams, only two longitudinal reinforcing bars
were detected. The rebars were positioned in both corners—approximately 0.06 m from the
beginning (and end) of tested beams. The distance between rebars ranged from 0.36 to 0.38 m.

NYBRIDGE_Book.indb 250 7/17/2013 8:44:04 AM

 GPR testing of reinforced concrete viaduct beams 251

Figure 8. Processed transverse radargrams: a) T20 and b) T45.

Figure 9. Comparison of cross-sections of the viaduct beams received from: a) the GPR tests and
b) the catalogue of typical beams of Gromnik type.

Additionally, it was discovered that the upper part of transverse reinforcing bars (stirrups)
was positioned at a depth of approximately 0.055 m. Spacing of stirrups was approximately
constant. However, they were more densely positioned at the ends of the beams (0.13 m)
and sparser in the middle part (0.20 m). The main bottom reinforcing bars were situated at a
level of approximately 0.50 m from the top edges of the beams. The numbers and spacing of
rebars could not be accurately determined due to the complex cross-section of the beams.
At the end of the analysis process, the location of the internal hole was identified inside
each of the beams. The top edge of the hole was at a depth of approximately 0.12 m and the
lower edge at approximately 0.40 m. In the transverse direction of the beam, the aperture
starts at approximately 0.18 m and ends at 0.37 m, counting from the beginning of the beam.
After analysis of all GPR results for both beams, it was concluded that Gromnik–type
beams were used to construct the viaduct. Figure 9 shows a comparison between the catalogue

NYBRIDGE_Book.indb 251 7/17/2013 8:44:05 AM

 252 Bridge monitoring

beam of Gromnik type and results of the GPR research. Some differences were detected, for
example in the number of upper reinforcement bars. In order to confirm the number of rebars,
it was decided to uncover the concrete layer in one of the damaged beams. It turned out that
only two longitudinal rebars existed in the top part of the beam. This confirms the necessity of
applying the GPR method to the detection of reinforcement bars, especially in the case of old
RC bridges, where differences between the project assumptions and reality may occur. These
differences should be taken into consideration in the repair project, especially in the calcula-
tions of the necessary rebars. Due to a complicated beam cross-section (the internal hole and
some edges), the number of the bottom reinforcing bars could not be determined accurately.

5 CONCLUSIONS

As a result of the tests conducted on viaduct beams using the GPR method, the following
conclusions can be drawn:
1. The performed research allowed us to identify the locations of: (i) the main upper and
bottom reinforcing bars, (ii) the stirrups and their spacing (also the differences in their
spacing in mid-span and support zones), and (iii) an internal hole in the main beams. The
test results showed quite good compatibility with the Gromnik–type beam. Disparities in
the results were due to the relative complexity of the beam cross-section (many layers and
edges causing reflection of the EM waves). This complexity had a negative influence on
the ability to scan efficiently through the whole depth of the beam.
2. The GPR results were affected by the complexity of the viaduct beam cross-section, which
consisted of various media (concrete, steel, hole filled by air). Generally, concrete is not a
homogeneous material and should be considered as a composite material. Therefore, the
velocities of the wave propagation through the concrete elements vary according to their
physical and chemical properties. In this case study, the application of the wave velocity on
the level of 0.10 m/ns allowed us to obtain quite acceptable results.
3. Taking into account the advantages and disadvantages of non-destructive methods, pre-
sented by McCann & Forde (2001), it can be concluded that the GPR technique is the
most suitable method for bridge testing in order to detect the cross-section of the beams.
GPR measurements using a bipolar antenna with a high frequency of 2000 MHz proved
to be a quite effective method of determining the parameters of the RC viaduct beams.
The GPR technique also allowed detection of discontinuities in the concrete structure
(e.g., the hole in the beam). It should be noted that the analysis of radargrams requires
skill to interpret the data. Research on accurate ways of determining the diameter of
rebars occurring in concrete should be continued.

REFERENCES

Annan, A.P. 2001. Ground Penetrating Radar. Workshop Notes. Mississauga: Sensors & Software Inc.
Drobiec, L., Jasinski, R. & Piekarczyk, A. 2010. Diagnosis of Reinforced Concrete Structures. Warsaw:
Scientific Publisher PWN.
Hugenschmidt, J. 2002. Concrete bridge inspection with a mobile GPR system. Construction and Build-
ing Materials 16(3): 147–154.
McCann, D.M. & Forde, M.C. 2001. Review of NDT methods in the assessment of concrete and
masonry structures. NDT & Evaluation International 34(2): 71–84.

NYBRIDGE_Book.indb 252 7/17/2013 8:44:07 AM

 Chapter 23

Monitoring of soil-steel railway bridge under service loads

D. Beben & W. Anigacz

Faculty of Civil Engineering, Opole University of Technology, Opole, Poland

ABSTRACT: The paper presents the results of experimental tests under dynamic loads
that were conducted on a soil-steel railway bridge. On the basis of the measured displace-
ments, use of the Frequency Domain Decomposition method determined the frequencies of
this bridge. The microwave interferometric radar and inductive gauge were used for monitor-
ing of displacements of this railway bridge. The main aim of this paper is the evaluation of
possibility of the radar application for the measurements of soil-steel bridge under dynamic
loads. The measurements were made for all trains which had been running over the bridge
during a 24 hour period. The dynamic amplification factors from tests were in the range of
1.10 to 1.375. The frequencies of bridge did not exceed 5.5 Hz. The maximum displacements
of this bridge did not exceed 0.65 × 10–3 m. Conclusions drawn from the tests can be helpful
in the measurements and assessment of dynamic behavior of such bridges.

1 INTRODUCTION

The dynamic and static tests of bridges and culverts are conducted in order to confirm struc-
tural specifications, or to provide diagnostic surveys for planning maintenance and mod-
ernization. Typical experimental tests are usually carried out using accelerometers, strain and
inductive (or dial) gauges. These gauges are quite accurate and reliable. The traditional moni-
toring methods for bridges are presented by (Lee et al. 2002), (Heieh et al. 2006). Moreover,
testing of structures can give rise to accessibility problems, often requiring the use of scaf-
folds. For these reasons, the application of radar measurements is a quite good alternative.
The IBIS-S interferometric radar applied in tests, is a precision microwave instrument
suitable to the non-contact vibrations and displacements monitoring of various engineering
structures. The high accuracy and quickness of measurement are among of the main charac-
teristic features of the radar. The interferometry method for non-contact dynamic and static
measurement of the vibration of various engineering structures (bridges, towers, dams) was
described by (Dei et al. 2009), (Gentile 2010), (Fratini et al. 2009), (Pieraccini et al. 2007).
The main aim of this paper is the evaluation of possibility of the interferometric radar
application for the measurements of soil-steel bridge under dynamic loads. The vertical dis-
placements were measured during the field load tests. On the basis of received displacements,
the frequencies of this bridge were determined using the Frequency Domain Decomposition
(FDD) method. In order to verify the measurements of the bridge deflections conducted using
the radar, the inductive gauge was applied. Because the standard conditions for the radar
measurements could not be complied, a special “transmission gear” was used. The monitoring
was made for all trains which had been running over the bridge during a 24 hour period.

2 SHORT BRIDGE DESCRIPTION

The tested railway bridge in the cross section has two spans in shape of closed arch. The
effective length of shells is L1 = L2 = 4.40 m that are placed directly on a special profiled layer

253

NYBRIDGE_Book.indb 253 7/17/2013 8:44:07 AM

 254 Bridge monitoring

Figure 1. The CSP railway bridge: a) cross-section, b) longitudinal section A-A.

of soil substructure of approximately 0.20 m thickness and compacted to reach the indicator
density ID = 0.98–0.95 according to the Proctor Normal scale (Figure 1).
The load bearing structure was constructed as two shells assembled from the corrugated
steel plate (CSP) sheets. The corrugation depth of a = 0.05 m, pitch of b = 0.15 m and plate
thickness of t = 0.003 m were designed. The individual sheets were connected together using
high strength bolts ø = 20 mm, and covered with the layers of soil (approximately 0.20–0.30 m
thick) properly compacted (in the Proctor Normal scale ID = 0.95 for the soil connected
directly with the steel structure and ID = 0.98 for the remaining part of the backfill – blanket).
The soil cover over the CSP bridge (including ballast, blanket and backfill) equals 2.40 m.
The width of the bridge shell at the top is Bt = 16.00 m, whereas at the bottom is Bb =
21.80 m. The height of shell is H1 = H2 = 2.80 m. In the plan view, the object is situated per-
pendicularly to the railway line. On the bridge two rail tracks are arranged, first is electrified
and the second is designed for non-electrified rolling stocks (Figure 1).

3 BACKGROUND OF RADAR TECHNIQUE AND MEASUREMENTS

METHODOLOGY

3.1 General overview

The radar can be operated in two different modalities: providing only range resolution or pro-
viding both range and cross-range resolution. The measurement works were made by appli-
cation an industrially engineered microwave interferometer (model IBIS-S – Figure 2), which
consists of a sensor module, a control PC and a power supply unit. Detailed description of
microwave interferometer is given by (Dei et al. 2009), (Gentile 2010). The measurement

NYBRIDGE_Book.indb 254 7/17/2013 8:44:07 AM

 Monitoring of soil-steel railway bridge under service loads 255

Figure 2. View of the CSP railway bridge during experimental tests using the IBIS-S interferometric
radar.

system consists of specialized software that allows to control the measurement process, and
preliminary analysis of the results in real time. It also allows to conduct a comprehensive
analysis of the results during post-processing stage.
The IBIS-S interferometric radar with a range of 1000 m is designed to measure and ana-
lyze fast-changing movements and vibrations of the structure. The interferometric system
performs the precision measurement of changes of the reflected signal phase in relation to
the emitted signal. In typical measurement conditions, the radar has the following charac-
teristics: range resolution of 0.50 m, the displacement measurement accuracy of 0.01 mm,
the sampling rate up to 200 Hz. It should be noted that the tested object may be situated
10–4000 m from the instrument. The pixel distinguishability is achieved when the change of
distance is 0.50–0.75 m and azimuth is 4.5 mrad.
The basis of the instrument is the working radar with the Ku-band (17.1–17.3 GHz, the
wavelength of 17.2–17.4 mm), which uses the synthetic aperture method (In-SAR) – previ-
ously applied only in the satellite measurements. The radar distinguishes the source of the
reflected signals according to their distance from the instrument. It results from stepwise fre-
quency change (the range of 17.1–17.3 GHz is divided into 256 intervals with linearly increas-
ing frequency). Then, this distance is divided into sectors (bins) with a length of 0.50 m.
The sources located in one sector are indistinguishable. Each of the continuously repeated
sequences of 256 successive signals takes 0.005 s. Therefore, it is possible to register the loca-
tion of points with the dominant reflections in each sector with a frequency of 200 Hz, and
thereby to detect their vibrations with frequency of 100 Hz.
In the case of application of the microwave reflectors which give an unambiguous reflec-
tion, the high frequency of emitted signal enables to detect displacements even of the order
of 1 μm. However, from a practical point of view, it is assumed that the accuracy of measured
displacements is approximately 0.01 mm (with a distance not exceeding 500 m).

3.2 Measurements methodology

A coherent microwave beam with a very low power and variable frequencies is emitted by the
microwave interferometeric radar in direction of the tested object. Next, the reflected signal
is received and analyzed.
As results from the above-mentioned description, displacements of the reflector are
detected and recorded only in the radial direction. The high accuracy of measurements makes

NYBRIDGE_Book.indb 255 7/17/2013 8:44:07 AM

 256 Bridge monitoring

possible to carry out tests under conditions in which the displacements are not parallel to the
axis of the microwaves beam. It is possible in the case when the change of radial distance is
not less than 10% of the real displacements. It means that the angle between the direction
of the microwaves beam axis and the direction of displacements should not exceed 84°. In
these experimental studies, the distance from the radar to the microwave reflector was 14.5 m
(bin = 34). The instrument performed recording with a frequency of 94.26 Hz in all sectors
up to a distance of 30 m, i.e. for 60 sectors. It was a compromise between the expected range
of vibration frequencies and the amount of data recorded. At the stage of results working
out it turned out that frequencies had not exceeded 15 Hz. Typical conditions for the radar
measurements could not be fulfilled in the case of this bridge. Thus, a special “transmis-
sion gear” (Figure 3) must be used. Thanks to this mechanism, the vertical displacements of
the bridge were transformed into the horizontal displacements. The mechanism operation is
based on the assumption that “transmission gear” acts as a lever with one axis of rotation.
Figure 4 shows instrumentation and idea of the radar measurements using the “transmis-
sion gear”. As a result of preliminary control testing in the lab and field it was found that
the specially constructed mechanism was operated in a satisfactory manner. Details can be
found in (Beben 2011).
The tested railway bridge was made from corrugated steel plates (pitch of 0.15 m). Such
shape of the structure could potentially result in the signal reflections from the multi-sur-
faces. As a result of this phenomenon, the measurement signals reflected from the reflector
might be drowned out. In order to reduce the risk of drowning out the signals, the measure-
ment beam of microwaves was placed horizontally (close to the axis of the bridge). Moreover,
a special microwave horns IBIS-H23 type with a maximum gain of 23 dBi were used as the
interferometer antennas (transmitting and receiving). The measurement beam with the small-
est obtuse angle is created: for –3 dB is 10° vertically and 11° horizontally, which corresponds
to the width of the beam of 2.9 m on a distance of 16.5 m (sector 34) from the interferometer
to the reflector. During the measurements, it turned out that the signals reflected from the
reflector significantly dominated by the noise caused by other sources.
The dynamic characteristics of “transmission gear” and the reflector were also deter-
mined by (Beben 2011). It was an extremely important problem, because it was necessary
to distinguish the displacements spectra between the “transmission gear” and the bridge.
For this purpose, the pre-tests in several trials were conducted. The artificial creations of
vibrations (few millimeters) by pushing the vertical rod were made. In all considered cases,
the almost identical results were obtained, i.e. the occurrence of the same frequencies and a

Figure 3. View of the “transmission gear” used to tests.

NYBRIDGE_Book.indb 256 7/17/2013 8:44:07 AM

 Monitoring of soil-steel railway bridge under service loads 257

Figure 4. Instrumentation and idea of the radar measurements using the “transmission gear”.

similar course of damping. The experimental tests were conducted in one cross section (in
crown directly under the railway line) of the CSP bridge where the maximum displacements
and vibration frequencies were expected. The field tests were conducted continuously for
24 hours. The forty-one various trains were noticed (eighteen – passenger, eighteen – freight,
five – others). Details of passing trains are presented by (Beben 2011). In order to evaluate
the possibility of the interferometric radar application, at the same time, the inductive gauge
was used. The inductive gauge was placed in the same cross section of the soil-steel bridge in
which the special “transmission gear” was fixed. So, two independent techniques for deflec-
tion measurements of the CSP bridge were applied.

4 RESULTS OF MEASUREMENTS AND THEIR ANALYSES

4.1 Bridge displacements and frequencies

The vertical displacements of shell structure of the soil-steel bridge were measured during con-
ducted experimental tests using interferometric radar (IR) and inductive gauge (IG). The vibra-
tion frequencies of the CSP bridge were also recorded using the radar. In order to prove the
possibility of using the interferometric radar for measurements of these specific CSP bridge,
two passages of trains are analyzed in detail. The passages of trains with various weights of
particular wagons were selected. This choice of trains was determined for the verification of
possibilities of measurements with the radar for the rapidly changing various loads.

NYBRIDGE_Book.indb 257 7/17/2013 8:44:08 AM

 258 Bridge monitoring

Figure 5 shows example of displacements of the crown of CSP bridge. The effect of each
axle of the freight train (no. 29) can actually be also observed from the displacement versus
time plots. The total weight of passing train was 11846 kN. Almost identical vertical displace-
ment courses were registered using the interferometric radar (IR) and the inductive gauge
(IG). During train rides, four main phases of displacements were emphasized. The first
phase represents impact of locomotive and two heavy wagons. The maximum displacements
amounted to fIR = 0.53 × 10–3 m (fIG = 0.55 × 10–3 m). In this stage the highest frequencies were
1.2 and 1.7 Hz (Figure 6a). The second phase relates to passage of four lighter wagons. In this
stage the maximum displacements equaled almost fIR = 0.15 × 10–3 m (fIG = 0.16 × 10–3 m) with
frequency of 0.6 Hz. The third phase represents impact of eight heavy wagons. The largest
displacements amounted to fIR = 0.22 × 10–3 m (fIG = 0.24 × 10–3 m) with frequencies of 0.6 Hz
(less distinct frequencies were 0.9, 1.8 and 2.5 Hz – Figure 6b). The fourth phase relates to
vibration reduction (damping) after passage of the freight train. This stage was dominated
by frequency of 8.9 Hz which was identified, according to the preliminary tests, as a natural
frequency of the “transmission gear”.
The typical response of the CSP railway bridge to passing trains is usually the collection of
three frequencies in the range of 0.6 to 3.0 Hz. The highest values of frequencies were caused

Figure 5. The courses of vertical displacements measured using the radar (continuous line) and the
inductive gauge (dotted line) during passage of the freight train no. 29.

Figure 6. Spectra of displacements measured using radar at the crown of CSP bridge in phase: a) I and
b) III, for passage of the freight train no. 29.

NYBRIDGE_Book.indb 258 7/17/2013 8:44:08 AM

 Monitoring of soil-steel railway bridge under service loads 259

by passage of the express trains (speed was varying from 70 to 120 km/h). In turn, the maxi-
mum vertical displacements of the crown of CSP bridge do not exceed fIR = 0.65 × 10–3 m (fIG =
0.67 × 10–3 m). They were obtained during passage of the heavy freight trains. Then, the small-
est deflection of the bridge registered during tests was fIR = 0.08 × 10–3 m (fIG = 0.08 × 10–3 m).
The CSP bridge responds to change of loads very fast. The displacements decrease imme-
diately after passages of trains. The maximum measurement results of the CSP bridge using
the interferometric radar were quite similar to results received using the inductive gauge (the
differences did not exceed 5%). Such differences can be resulted in applying the “transmission
gear”. The interferometric radar correctly responds to rapid and frequent change of loads.
The courses of displacements in time of the CSP bridge using the both measurement tech-
niques were almost identical. The natural vibrations of shell structure were not observed. It
is probably caused by large thickness of the backfill placed around steel shell structure which
in this case is a damper of vibration.

4.2 Dynamic amplification factor

There are many factors which can influence the number of dynamic effects. Speed of the
train, as well as the span and the mass of the bridge structure are known to be important
factors that have a great influence on the dynamic response. The number of axles and axle
loads, as well as the spacing of the axles, can also influence the dynamic behavior. The track
can have varying dynamic characteristics owing to differences in ballast, sleepers and other
track components. Vertical irregularities on the track, as well as wheel defects, also result in
increased dynamic effects (Flaner & Karoumi 2009).
Traditionally, the dynamic amplification factor (DAF) is determined in experimental tests
through movement of the same load at different speeds across the bridge. Such tests for the
dynamic analysis of the bridge were presented by (Manko & Beben 2008). In this case, owing
to the character of conducted tests (passages of miscellaneous trains, at various speeds), the
DAF factors were estimated for all passing trains. Only dynamic testing under traffic loading
is suitable for the evaluation of the DAF (Bakht & Pinjarkar 1989).
The DAFs on vertical displacement of the crown (φdisp.) are calculated according to Equa-
tion (1):
max (i )
fddyn
φdisp.(i) = max (i )
(1)
fsstat

where φdisp.(i) is the DAF for each passing train i, while fdyn
max(i)
is the maximum dynamic dis-
max(i)
placement for each passing train i and fstat is the maximum static displacements obtained
(for each passing train i) by filtering the dynamic displacements, taken at the time when the
maximum dynamic displacement occurs.
The DAFs for vertical displacements of the crown of CSP bridge in relation to the weight
of trains and their speeds are shown in Figure 7. As mentioned earlier, the DAF takes into
consideration many various factors, e.g. the mass of trains, the speed of trains, the arrange-
ment of axles and their loads, the technical quality of the track, etc. In turn, the dynamic
coefficient allows the calculation of the dynamic impact, used as a partial safety factor in
the calculations at the design stage. In other words, the internal forces and moments in the
critical bridge elements, received from the static loads, are multiplied by the DAF to receive
the dynamic impact. The significant differences in the definition of these factors can be seen,
and, from a practical point of view, can be compared. The DAFs also enable the evaluation
of current design standards in regard to dynamic impacts.
For small-to-medium traditional bridges, the impact factor increases with the decreased
weight of the vehicle (McLean & Marsh 1998). Therefore, data obtained from lightly loaded
trains can lead to an overestimation of the DAF, and should not be considered when comput-
ing this factor. In the case of the CSP bridges, the dynamic impact is more complicated, and

NYBRIDGE_Book.indb 259 7/17/2013 8:44:08 AM

 260 Bridge monitoring

Figure 7. The DAFs of CSP railway bridge in relation to: a) weight of trains, b) speeds of trains.

is probably a result of the flexible character of the bridges, which are sensitive to rapid and
frequent changes in load. The influence of train speeds was clearly visible during the passage
of express trains (speed v = ca. 120 km/h, weight 3236 kN). In this case, the DAFs were much
higher (ca. 1.375) than during the passage of similar loads, but at significantly smaller speeds
(ca. 40–45 km/h, DAFs = ca. 1.24; Figure 7a). However, the influence of train weight on the
DAF was visible for heavy freight trains (3236–16867 kN) passing at relatively low speed v =
30–40 km/h (max DAFs = 1.32). In these cases, the DAFs were significantly higher than the
DAFs received during the passages of other trains (maintenance trains and locomotives) at
speeds v = ca. 40–55 km/h (Figure 7b). Besides, in the case of the CSP bridge, passing light
trains (weight 1353 kN) at medium-high speeds (v = ca. 60–75 km/h) did not cause increased
values of the DAFs (ca. 1.235), as in the case of the traditional steel or RC bridges.
So, it can be concluded that in the case of the soil-steel bridge, for small train speeds (v <
45 km/h), the DAF is more related to the weight than the speed, whereas for higher speeds (v
> 60 km/h), the weight has a smaller influence on the DAFs (Figure 7b). In these cases, the
train speeds had a significant impact on the DAF, which is probably related to greater impact
of loads caused by the trains moving at high speeds. However, the impact of train weight
(axle loads) cannot be completely neglected.
The obtained DAFs, based on the experimental tests, were compared with values calcu-
lated from the Eurocode, (CEN EN 2002) according to Equation (2):

1 44
φEurocode 0 82 i h 1 0 φEurocode ≤ 1.67 (2)
lφ − 0.2

NYBRIDGE_Book.indb 260 7/17/2013 8:44:09 AM

 Monitoring of soil-steel railway bridge under service loads 261

where lφ = L1 is the shell span (determinant length) in meters.

It should be added that the Eurocode, (CEN EN 2002) does not provide determinant
length specific to CSP bridges. In the case of arch bridges and reinforced concrete (RC)
bridges, of all types with a depth of soil cover (hc) more than 1.0 m, the dynamic coefficient
may be reduced according to Equation (3):

hc − 1.0
φEurocodee (reduce) φEurocode
d − ≥ 1.0 (3)
10

where hc is the depth of soil cover.

The DAF received from the Eurocode, (CEN EN 2002) equals to 1.578, and the reduced
DAF amounted to 1.438. So, these DAFs are higher than those obtained from experimental
tests.
Based on the present study, DAF may be helpful in the estimation of the minimum depth
of soil cover. The DAFs considered can be applied to the evaluation of the load rating of the
CSP bridges through the estimation of the dynamic impact of the live loads. Currently, the
appropriate bridge regulations are used, which do not fully reflect the real dynamic behavior
of the CSP bridges.

5 CONCLUSIONS

As a result of the 24-hours monitoring of the soil-steel railway bridge during service loads,
the following general conclusions can be drawn:
1. The largest vertical displacements of the crown of CSP bridge did not exceed 0.65 × 10–3 m.
They were obtained during the passage of heavy freight trains. The vertical displacements
of the CSP bridge seem to depend more on the weight of trains than their speed. The
dominant frequencies of the CSP bridge caused by kinematic excitations were in the range
of 1.1 to 5.5 Hz. The highest values of frequencies were caused by the passage of the trains
with speeds varying from 70 to 120 km/h (regardless of their weight). Whereas for slower
train speeds (v ≤ 40 km/h), their weights had a greater influence on the frequency level of
the CSP bridge.
2. The DAFs from tests were in the range of 1.10 to 1.375. The largest DAF was obtained
for express trains (the speed was almost 120 km/h). The DAFs caused by the passage of
the heavy freight trains were a little bit over 1.30, whereas, for the slow passenger trains
(1353 kN), the DAFs did not exceed 1.235. The obtained results may be helpful to estima-
tion of load rating and maintenance of the CSP bridges.
3. The DAF factors received from tests were smaller than the values calculated according
to Eurocode, (CEN EN 2002). This is probably owing to the use of simple formulas that
depend only on a function of the shell span. The reduced impact factor of the Eurocode,
(CEN EN 2002) was the most similar to the DAF received from the experimental tests
(ca. 5% higher). It seems that the realistic DAF for design calculations, besides the bridge
span, should also take into account the maximum speed of the vehicle and the depth of
soil cover, as well as the axle loads.
4. The coherent interferometric radar is a suitable measurement instrument to long-term
monitoring of small-to-medium culverts or bridges. It allowed to get the displacements
and frequencies of this specific type of the soil-steel bridges very fast and with a high
accuracy level. The interferometric radar registered even very small values of displace-
ments and frequencies of the bridge with an accuracy in the order of 0.01 mm and 0.1 Hz,
respectively. The special “transmission gear” need to be used during testing of small
bridges and culverts made from the CSP elements. This necessity results from the possibil-
ity of drowning out the signal by the corrugation surfaces.

NYBRIDGE_Book.indb 261 7/17/2013 8:44:09 AM

 262 Bridge monitoring

5. Almost all of the displacement results received using the interferometric radar were as
accurate as (mostly smaller) those obtained with the conventional inductive gauge. Gen-
erally, the application of the specially constructed “transmission gear” did not influence
measurement results significantly.

REFERENCES

Bakht, B. & Pinjarkar, S.G. 1989. Review of Dynamic Testing of Highway Bridges. Structural Research
Report SRR-89–01. Downsview, Ontario: Ministry of Transportation of Ontario.
Beben, D. 2011. Application of the interferometric radar for dynamic tests of corrugated steel plate
(CSP) culvert. NDT & Evaluation International 44(5): 405–412.
CEN, EN 1991–2. 2002. Eurocode 1: Actions on Structures – Part 2: Traffic Loads on Bridges. Brussels:
European Committee for Standardization.
Dei, D., Pieraccini, M., Fratini, M., Atzeni, C. & Bartoli, G. 2009. Detection of vertical bending and
torsional movements of a bridge using a coherent radar. NDT & Evaluation International 42(8):
741–747.
Flener, E.B. & Karoumi, R. 2009. Dynamic testing of a soil-steel composite railway bridge. Engineering
Structures 31(12): 2803–2811.
Fratini, M., Parrini, F., Pieraccini, M., Borri, C. & Atzeni, C. 2009. Structural oscillation modes identi-
fication by applying controlled loads and using microwave interferometry. NDT & Evaluation Inter-
national 42(8): 748–752.
Gentile, C. 2010. Deflection measurement on vibrating stay cables by non-contact microwave interfer-
ometer. NDT & Evaluation International 43(3): 231–240.
Heieh, K.M., Halling, H.W. & Barr, P.J. 2006. Overview of vibrational structural health monitoring with
representative case studies. Journal of Bridge Engineering 11(6): 707–715.
Lee, J.W., Kim, J.D., Yun, C.B., Yi, J.H. & Shim, J.M. 2002. Health-monitoring method for bridges
under ordinary traffic loading. Journal of Sound Vibration 257(2): 247–264.
Manko, Z. & Beben, D. 2008. Dynamic testing of a corrugated steel arch bridge. Canadian Journal of
Civil Engineering 35(3): 246–257.
McLean, D.I. & Marsh, M.L. 1998. Dynamic Impact Factors for Bridges. National Cooperative High-
way Research Program, NCHRP Synthesis 266. Washington: Transportation Research Board.
Pieraccini, M., Parrini, F., Fratini, M., Atzeni, C., Spinelli, P. & Micheloni, M. 2007. Static and
dynamic testing of bridges through microwave interferometry. NDT & Evaluation International 40(3):
208–214.

NYBRIDGE_Book.indb 262 7/17/2013 8:44:10 AM

 6 Bridge aesthetics & history

NYBRIDGE_Book.indb 263 7/17/2013 8:44:10 AM

This page intentionally left blank
 Chapter 24

Artists’ portraits of the Sydney Harbour Bridge

S. Rothwell
Stuart Rothwell & Associates, Brisbane, Queensland, Australia

ABSTRACT: The Sydney Harbour Bridge in Australia was a spectacular culmination for
the rapid development of modern steel arch bridges, which began with the construction of
Eads’ bridge over the Mississippi at St Louis in 1874, a year synonymous with the birth
of modern art. In turn, the construction of the Sydney Harbour Bridge had a consider-
able impact on the development of a seminal modern art movement in Australia. When the
harbour bridge was being built, the art produced represented the hopes and aspirations of
a fledgling nation, and reflected the construction methodology adopted. Selected art works
inspired by the bridge are reviewed and the development of the modern steel arch bridge is
described to illustrate the intellectual and aesthetic progress that the Sydney Harbour Bridge
represented.

1 INTRODUCTION

Life without industry is guilt

Industry without art is brutality
John Ruskin 1870
The Australian historian Bernard Smith (1916–2011) has remarked that modern art
‘developed out of French impressionism’ (Smith 1992), the origins of which can be traced
to the work of Edouard Manet (1832–1883), who commenced his art training in 1850.
Ironically, modernity—as a concept that today’s structural engineer can appreciate—
commenced around the same time, with the construction of large span railway bridges in
wrought iron, designed on the basis of rational, mathematically based analysis and rigor-
ous material testing. In the succeeding decade the arrival of steel rapidly extended the
tentacles of the permanent way, and bridge engineering as practiced today emerged from
the smoke and steam of that era. The development of modern art in France shadowed the
emergence of the modern bridge, and was both influenced and inspired by it (Rothwell
2011). Similarly, the germination of modern art in Australia coincided with the successful
promotion of the scheme to construct a bridge across Sydney Harbour, which provided
both inspiration and opportunity for pioneer modernists, like Grace Cossington Smith
(1892–1984), to forge a new art movement in Australia, with paintings such as Before the
arches met, shown in Figure 1.
Although the arch bridge is a structural idea rooted firmly in antiquity, the large metal
arch is a modern creation both in terms of design and construction. While Sydney Harbour
Bridge was never the longest steel arch bridge in the world (Smith 1953), it was always the
most audacious in concept, the most daring in its execution, and the most imposing in-situ.
Historically, the bridge can be viewed as a series of stages; a struggle from concept to frui-
tion; the initial symbolism of its construction for a fledgling nation and a fading empire; its
rise to iconic significance for Sydney and Australia; and, its eventual relaxation into the eve-
ryday life of the city and the general vista of the harbour.

265

NYBRIDGE_Book.indb 265 7/17/2013 8:44:10 AM

 266 Bridge aesthetics & history

Figure 1. Grace Cossington Smith, Australia, 1892–1984, Before the arches met, 1930 Crayon and coloured
pencils on cream wove paper, 37.8 × 43.4 cm, Queensland Art Gallery Godfrey Rivers Trust, 1976.

2 THE STRUGGLE

A bridge across Sydney Harbour was mooted quite regularly from 1815 (Smith 1953). Nev-
ertheless, it would be another century before the idea finally commenced its fateful trajectory
when, in 1912, the Queenslander J.J.C. Bradfield (1867–1943), a passionate advocate for a
harbour crossing, was appointed Engineer-in-Chief for the Sydney Harbour Bridge. The fol-
lowing year Bradfield’s proposed cantilever bridge for all traffic modes, including railways,
was accepted by the Public Works Committee (Raxworthy 1989).
Coincidentally, in 1913, a young art student, Nora Simpson (1895–1974), returned from
a trip to Europe, where she had collected colour prints, books and magazines about post-
impressionism and cubism, which she was able to share with her isolated student friends,
including Grace Cossington Smith and Roland Wakelin (1887–1971) (Smith 1992). The onset
off the Great War (1914–18) interrupted Bradfield’s dream and delayed youthful artistic
experimentation, but the seeds of modernity had been planted in Australia.
Meanwhile, in the USA, Gustave Lindenthal (1850–1935) constructed the Hell Gate
Bridge, a large steel arch structure over the East River in New York, shown in Figure 2, and
completed in 1916 (Smith 1953).
Tenders were invited for the Sydney Harbour Bridge in December 1921. The invitation
however, only called for the detailed design and construction of the cantilever bridge, and a
concept design for that form was included in the invitation documents. After an overseas visit
in 1922, involving discussions with prospective tenderers in the USA and England, Bradfield
revised the tender documents to include an arch design (Lalor 2005). Herein lay the genesis
for the eventual acrimonious controversy over design provenance. Bradfield contended that
the final design for the harbour bridge was based on the concept he developed for inclusion
in the amended tender documents. The British contractor’s consultant, R. Freeman (1880–
1950)—later Sir Ralph—and his team executed the detail design. Freeman maintained that

NYBRIDGE_Book.indb 266 7/17/2013 8:44:10 AM

 Artists’ portraits of the Sydney Harbour Bridge 267

Figure 2. Hell Gate Bridge, New York (Library of Congress, Washington).

his concept for an arch predated the tender revision. Gordon Stuckey, Bradfield’s assistant,
later dryly remarked that the bridge ‘was big enough for both of them’ (Raxworthy 1989).
Lindenthal poignantly observed that it looked a lot like his (Lalor 2005), which it certainly
does. In fact, the two bridges are so similar that the Sydney bridge must be considered as
derivative of Lindelthal’s masterpiece.
Like in Sydney, the Hell Gate Bridge is a two hinge, spandrel braced arch, and was con-
structed using a cantilever technique (Smith 1953), shown in Figure 3. Note the travelling
cranes, also part of the construction technique at Sydney, and discernable in Grace Cossing-
ton Smith’s painting (Fig. 1).
The anchoring technique at Sydney, however, was different—utilizing cables—evident too
in Figure 1. Also, the two bridges are dissimilar in scale; at 298 m, the New York structure is
considerably shorter than the Sydney bridge, which spans 503 m (Dupré 1998).
Both the Hell Gate and Sydney Harbour bridges were born of a big iron/steel arch bridge
tradition that commenced with the St Louis Bridge over the Mississippi River. The bridge,
shown in Figure 4, was completed by James Buchanan Eads (1820–1887), in 1874 (Dupré
1998), the year in which Claude Monet (1840–1926) exhibited Impression—Soleil Levant and
so defined an art movement. Eads’ was the first major bridge constructed of steel and the
first large arch constructed by a cantilever method (Smith 1953).
Soon afterwards, Gustave Eiffel (1832–1923), now more famous for his tower in Paris, pro-
duced two notable wrought iron arch bridges; the Maria Pia Viaduct in Porto (1877), shown
in Figure 5, and the Garabit Viaduct near Saint Fleur (1885), with arch spans of 160 m
and 165 m respectively, both constructed by a cantilever technique (Dupré 1998). Also con-
structed in Porto, in 1885, was the Luis I bridge by Eiffel’s former associate Théophile Seyrig
(1843–1923). At the time its arch span of 172 m was the longest in the world (Smith 1953).
As the nineteenth century drew to a close Leffert L Buck (1837–1909) designed two
important steel arch bridges across the gorge at Niagara Falls. In 1897 the Niagara Railway
(Whirlpool) Bridge was completed with an arch span of 168 m. Simultaneously, Buck was
also working on a design for the Falls View (Honeymoon) Bridge, which—with a new record
span of 256 m—was completed in 1898. Shortly afterwards, the first large steel bridge built
in France was the Viaur Viaduct, a 220 m arch at Aveyron, completed in 1902 under the
supervision of Paul Joseph Boden (1848–1926).
A young Ralph Freeman was also involved in design of the Victoria Falls Bridge (1905).
The construction of that bridge over the Zambesi River, with an arch span of 157 m, pres-
aged the technique used at Sydney, where anchor cables—looped through rock tunnels—
supported the half arches, which were progressively cantilevered from each abutment until
they eventually met (Smith 1953).

NYBRIDGE_Book.indb 267 7/17/2013 8:44:11 AM

 268 Bridge aesthetics & history

Figure 3. Constructing Hell Gate Bridge, New York (Library of Congress, Washington).

Figure 4. Constructing Eads’ Bridge, St Louis (Library of Congress, Washington).

Figure 5. Maria Pia Viaduct, Porto (Photo by Jose Goncalves).

While construction of the Sydney project commenced in mid 1923, the contract for the
bridge was not let until March 1924, and erection of the arch did not begin until 1929.
Given the approaching grim economic times, and a setting imperial sun, the Sydney Harbour
Bridge soon became a potent symbol for both an aspirational Australia and a fading British
Empire.

NYBRIDGE_Book.indb 268 7/17/2013 8:44:11 AM

 Artists’ portraits of the Sydney Harbour Bridge 269

3 THE SYMBOL

From the start of construction, Sydney’s art community was attracted to the harbour to
draw, paint and photograph (Dutton 1986). The etchings of Jessie Traille (1881–1967), pro-
duced between 1927 and 1931, are an emotive collection of snatched moments in construc-
tion (Prunster 1982).
The key features of the Sydney Harbour Bridge’s construction method can also be dis-
cerned in a view from The Rocks painted by Roland Wakelin (Fig. 6), who—after his friend-
ship with Norah Simpson—studied in Paris in the mid 1920’s, and during that time was heavily
influenced by the works of Paul Cezanne (1839–1906) (Prunster 1982). Wakelin, originally
from New Zealand, had become interested in modern art when, after arriving in Australia,
he saw a reproduction of Nude Descending a Staircase, by Marcel Duchamp (1887–1961), in
a Sydney newspaper report of the controversial 1913 New York Armory Show (Smith 1992).
The simplification and reduction of form in his painting, shown in Figure 6, demonstrate
how forcefully he had been affected by the various outside styles he had encountered as a
young man.
The salient features of construction are also evident in the Japanese influenced print by the
Victorian, Frances (Frankie) Derham (1894–1967), shown in Figure 7. Her father was an engi-
neer and her drawing skills, which are obvious in this work, were noticed when she began to assist
with his draughting. That connection may explain her desire to catch the scene, and her instinctive
sense of scale that captures the enormity of the span being attempted in September 1929.
With The Bridge, shown in Figure 8, Dorrit Black (1891–1951) produced a de-urbanised,
serene image with a flattened, almost abstract, post-impressionist landscape, with cubist over-
tones. Rather than show a scene of industrial activity, she has chosen, seemingly, to emphasise
the nation’s colonial origins, with a sailing ship in the harbour. But, like the others, she has
recorded the anchor cables and cranes, those essential features of construction. Black was
a South Australian, who in 1927 travelled to London and studied at the Grosvenor School
of Modern Art. Later in the same year she enrolled in the Paris Academy of the French

Figure 6. Roland Wakelin, The Bridge under construction, 1928 Oil on composition board,
96.5 × 118 cm, Queensland Art Gallery, Brisbane.

NYBRIDGE_Book.indb 269 7/17/2013 8:44:13 AM

 270 Bridge aesthetics & history

Figure 7. Frances Derham, Australia, 1894–1967, Sydney bridge, 1929 Linocut, printed in colour, from
three blocks, 17.8 × 25.9 cm, National Gallery of Art, Canberra.

pioneer modernist Andre Lhote (1885–1962), and in 1928 attended his summer school at
Mirmande in the south of France. She also took lessons from Albert Gleizes (1881–1953),
another influential French modernist—now remembered in the USA for his paintings of
New York’s Brooklyn Bridge—and travelled extensively in Europe (Burke 1980). She arrived
back in Sydney in late 1929 and The Bridge, shown in Figure 8, was painted the following
year, encapsulating the many influences from her European sojourn. Although Dorrit Black
mixed in the modernist circle of the early 1930’s in Sydney, she is said to have been something
of a loner and her painting does contain an undercurrent of melancholy.
There is, however, no shortage of vitality and structural intensity in the works of Grace
Cossington Smith, who seemed to capture the dynamic possibilities of the bridge under con-
struction in a quite unique way. Her The bridge in-curve, shown in Figure 9, and Before the
arches met (Fig. 1) exhibit a kaleidoscope of influences from impressionism to art deco.
Yet, for all their artistic modernity, her works are also a faithful representation of the
essential characteristics of the bridge and the manner in which it was built. In Figures 9, as
in Figure 1, we can see the two arches reaching out from opposite sides of the harbour, the
travelling cranes, and the temporary cables that restrained the end of each cantilevered arch
until they finally met in 1932.
Cossington Smith’s numerous paintings of Sydney’s bridge—now spread around various
Australian State galleries—are exemplars of the modernist style that split the artistic world
in Australia between the two world wars. With any revolution there is always a rearguard of
traditionalists protecting the status quo and many talented artists, notably the Lindsay broth-
ers, resisted the modernists (Dutton 1986). In 1927, Percy Lindsay (1870–1952) painted the
bridge during the early stages of construction. His The Bridge Builders, shown in Figure 10,
acknowledges the contribution of the unsung individual and, like Jessie Traille’s work, illus-
trates the enormity of the project by introducing ant like workers into the foreground of his
scene, with a distant North Shore.
When the bridge was opened on 19 March 1932, Francis de Groot (1888–1969) almost
upstaged the Premier, Jack Lang (1876–1975), by prematurely cutting the ribbon. Fifty years
later the Australian artist, Sydney Nolan (1917–1992), always fascinated by anti-heroes, drew
a childlike image of the scene (Prunster 1982), now in a private collection, and his caricature
of the unimpressed Lang is a model of spare portraiture.

NYBRIDGE_Book.indb 270 7/17/2013 8:44:15 AM

 Artists’ portraits of the Sydney Harbour Bridge 271

Figure 8. Dorrit Black, Australia, 1891–1951, The Bridge, 1930 Oil on canvas on board, 60.0 × 81.0 cm,
Art Gallery of South Australia, Adelaide Bequest of the artist.

Figure 9. Grace Cossington Smith, Australia, 1892–1984, The bridge in-curve, 1930 Tempera on card-
board, 83.6 × 111.8 cm, National Gallery of Victoria, Melbourne presented by the National Gallery
Society of Victoria, 1967.

NYBRIDGE_Book.indb 271 7/17/2013 8:44:18 AM

 272 Bridge aesthetics & history

Coincidentally, another significant Australian arch bridge was opened in Brisbane a week
later. The Grey Street (William Jolly) Bridge across the Brisbane River is a three span rain-
bow arch structure openly derived from the single rainbow arch span of a bridge across the
Mississippi River at St Paul in Minnesota. The Grey Street Bridge was also painted sev-
eral times during construction, most notably by Queensland artist Vida Lahey (1882–1968),
another important early Australian modernist. One of the three paintings she produced is
shown in Figure 11, and depicts the steelwork of one span of the bridge before it was encased
in concrete (Rothwell 2011).

Figure 10. Percy Lindsay, Australia, 1870–1952, The bridge builders, 1927 Oil on canvas, 51.3 × 68.9 cm,
National Gallery of Victoria, Melbourne Felton Bequest, 1928.

Figure 11. Vida Lahey, Building the bridge, 1931 Watercolour and gouache over pencil on wove paper
on cardboard, 24.5 × 28.5 cm, Queensland Art Gallery, Brisbane, Gift from the estate of Mrs Gladys
Powell, 1941.

NYBRIDGE_Book.indb 272 7/17/2013 8:44:22 AM

 Artists’ portraits of the Sydney Harbour Bridge 273

4 THE ICON

Once the Sydney Harbour Bridge was operational its symbolism merged into iconography, for
both the city and the nation. At the time, Sydney Ure Smith (1887–1949) was an influential
promoter of the visual arts through his magazine Art in Australia. Though not a dedicated
modernist, he was not totally unsympathetic (Smith 1992). He was quite an accomplished
draughtsman himself and in a watercolour, shown in Figure 12, he captured the continuing
construction of the granite pylons after the arch was completed. The pylons do not serve an
ambiguous structural purpose for the arch—as at Hell Gate (Dupré 1998), although they do
steepen the thrust angle at the abutment and so minimised the skewback. They were also used
to assist construction, but were not essential, so their inclusion has been vigorously debated
ever since. It has been strongly argued, however—most notably by the celebrated Australian
art critic, Robert Hughes (1938–2012)—that the iconic appeal of the bridge would be dimin-
ished without them (Dutton 1986).

Figure 12. Sydney Ure Smith, Australia, 1887–1949, Sydney Harbour Bridge from West Circular
Quay, 1931, watercolour and pencil, 36.6 × 29.4 cm National Gallery of Victoria, Melbourne, Felton
Bequest, 1931.

NYBRIDGE_Book.indb 273 7/17/2013 8:44:26 AM

 274 Bridge aesthetics & history

Figure 13. Margaret Olley, Australia, 1923–2011, Sydney Harbour Bridge, late afternoon, 2011 Oil on
board, 46 × 276 cm.

5 THE VISTA

Gradually, Sydney became accustomed to its bridge. While it can still command symbolic and
iconic significance on the right occasion, it has eventually settled into its everyday role as part
of the city vista and the nation’s subconscious, where it remains a source of inspiration for
Australian artists, most recently the late Margaret Olley (1923–2011) whose swansong work
Sydney Harbour, late afternoon, shown in Figure 13, was completed in the year she died.

6 CONCLUSIONS

The modern steel arch bridge can be traced to Eads bridge over the Mississippi at St Louis
in 1874, coincidentally a year commonly associated with the birth, in France, of modern
art. In Australia, modernism—both in art and steel arch bridge construction—seemed to
arrive coincidentally, in 1913, and the result was the Sydney Harbour Bridge as a subject for
Australian modernist painters.

ACKNOWLEDGEMENTS

The author is grateful to the galleries that gave permission to reproduce the images used in
this paper. He is also indebted to numerous people at those institutions who were so generous
with their time and support, namely Kirsty Burow, Tracy Dall, Nicole Klux, Jennie Moloney
and Nick Nicholson. Further, the author acknowledges the following copyright holders who
have kindly given permission to reproduce the work of certain artists—Ann Mills for her
late aunt, Grace Cossington Smith (who ‘loved the bridge’), Judith Murray on behalf of her
father, the late Roland Wakelin, Mark Derham for his late grandmother, Frances Derham,
Shirley Lahey, niece of the late Vida Lahey, and Philip Bacon on behalf of the Estate of
Margaret Olley.

REFERENCES

Burke, J. 1980. Australian Woman Artists 1840−1940. Melbourne, Australia: Greenhouse Publications.
Dupré, J. 1998. Bridges. Köln, Germany: Könemann.
Dutton, G. 1986. The Innovators. Crows Nest, NSW, Australia: The MacMillan Company of
Australia.
Lalor, P. 2005. The Bridge. Crows Nest, NSW, Australia: Allen & Unwin.
Prunster, U. 1982. The Sydney Harbour Bridge 1932–1982. Sydney, Australia: Angus & Robertson.
Raxworthy, R. 1989. The Unreasonable Man. Marrickville, NSW, Australia: Southwood Press.
Rothwell, S. 2011. The Bridge Painters. Engineering World, 21(1), February / March 2011, 14–25. Crows
Nest, NSW, Australia: Engineers Media.
Smith, B. 1992. Australian Painting 1788–1990. South Melbourne, Australia: Oxford University Press.
Smith, H.S. 1953. The World’s Great Bridges. London, Great Britain: Phoenix House.

NYBRIDGE_Book.indb 274 7/17/2013 8:44:28 AM

 Chapter 25

Measures of structural art: A case study using Streicker Bridge

D.H. Sigurdardottir & B. Glisic

Princeton University, NJ, USA

ABSTRACT: Bridge designs are studied through the lens of the three measures of structural
art. The three measures, introduced by the Scottish engineer Thomas Telford in the 19th
century and further developed by Princeton Professor David Billington, are the scientific,
social, and symbolic aspects of the structure, referring to structural efficiency, economy, and
elegance, respectively. Structural art is achieved through an elegant, still cost-effective struc-
ture, which displays its function clearly with its form. A case study using Streicker Bridge on
Princeton University Campus is presented. The structural system is shown to be efficient and
then the structural behavior is verified using Structural Health Monitoring based on fiber
optic strain sensors and a Finite Element Model. A cost-comparison with similar bridges is
used to assess the cost-effectiveness of the bridge, with emphasis on the life-cycle cost. The
elegance of the bridge is evaluated through the choice of form for structural purposes and
the bridge interaction with the users and its surroundings. The discussion shows that new
technology and approaches are applicable to the concept of structural art and they are dem-
onstrated using Streicker Bridge as a case study.

1 INTRODUCTION

Bridges are the links in the road network, which provide a safe path over major obstacles
such as rivers, valleys, or motorways. They are therefore important works of structural engi-
neering. Through the centuries bridges have been considered regional landmarks. In modern
days, this symbolic role of bridges has grown, sometimes to the point where the bridge form
is conceived by sculptors, not engineers. This evolution has led to many irrational structural
forms and unnecessary spending of funds through expensive construction or costly retrofits.
Some bridge owners seem to be under the illusion that elegant and beautiful bridge forms
cannot be envisioned by structural engineers.
This illusion is far from the truth. Bridges by structural engineers such as Maillart, Menn
and Virlogeux, among others, show that rational and functionally feasible bridge forms can
also display beauty and elegance. Even mass-produced high-way bridges can, and should be
attractive structures (Gottemoeller 2004).
How can we evaluate attractiveness of bridge designs? Great engineers have reflected upon
this question when discussing their own and others work (e.g., Menn 1991, Virlogeux 1996,
Baus & Schlaich 2008 among others). One approach is to use the measures of structural art.
Thomas Telford first presents the basis of structural art which has since then been adopted
and further developed by David Billington (Telford 1814, Billington 1983). The three meas-
ures of structural art are the scientific, social, and symbolic aspects of the structure, refer-
ring to structural efficiency, economy, and elegance, respectively. Structural art is achieved
through an elegant, still cost-effective structure which displays its function clearly with its
form (Billington 1983, 2003). New technologies, design, analysis, and construction methods
have emerged within structural engineering since these measures were developed. It is not the

275

NYBRIDGE_Book.indb 275 7/17/2013 8:44:28 AM

 276 Bridge aesthetics & history

intent to challenge the structural art concept in this discussion, but rather to adapt it by tak-
ing advantage of current technologies and approaches.
The scientific aspect is measured by efficiency of structural form and use of construc-
tion materials. Conventionally, this analysis has been performed through numerical analy-
sis based on design values or by lab tests of scale models. The evaluation of structural
behavior can for example be performed through Finite Element Analysis (FEA) of the
structure. However, true structural behavior does not only depend on design, the construc-
tion also plays a significant role. Connection details, actual material strength and other
construction aspects can affect the true structural behavior. Numerical analysis assumes
that the structure truly behaves in accordance with design. Many of the structural art-
ist used testing and deflection measurements to verify that their structures behaved as
they intended (Garlock & Billington 2008). In recent years Structural Health Monitoring
(SHM) has been gaining momentum and it provides an excellent way of assessing true
structural behavior. It is therefore argued here that using FEA combined with SHM pro-
vides an effective and accurate way to evaluate the scientific measure; although this is not
the primary aim of SHM.
The social aspect is related to economy and the transformation of society through sav-
ings of (often public) funds and the social benefits of the use of the structure. Construc-
tion cost-comparisons comprise the core of the social measure. Bridges are often built using
public funds emphasizing the importance of keeping costs at a responsible level. However,
construction cost does not provide the full picture of the cost of the structure to the owner.
Life-cycle cost has become a recognized way of evaluating the cost of a construction over
its service time from construction to demolition. A structure with low construction cost due
to minimized or cheap material usage may need costly retrofits or repeated maintenance
increasing the life-cycle cost significantly whereas sustainable designs which use low main-
tenance materials might result in higher construction costs. Life-cycle cost analysis therefore
could provide a long-term evaluation of the cost of the structure which is what the society
will pay, eventually.
Finally, the symbolic measure is a result of aspiration to elegance, i.e., to artistic appear-
ance and meaning of the structure. “This idea includes, along with the visual expression
of efficiency and economy through thinness and integration, the expression of contrast”
(Billington 1983). The symbolic measure is inherently the most qualitative of the three meas-
ures of structural art. It can be evaluated in part through how well the structure interacts
with its surrounding environment and its iconic importance for the region. The aspiration to
slenderness and integration favors logical and rational forms which have function as well as
aesthetics.
These three aspects of structural art are often inter-correlated and considerations of all of
them affect the choices made by the designer. Which ideal is the most prominent during the
design process depends on the designer’s approach, and also on the functional, environmen-
tal, and social constraints imposed on the structure, including construction and life-cycle cost
requirements. A large scale structure, for example, will need more emphasis on the scientific
aspect than smaller structures in order to span larger distances and overcome greater forces.
Contrary, the social and symbolic ideals can have more influence on the design of smaller
structures, since there is no major scientific challenge. This is especially emphasized in an
urban area where the integration of the structure into its environment and its interaction
with the user is very important.
The objective of the paper is to discuss structural art (a concept first introduced in
1814, and which has been developed and adapted since then) and display how we can
take advantage of current technologies and approaches in evaluating structures as art. By
identifying works of structural art and making the engineering community aware of them,
new designs and constructions may be positively influenced: by understanding the real
behavior and performance of structures as built some less rational forms may be avoided
in the future.

NYBRIDGE_Book.indb 276 7/17/2013 8:44:28 AM

 Measures of structural art: A case study using Streicker Bridge 277

2 CASE STUDY: STREICKER BRIDGE

This chapter aims to study Streicker Bridge, a pedestrian bridge on Princeton University
Campus, through the lens of structural art as discussed above. Streicker Bridge is a recently
built pedestrian bridge at Princeton University Campus (open in June 2010). The bridge is
104 meters long and crosses Washington Road. Streicker Bridge was designed by Swiss engi-
neer Christian Menn in collaboration with HNTB architecture and engineering firm, whose
lead engineers for the project were Theodore P. Zoli and Ryan Woodward.
Streicker Bridge has a main span and four approaching ramps, so-called “legs”. Struc-
turally, the main span is a deck-stiffened arch and the legs are curved continuous girders
supported by steel columns thus the deck carries all significant bending moments, not trans-
ferring them down to the arch or columns. The legs are horizontally curved and the shape
of the main span follows this curvature. The arch and columns are weathering steel while
the main deck and legs are reinforced post-tensioned concrete. A bridge rendering, photo-
graph, and map are shown in Figure 1. Besides its primary aim, to provide and facilitate safe
pedestrian crossing over Washington Road, the bridge has strong symbolic and aesthetic
significance for Princeton University. In the three following sections the three measures of
structural art are discussed for Streicker Bridge.

2.1 Scientific measure

Streicker Bridge is equipped with an SHM system based on fiber-optic technology for research
and educational purposes. Only the part of the monitoring system relevant to this paper will be
presented here, see Glisic 2011, Glisic et al., 2011, Hubbell 2011, and Sigurdardottir et al., 2011
for more detailed description and analysis. Discrete long-gauge Fiber Bragg-Grating strain and
temperature sensors where embedded in the concrete during the construction of the bridge. The
SHM system was designed using a loose structural analysis approach (Glisic & Inaudi 2007).
Half of the main span and one of the approach legs were equipped. Figure 2, shows the sensor

Figure 1. (a) Rendering (Princeton University 2012a), (b) photograph of Streicker Bridge, (c) map of
Science Neighborhood on Princeton Campus, adopted from (Princeton University 2012b).

Figure 2. Elevation of Streicker Bridge. Sensor locations are indicated with black lines. Each cross-
section typically has two sensors parallel to the centroid line of the deck.

NYBRIDGE_Book.indb 277 7/17/2013 8:44:28 AM

 278 Bridge aesthetics & history

locations. The cross-sections above columns and in the middle of spans are monitored. Every
monitored cross-section has two sensor locations, at the top and at the bottom; the sensors are
close to the axis of symmetry of the cross-sections and parallel to the centroid line of the bridge.
The design of Streicker Bridge strives to achieve global efficiency through its form. The
bridge was derived using graphic statics, minimizing (cancelling) the bending moments in the
deck under dead load. The dead load attributed to each column was calculated and represented
with a resultant vector (P in Figure 3). The funicular polygon shown in Figure 3 was then used
to determine the magnitude and direction of the normal forces, N, in the arch segments.
Similar methods are often used to determine the shape of arches. What sets Streicker
Bridge apart from most typical arch bridges is that it is a deck-stiffened arch and the arch is
truly segmented, following the shape of the funicular polygon exactly. This is possible since
the arch is very slender and its own weight is negligible in comparison to the weight of the
deck. The arch segments are welded together, providing moment rigidity, which can lead to
undesirable effects since the arch is assumed to carry only compressive forces. An evaluation
of the true structural behavior will reveal if this connection detail compromises the global
efficiency of the design. In summary, the aim of the design and form of Streicker Bridge is to
let the deck carry the bending moments keeping the arch in pure compression, thus minimiz-
ing the amount of material needed to construct the arch. Furthermore, the shape of the arch
is such that the bending moments due to dead loads are cancelled in the deck, facilitating its
slenderness and in turn minimizing the use of materials.
In order to evaluate the efficiency of Streicker Bridge the SHM system was used to con-
firm that the bridge behaves according to its design. If the deck does not transfer the bending
moments down to the arch and columns, then the system is considered efficient. A more detailed
analysis of this test and its results are presented in the literature (Sigurdardottir et al., 2012).
A test of the bridge was performed on March 18, 2011. Four golf carts were used as a mov-
ing static load, see Figure 4. The carts were positioned sequentially at 13 different locations

Figure 3. Dead load funicular polygon and half span of the bridge showing that the shape of the
arch is based on the diagram. Note that the arch is truly polygonal, following the shape of the funicular
polygon exactly.

Figure 4. Testing of Streicker Bridge. Golf carts used as moving static load.

NYBRIDGE_Book.indb 278 7/17/2013 8:44:29 AM

 Measures of structural art: A case study using Streicker Bridge 279

along the bridge, and measurements were taken at every sensor on the bridge simultaneously.
The bending-moment was calculated for each sensor pair. Finally the influence lines for
bending moments were created. A set of influence lines provides information about moment
distribution in the structure and the influence lines are therefore an excellent tool to deter-
mine if any moment is transferred to the arch and columns.
A Finite Element Model (FEM) was created using the analysis program SAP2000 (CSI
Berkeley). Frame elements (CSI Berkeley, 2007) were used to simulate the arch, the support-
ing columns, the main deck and the southeast deck. Geometric complexity required some
simplifications, including incremental change of cross-sections, stiffness modifications for
the V-columns and connection details between the arch and legs. The golf carts were modeled
as four point loads, each acting at the approximate axel location of the carts. The loads were
moved over the bridge model and the bending moments calculated at the sensor locations,
creating influence lines similar to those obtained by measurements.
The influence lines for bending moment generated by the load test were compared with
those obtained by FE modeling. The two sets of influence lines were, in general, in a very
good qualitative and quantitative agreement and gave a good idea about the static behavior
of the bridge. This confirms that the FEM is suitable to capture the true structural behavior
and can be used to study the moment distribution between the arch and the deck. This is
done since no sensors were installed on the arch due to aesthetic reasons. Figure 5 shows that
the moments in the deck are in general 100 times larger than the bending moments in the
arch. This confirms that the arch does not carry any significant bending moments. A similar
analysis shows that the southeast leg is a continuous girder, transferring no significant bend-
ing moments down to the columns.
The FE model describes the southeast leg of Streicker Bridge as a continuous girder and the
main span as deck-stiffened arch. The influence lines derived from measurements are in a very
good agreement with the FE model, they confirm the structural system of the bridge, and conse-
quently they confirm that Streicker Bridge is an efficient structure, which behaves as designed.

Figure 5. Influence lines for three cross-sections in the deck and for one point on the arch. Note the
difference in scale on the y-axis, kN-m vs. N-m. The bending moment is primarily carried by the deck.

NYBRIDGE_Book.indb 279 7/17/2013 8:44:29 AM

 280 Bridge aesthetics & history

2.2 Social measure

The bridge facilitates mobility across the busy Washington Road and provides a safe
connection between Princeton University communities. Created as a part of Princeton
University’s Natural Sciences neighborhood, the bridge connects the Icahn Laboratory
(Lewis-Sigler Institute for Integrative Genomics) and new neuroscience and psychology
building (under construction) on the west with Jadwin Hall (physics), Fine Hall (mathemat-
ics), Frick Chemistry Laboratory, and Lewis library on the east. In addition, sport facilities
(Roberts Soccer Stadium, Lenz Tennis Center, Poe Field, etc.) on the west are connected with
the sport facilities on the east (Princeton Football Stadium, Clarke Baseball Field, Finney
Field, Weaver Track Stadium, etc.), see Figure 1c. Streicker Bridge has become an integral
part of the university, changing the way people move around the campus.
As discussed in the introduction the social (economic) ideal is assessed through a cost com-
parison. Table 1 lists the construction costs of three pedestrian bridges from the cheapest to
the most expensive per square meter. All three bridges are steel pipe arch bridges.
The three bridges are all within the same cost range (2–5% difference), Streicker Bridge being
the least expensive per square meter but the most expensive in total cost. From this comparison
Streicker Bridge can be considered to have a reasonable cost. Menn argues that increasing con-
struction cost by 5–7% to achieve a visually interesting structure is acceptable (Menn 1985).
A more detailed analysis of the cost and construction of Streicker Bridge reveals why Stre-
icker Bridge can be considered cost-effective even if it is the most expensive of the bridges
listed. Due to relatively strict safety conditions imposed by particular site requirements and
owner’s responsibilities a high fence was installed in addition to the traditional railing. These
requirements increase the cost in comparison with the other bridges. The arch is made from
Yoloy, high-strength low-alloy steel which corrodes on the exterior to protect itself. This elim-
inates need for maintenance, such as painting. This material does increase the initial material
cost, but this is compensated for by the life-cycle cost savings. Similar analysis applies for the
ornamental metals, i.e. railing, fence, and soffits, which are all stainless steel minimizing the
maintenance cost. The authors do not foresee any need for maintenance for years to come.
On the other hand, the Solférino Bridge had dynamic problems due to transverse motions
on its inauguration day and had to be closed for further reinforcement (Blekherman 2007).
Costly retrofits are not reflected in the construction cost, further emphasizing the impor-
tance of using life-cycle cost evaluations for the economic measure. Taking these factors into
consideration Streicker Bridge can be considered as a cost-effective structure in long-term
although it is not the least expensive bridge of its kind by construction cost. Most impor-
tantly, the total cost of the bridge is likely to be close to its life-cycle cost.

2.3 Symbolic measure

Menn’s design approach can be illustrated with a target, shown in Figure 6 (Menn 1985).
Contrary to the “classical” procedure where accent is given to serviceability and structural
safety, Menn believes that these two aims are easy to achieve (they are mostly driven by the

Table 1. Construction costs of steel pipe arch pedestrian bridges, to the best of the authors’
knowledge.

Bridge Year Total length [m] Cost [USD] Cost/m2

Streicker Bridge 2010 104 $5,866,085 $8,321

Solférino Bridge 1999** 140 $5,765,641 $8,643
SC Route 27, SC 2003* 227 $5,627,247 $10,656

* Construction of Streicker Bridge started in 2008 therefore costs are projected to March 2008 using
ENR Construction Cost Index.
** International project, projected to March 2008 using ENR Construction Cost Index.

NYBRIDGE_Book.indb 280 7/17/2013 8:44:31 AM

 Measures of structural art: A case study using Streicker Bridge 281

codes); therefore, for him, the sequence in design starts with the aesthetics (symbolic aspect)
and the economy (social aspect), which depend on the engineer’s talent and creativity. He
emphasizes three main features related to the visual aspect of bridges; transparency and slen-
derness, simplicity and regularity, and artistic shaping, highlighting that “All possible angles
of view of the bridge user are certainly important” (Menn 1985).
The symbolic ideal is based on an evaluation of the visual appearance of the bridge. Four
main features are most prominent, 1) “X”-shape in plan, 2) curved legs, 3) branch-like col-
umns, and 4) the arch. Each feature is discussed separately, as follows, emphasizing the func-
tion and the user perception.

2.3.1 “X”-shape in plan

The geometrical shape of Streicker Bridge is particular because it connects four different
points in plan due to its “X”-shape, whereas most other bridges only connect two points. This
characteristic increases its accessibility and adds to the function of the bridge; as it essentially
serves as two bridges, see Figure 7.
The symmetric shape conforms to the aspiration for simplicity and regularity. Symbolically,
Streicker Bridge has great significance and meaning for Princeton’s campus. The bridge’s
“X”-shape in plan symbolizes cross-disciplinary collaborations within the natural sciences.
Even though this feature is not clearly visible for users outside of the campus community the
bridge will “stand as a tangible symbol of the cross-disciplinary collaborations that are cen-
tral to scientific research and teaching today” (President S.M. Tilghman in News at Princeton
2006) (Tilghman 2006).

2.3.2 Curved legs

In pedestrian bridges, it is particularly important to account for lateral movement (Roberts
2005), and the curved legs of Streicker Bridge have a structural function in that regard as well

Figure 6. Menn’s design philosophy (Menn (1985).

Figure 7. Streicker Bridge, picture taken from above. Showing the “X”-shape in plan.

NYBRIDGE_Book.indb 281 7/17/2013 8:44:31 AM

 282 Bridge aesthetics & history

as serving as the walkway. The four-approach ramps merge to form two leaning arch-pathways
and a single main span that laterally brace the slender pipe arch underneath, see Figure 8. The
bridge’s configuration takes care of lateral vibration just by its elegant structural form. Apart
from serving a structural function the leaning arches create a visually interesting structure as
the pedestrian approaches the bridge. Once on the bridge the pedestrian travels on a curved
path creating a more intriguing experience than on a straight path; see Figure 8.

2.3.3 Branch-like columns

Streicker Bridge is elegant; in its interaction and integration with the surrounding environ-
ment, see Figure 9.
While the bridge may contrast sharply with both the natural woodland backdrop and the
campus buildings around it, it is at the same time unobtrusive as its legs naturally evolve
from the flat pedestrian paths to rise over the Washington Road. Moreover, the decision to
use weathering steel is consistent with the bridge’s surroundings. The deep red, brown of
the rusted steel fits into the woody background. The “V”-shaped columns provide lateral
stability in addition to blending with the neighboring tree branches. It is this careful balance
between contrast and harmony, which contributes to the visual importance of the structure.

2.3.4 The arch

For vehicular traffic the most prominent feature of the bridge is the arch itself, see Figure 10.
Its slender and transparent profile creates a gate at the south entrance of Princeton Univer-
sity Campus.
According to Menn, “The visual expression of efficient structural function is…a funda-
mental criterion of elegance in bridge design,” and in designing Streicker Bridge, he sought to
promote a clear structural expressionism that followed the flow of forces (Menn 1990). The
thinness not only allows for the segmented construction but also enhances the visual expres-
sion of the polygonal shape.
Streicker Bridge integrates well with its natural and built surroundings. The features men-
tioned above cannot be experienced all at once; the vehicular traffic will have a different

Figure 8. Two leaning arches create the curved approach legs and merge in the main span.

Figure 9. Branch-like columns in harmony with the surrounding woodland.

NYBRIDGE_Book.indb 282 7/17/2013 8:44:32 AM

 Measures of structural art: A case study using Streicker Bridge 283

Figure 10. The truly polygonal arch is the gate to the south-entrance of Princeton University Campus.

perception than the pedestrian crossing over the bridge or the bicyclist biking under it. The
bridge provides a visually intriguing experience and perception from different angles allowing
a variety of sensations.

3 CONCLUSIONS

The Scottish engineer Thomas Telford first introduced structural art in the beginning of the
19th century (Telford 1814). Professor David Billington has developed the three measures,
the scientific, social and symbolic (Billington 1983). New technology and increased environ-
mental awareness provide new opportunities in the evaluation of structural art. Structural
Health Monitoring (SHM) in combination with Finite Element Analysis (FEA) provides an
excellent tool for evaluation of the scientific measure. The social/economic aspect is tradi-
tionally studied through construction cost-comparisons; bringing in life-cycle cost considera-
tions gives an even better view of the cost-effectiveness of the structure. The importance of
studying bridge designs through the measures of structural art is ever growing as the iconic
role attracts sculptors to the design of bridges, often leading to ill-designed bridge forms.
The symbolic aspect of structures will always be subject to individual opinions and views but
using criteria of for example, slenderness and regularity, as well as functionality of form one
can argue for aesthetically pleasing or unpleasing features of structure.
Streicker Bridge on Princeton University Campus was used to demonstrate how new tech-
nologies and approaches can be used to evaluate bridges in the framework of structural art. The
three ideals, scientific, social, symbolic were studied. The bridge was shown to behave according
to design, using a combination of SHM and FEA. The main span was shown to truly behave
as a deck-stiffened arch and the approach legs as continuous girders. This structural system
argued to be efficient as it cancels the bending moments in the deck due to dead loads. A cost-
comparison between similar bridge types showed that constructing the bridge was expensive, but
cost-effective in long-term, especially through its use of maintenance free materials. The bridge
has a large iconic impact on the Princeton University Campus; it blends in with its woodland
environment and symbolizes the cross-disciplinary atmosphere of the science neighborhood.
This paper shows the potential of new approaches in determining whether a structure
meets the efficiency and economy criteria of structural art. The elegant Streicker Bridge
meets these criteria and can be argued to qualify as a work of structural art.

ACKNOWLEDGEMENTS

This project has been realized with great help, and kind collaboration of several profession-
als and companies. We would like to thank Steve Hancock and Turner Construction Com-
pany; Ryan Woodward and Ted Zoli, HNTB Corporation; Dong Lee and A.G. Construction

NYBRIDGE_Book.indb 283 7/17/2013 8:44:34 AM

 284 Bridge aesthetics & history

Corporation; Steven Mancini and Timothy R. Wintermute, Vollers Excavating & Construction,
Inc.; SMARTEC SA, Switzerland; Micron Optics, Inc., Atlanta, GA. In addition the fol-
lowing personnel, departments, and offices from Princeton University supported and helped
realization of the project: Geoffrey Gettelfinger, James P. Wallace, Miles Hersey, Paul Pruc-
nal, Yanhua Deng, Mable Fok; Faculty and staff of Department of Civil and Environmental
Engineering and our students: Maryanne Wachter, Jessica Hsu, George Lederman, Jeremy
Chen, Chienchuan Chen, Thomas Mbise, Peter Szerzo, Allison Halpern, Morgan Neal, Dan-
iel Reynolds, Pedro Afonso, Daniel Schiffner, and David Hubbell.

REFERENCES

Baus, U. and Shclaich M. (2008). Footbridges: structure, design, history. Birkhäuser, Basel, Boston,
Berlin.
Billington, D.P. (1983). The Tower and the Bridge. New York: Basic Books, Inc., Publishers.
Billington, D.P. (1997). Robert Maillart: Builder, Designer, Artist. New York: Cambridge University
Press.
Billington, D.P. (2003). The Art of Structural Design: A Swiss Legacy. Princeton: Princeton University
Art Museum.
Blekherman, A.N. (2007). “Autoparametric Resonance in a Pedestrian Steel Arch Bridge: Solferino
Bridge, Paris.” Journal of Bridge Engineering, 12(6), 669–676.
CSI Berkeley. SAP2000.
CSI Berkeley. (2007). CSI Analysis Reference Manual for SAP2000.
Garlock, M.E.M. & Billington, D.P. (2008). Félix Candela. Engineer, Builder, Structural Artist.
Princeton University Art Museum in association with Yale University Press.
Glisic, B. (2011). “Streicker Bridge: an on-site SHM laboratory at Princeton University Campus”,
SMAR-2011, February 8–10, 2011, Dubai, UAE.
Glisic, B., and Inaudi, D. (2007). Fibre Optic Methods for Structural Health Monitoring. Chichester:
John Wiley & Sons.
Glisic, B., Chen, J., and Hubbell, D. (2011). “Streicker Bridge: a comparison between Bragg gratings
long-gauge strain and temperature sensors and Brillouin scattering-based distributed strain and tem-
perature sensors”. SPIE Smart Structures/NDE, March 6–10, 2011, San Diego, USA.
Gottemoeller, F. (2004). Bridgescape. The Art of Designing Bridges, 2nd Edition. John Wiley & Sons,
Inc.
Hubbell, D. (2011). Determining the structural behavior of Streicker Bridge using fiber optic sensors.
MSE Thesis, Princeton University; 2011.
Liew, K. (2010). Bridging art and engineering: structural art and performance analysis of the Streicker
Bridge. Senior Thesis, Princeton University.
Menn, C. (1985). “Aesthetics in Bridge Design.” Bulletin of International Association for Shell and
Spatial Structures, 26–2(88), 53–62.
Menn, C. (1990). Prestressed Concrete Bridges. Boston: Birkhauser Verlag.
Menn, C. (1991). “An approach to bridge design”. Engineering Structures, 13(2), 106–112.
Princeton University. (2012a). Facilities of Princeton University. Web. 26. Nov. 2012: http://www.
princeton.edu/facilities/info/major_projects/PedBridge/
Princeton University. (2012b). Campus Map. Web. 13. Dec. 2012: http://etcweb.princeton.edu/pumap/
Roberts, T.M. (2005). “Lateral Pedestrian Excitation of Footbridges”. Journal of Bridge Engineering,
ASCE, 10(1), 107–112.
Sigurdardottir, D., Hubbell, D., Afonso, J., and Glisic, B. (2011). “Streicker Bridge: assessment of struc-
tural health condition through static and dynamic monitoring”. SHMII-5, December 11–15, 2011,
Cancun, Mexico.
Sigurdardottir, D., Afonso, J., Hubbell, D., and Glisic, B. (2012). “Streicker Bridge: A two-year moni-
toring overview”. IABMAS 2012, July 8–12, 2012, Streso, Lake Maggiore, Italy.
Telford, T. 1814. ”Bridge”. New Edinburgh Encyclopedia. Whiting and Watson, New York. Second
American edition, vol. IV, 470–532.
Tilghman, S. (2006). “Alumnus funds pedestrian bridge to link two sides of campus.” Princeton Weekly
Bulletin 96.12: Web. 12 Jan 2010. http://www.princeton.edu/pr/pwb/06/1211/3a.shtml
Virlogeux, M. (1996). “Structural and Architectural Design of Bridges”. Structural Engineering Inter-
national. 6(2), 80–83.

NYBRIDGE_Book.indb 284 7/17/2013 8:44:35 AM

 Chapter 26

Lindenthal and the Manhattan Bridge eyebar chain controversy

K. Gandhi
Gandhi Engineering, Inc., New York, NY, USA

ABSTRACT: Manhattan Bridge was the third East River bridge planned in 1899 for
construction, alongside the Brooklyn and Williamsburg Bridges. On January 1, 1902 Seth
Low became Mayor of New York City, and appointed Gustav Lindenthal as Commissioner
of the Bridge Department. Design plans of the Manhattan Bridge with wire cables as the
main supporting elements were already at an advanced stage, and the tower foundations were
practically completed. Lindenthal rejected the old plans and presented a new design, which
instead employed eyebar chains. There was an immediate strong opposition to this move. The
Board of Aldermen refused to fund this construction scheme. When Seth Low lost reelection
in 1904, Lindenthal also lost his job. The bridge was built as a wire cable suspension bridge
and opened to traffic on December 31, 1909. This paper covers the controversy between eye-
bar chain and wire cable, and Lindenthal’s performance as Commissioner.

1 INTRODUCTION

Although this paper is about the Manhattan Bridge and the eyebar chain controversy, it is
very difficult to separate Gustav Lindenthal, the person who created this controversy, from
his impact on the major bridge construction projects in New York City years after his depar-
ture as Commissioner of Bridge Department from January 1, 1902 to December 31, 1903.
He was one of the visionary and creative bridge engineers who was very active for almost
half a century, straddling the last 15 years of the 19th and first 35 years of the 20th centuries,
until his death in 1935. Lindenthal had obtained half-a-dozen patents related to the use of
wirelink chain as a supporting member of a suspension bridge before he assumed office as
Commissioner of the Bridge Department; however, none of them was ever used for a bridge
project (Griggs 2012).
When Lindenthal came into office, the foundations of the Manhattan Bridge towers were
practically completed, and the design plans were at an advanced stage for a wire cable suspen-
sion bridge. Richard S. Buck was in charge of the Manhattan and Blackwell’s Island (now
Ed Koch Queensboro) Bridges, and Leffert L. Buck (no relation to R.S. Buck) was in charge
of the Williamsburg Bridge. The Williamsburg Bridge was expected to be completed in about
a year.
Based on Lindenthal’s actions in the first few months in office, which were published in the
newspapers and technical journals of that period, it became apparent that he had planned to
1. Use his patented wirelink chains for the Manhattan Bridge,
2. Take credit for the successful completion of the Williamsburg Bridge,
3. Use delays by John A. Roebling’s Sons Co. (Roebling Co.) in completion of the Williamsburg
Bridge to promote his wirelink chain scheme for the Manhattan Bridge, and
4. Reorganize the Bridge Department to paralyze it.

285

NYBRIDGE_Book.indb 285 7/17/2013 8:44:35 AM

 286 Bridge aesthetics & history

2 LINDENTHAL’S PATENTS RELATED TO WIRELINK CHAIN

The details and description of Lindenthal’s patents related to his wirelink chain are taken from
the paper by Griggs (2012). Lindenthal’s second patent (#311,338) is shown in Figure 1.
Figure 1A shows “a braced chain suspension bridge for a single span with two forms of
bracing; Figure 1B is for multiple span braced chain suspension and arch bridges. In Fig-
ure 1B he also provided details to balance loads on pier tops by connecting top and bottom
members of adjacent spans (Griggs 2012). Lindenthal’s fifth patent (#500,267) was for chain
cable links made of looped wire. His sixth patent (#542,829) “incorporated his wire link
cables but this time had three rows of links connected together vertically with eyebar links”
(Griggs 2012).

3 PROPOSAL TO USE EYEBAR CHAINS FOR MANHATTAN BRIDGE

On his fourth day in office (January 4, 1902), Lindenthal called R.S. Buck and others in his
office and strongly expressed his belief that wire link chains were the proper form of construc-
tion for the Manhattan Bridge and had to be adopted. Again, on March 20, 1902, he asked
R.S. Buck and his assistant Henry A. LaChicotte to prepare the plans for the Manhattan
Bridge with wire links for cables (Buck, R.S. 1904). Lindenthal even loaned his set of draw-
ings to R.S. Buck that he had prepared for the design of the Quebec Bridge Competition
using his patented wire link chains (Patent #542,829).
Lindenthal denied that he ever forced his subordinates to use his patented wirelink chains.
Once it became publically known that Lindenthal was promoting his patented wirelink chain
system for the Manhattan Bridge, he said that what he was suggesting to use was the common
eyebar chain system in which the eyebars were connected using a pin of the same diameter as
the eyebars. He further implied that R.S. Buck had a hand in publication of newspaper articles
critical of his performance (Lindenthal 1904). In early March of 1902, R.S. Buck was demoted
to the position of Principal Assistant Resident Engineer (Railroad Gazette 1902c), and his
assistant, Henry A. LaChicotte, was promoted to Engineer-in-Charge of the Manhattan and

Figure 1A. Lindenthal Patent #311,338, Figure 1B. Lindenthal Patent #311,338, Multiple
Single Span Suspension Bridge. Span Arch/Suspension Bridge.

CH26_Gandhi-Final.indd 286 7/17/2013 2:41:34 PM

 Lindenthal and the Manhattan Bridge eyebar chain controversy 287

the Blackwell’s Island Bridges (Engineering News Supplement 1902a). R.S. Buck resigned, and
accepted a job as the Chief Engineer of Dominion Bridge Company of Montreal, Canada.

4 CREDIT FOR SUCCESSFUL COMPLETION OF WILLIAMSBURG BRIDGE

Lindenthal knew that the Williamsburg Bridge would be completed within the two years of
his term as the Commissioner of the Bridge Department. Leffert L. Buck was the Chief Engi-
neer of the Williamsburg Bridge since 1895. As long as L.L. Buck was in charge of the Wil-
liamsburg Bridge, the credit for its successful completion would go to L.L. Buck. Lindenthal
appointed L.L. Buck as the Chief Engineer of the Bridge Department, and then planned to
fire him and offered his position of Chief Engineer to Foster Crowell, a prominent bridge
engineer of that period, who declined the offer. Word leaked out about Lindenthal’s plans.
Calls came for the removal of Lindenthal, which forced Mayor Seth Low to act. The mayor
held a meeting in May 1902 with L.L. Buck and persuaded Buck to return to the Williamsburg
Bridge as its Chief Engineer (Railroad Gazette 1902e). Lindenthal then assumed the dual
titles of the Commissioner and the Chief Engineer of the Bridge Department.

5 USE DELAYS BY ROEBLING CO. TO PROMOTE EYEBAR SCHEME

In 1901, the Roebling Co. asked for a time extension of 10 months for wire cable making of the
Williamsburg Bridge, but was given 6 months by the Williamsburg Bridge Commission. Roebling
Co. was behind schedule when Lindenthal came into office on January 1, 1902. He used this
delay to promote his eyebar scheme, promising that it would be speedier and less costly.
Lindenthal assessed a liquidated damage of $1,000 per day against the Roeblings and
deducted a penal sum of more than $175,000 from the payment due to the Roebling Co.
(Hungerford 1903). The Roeblings went to court to recover the money withheld by the Bridge
Department. The case was tried before Judge Alton B. Parker, who ruled in favor of the
Roeblings and pointed out that “there is nothing in the contract which says the Commis-
sioners are clothed with power to determine whether there was any delay on the part of the
plaintiff ” (Engineering News 1905).

6 REORGANIZATION OF BRIDGE DEPARTMENT

Lindenthal’s moves were meant to sideline or remove all senior competent engineers in the
Bridge Department who disagreed with him on technical issues. Lindenthal made the follow-
ing additional changes:
Charles C. Martin, involved in the construction of the Brooklyn Bridge with the Roeblings
since 1869, was the Chief Engineer and Superintendent of the Brooklyn Bridge since its
opening in 1883, was appointed Consulting Engineer of the Bridge Department to advise
the Commissioner. The reason was that Martin approved a plan, which became known as
the Martin plan, to relieve congestion on the Brooklyn Bridge for commuters who crossed the
bridge during rush hour. Samuel R. Probasco, who was the Assistant Engineer of the Brook-
lyn Bridge since its beginning, was appointed Assistant Engineer in Charge of machine shops,
repair plants, and repair forces (Railroad Gazette 1902a).
Othniel F. Nichols, who was Engineer-in-Charge of the Williamsburg Bridge, was fired.
Leon S. Moisseiff, who was an Assistant Engineer on Computation Design of the Blackwell’s
Island and Manhattan Bridges, was appointed Assistant to Lindenthal.
Two engineers, realizing that talk of redesigning of the Manhattan Bridge would delay fur-
ther the construction, resigned. They were Francis L. Pruyn, in charge of construction on the
Brooklyn side (Engineering News Supplement 1902b), and Frank M. Sylvester in charge of

NYBRIDGE_Book.indb 287 7/17/2013 8:44:36 AM

 288 Bridge aesthetics & history

construction on the Manhattan side. Oscar Erlandsen was promoted from Assistant Engineer
to Principal Assistant Engineer (Engineering Record 1902a).

7 LINDENTHAL’S LETTER TO THE BROOKLYN EAGLE

Lindenthal, in an open letter to the Editor of the Brooklyn Eagle, accused his staff of cor-
ruption, incompetence, and mismanagement—charges his staff could not respond to. The
irony of the whole matter was that the engineers Lindenthal was criticizing were far more
experienced than he was, and what he claimed as facts were either distorted or plain lies. The
letter is reprinted in its entirety below (Lindenthal 1902):
To the Editor of the Brooklyn Eagle:
Some time ago I had received a well-meaning but unnecessary, warning, that a cer-
tain aggrieved firm of Tammany contractors were conspiring with some disaffected
underlings in my department, to attack me through the newspapers for purposes of
their own and which are not hard to guess.
The ‘twaddle about “change of plans” and “wire links” and “high engineering
authorities and other momentous nothings, which broke loose of a sudden, did not
come, therefore, as a surprise to me, and it would not be worth my notice except that
insinuations are being made against my character. As you have given space in your
paper to such attacks upon me I trust you will have the fairness to publish with the
same prominence what I have to say on this matter:
It is true, and nothing new, that among many patents on various inventions I hap-
pen to have some relating to wire links as is well known to engineers from technical
publications here and abroad. It is not a fact, however, that I had asked for a change
of plans in bridges Nos. 3 and 4 over the East River so that they may be used in them.
Any assertion to the contrary is a lie, as stupid as it is malicious.
It appears needful, however, to state that I found much room for improvement in the
plans for those bridges on the score of greater economy, better aesthetic appearance
and for making them fireproof. The plans for Blackwell’s Island Bridge provided no
access to the island, etc., etc. The plans showed immaturity throughout and had never
been examined or passed upon by expert engineers. It is the business of the Commis-
sioner of Bridges to see that the city shall have the best and most economical plans that
can be had whether the engineers of the department like it or not. I am not disposed
to take these engineers at their own valuation. I hold that the plans and specifications
shall be subjected to the scrutiny and criticism of the best experts before I approve
them. For this course I believe I have ample justification the Brooklyn Bridge, twenty
years ago, when steel was nearly twice as costly, cost only 10 cents per pound, finished
in place. That is, after twenty years of progress, when every other form of steel product
has cheapened, the city is paying 40 per cent more for its cables, and at a time when
steel wire cables for some recent Western suspension bridges were erected for less than
8 cents per pound. Another experience in Bridge No. 2 was furnished, when the specifi-
cations for a large amount of steel were needlessly so drawn that only one steel works
in the country could furnish it, at prices about 50 per cent higher than the best struc-
tural steel in first class railroad bridges elsewhere. The extra cost on these items alone
represents in the aggregate over $500,000 taken out of the city treasury for whimsical
engineering, to say the least. It would seem that these few illustrations should be suf-
ficient to warrant my caution with plans and specifications.
I have only to add that the earlier cause of my dissatisfaction with the engineering
department was too great a willingness to allow delay in the work and to pass contrac-
tors’ claims, which had no basis in the agreements for same.
GUSTAV LINDENTHAL
Commissioner of Bridges, Department of Bridges, New York, April 14, 1902.

NYBRIDGE_Book.indb 288 7/17/2013 8:44:36 AM

 Lindenthal and the Manhattan Bridge eyebar chain controversy 289

8 FACT CHECK ON LINDENTHAL’S LETTER TO THE BROOKLYN EAGLE

In his one page letter, Commissioner Lindenthal accused his perceived enemies, a particular
contractor (Roebling Co.) and certain individuals in his department (L. L. Buck and R. S.
Buck) of corruption, collusion, incompetence, lack of thoroughness, and wastefulness. Fol-
lowing is an analysis of the four key issues raised by Lindenthal in his letter.

8.1 Tammany contractor and underlings in bridge department

Mayor Low came to power as a reform candidate, and using that as an excuse Lindenthal
blamed all contractors (specifically the Roeblings) who did business with the previous admin-
istration. Roebling had a contract to furnish wire cables for the Williamsburg Bridge, and
the main person who was dealing with the Roeblings was L. L. Buck. Lindenthal tarred the
reputation of L. L. Buck, whose record was spotless.
L.L. Buck’s major experience included the construction of the Verrugas Viaduct high up
in the Andes in Peru, South America (Gandhi 2009); repairs to the first railway suspension
bridge over the Niagara Gorge originally constructed by John A. Roebling (Buck L.L. 1881);
replacement of the stone towers of the same bridge with iron towers without interrupting
the railroad traffic except for 8-1/2 hours during which time the cables were transferred from
the stone to the iron towers (Buck L.L. 1887). He also designed the replacement of the sus-
pension bridge by an arch bridge of record span of 550 ft. and a record combined loading
of 10,000 lbs. of live load per foot for the two level arch bridge. In 1895, L. L. Buck was the
unanimous choice to design and build the Williamsburg Bridge (Gandhi 2006). There was no
inappropriate relationship found between the Roeblings and L. L. Buck.

8.2 Making Manhattan and Blackwell Island’s Bridges more economical

and aesthetically appealing
Since making the Manhattan Bridge more economical and aesthetically pleasing is covered in
this paper, we have discussed only the Blackwell’s Island Bridge in this section.
The Blackwell’s Island Bridge was designed by Richard S. Buck as a double cantilever
bridge with two suspended spans, each of 330 ft. Lindenthal revised the design by eliminating
the two suspended spans and increased the lengths of the cantilever arms so that they met at
mid-span. In order to respond to criticism generated as a result of this change, Mayor Low
appointed a Commission of Engineers on November 3, 1902 consisting of William H. Burr,
Henry W. Hodge, and Palmer C. Ricketts. The Commission gave its report comparing the
original plan of R. S. Buck with the recent plan by Lindenthal and “the Board determined
that if they had to pick one of the designs, they would pick the Buck design” (Griggs, 2012)
The Board also appended a third plan, widening the deck width, which was approved by
Mayor Low.
In his letter to the Brooklyn Eagle, Lindenthal had emphasized “greater economy” in his
plans for the Manhattan and Blackwell’s Island Bridges. After revising the plans for the
Blackwell’s Island Bridge, Lindenthal wrote to Mayor Low (Griggs 2012) as follows: “It may
be of interest to your Honor to learn that a cost of bridge structure (from anchor pier to
anchor pier) under the old plans, fire-proof flooring, would have been $4,793,000.00. The
cost under the revised plans is estimated to be $3,956,000, a saving of $837,000, or about 17%
(City 1903f).”
When the bids for the superstructure of the Blackwell’s Island Bridge were opened on
September 24, 1903, there was only one bid for $5,255,514 submitted by Pennsylvania Steel
Co. which was about 33% higher than Lindenthal’s estimate. This single bid was thrown out,
and the project was rebid.
The second time, two bids were received on November 5, 1903: one of $5,132,985 from
the Pennsylvania Steel Co. and the second of $5,188,850 from Milliken Brothers. It was no
secret among the construction community that the cost of erection was inflating the bid

NYBRIDGE_Book.indb 289 7/17/2013 8:44:36 AM

 290 Bridge aesthetics & history

prices (Engineering Record 1903f). This proved that Lindenthal’s design was not economical.
The original cantilever design with suspended spans, which most contractors were familiar
with, would have been lower, close to $4,793,000 as estimated by R.S. Buck. The contract
for the superstructure of the Manhattan Bridge was awarded to Pennsylvania Steel Co. for
$5,132,985 at 30% higher than Lindenthal’s estimate.

8.3 Immaturity of staff and plans never been examined by expert engineers
Nothing could be further from the truth when Lindenthal made this statement. In his paper,
Griggs presents line diagrams of 8 different cantilever and suspension bridge schemes pre-
pared by Thomas C. Clarke and Richard S. Buck for the Manhattan and Blackwell’s Island
Bridges (Griggs 2012). R.S. Buck’s design included the experience gained on the Brooklyn
and Williamsburg Bridges in presenting his plans. The selection of the final designs for
each of these two bridges were arrived at after careful consideration of locations, techni-
cal advances in the art and science of bridge construction, loads imposed on these bridges,
and budgets. Lindenthal did not ask R.S. Buck about his alternative schemes for these two
bridges.
R.S. Buck assisted L. L. Buck during the design of the Railway Arch Bridge at Niagara
Falls and later became the Resident Engineer for the same bridge (Buck R.S. 1898). After
the demolition of the Niagara Suspension Bridge, originally built by John A. Roebling, the
cables were salvaged. Each cable weighed about 170 tons and was about 1,400 long. R.S.
Buck used these cables to build a 25 ft. wide suspension bridge connecting Lewiston, NY
and Queenston, Ontario, Canada with a clear span of 800 ft. for electric cars, vehicles, and
pedestrians. R.S. Buck also designed rigid frame reinforced concrete Green Island and Goat
Island Bridges at Niagara Falls, NY. He was appointed Chief Engineer in July 1899 for the
design of the Manhattan and Blackwell’s Island Bridges.
R.S. Buck was the first to develop the specification for the use of nickel steel for the
Manhattan Bridge. This was also the first time in the history of bridge engineering that nickel
steel was specified for a major bridge. These two bridges retained most of the structural fea-
tures originally designed by R. S. Buck.

8.4 Price of steel wires

This is another subject where Lindenthal tried to mix up facts with figures. Lindenthal stated
that steel (wires) cost 10 cents a pound, finished in place for the Brooklyn Bridge 20 years
ago, and the City was paying 40% more for the cables of the Williamsburg Bridge. He accused
L. L. Buck of developing specifications for the steel wires in such a way that only one steel
mill in the entire U.S. (Roeblings) could meet these tight specifications; as a result the City of
New York paid half-a-million dollars more just for this one item.
Colonel Washington Roebling, who was the Chief Engineer of the Brooklyn Bridge recalled
a price of 16 cents a pound, whereas Charles C. Martin who was the so called Project Manager
recalled the price being close to 20 cents a pound (Railroad Gazette 1902d). The wires in Brook-
lyn Bridge had an ultimate strength of 160,000 psi whereas those in the Williamsburg Bridge had
an average strength of 215,000 psi. If one assumes that the galvanized wires for the Brooklyn
Bridge were delivered on the ground for 10 cents a pound, they still had to be cleaned, oiled,
and rewound, and threaded up into cables. The old wire for the Brooklyn Bridge was delivered
in 800 ft. lengths, whereas the new wire for the Williamsburg Bridge was delivered in 3,800 ft.
lengths, requiring only one-fourth the number of splices. Also L. L. Buck required that the wire be
produced from the open-hearth steel as opposed to the old steel produced by the Bessemer proc-
ess. It was calculated that the new cables in position would cost 14–1/4 cents per pound of wire.
In summary, the new steel wire for the Williamsburg Bridge had 34% more tensile strength, fewer
splices, and was still cheaper than the steel for the Brooklyn Bridge. Lindenthal had distorted
his figures. There was no collusion found between the Roeblings and the engineers of the Bridge

NYBRIDGE_Book.indb 290 7/17/2013 8:44:36 AM

 Lindenthal and the Manhattan Bridge eyebar chain controversy 291

Department. There was no overpayment of $500,000 for the cables of the Williamsburg Bridge
as claimed by Lindenthal.
The “Brooklyn Eagle” wrote a scathing editorial in response to Lindenthal’s letter (Engi-
neering Record 1902b).

9 LINDENTHAL’S PROPOSAL TO USE EYE-BAR CHAINS

In his report for the six months ending on June 30, 1902, Lindenthal discussed the progress
of the three East River bridges under construction (Engineering News 1902). As for the
Manhattan Bridge, Lindenthal stated that the original plans contemplated the use of four
wire cables in combination with four stiffening trusses. However, because wire cables were
a source of delay in the on-going construction of the Williamsburg Bridge and for greater
economy, the plans were revised. Under the new plans, the wire cables were replaced by
forged eyebars of high grade steel in combination with a more effective stiffening system,
thereby reducing the construction and future maintenance costs.
Learning from the experience gained with passenger transportation over the Brooklyn
Bridge, Lindenthal had proposed a two level bridge with four tracks, instead of the former
two on the upper levels for the elevated railroads, and on the lower level four tracks for trolley
cars or moving platforms, as later exigencies may decide he proposed a fireproof bridge with
wide roadway and two promenades.
The plans also included new architectural and aesthetic features that were subject to review
and approval by the Municipal Art Commission under Section 637 of the Revised City
Charter. Lindenthal recommended awarding a single contract that would include both the
steel towers and the entire superstructure (including the chain cables), by pointing out that it
would eliminate delays resulting from separate contracts for different parts of the bridge.
The details of the main span, typical cross section of the roadway, and the main tower are
shown in Figures 2 through 4, respectively (Engineering Record 1903a). These are the details
of the bridge proposed by Lindenthal and not of the bridge that was finally built.
Figure 4 shows the details of the main tower with a hinge at the base and connections for
eyebar chains (Engineering Record 1903g).

Figure 2. Main spans of Manhattan Bridge.

Figure 3. Typical section of roadway.

NYBRIDGE_Book.indb 291 7/17/2013 8:44:36 AM

 292 Bridge aesthetics & history

Figure 4. Details of one of the main towers and connections for eyebar chains, Manhattan Suspension
Bridge, New York City proposed by Lindenthal.

10 HILDENBRAND’S RESPONSE TO LINDENTHAL PROPOSAL

Wilhelm Hildenbrand, who was Engineer of Cable Construction for Roebling Co., responded
to Lindenthal’s reasons for the selection of the eyebar chain system point by point, and made
a comparison between wire cables and eyebar chains under exactly similar conditions of loads,
stiffness, unit stresses, pressure on foundations, and margin of safety (Hildenbrand 1903a).
Hildenbrand demonstrated that, if the Williamsburg Bridge were to be built using the eye-bar
chain, it would be about 50 percent heavier than the wire cable system. He calculated the addi-
tional costs in detail for various major components of the Williamsburg Bridge as follows:

Anchor Chains $77,080

Anchorages $432,040
Towers $186,200
Tower Foundations $153,320
Cables $2,514,891
Total $3,363,531

Hildenbrand concluded that the eyebar chain system would take longer to build and cost
between $2 million and $3 million more if the Manhattan Bridge adopted the eyebar cable
system.

11 SELECTION OF A COMMISSION OF ENGINEERS BY MAYOR LOW

On February 10, 1903, Mayor Seth Low selected a Commission of five engineers and submit-
ted the revised plans of the Manhattan Bridge incorporating the eyebar cables to the Com-
mission, whose members were: (1) Lieut.-Col. Charles W. Raymond, U.S. Engineer Corp.,
(2) Mr. George S. Morison, (3) Mr. Charles C. Schneider, (4) Mr. Henry W. Hodge, and
(5) Prof. Mansfield Merriman (Engineering News Supplement 1903a).
This Commission was charged by Mayor Low to provide answers to the following four
questions (Engineering News 1903a):

NYBRIDGE_Book.indb 292 7/17/2013 8:44:37 AM

 Lindenthal and the Manhattan Bridge eyebar chain controversy 293

1. Are the plans in accordance with advanced knowledge of suspension bridge designing
with a view to economy of construction, provision for temperature stresses, rigidity under
concentrated loads, and resistance to wind pressures; also as regards quality of steel and
its protection against corrosion?
2. Will the strength, stability, and carrying capacity of the bridge be adequate for any con-
gestion of traffic that may occur on the railroad tracks, roadways, and promenades?
3. Will the structure, as designed, be fireproof ?
4. Do the plans permit a speedy erection of the superstructure after the completion of the
anchorages and tower foundations?

12 PRELIMINARY REPORT OF THE COMMISSION

The Commission of five engineers submitted its Preliminary Report on March 9, 1903. This
report covered all features important to the Art Commission such as general description of
the bridge and the design changes, loads, cables, towers, and anchorages, but excluded the
matters dependent on the quality of the steel. The Commission responded to the four ques-
tions posed by Mayor Low as follows (Engineering Record 1903b):
1. “The plans are in accordance with advanced knowledge of suspension bridge design-
ing. They are likely to be as economical in construction as other forms of suspension
bridges. They provide for a structure of unusual rigidity under concentrated loads. Ample
provision is made for wind pressure. They are consistent with the best protection from cor-
rosion. Our report on the quality of the steel must be deferred to await the results of tests
not yet made.
2. The strength, stability, and carrying capacity of the bridge will be adequate for any con-
gestion of traffic which may occur on the railroad tracks, roadway, and promenades if the
provisions for loads laid down heretofore are followed.
3. The structure as designed will be incombustible.
4. The design favors speedy erection of the superstructure after the masonry is ready.”

13 LINDENTHAL’S COST COMPARISON OF EYEBAR AND WIRE CABLE

The final report of the Commission of five engineers was delayed due to serious illness of the
Chairman, Mr. George S. Morison. Meanwhile, Commissioner Lindenthal prepared a com-
parison table between the eyebar chain and steel wire cable. The written text accompanying
Lindenthal’s comparison tried to respond to the points raised by Hildenbrand. The answers
were in favor of the eyebar system, and showed that the eyebar system would be $142,000
cheaper than the wire cable system (Engineering News 1903c; Engineering Record 1903c).

14 REBUTTALS TO LINDENTHAL’S COST COMPARISON

14.1 Wilhelm Hildenbrand

Hildenbrand contradicted all assumptions that Lindenthal had made in arriving at his con-
clusion that his eyebar design was cheaper and faster than the wire cable design. Hildenbrand
implied that Lindenthal had manipulated his numbers, and was pushing the eyebar scheme
with all the force that his official power as bridge commissioner gave him (Hildenbrand
1903c). Hildenbrand went back to reanalyze the design and came back with the following
conclusions (Hildenbrand 1903b):
1. The eyebar chains, in order to sustain the same loads with the same factor of safety, must
be 7-1/2 times heavier than the wire cables, and the total dead load of a chain bridge must
be 82% greater than that of a wire cable bridge.

NYBRIDGE_Book.indb 293 7/17/2013 8:44:37 AM

 294 Bridge aesthetics & history

2. The total cost of a chain bridge would be $5,280,000 more than the cost of a wire cable
bridge.

14.2 Gus Henning

Gus Henning, a prominent bridge engineer with wide experience in the design, fabrication,
erection of large suspension bridges and familiarity with the big bridge fabrication shops in
the U.S. and in Europe, questioned most of the assumptions, unit prices, and unit stresses
distorted by Lindenthal in favor of the eyebar cables (Henning 1903a).
For example, the Department of Bridges had the records of more than 10,000 tests on steel
wires that showed an average result of more than 185,000 psi (lbs. per sq. inch) yield point
and more than 225,000 psi ultimate capacity. However, Lindenthal chose to use 150,000 psi
for yield point and 190,000 psi for ultimate stress for steel wires. Similarly, he used one-third
of the yield stress as working or allowable stress for steel wires, and two-third of the yield
stress as a working stress for the eyebar. Lindenthal artificially inflated the price of steel wire
and lowered the cost per pound of nickel steel.
Henning also pointed out the additional and unknown erection problems and costs due
to the use of heavy eyebars, which had never been used before. Henning estimated that the
eyebar bridge would cost $1,465,700 more than the wire cable Manhattan Bridge.

15 FINAL REPORT OF THE COMMISSION

On June 29, 1903, the Special Commission of five engineers appointed by Mayor Seth Low
submitted its Final Report (Engineering News 1903d). Lieut. Col. Charles W. Raymond was
replaced by Theodore Cooper. The members reviewed the test results of nickel steel bars
manufactured and tested by different steel plants.
They stated that the results of their tests indicated satisfactory material would be obtained
with steel having the following properties in full-size annealed eyebars: ultimate strength,
85,000 lbs.; actual elastic limit, 48,000 lbs.; percentage of elongation in 18 ft., 9; and per-
centage of reduction at fracture, 40. The maximum unit stress under the actual dead load
combined with a working load of 8,000 lbs. per lineal foot, together with temperature effects,
should not exceed 30,000 lbs., according to the Commission. The tests indicated that these
properties could be obtained with open hearth steel with 3.25% to 3.5% of nickel, not over
0.05% of sulphur, and not over 0.06% phosphorous for acid steel or 0.04% for basic steel.
Three steel makers were said to be ready to furnish this material, and others would arrange to
do so in a reasonable time if the demand warranted it (Engineering Record 1903d).
It was the Commission’s conclusion that the strength of the steel eyebars having the above
properties would be 50% greater than the standard bars now used; and the working stress,
therefore, could be increased by the same proportion. Having satisfied that the high strength
nickel steel would be available for the making of the eyebars, the Commission unanimously
recommended the adoption and execution of the proposed design of the Manhattan Bridge
as submitted by Commissioner Lindenthal.

16 RESPONSE TO THE COMMISSION’S FINAL REPORT

The key responses came from the following individuals:

16.1 Wilhelm Hildenbrand

Hildenbrand had met with the Commission to present his arguments but failed to con-
vince the Commission that the wire cable design would be cheaper and faster. He summed

NYBRIDGE_Book.indb 294 7/17/2013 8:44:37 AM

 Lindenthal and the Manhattan Bridge eyebar chain controversy 295

up the substance of the two Reports in two sentences that “(1) It is possible to build a
bridge over the East River on Mr. Lindenthal’s plans; (2) It is possible to manufacture, at
a reasonable price, steel which in small test pieces shows a tenacity of 85,000 lbs. and an
elastic limit of 48,000 lbs. per sq. in.” He was disappointed that the Commission didn’t
do a thorough job by verifying facts and questioning all assumptions for their validity
(Hildenbrand 1903d).

16.2 George W. Colles

This Canadian engineer from Montreal commented on the idea of an eyebar chain bridge in
New York City by saying, “Surely New York is ugly enough already, and it is bad enough to
have skeleton-bridge towers, without the added eyesore of a chain bridge” (Colles 1903a). On
his second set of comments, Colles added that the Manhattan Bridge with eyebars would be
the ugliest of the three East River bridges just as the chain bridge over the Allegheny River
(designed by Lindenthal) at Pittsburgh is the ugliest of the three (Colles 1903b).
On the serious side, Colles thought the Commission had failed in its mission and consid-
ered it a folly to build a ten million dollar bridge without any information on its final costs.
He questioned most of the reasons given by the Commission in support of the eyebar bridge,
thereby exposing its lack of experience in the subjects for which the Commission was sup-
posed to give its expert opinion.

16.3 Gus C. Henning

The thrust of Henning’s comments was that the Commission had failed in answering the
four questions that it was asked to resolve, and the basic reason was that none of the five
members, all of them well-known engineers, had any experience with suspension bridge con-
struction (Henning 1903b). These two reports did not answer (1) the economy of the design;
(2) corrosion of the pin-connected bridges with excessive wear and corrosion of pins and
holes in the eyebars; he observed that it was impossible to prevent corrosion between pin
and eyebars; (3) behavior of the chain and changes in its shape under load and temperature
changes; does this change take place in the pin joints or by bending of the eyebar; (4) how the
test results of small nickel eyebars, 6-inches wide, correlate with the strength of an 18-in. wide
eyebar?; (5) the effect of friction losses of at least 15% in testing machines which would lower
the yield stress of nickel eyebars, from 48,000 psi to 42,000 psi and the ultimate stress from
85,000 psi to 72,250 psi; and (6) erection time; the report stated that the design favors speedy
erection of the superstructure once the masonry is ready. It was the unanimous opinion of all
engineers experienced in the erection of suspension bridges that the erection time would be
considerably longer for an eyebar cable.

16.4 Henry Szlapka

Szlapka raised the practical questions of painting multiple eyebars and the number of eye-
bars that can be packed at one joint and whether they would share the applied load equally.
He mentioned about the difficulty of heating uniformly large bars during the annealing
process. He questioned the applicability of test results obtained from small size eyebars to
large prototype eyebars. Szlapka believed that the only solution was to manufacture full-size
nickel eyebars and test them to learn about their performance (Szlapka 1903).

16.5 Leon S. Moisseiff

Moisseiff wasn’t responding to the Commission’s Final Report. As an Assistant to Com-
missioner Lindenthal, he was disingenuous in answering various points raised by Henry
Szlapka. While calling the various test results satisfactory for the nickel eyebars, he was with-

NYBRIDGE_Book.indb 295 7/17/2013 8:44:37 AM

 296 Bridge aesthetics & history

holding the technical information. Moisseiff was promoting the eyebar chain design for the
Manhattan Bridge (Moisseiff 1903).

16.6 Richard S. Buck

R. S. Buck, in his letter to the Railroad Gazette (Buck R.S. 1903), reiterated that his design
of the wire cable suspension bridge was fully developed before Lindenthal became the Com-
missioner. Buck compared the old design with the new design considering:
1. Comparative cost and speed of eyebar and wire cable erection;
2. Rigidity under partial loading;
3. Comparative capacity of the two designs;
4. Accessibility for protection and inspection;
5. Pin connections and suspender bands;
6. Fire-proof floor system; and
7. Rocker-bent towers.
Buck’s conclusion was that the wire cable design was superior to the eyebar chain design.
Lindenthal, not to be outdone, submitted his response on his last week in office that he was
correct in selecting the eyebar chain design (Lindenthal 1903d).

17 REJECTION OF FUNDING BY THE BOARD OF ALDERMEN

In the first week of June 1903, the Board of Estimates appropriated $6,533,000 for the super-
structure of the Manhattan Bridge. Commissioner Lindenthal announced his plan to award
a single contract covering all iron work including the eyebar chain cables.
On June 19, 1903, four of the most prominent bridge engineers of that period, Othniel F.
Nichols, Gus C. Henning, Richard S. Buck, and Foster Crowell, were invited by the Finance
Committee of the Board of Aldermen to express their views. They unanimously opposed the
substitution of wire cables by eyebar chains for the Manhattan Bridge (Engineering News
1903b). They protested against the appropriation of money for construction of the bridge
as planned by Commissioner Lindenthal on the ground that the cost of such a bridge would
exceed by several million dollars the cost of an equivalent bridge built with wire cables. As a
result, the Board of Aldermen on July 14, 1903 voted against the resolution to appropriate
$6,533,000 for the superstructure of the Manhattan Bridge (Engineering News 1903g).
After the first defeat of the appropriation by the Board of Aldermen, the Board of Esti-
mate approved the same amount, $6,533,000, for the superstructure of the Manhattan
Bridge. This last resolution of the Board of Estimate was taken up once more by the Board
of Aldermen, and was, for the second time, defeated by a vote of 41 to 25 on July 31, 1903,
when the Board of Aldermen refused to authorize a bond issue for the eyebar superstructure
of the Manhattan Bridge.
A large sum appropriated for the Manhattan Bridge still remained unexpended, and it
was possible for Lindenthal to continue work on the superstructure by splitting the work
into smaller contracts without interference from the Aldermen (Engineering Record 1903e).
Lindenthal threatened to use the available money “to build the tower foundations in such a
way to make it impossible to build any bridge but his own on them” (Griggs 2012).

18 EVENTS AFTER THE DEPARTURE OF LINDENTHAL

On January 1, 1904, George B. McClellan became the new mayor of New York City, and
he appointed George E. Best as his Bridge Commissioner. He selected Othniel F. Nichols
as his Chief Engineer. Richard S. Buck came back to the Bridge Department in charge of

CH26_Gandhi-Final.indd 296 7/17/2013 3:51:14 PM

 Lindenthal and the Manhattan Bridge eyebar chain controversy 297

the Manhattan and Blackwell’s Island Bridges. The new Commissioner threw out the plans
prepared by Lindenthal and went back to the original design for the Manhattan Bridge with
wire cables. The architectural ornamentations for the Manhattan Bridge recommended by
Henry F. Hornbostel, the architect appointed by Lindenthal, were also rejected (Engineering
News 1904a). George Best selected Carrere and Hastings, the architects who designed the
New York City’s Main Public Library on 42nd Street, as the new consultants to recommend
architectural treatment for the Manhattan Bridge. The new design incorporated the following
changes (Engineering News 1904b):
1. Wire cables were used for the main suspension members. The Lindenthal design had eye-
bar chains.
2. The towers consisted each of four steel columns of compact design, one under each cable,
set in a transverse row and braced together. They were fixed at the bottom, and support
cable saddles on top, the cables being continuous over the saddles. The Lindenthal design
provided a pin-bearing at the bottom of each tower column and had the suspension chains
connected by a pin at the top.
3. The stiffening trusses were parallel-chord trusses at the level of the roadway. The Lindenthal
design had its stiffening trusses along the suspension chains.
4. The large halls in the anchorage of the Lindenthal design were omitted.
The new design with wire cables was submitted to the Municipal Art Commission for the
meeting of July 12, 1904 for approval. This Commission had previously approved the eye-
bar chain design submitted by Lindenthal (Engineering News 1904d). Although Lindenthal
was out of office, he was still interested in the project. Lindenthal sent a letter to the Art
Commission criticizing the new design and insisting that the eyebar chain design was supe-
rior. He predicted that, under the new design, the deflections and vibrations under the
action of live load will be more than five times greater (Engineering News 1904c). The
meeting adjourned without any decision. However, at the Art Commission meeting held on
September 15, 1904, the new bridge design was approved by a vote of 7 to 1 (Engineering
News 1904e).
There were several delays in building the Manhattan Bridge. The stringing of wires for
permanent cables of the Manhattan Bridge began on August 10, 1908. The last of the 37,888
wires was strung on the afternoon of December 10, 1908. The four cables were completed in
four months. The Manhattan Bridge was officially opened on December 31, 1909 by Mayor
McClellan. This was also his last day in office. The entire superstructure was designed and
constructed in his administration starting in January 1904.

19 CONCLUSIONS

When recommendations were made about a head-on competition between the eyebar and
wire cables for the Manhattan Bridge (Engineering News 1903e and 1903f), Lindenthal
seemed reluctant and gave multiple reasons why it was not feasible (Lindenthal 1903a). He
did not respond to the analyses performed by his critics who showed the superiority of the
wire cable system in terms of both cost and erection time, perhaps because of his lack of
experience in fabrication and erection of large suspension bridges. However, it must be said
in Lindenthal’s defense that the blue ribbon panel of 5 prominent engineers concurred with
his system, and he was entitled to hold his ground against his critics.
The opportunity to provide contractors with an option to select either the eyebar chain or the
wire cable for a given suspension bridge came years later when the Port of New York Author-
ity (now Port Authority of New York and New Jersey) invited competitive bids on alternative
designs for parallel wire or eyebar chains for the George Washington Bridge. Proposals were
received from five bidders on October 3, 1927, resulting in 14 different combinations cover-
ing the steelwork. The contract for wire cables was awarded to the John A. Roebling’s Sons

CH26_Gandhi-Final.indd 297 7/17/2013 3:51:14 PM

 298 Bridge aesthetics & history

Co. for $12,339,977.00. This was the end of the eyebar chain controversy; it was established
from then on the superiority of the wire cable system.

20 LINDENTHAL’S PERFORMANCE AS BRIDGE COMMISSIONER

It is the author’s opinion that Lindenthal’s performance as a Bridge Commissioner was poor
based on the questions arising from the following:
1. Lindenthal’s presence in the Bridge Department had a corrosive effect on the morale of
the staff because of his published criticism of the most respected engineers in the depart-
ment, and demotion and cut in salaries of many staff members.
2. At the time Lindenthal took office, New York City had the best bridge engineers in the coun-
try—and possibly in the world—with experience in the design and construction of long-span
bridges. The basic design parameters of the Manhattan and Blackwell’s Island Bridges were
already decided, and the designs had progressed to an advanced state. Had Lindenthal not
intervened, these two bridges would have been built around 1905 or 1906. The delays caused by
his intervention cost the City millions of dollars by redesigning these bridges multiple times.
3. By removing L. L. Buck from the Williamsburg Bridge to make him the Chief Engineer,
Lindenthal delayed the job by more than 6 months. Even after the Williamsburg Bridge
was completed, there was a long delay in opening the bridge to traffic because Lindenthal
had failed to make arrangements with the street railway companies for crossing the bridge.
The City lost revenue during this period.
4. Practically no new construction took place during Lindenthal’s two years in office. The
work that was performed was either planned or on-going before he joined the Bridge
Department. Lindenthal’s combative style and “know-it-all” approach alienated his staff
and members in other branches of City government whose support was vital to get the
funding in place for the City’s major bridge projects. The Board of Aldermen refused twice
to fund the Manhattan Bridge project because of Lindenthal’s insistence that the eyebar
chain design was cheaper and faster, although he had no basis to prove it. Any tactful Com-
missioner would have made peace with the Board of Aldermen to build these bridges and
reduce congestion on Brooklyn Bridge, the only East River Bridge existing at that time.
5. During his two years in the office (and thereafter), Lindenthal was mired into petty and
long arguments with individuals who thought differently than he did. He was wasting an
inordinate amount of time in letter writing and responding activities. This was not the role
of the Commissioner of the Bridge Department.
6. When L. L. Buck gave up his position as the Chief Engineer and moved back to the
Williamsburg Bridge with the intervention of Mayor Low, Lindenthal appointed himself
as the Chief Engineer. This is akin to the statement President Lincoln made that a “person
who acts as his own attorney has a fool for a client.” Lindenthal lost the opportunity to
get frank and confidential opinions and advice from a trusted subordinate by becoming
the Chief Engineer of his own Department.
7. Due to his perceived insensitivity to the plight of commuters crossing the Brooklyn
Bridge during the rush hours, the Brooklyn Eagle wrote an article which characterized
Mr. Lindenthal as follows (Brooklyn Eagle 1902):
One of the difficulties with Mr. Lindenthal is that he knows too much. It follows that
those who know everything can be taught nothing. That is the Lindenthal predicament.
There is no necessity for adding to this bill of particulars. There is one way out of the
trouble which time aggravates rather than subdues. Let us have a new Commissioner.
This demand for a change is made in all seriousness. It is addressed to the mayor,
upon whom the real responsibility for Lindenthal has been imposed.
The author concurs with this assessment of Commissioner Lindenthal.

NYBRIDGE_Book.indb 298 7/17/2013 8:44:37 AM

 Lindenthal and the Manhattan Bridge eyebar chain controversy 299

ACKNOWLEDGMENTS

The author thanks Robert Mayo, Director of Folsom Library of Rensselaer Polytechnic
Institute (RPI), Troy, NY, and Brenda Hill of Gandhi Engineering, Inc. for her help in the
preparation of the paper.

REFERENCES

Brooklyn Eagle 1902. “Swanstrom and Lindenthal,” April 17, 1902, p. 4.

Buck, L.L. 1881. The re-enforcement of the anchorage and renewal of the suspended superstructure of
the Niagara Railroad Suspension Bridge. ASCE Transactions X(CCXXIII): 195–224.
Buck, L.L. 1887. Replacing the stone towers of the Niagara Railway Suspension Bridge, with iron tow-
ers. ASCE Transactions XVII(368): 204–212.
Buck, R.S. 1898. The Niagara railway arch. ASCE Transactions XL(836): 125–177.
Buck, R.S. 1903. Eye-Bars or Cables for Manhattan Bridge. Railroad Gazette 35(49), December 4, 1903:
859–862.
Buck, R.S. 1904. The Manhattan Bridge Design. Railroad Gazette 36(2), January 8, 1904: 19.
Colles, George W. 1903a. Engineering Impressionism; and the Manhattan Bridge. Engineering News
50(4), July 23, 1903: 79.
Colles, George W. 1903b. Further Remarks on the Aesthetics of the Manhattan Bridge and Engineering
Impressionism. Engineering News 50(6), August 6, 1903: 124.
Engineering News 1902. Progress on the New East River Bridges at New York City. 48(8), August 21,
1902: 124.
Engineering News 1903a. Report of Special Commission of Engineers on the Plans for the Manhattan
Bridge, New York City. 49(11), March 12, 1903: 243–244.
Engineering News 1903b. The Proposed Substitution of Eye-Bars. 49(26), June 25, 1903: 557.
Engineering News 1903c. Comparison of Eye-Bar Chains with Steel Wire Cables for the Manhattan
Bridge. 49(26), June 25, 1903: 576.
Engineering News 1903d. Final Report of Special Commission Plans for the Manhattan Bridge. 50(1),
July 2, 1903: 24.
Engineering News 1903e. Editorial. 50(2), July 9, 1903: 36.
Engineering News 1903f. Editorial. 50 (3), July 16, 1903: 49.
Engineering News 1903g. The Manhattan Suspension Bridge. 50(6), August 6, 1903: 113.
Engineering News 1904a. The Architectural Decoration of the Manhattan Bridge. 51(14), April 7,
1904: 327.
Engineering News 1904b. The Manhattan Bridge across the East River, New York City. Revised Plans.
52(1), July 7, 1904: 1–3.
Engineering News 1904c. The New Designs for the Manhattan Bridge over the East River. 52(3), July
21, 1904: 70.
Engineering News 1904d. Editorial. 52(11), September 15, 1904: 239.
Engineering News 1904e. The Design for the Manhattan Bridge. 52(12), September 22, 1904: 270.
Engineering News 1905. Editorial. 54(14), October 5, 1905: 356.
Engineering News Supplement 1902a. Personal. 47(11), March 13, 1902: 93.
Engineering News Supplement 1902b. Personal. 47(13), March 27, 1902: 113.
Engineering News Supplement 1903a. Personal. 49(8), February 19, 1903: 89.
Engineering Record 1902a. Personal and Obituary Notes. 45 (2), January 11, 1902: 41–42.
Engineering Record 1902b. Amenities in the New York Bridge Department. 45(16), April 19, 1902:
361–362.
Engineering Record 1903a. The Manhattan Bridge, New York. 47(8), February 21, 1903: 196–197.
Engineering Record 1903b. Preliminary Report of Manhattan Bridge Commission. 47(12), March 21,
1903: 296.
Engineering Record 1903c. A Comparison of Eyebar Chains with Steel Wire Cables for the Manhattan
Bridge. 47(26), June 27, 1903: 696–697.
Engineering Record 1903d. Notes and Comments. 48(3), July 18, 1903: 63.
Engineering Record 1903e. Notes and Comments. 48(6), August 8, 1903: 151.
Engineering Record 1903f. The Blackwell Island Bridge Bids. 48(19), November 7, 1903: 547.

NYBRIDGE_Book.indb 299 7/17/2013 8:44:37 AM

 300 Bridge aesthetics & history

Engineering Record 1903g. Towers and Anchorages of the Manhattan Bridge. 48(24), December 12,
1903: 730–732.
Gandhi, K. 2009. Verrugas Viaduct and its reconstruction, Peru, South America. In Khaled M.
Mahmoud (ed.), Safety and Reliability of Bridge Structures, CRC Press, 2009: 461–479.
Gandhi, K. 2006. Roebling’s railway suspension bridge over Niagara Gorge. 5th International Cable
Supported Bridge Operators’ Conference, New York City, 2006.
Griggs, Jr., Francis E. 2012. Gustav Lindenthal, Dean of American Bridge Builders, Paper 568. The
Municipal Engineers Journal 98(1): 22–70.
Henning, Gus C. 1903a. Wire Cables or Eye-Bar Cables for the Manhattan Bridge. Engineering News
50(1), July 2, 1903: 14–15.
Henning, Gus C. 1903b. Criticisms of the Report of the Special Committee of Engineers on the Plans
for the Manhattan Bridge. Engineering News 50(5), July 30, 1903: 102.
Hildenbrand, Wilhelm 1903a. A Comparison between Eye-Bar Chain and Wire Cables for Suspension
Bridges. Engineering News 49(11), March 12, 1903: 229–231.
Hildenbrand, Wilhelm 1903b. The Comparison between Eye-Bar Chains and Wire Cables for the
Manhattan Suspension Bridge. Engineering News 50(1), July 2, 1903: 14.
Hildenbrand, Wilhelm 1903c. Wire Cables Versus Eyebar Chains for Suspension Bridges. Engineering
Record 48(1), July 4, 1903: 22–23.
Hildenbrand, Wilhelm 1903d. The Board of Experts’ Report on Eye-Bar Cables for the Manhattan
Bridge. Engineering News 50(2), July 9, 1903: 38.
Hildenbrand, Wilhelm 1903e. The Design of Long Span Suspension Bridges. Engineering News 50(24),
December 10, 1903: 522–523.
Hungerford, Edward 1903. The Williamsburg Bridge. Eagle Press, Brooklyn, NY: 117–118.
Lindenthal, G. 1902, Letter to the Editor, Brooklyn Eagle, April 15, 1902.
Lindenthal, G. 1903a. Some Comparisons between the Original and the New Designs for the Manhattan
Bridge. Engineering News 50(5), July 30, 1903: 102.
Lindenthal, G. 1903d. The Manhattan Bridge Designs. Railroad Gazette 35(52), December 25,
1903: 913.
Lindenthal, G. 1904. The Manhattan Bridge Designs. Railroad Gazette 36(3), January 15, 1904: 35.
Moisseiff, Leon S. 1903. Eye-Bar Chain Bridge and Tests of Nickel Steel Therefor. Engineering News
50(11), September 10, 1903: 225.
Railroad Gazette 1902a. The Bridge Department of New York City. 34(4), January 24, 1902: 64.
Railroad Gazette 1902b. The Architecture of Bridge Design. 34 (6), February 7, 1902: 93.
Railroad Gazette 1902c. Personal. 34(11), March 14, 1902: 199.
Railroad Gazette 1902d. A Brooklyn Bridge Episode. 34(16), April 18, 1902: 290.
Railroad Gazette 1902e. Personal. 34 (19), May 9, 1902: 353.
Szlapka, Henry 1903. Engineers’ Reports and Chain Cable for Large Suspension Bridges. Engineering
News 50(7), August 13, 1903: 143–144.

NYBRIDGE_Book.indb 300 7/17/2013 8:44:37 AM

This page intentionally left blank
 Edited by
Mahmoud

On Thursday evening, May 23, 2013, the Interstate 5 Bridge over the Skagit
River in Washington state collapsed due to impact by an oversize truck, dumping
vehicles and people into the water. Fortunately, the bridge is located in a rural
area and nobody was killed in the accident, but three people were rescued after
their cars plunged into the frigid water of the Skagit River. According to
Washington state officials, the bridge was inspected last year and was not
structurally deficient, but collapsed because of apparent impact from an oversize
truck. Nevertheless, the collapse of the steel truss bridge renewed appeals for

Durability of Bridge Structures

greater investment in the nation’s aging infrastructure. These appeals are echoed
throughout the bridge engineering community worldwide, as the condition of
deteriorated bridges worsens with increasing traffic loads combined with lack
of proper maintenance.

Bridge engineers from different countries shared their experience toward

achieving durable bridge structures, during the 7th New York City Bridge
Conference, held on August 26-27, 2013. This book contains select papers that
were presented at the conference. These peer-reviewed papers are valuable
contributions and of archival quality in bridge engineering.

Durability of Bridge Structures

Edited by Khaled M. Mahmoud

an informa business